Sage Journals: Discover world-class research

Abstract

Systemic preconditioning with the TLR9 ligand CpG induces neuroprotection against brain ischemic injury through a tumor necrosis factor (TNF)-dependent mechanism. It is unclear how systemic administration of CpG engages the brain to induce the protective phenotype. To address this, we created TLR9-deficient reciprocal bone marrow chimeric mice lacking TLR9 on either hematopoietic cells or radiation-resistant cells of nonhematopoietic origin. We report that wild-type mice reconstituted with TLR9-deficient hematopoietic cells failed to show neuroprotection after systemic CpG preconditioning. Further, while hematopoietic expression of TLR9 is required for CpG-induced neuroprotection it is not sufficient to restore protection to TLR9-deficient mice that are reconstituted with hematopoietic cells bearing TLR9. To determine whether the absence of protection was associated with TNF, we examined TNF levels in the systemic circulation and the brain. We found that although TNF is required for CpG preconditioning, systemic TNF levels did not correlate with the protective phenotype. However, induction of cerebral TNF mRNA required expression of TLR9 on both hematopoietic and nonhematopoietic cells and correlated with neuroprotection. In accordance with these results, we show the therapeutic potential of intranasal CpG preconditioning, which induces brain TNF mRNA and robust neuroprotection with no concomitant increase in systemic levels of TNF.

Keywords

acute stroke brain ischemia focal ischemia ischemic preconditioning and induced neuroprotection neuroprotection

INTRODUCTION

Despite progress in endovascular and surgical procedures for patients with coronary-artery disease and atherosclerosis, there is significant peri-procedural morbidity because of ischemic brain injury. The use of diffusion-weighted imaging has recently confirmed that a substantial number of patients undergoing cardiac surgeries such as coronary artery bypass grafting, cardiac valve replacement, and aortic repair, display evidence of peri- or post-operative cerebral ischemic lesions.^1–4 These incidences of brain ischemia-reperfusion can be anticipated before they occur; thus, prophylactic treatment would be highly beneficial for these patients. Research investigating prophylactic stroke therapies has identified Toll-like receptor ligands as preconditioning agents that provide robust protection against ischemic injury.

Systemic low-dose administration of the TLR9 agonist unmethylated cytosine-phosphate-guanine-rich DNA oligonucleotides (CpG) before stroke induces neuroprotection against subsequent ischemic damage in rodent and nonhuman primate models.^5,6 CpG preconditioning requires tumor necrosis factor (TNF) to protect against stroke injury, as shown by the failure of TNF-deficient mice to show neuroprotection after CpG preconditioning.⁶ However, the target of TLR9 engagement and subsequent role of TNF is unclear. TLR9 is expressed on a wide variety of cell types both in the brain (microglia, neurons, astrocytes, cerebral vascular endothelium) and on hematopoietic cells in the periphery.^7–11 Thus, it is unknown whether CpG preconditioning activates TLR9 on cells within the brain or whether TLR9 signaling in systemic circulating cells is required to elicit the neuroprotective phenotype.

After systemic administration of CpG, circulating hematopoietic cells are activated resulting in cell state changes that affect cytokine production^12–14 and endothelial reactivity.⁷ Hematopoietic cell activation can induce changes in parenchymal tissues including the brain, resulting in a secondary cascade of events.¹⁵ In addition, it is possible that CpG administered systemically crosses the blood–brain barrier (BBB) to directly affect the brain. We have previously shown that mouse mixed cortical cells pretreated with CpG have significantly decreased cell death in an in vitro model of ischemia, indicating that direct activation of brain cells in the absence of hematopoietic cells induces neuroprotection against ischemic insult.⁶ Further, a study showed that TLR4 expression on hematopoietic cells is not required to elicit an inflammatory response in the central nervous system (CNS).¹⁶ Thus, systemic CpG preconditioning may induce protection through the direct activation of TLR9 on resident brain cells.

To determine whether hematopoietic cells are required for CpG preconditioning, we used TLR9-deficient (TLR9KO) and wild-type (WT) mice to generate bone marrow chimeric mice expressing TLR9 exclusively on hematopoietic cells (WT → TLR9KO) or parenchymal cells (TLR9KO → WT). The term ‘parenchymal cells’ refers to cells of the recipient that are radiation resistant and of nonhematopoietic origin, e.g., endothelial and epithelial cells, neurons, astrocytes, microglia, and smooth muscle cells. We show that hematopoietic expression of TLR9 is required, but not sufficient for systemic CpG preconditioning-induced neuroprotection. In addition, we provide evidence that although TNF is required for CpG preconditioning, systemic TNF levels did not correlate with the protective phenotype. However, induction of cerebral TNF mRNA required expression of TLR9 on both hematopoietic and parenchymal cells and correlated with neuroprotection. This suggests that induction of TNF in the brain is associated with CpG-induced neuroprotection. These results underscore the importance of understanding the mechanism associated with CpG-induced neuroprotection to further the development of candidate therapeutics for ischemic injury to the brain.

MATERIALS AND METHODS

Animal Care and Criteria

C57Bl/6J mice and the congenic mouse strain B6.SJL-Ptprc^a Pep3^b/BoyJ were obtained from Jackson Laboratories (West Sacramento, CA, USA). TLR9KO mice were purchased from Oriental BioService (Osaka, Japan) and were backcrossed eight generations onto C57Bl/6J. All stroke studies were performed with male mice between 10 and 14 weeks of age. All mice were randomized at time of drug treatment with surgeon blinded to treatment and genotype. All mice were given free access to food and water and housed in a facility approved by the Association for Assessment and Accreditation of Laboratory Animal Care International. Animal protocols were approved by the Oregon Health and Science University Institutional Animal Care and Use Committee (OWLAW# A3304–01) and met the guidelines set forth by the National Institutes of Health.

Generating Chimeric Mice

The TLR9KO mice backcrossed onto C57Bl/6 have the CD45.2 allelic version of the CD45 leukocyte common antigen expressed on all leukocytes. To distinguish between recipient and donor cells in the chimeric animals, we used the congenic strain B6.SJL-Ptprc^a Pep3^b/BoyJ, which have the CD45.1 allelic version of the CD45 leukocyte common antigen, as our TLR9 replete strain (WT). Recipient mice were irradiated with a split dose of 12 grays (6 grays each dose, separated by 4 hours). Bone marrow was flushed from the femurs and the tibias of donor mice, and 1 × 10⁶ cells were injected intravenously into irradiated recipients. Chimeras were administered 250 mg/L ciprofloxacin antibiotic added to drinking water for 2 weeks. Three bone marrow chimera groups were produced (donor → recipient): (1) WT recipients receiving bone marrow from TLR9KO mice (TLR9KO → WT), (2) TLR9KO recipients receiving bone marrow from WT mice (WT → TLR9KO), and (3) WT recipients receiving bone marrow from WT mice (WT → WT).

Determination of Reconstitution Efficiency

After 4 to 6 weeks of reconstitution, chimerism was verified using flow cytometry to determine the percent blood leukocytes positive for CD45.1 (TLR9 replete) versus CD45.2 (TLR9KO). Mice were deeply anesthetized with isoflurane, and blood was collected via cardiac puncture in sodium heparin and red blood cells were lysed using two volumes of 1 × RBC lysis buffer (Biolegend, San Diego, CA, USA). Samples were washed using cell staining buffer (Biolegend) and were incubated with Fc-Block (BD Biosciences, Franklin Lakes, NJ, USA) for 10 minutes. Cells were stained with mouse anti-mouse CD45.1 (eBiosicences, San Diego, CA, USA) and CD45.2 antibodies (BD Biosciences) conjugated with PeCy7 and FITC, respectively, for 20 minutes at 4°C. Results were acquired using an LSRII cytometer (BD Biosciences) with no fewer than 10,000 total events collected per sample. Post acquisition analysis was performed using FlowJo 9.1 (Tree Star Inc., San Carlos, CA, USA).

Drug Treatments

Mice were administered a subcutaneous injection of CpG ODN 1826 (CpG; Invivogen, San Diego, CA, USA), lipopolysaccharide (LPS) (Escherichia coli serotype 0111:B4; phenol extraction purified, protein content 3%; Sigma-Aldrich, St. Louis, MO, USA), or saline vehicle in a total volume of 100 μL. For intranasal administration, mice were administered CpG or vehicle in a total volume of 20 μL as described previously.¹⁷ Temperature, weight, and general appearance were assessed in a subset of animals at 2, 7, 24, 48, and 72 hours after drug treatment via both routes of administration. There was no effect on temperature, weight, or sickness behavior in CpG treated animals compared with vehicle treated. For ischemia experiments, mice were treated 72 hours before the induction of ischemia. For cytokine and quantitative PCR studies, tissues were collected at 1 and 3 hours after injection.

Ischemia-Reperfusion Model

Focal cerebral ischemia was induced by middle cerebral artery occlusion (MCAO) as described previously.⁶ In brief, MCAO was performed in anesthetized mice (1.5% to 2% isoflurane) by threading a 7–0 silicon-coated nylon surgical filament (Docol, Redlands, CA, USA) through the external right carotid artery to the internal carotid artery, blocking blood flow at the bifurcation into the middle cerebral artery and anterior cerebral artery. After 45 or 60 minutes of occlusion, the monofilament was removed and blood flow was restored. The duration of MCAO was optimized to obtain consistent baseline infarct sizes across studies between surgeons. Cerebral blood flow was monitored throughout the procedure by laser Doppler flowmetry (Transonic System Inc., Ithaca, NY, USA) and animals were excluded if blood flow was not reduced by 80% or greater during occlusion. Body temperature was maintained at 37°C during the surgery.

Evaluation of Infarct Size

Twenty-four hours after MCAO, mice were deeply anesthetized with isoflurane and perfused with ice-cold saline containing 2 U/mL sodium heparin. Brains with olfactory bulbs removed were sectioned into 1 mm slices beginning from the rostral end, for a total of seven slices. The infarct area was visualized by incubating the sections in 1.5% 2,3,5-triphenyltetrazolium chloride (TTC; Sigma-Aldrich) in phosphate-buffered saline for 10 minutes at 37°C before transferring to 10% formalin (Sigma-Aldrich). Sections were then imaged and the infarct was measured by an investigator blinded to the experimental conditions using ImageJ software (NIH Image, Bethesda, MD, USA). Infarct volume was calculated using the indirect method ((contralateral live – ipsilateral live)/contralateral live × 100) to account for the effects of edema, and the results are normalized to genotype-matched vehicle-treated controls.

Analysis of Plasma Tumor Necrosis Factor Levels

Mice were deeply anesthetized with isoflurane and blood was collected via cardiac puncture in sodium heparin (2 U/mL). The blood was then centrifuged (380 × g for 20 minutes) and the plasma was removed and stored at −80°C. ELISA was used to quantify plasma TNF (R&D Systems, Minneapolis, MN, USA). Samples were run in duplicate.

Tissue Processing and Quantitative Real-Time PCR

Total RNA was isolated from the cerebral cortex using the Qiagen Rneasy Lipid Mini Kit (Qiagen, Valencia, CA, USA). RNA was reverse transcribed using an Omniscript Reverse Transcription kit (Qiagen). Quantitative PCR was performed using the TaqMan Gene Expression Assays (Applied Biosystems, Carlsbad, CA, USA) on the Applied Biosystems Step One Plus PCR system. Results were normalized to β-actin expression and are represented as delta cycle threshold (dCT) values.

Statistical Analysis

Statistical testing was performed using Prism 5 software (GraphPad, LaJolla, CA, USA). Values are given as group means ± standard error of the mean (s.e.m.). Data were analyzed by Student's T-test, one-way ANOVA or two-way ANOVA, and Bonferroni's post hoc test as denoted in the figure legends. Significance was defined as P < 0.05.

RESULTS

CpG Preconditioning Requires TLR9

While CpG is known to activate TLR9 signaling, prior studies had not examined whether CpG preconditioning in vivo required the presence of TLR9 for neuroprotection against ischemic injury. To evaluate the requirement for TLR9 in CpG-induced protection, TLR9KO mice were preconditioned with CpG (1.6 mg/kg; subcutaneously) or vehicle 72 hours before MCAO (45 minutes). As anticipated, CpG preconditioning failed to protect TLR9KO mice (infarct size 93.55 ± 11.73% of vehicle; n = 6 to 11; P > 0.05;

Figure 1). Thus, CpG-induced neuroprotection in vivo is mediated through TLR9. Of note, comparison of vehicle-treated WT and TLR9KO mice showed no difference in infarct size as percent of hemisphere (42.6 ± 5.4% versus 38.2 ± 4.5%, respectively) showing that TLR9 does not have an endogenous role in ischemic injury.

Figure 1.

CpG preconditioning requires TLR9. C57Bl/6J or TLR9KO mice were administered subcutaneous injection of CpG (1.6 mg/kg; n = 11 C57Bl/6J and n = 8 TLR9KO) or vehicle (n = 6 C57Bl/6J and n = 9 TLR9KO) 72 hours before 45 minutes middle cerebral artery occlusion (MCAO). Infarct volume was determined by 2,3,5-triphenyltetrazolium chloride (TTC) staining 24 hours after MCAO. Preconditioning with CpG reduced infarct damage after MCAO for C57Bl6/J mice but not for TLR9KO mice. Values are group means ± s.e.m.; **P < 0.01 versus vehicle control by two-way ANOVA and Bonferroni post hoc test.

TLR9 Is Required in Both the Hematopoietic and Parenchymal Compartments for CpG-Induced Neuroprotection Against Brain Ischemia

TLR9 is expressed on multiple cell types within the body including circulating hematopoietic cells and brain cells.^7–11 To address the role of hematopoietic cells in response to systemic CpG preconditioning, we generated radiation-induced reciprocal bone marrow chimeras from WT (CD45.1 haplotype) and TLR9KO (CD45.2 haplotype) mice. We generated three groups of chimeric mice: (1) mice lacking TLR9 on hematopoietic cells (TLR9KO → WT), (2) mice lacking TLR9 on parenchymal (nonhematopoietic) cells (WT → TLR9KO), and (3) control mice replete with TLR9 (WT → WT). To confirm chimerism, we stained peripheral blood from these animals with antibodies specific to the CD45.1 (WT) and CD45.2 (TLR9KO) leukocyte common antigens to distinguish between donor and recipient cells, and then analyzed these samples using flow cytometry. Four to six weeks after bone marrow transfer, >80% of the hematopoietic cells were donor derived in the chimeric mice (Figure 2).

Figure 2.

Generation of TLR9 reciprocal bone marrow chimeric mice. Mice were irradiated with a split dose of 12 grays then given intravenous injection of donor bone marrow cells. TLR9KO → WTand WT → TLR9KO chimeric mice were generated. To confirm chimerism, different allelic forms of the CD45 leukocyte common antigen are used to distinguish WT cells (CD45.1 positive; y axis) and TLR9KO cells (CD45.2 positive; x axis) in the systemic circulation by flow cytometry. Four to six weeks after bone marrow transfer, >80% of the hematopoietic cells were donor derived in the chimeric mice. Representative plots for each group are shown. WT, wild type.

To determine whether TLR9 expression in the hematopoietic compartment has a role in CpG-induced neuroprotection, chimeric mice were preconditioned with CpG (1.6 mg/kg; subcutaneously) or vehicle 72 hours before MCAO, and ischemic damage was evaluated 24 hours after occlusion. As expected, WT → WT mice administered CpG exhibited significant protection from ischemic injury compared with vehicle-treated controls (infarct size 65.66 ± 8.34% of vehicle; n = 11/group; Figure 3). In contrast, TLR9KO → WT mice administered CpG failed to show a significant reduction in infarct size (91.71 → 4.63% of vehicle; n = 9/group; Figure 3), indicating that TLR9 expression on hematopoietic cells is required for neuroprotection. Furthermore, WT → TLR9KO mice administered CpG also failed to show a significant reduction in infarct size (90.29 ± 6.82% of vehicle; n = 18 and 16; Figure 3), indicating a critical role for TLR9 expression on parenchymal cells as well as on hematopoietic cells. To confirm that the result was specific for TLR9 activation, we preconditioned WT → TLR9KO and TLR9KO → WT chimeric mice with a TLR4 ligand LPS (1 mg/kg), and achieved significant protection in both chimeric animal models (infarct size 70.01 ± 12.96% of vehicle and 33.92 ± 11.19% of vehicle, respectively; n = 6 to 7/group; P < 0.05). Thus, lack of TLR9 expression by parenchymal or hematopoietic compartments results in failure to achieve neuroprotection after CpG preconditioning. These data suggest that ischemic neuroprotection induced by CpG preconditioning requires expression of TLR9 on cells within hematopoietic and parenchymal compartments. Moreover, TLR9 expression in either compartment alone is not sufficient for CpG-induced neuroprotection.

Figure 3.

TLR9 is required in hematopoietic and parenchymal compartments to achieve CpG-induced neuroprotection against brain ischemic injury. TLR9KO → WT, WT → TLR9KO, and WT → WT mice were injected subcutaneously with CpG (1.6 mg/kg; subcutaneously) or vehicle 72 hours before middle cerebral artery occlusion (MCAO) (n = 9 to 18/group), and ischemic damage was evaluated by 2,3,5-triphenyltetrazolium chloride (TTC) staining 24 hours after occlusion. CpG preconditioning of TLR9KO chimeric mice did not reduce infarct volume after MCAO. Values are group means ± s.e.m.; **P < 0.01, versus vehicle control by two-way ANOVA and Bonferroni post hoc test. WT, wild type.

TLR9 on Hematopoietic Cells Is Not Required to Elicit a Peripheral Tumor Necrosis Factor Response to CpG

To further understand the absence of a CpG preconditioning effect in mice that do not express TLR9, we measured circulating plasma levels of TNF, a critical mediator of CpG-induced neuroprotection. We have shown previously that systemic CpG preconditioning fails to protect TNF-deficient mice against ischemic brain injury and that CpG induces an increase in plasma TNF within 1 hour after administration.⁶ These results suggest that systemic induction of TNF may be required for the neuroprotective effect of CpG preconditioning. Thus, we tested whether TLR9 expression on hematopoietic cells was required for CpG-mediated induction of systemic TNF levels using mice deficient in hematopoietic TLR9 (TLR9KO → WT). We found that TNF was induced in the plasma in response to CpG (1.6 mg/kg; subcutaneously) in the TLR9KO → WT mice (442.39 ± 63.8 pg/mL; n = 8; Figure 4), suggesting that hematopoietic expression of TLR9 is not required for plasma TNF induction. However, the TNF response was attenuated compared with WT → WT controls (821.44 ± 134.27 pg/mL; n = 5; Figure 4), showing that TLR9-expressing hematopoietic cells contribute to CpG-induced production of TNF in the plasma. The attenuated induction of TNF may stem from the ~10% of recipient hematopoietic cells that still express TLR9 in the TLR9KO → WT mice (Figure 2, upper right panel). This may provide enough cross-talk with the TLR9-expressing parenchyma to induce the limited TNF response observed in these mice. However, the ~10% of recipient cells expressing TLR9 were not sufficient to induce neuroprotection in the TLR9KO → WT mice.

Figure 4.

Expression of TLR9 on hematopoietic cells is not necessary to induce tumor necrosis factor (TNF) in the systemic circulation. TLR9KO → WT, WT → TLR9KO, and WT → WT mice were injected subcutaneously with CpG (1.6 mg/kg), lipopolysaccharide (LPS) (1 mg/kg), or vehicle (n = 4 to 8/group). TNF was measured by ELISA in plasma samples collected 1 hour after injection. Values are group means ± s.e.m.; ***P < 0.001 versus vehicle for respective chimera; ^###P < 0.001 compared with WT → WT for respective treatment by two-way ANOVA and Bonferroni post hoc test. WT, wild type.

In addition, we examined whether TLR9 expression on hematopoietic cells was sufficient to produce TNF in response to CpG using WT → TLR9KO chimeric mice, which only express TLR9 on hematopoietic cells. We found that TNF levels failed to increase in response to CpG administration despite the presence of TLR9 on the hematopoietic cells (Figure 4). Together, these data suggest that parenchymal cells may be major contributors to the CpG-mediated induction of plasma TNF, and that TLR9 expression on hematopoietic and parenchymal cells is necessary for optimal induction of TNF. Of note, all chimeras, irrespective of genotype, induced TNF levels that were equivalent to WT mice in response to the TLR4 ligand, LPS (1 mg/kg; Figure 4), showing that the differences were specific to CpG/TLR9-mediated TNF induction.

CpG Induction of Brain Tumor Necrosis Factor Requires TLR9 Expression in Both the Hematopoietic and Parenchymal Compartments

Systemic LPS administration has been shown to induce TNF in the brains of chimeric mice that have been reconstituted with TLR4-deficient hematopoietic cells¹⁶; thus, we hypothesized that systemic CpG preconditioning would also induce TNF in the brains of mice deficient in TLR9 expression on hematopoietic cells. Therefore, we examined TNF expression in the brains of mice lacking TLR9 on hematopoietic cells (TLRKO → WT) or parenchymal cells (WT → TLR9KO) 3 hours after CpG (1.6 mg/kg; subcutaneously) or vehicle administration. We found that TNF mRNA in the brain was increased in WT → WT control mice in response to systemically administered CpG (13.9 ± 0.4 dCT; n = 4; Table 1) while gene induction was absent in both the TLR9KO → WT and WT → TLR9KO mice (not detected, n = 6/group; Table 1). TLR9KO chimeras showed induction of TNF mRNA in response to the TLR4 ligand, LPS (1 mg/mL; subcutaneously; Table 1), showing that the differences observed were specific to the CpG-mediated TLR9 response. Lack of CpG-induced TNF expression in the brains of TLR9KO → WT and WT → TLR9KO indicates TLR9 expression is required on cells in hematopoietic and parenchymal compartments to induce TNF expression. In addition, the loss of TNF expression in the brain correlated with the loss of protection observed in the TLR9KO → WT and WT → TLR9KO mice, suggesting that TNF expression in the brain parenchyma, but not plasma, is associated with the neuroprotective phenotype.

Table 1

Brain TNF mRNA expression 3 hours after drug administration

Chimera	Unhandled	CpG	LPS
Hematopoietic→Parenchymal
WT	Not detected	13.8±0.6	12.6±0.9
WT→WT	Not detected	14.1±0.2	Not determined
WT→TLR9 KO	Not detected	Not detected	13.2±0.7
TLR9KO→WT	Not detected	Not detected	12.1±0.04

dCT, delta cycle threshold; LPS, lipopolysaccharide; TNF, tumor necrosis factor; WT, wild-type.

Values: calculated dCT; Not detected: Ct >38; n=4 to 6/group.

Intranasal CpG Preconditioning Provides Robust Neuroprotection in the Absence of Increased Plasma Tumor Necrosis Factor

We next determined whether CpG preconditioning by intranasal administration could elicit brain TNF expression and subsequent neuroprotection without the induction of circulating plasma TNF. Ross et al¹⁸ showed that higher intranasal doses of IFNβ-b1 were required to obtain efficient brain distribution compared with an intravenous route of administration. However, even administering IFNβ-b1 at a 10-fold higher concentration resulted in minimal systemic exposure. Therefore, for intranasal administration, we chose to increase the dose of CpG 4-fold compared with what we used for subcutaneous administration. C57Bl/6J mice were treated intranasally with CpG (3.2 mg/kg; n = 6) and TNF mRNA was measured in the brain 3 hours later. We found that TNF mRNA was induced in the brain (dCT = 15.0 ± 0.4) to a similar extent seen with systemic administration of CpG (dCT = 13.8 ± 0.6). To establish whether intranasal CpG preconditioning was protective, mice were administered CpG (3.2 mg/kg; n = 8) or vehicle (n = 3) intranasally 72 hours before MCAO (60 minutes). Subcutaneous preconditioning with CpG (0.8 mg/kg; n = 8) or vehicle (n = 6) was used as a positive control. We found that infarct volume was significantly decreased in mice preconditioned with CpG regardless of route of administration when compared with vehicle-treated controls (intranasal: 53.05 ± 9.17% of vehicle; subcutaneous: 28.11 ± 3.83% of vehicle; P < 0.001; Figure 5). To ascertain whether intranasal administration resulted in TNF release into the plasma, TNF was measured at 1, 3, and 6 hours after CpG administration (3.2 mg/kg; n = 6/time point). We found that TNF was below the level of detection (23.4 pg/mL) at all time points measured. Thus, CpG preconditioning via intranasal administration provides robust protection against stroke injury and induces TNF in the brain without causing changes in plasma TNF levels.

Figure 5.

Intranasal administration of CpG preconditions mice against subsequent ischemic injury. Mice were administered CpG or vehicle via intranasal (IN) (CpG 3.2 mg/kg; n = 8 and vehicle n = 3) or subcutaneous (SC) administration (CpG 0.8 mg/kg; n = 8 and vehicle n = 6) 72 hours before 60 minutes middle cerebral artery occlusion (MCAO). Infarct volume was determined by 2,3,5-triphenyltetrazolium chloride (TTC) staining 24 hours after MCAO. Mice administered vehicle via intranasal or subcutaneous routes have comparable infarcts; groups were combined for analysis. Preconditioning with intranasal administration of CpG reduced infarct damage after MCAO. Values are group means ± s.e.m.; ^***P < 0.001 versus vehicle control, ^#P < 0.05 versus CpG subcutaneous administration by one-way ANOVA and Bonferroni post hoc test.

DISCUSSION

Preconditioning via systemic administration of CpG induces robust neuroprotection against subsequent cerebral ischemia. We show that the protective effect of CpG requires TLR9, as CpG preconditioning in mice deficient in TLR9 fails to protect against stroke injury. We created TLR9KO bone marrow chimeric mice to evaluate the target for CpG-induced neuroprotection against stroke. We found that the neuroprotective effect of systemic CpG preconditioning was absent in mice deficient in either hematopoietic or parenchymal expression of TLR9. This indicates that the cells from both compartments interact to orchestrate the neuroprotective response to CpG that limits ischemic damage. The requirement for both hematopoietic and parenchymal cells to respond to a stimulus was also reported in adjuvant studies using the immune modulator poly-IC. Reciprocal bone marrow chimeric mice created with mice deficient in the poly-IC receptors, TLR3 and Melanoma differentiation associated protein 5 did not mount a primary T-cell adjuvant response to poly-IC.¹⁹ In this case, receptor expression was needed in both compartments to produce a response to poly-IC that was equivalent to WT → WT controls. These results indicate that hematopoietic cells alone are not sufficient for immune activation and that other cell types must interact to elicit a WT response.

We have shown previously that neuroprotection induced by systemic CpG preconditioning requires TNF and that TNF is induced in the plasma after CpG preconditioning.⁶ Here, for the first time, we show that CpG preconditioning also induces TNF mRNA in the brain after CpG preconditioning. Using the TLR9 bone marrow chimeric strategy, we investigated the source and importance of CpG-mediated induction of TNF in the plasma and in the brain. We found that CpG treatment in mice with TLR9 only on parenchymal cells (TLR9KO → WT), but not on hematopoietic cells, produced a significant yet attenuated peripheral TNF response compared with WT mice. However, the systemic TNF response was completely absent when TLR9 was expressed only on hematopoietic cells (WT → TLR9KO) and not on parenchymal cells. This finding was not anticipated given that hematopoietic cells are known to express TLR9.^12,13 In support of our finding, another study showed that primary macrophages expressing TLR9 did not secrete TNF or IL-6 in response to CpG uptake in culture.²⁰ Several recent studies support the concept that nonhematopoietic cells are critical to inflammatory responses. For example, a study that used TLR4 chimeric mice in a model of LPS-induced uveitis found that, although macrophages and dendritic cells are prevalent in the eye, TLR4-mediated responses on parenchymal cells (e.g., endothelial and epithelial cells) were responsible for the induction of the inflammatory disease state.²¹ In addition, Longhi et al¹⁹ administered the immune activator poly-IC to Melanoma differentiation associated protein 5-deficient chimeric mice and showed that cells of nonhematopoietic origin, but not hematopoietic cells, were required for the serum production of the type I IFN-α and β even though dendritic cells and monocytes secrete these cytokines in vitro in response to poly-IC. This suggests that there may be a dominant role for nonhematopoietic cells as instigators of inflammation caused by innate immune activators. Endothelial cells act as the major interface between circulating hematopoietic cells and the rest of the body and thus are a plausible source of inflammatory cytokines produced in response to systemic CpG. CpG is known to activate endothelial cells leading to increased expression of adhesion molecules (e-selectin, p-selectin, ICAM, VCAM) and chemokines (IL-8 and MCP-1).⁷ Additionally, activated endothelial cells are known to secrete and transport TNF.²² LPS-stimulated brain endothelial cells in vitro have been shown to secrete TNF in a polarized manner preferentially releasing TNF on the luminal side and into the peripheral circulation.¹⁵ However, our studies show that restriction of TLR9 expression to the parenchyma attenuated the systemic TNF response to CpG stimulation, and thus, the interaction between hematopoietic and endothelial cells in vivo may enhance TNF secretion.

In contrast, induction of TNF mRNA in the brain was absent in both the WT → TLR9KO and TLR9KO → WT mice, suggesting that both hematopoietic and parenchymal cells are required for the cerebral induction of TNF via systemic CpG preconditioning. Of note, WT → WT controls preconditioned with CpG-induced cerebral and systemic TNF levels equivalent to C57BL/6 mice and were protected against cerebral ischemia, validating that this irradiation model does not compromise the innate immune response in the brain, blood, nor the response to ischemia-reperfusion injury.^23,24 Our findings show that TLR9 expression in both the hematopoietic and parenchymal compartments is required for CpG preconditioning-induced expression of TNF in the brain and induction of neuroprotection. Thus, CpG-mediated neuroprotection correlates with TNF induction in the brain before stroke. We failed to see a similar correlation between systemic levels of TNF and neuroprotection. Although TNF is a proinflammatory mediator several studies have shown that TNF can protect neurons from excitotoxic and ischemic injury.^25–30 Of particular importance to our findings is that preconditioning with TNF via intercisternal administration protects against stroke; however, systemic preconditioning with TNF does not provide protection.²⁸ In addition, Lambertsen et al²⁶ recently used TNF chimeric mice to show that microglia, not infiltrating macrophages, are the source of cerebral TNF in response to stroke and that TNF signals through TNFR1 (TNF-p55R) to promote protection. Another robust preconditioning paradigm, brief ischemia, has been shown to upregulate TNFR1 before stroke and the loss of TNFR1 inhibits the neuroprotective effects against the stroke.^29,30 TNFR1 is known to promote cell survival by signaling TRAF-2 to associate with the cytoplasmic inhibitor of apoptosis protein and activate NFκB and cFos/cJun.²⁹ Thus, the induction of TNF before stroke may activate prosurvival signaling pathways by engaging TNFR1.

Taken together, our findings suggest that systemic administration of CpG requires the interaction between TLR9-expressing hematopoietic and parenchymal cells for neuroprotection and induction of brain TNF mRNA. Communication between immune signals and the brain is referred to as the neuroimmune axis and can occur via multiple pathways including afferent vagal nerve fibers, circumventricular organs, the BBB, and the hypothalamicpituitary–adrenal axis.²² Since it has been shown previously that endothelial cells become activated by CpG and interact with hematopoietic cells,⁷ we propose that the BBB serves as the critical site of communication for systemic CpG preconditioning and the subsequent induction of TNF in the brain based on the following models (Figure 6). Systemic CpG activates endothelial cells and circulating hematopoietic cells, such as leukocytes, causing them to adhere to the endothelium. These cell–cell interactions propagate endothelial and hematopoietic cell activation and cause the release of inflammatory cytokines, which produces the rise in systemic TNF levels. The endothelial activation may increase BBB permeability, allowing for CpG to cross the BBB and directly activate TLR9-expressing CNS cells, such as microglia, astrocytes, and neurons. In addition, evidence suggests that oligodeoxynucleotides of similar size can cross the BBB via saturable transporters and enter the CNS.³¹ We have shown previously that systemic CpG preconditioning causes upregulation of genes related to TLR signaling in the brain.³² Further, direct stimulation of astrocytes and microglia with CpG has been shown to produce TNF.^11,33,34 The activated endothelium may also signal CNS cells indirectly, possibly by abluminal secretion or transport of inflammatory cytokines into the perivascular space or through interaction with adjacent cells such as pericytes and astrocytes, to promote an inflammatory response in the brain.^15,22,35,36 Once activated, neurons, microglia, and astrocytes can produce TNF.^37–39 These processes may be responsible for promoting TNF production in the CNS after preconditioning, which correlates with neuroprotection against stroke.

Figure 6.

Proposed cellular mechanism of cerebral tumor necrosis factor (TNF) induction by systemic CpG preconditioning. Systemic CpG preconditioning requires activation of both hematopoietic and parenchymal cells to induce cerebral TNF and neuroprotection against stroke. Systemic CpG preconditioning may elicit brain TNF via indirect or direct activation of central nervous system (CNS) cells. Indirect activation (left side): systemic CpG preconditioning activates leukocytes (green) and endothelial cells (red) via TLR9. Endothelial cells and leukocytes interact via adhesion molecules (CAMs, selectins, integrins), which signal the brain to induce TNF, either through adjacent cells such as pericytes or astrocytes or through proinflammatory cytokines. Direct activation (right side): CpG stimulation of leukocytes and endothelial cells increases the permeability of the blood–brain barrier (BBB) allowing CpG to cross and directly activate TLR9 in CNS cells to induce TNF. Both mechanisms likely occur simultaneously in vivo.

Our findings suggest that CpG-stimulated hematopoietic and parenchymal cells cooperate to elicit a TNF response in the brain before stroke to promote neuroprotection. Importantly, we found that the hematopoietic compartment can be bypassed by intranasal administration of CpG preconditioning, which induced TNF mRNA in the brain and resulted in robust reduction in infarct size without any detectable TNF in the plasma. This finding supports the importance of TNF expression in the brain before stroke and that neuroprotection does not require a systemic TNF response. Further, the absence of the systemic inflammatory response is therapeutically relevant because, in human patients, systemic TNF can lead to severe side effects such as hypotension, hepatotoxicity, and leukopenia, and TNF is known to have a deleterious role in septic shock.⁴⁰ The intranasal route of administration is a noninvasive method to target the CNS and has shown success in clinical trials for the treatment of cognitive diseases such as Alzheimer's and schizophrenia.⁴¹ Thus, from a therapeutic perspective, intranasal administration of CpG offers a dosing strategy with superior clinical relevance for the prophylactic treatment of patients at high risk of suffering from cerebral ischemia, such as patients undergoing cardiovascular surgery.

These studies are the first to address the role of hematopoietic versus parenchymal cells in neuroprotection induced by systemic CpG preconditioning and to delineate the role of TNF expression in the periphery versus the brain. We show that TLR9-dependent hematopoietic and parenchymal cell responses to CpG are required to induce TNF expression in the brain and to provide neuroprotection. Importantly, neuroprotection induced by CpG preconditioning is independent of peripheral TNF expression. Further, we show that intranasal CpG preconditioning provides robust protection against stroke and results in cerebral TNF expression without induction of peripheral TNF levels, thus providing a superior route of administration for clinical applications of CpG as a prophylactic stroke treatment. These results advance the understanding of neuroimmune signaling involved in preconditioning-induced neuroprotection and provide insight for the development of potential stroke therapeutics by establishing that circulating TNF levels are not requisite.

DISCLOSURE/CONFLICT OF INTEREST

The authors declare no conflict of interest.

Footnotes

ACKNOWLEDGEMENTS

The authors would like to thank Valerie Conrad, Tao Yang, and Delfina Homen for their superior technical support.

References

Barber

Hach

Tippett

Ross

Merry

Milsom

. Cerebral ischemic lesions on diffusion-weighted imaging are associated with neurocognitive decline after cardiac surgery. Stroke 2008; 39: 1427–1433.

Bendszus

Reents

Franke

Mullges

Babin-Ebell

Koltzenburg

Brain damage after coronary artery bypass grafting. Arch Neurol 2002; 59: 1090–1095.

Bendszus

Stoll

. Silent cerebral ischaemia: hidden fingerprints of invasive medical procedures. Lancet Neurol 2006; 5: 364–372.

Floyd

Shah

Price

Harris

Ratcliffe

Acker

Clinically silent cerebral ischemic events after cardiac surgery: their incidence, regional vascular occurrence, and procedural dependence. Ann Thorac Surg 2006; 81: 2160–2166.

Bahjat

Williams-Karnesky

Kohama

West

Doyle

Spector

Proof of concept: pharmacological preconditioning with a Toll-like receptor agonist protects against cerebrovascular injury in a primate model of stroke. J Cereb Blood Flow Metab 2011; 31: 1229–1242.

Stevens

Ciesielski

Marsh

Yang

Homen

Boule

Toll-like receptor 9: a new target of ischemic preconditioning in the brain. J Cereb Blood Flow Metab 2008; 28: 1040–1047.

El Kebir

Jozsef

Pan

Wang

Filep

. Bacterial DNA activates endothelial cells and promotes neutrophil adherence through TLR9 signaling. J Immunol 2009; 182: 4386–4394.

Constantin

Cordenier

Robinson

Ala'Aldeen

Murphy

. Neisseria meningitidis-induced death of cerebrovascular endothelium: mechanisms triggering transcriptional activation of inducible nitric oxide synthase. J Neurochem 2004; 89: 1166–1174.

Jack

Arbour

Manusow

Montgrain

Blain

McCrea

TLR signaling tailors innate immune responses in human microglia and astrocytes. J Immunol 2005; 175: 4320–4320.

10.

Tang

Arumugam

Cheng

Mughal

Pivotal role for neuronal Toll-like receptors in ischemic brain injury and functional deficits. Proc Natl Acad Sci USA 2007; 104: 13798–13803.

11.

Buzas

Fan

Cohen

Wang

Mont

Painful pathways induced by TLR stimulation of dorsal root ganglion neurons. J Immunol 2011; 186: 6417–6426.

12.

Nierkens

Den Brok

Garcia

Togher

Wagenaars

Wassink

Immune adjuvant efficacy of CpG oligonucleotide in cancer treatment is founded specifically upon TLR9 function in plasmacytoid dendritic cells. Cancer Res 2011; 71 : 6428–6437.

13.

Talati

Kim

English

. Role of bacterial DNA in macrophage activation by group B streptococci. Microbes Infect 2008; 10: 1106–1113.

14.

Welner

Pelayo

Nagai

Garrett

Wuest

Carr

Lymphoid, precursors are directed to produce dendritic cells as a result of TLR9 ligation during herpes infection. Blood 2008; 112: 3753–3761.

15.

Verma

Nakaoke

Dohgu

Banks

. Release of cytokines by brain endothelial cells: A polarized response to lipopolysaccharide. Brain Behav Immun 2006; 20: 449–455.

16.

Chakravarty

Herkenham

. Toll-like receptor 4 on nonhematopoietic cells sustains CNS inflammation during endotoxemia, independent of systemic cytokines. J Neurosci 2005; 25: 1788–1796.

17.

Doyle

Yang

Lessov

Ciesielski

Stevens

Simon

Nasal administration of osteopontin peptide mimetics confers neuroprotection in stroke. J Cereb Blood Flow Metab 2008; 28: 1235–1248.

18.

Ross

Martinez

Renner

Thorne

Hanson

Frey

2nd . Intranasal administration of interferon beta bypasses the blood-brain barrier to target the central nervous system and cervical lymph nodes: a non-invasive treatment strategy for multiple sclerosis. J Neuroimmunol 2004; 151 : 66–77.

19.

Longhi

Trumpfheller

Idoyaga

Caskey

Matos

Kluger

Dendritic cells require a systemic type I interferon response to mature and induce CD4 + Th1 immunity with poly IC as adjuvant. J Exp Med 2009; 206: 1589–1602.

20.

Yasuda

Kawano

Yamane

Ogawa

Yoshinaga

Nishikawa

Restricted cytokine production from mouse peritoneal macrophages in culture in spite of extensive uptake of plasmid DNA. Immunology 2004; 111: 282–290.

21.

Kezic

Taylor

Gupta

Planck

Rosenzweig

Rosenbaum

. Endotoxin-induced uveitis is primarily dependent on radiation-resistant cells and on MyD88 but not TRIF. J Leuk Biol 2011; 90: 305–311.

22.

Quan

Banks

. Brain-immune communication pathways. Brain Behav Immun 2007; 21: 727–735.

23.

Terao

Yilmaz

Stokes

Russell

Ishikawa

Kawase

Blood cell-derived RANTES mediates cerebral microvascular dysfunction, inflammation, and tissue injury after focal ischemia-reperfusion. Stroke 2008; 39: 2560–2570.

24.

Turrin

Plante

Lessard

Rivest

. Irradiation does not compromise or exacerbate the innate immune response in the brains of mice that were transplanted with bone marrow stem cells. Stem Cell 2007; 25: 3165–3172.

25.

Bruce

Boling

Kindy

Peschon

Kraemer

Carpenter

Altered neuronal and microglial responses to excitotoxic and ischemic brain injury in mice lacking TNF receptors. Nat Med 1996; 2: 788–794.

26.

Lambertsen

Clausen

Babcock

Gregersen

Fenger

Nielsen

Microglia protect neurons against ischemia by synthesis of tumor necrosis factor. J Neurosci 2009; 29: 1319–1330.

27.

Liu

Ginis

Spatz

Hallenbeck

. Hypoxic preconditioning protects cultured neurons against hypoxic stress via TNFa and ceramide. Am J Physiol Regul Integr Comp Physiol 2000; 278: C144–C153.

28.

Nawashiro

Tasaki

Ruetzler

Hallenbeck

. TNF-alpha pretreatment induces protective effects against focal cerebral ischemia in mice. J Cereb Blood Flow Metab 1997; 17: 483–490.

29.

Pang

Kastin

. Tumor necrosis factor and stroke: role of the blood-brain barrier. Prog Neurobiol 2007; 83: 363–374.

30.

Pradillo

Romera

Hurado

Cardenas

Moro

Leza

TNFR1 upregulation mediates tolerance after brain ischemic preconditioning. J Cereb Blood Flow Metab 2005; 25: 193–203.

31.

Banks

Jaeger

Urayama

Kumar

Hileman

Gaskin

Preproenkephalin targeted antisenses cross the blood-brain barrier to reduce brain methionine enkephalin levels and increase voluntary ethanol drinking. Peptides 2006; 27: 784–796.

32.

Stevens

Leung

Vartanian

Gopalan

Yang

Simon

Multiple preconditioning paradigms converge on interferon regulatory factor-dependent signaling to promote tolerance to ischemic brain injury. J Neurosci 2011; 31: 8456–8463.

33.

Lee

Hong

Choi

Park

Kim

CpG oligodeoxynucleotides induce expression of proinflammatory cytokines and chemokines in astrocytes: the role of c-Jun N-terminal kinase in CpG ODN-mediated NF-kappaB activation. J Neuroimmunol 2004; 153: 50–63.

34.

Olson

Miller

. Microglia initiate central nervous system innate and adaptive immune responses through multiple TLRs. J Immunol 2004; 173: 3916–3924.

35.

Descamps

Coisne

Dehouck

Cecchelli

Torpier

. Protective effect of glial cells against lipopolysacharide-mediated blood-brain barrier injury. Glia 2003; 42: 46–58.

36.

Qin

Block

Liu

Breese

Hong

Systemic LPS causes chronic neuroinflammation and progressive neurodegeneration. Glia 2007; 55: 453–462.

37.

Biermann

Lagreze

Dimitriu

Stoykow

Goebel

. Preconditioning with inhalative carbon monoxide protects rat retinal ganglion cells from ischemia/reperfusion injury. Invest Ophthalmol Vis Sci 2010; 51 : 3784–3791.

38.

Botchkina

Meistrell

MER

Botchkina

Tracey

. Expression of TNF and TNF receptors (p55 and p75) in the rat brain after focal cerebral ischemia. Mol Med 1997; 3: 765–781.

39.

Butchi

Pourciau

Morgan

Peterson

. Analysis of the neuroinflammatory response to TLR7 stimulation in the brain: comparison of multiple TLR7 and/or TLR8 agonists. J Immunol 2008; 180: 7604–7612.

40.

Roberts

Zhou

Diaz

Jr Holdhoff

. Systemic use of tumor necrosis factor alpha as an anticancer agent. Oncotarget 2011; 2: 739–751.

41.

Dhuria

Hanson

Frey

II . Intranasal delivery to the central nervous system: mechanisms and experimental considerations. J Pharm Sci 2010; 99: 1654–1673.

TLR9 Bone Marrow Chimeric Mice Define a Role for Cerebral TNF in Neuroprotection Induced by CpG Preconditioning

Abstract

Keywords

INTRODUCTION

MATERIALS AND METHODS

Animal Care and Criteria

Generating Chimeric Mice

Determination of Reconstitution Efficiency

Drug Treatments

Ischemia-Reperfusion Model

Evaluation of Infarct Size

Analysis of Plasma Tumor Necrosis Factor Levels

Tissue Processing and Quantitative Real-Time PCR

Statistical Analysis

RESULTS

CpG Preconditioning Requires TLR9

TLR9 Is Required in Both the Hematopoietic and Parenchymal Compartments for CpG-Induced Neuroprotection Against Brain Ischemia

TLR9 on Hematopoietic Cells Is Not Required to Elicit a Peripheral Tumor Necrosis Factor Response to CpG

CpG Induction of Brain Tumor Necrosis Factor Requires TLR9 Expression in Both the Hematopoietic and Parenchymal Compartments

Intranasal CpG Preconditioning Provides Robust Neuroprotection in the Absence of Increased Plasma Tumor Necrosis Factor

DISCUSSION

DISCLOSURE/CONFLICT OF INTEREST

Footnotes

ACKNOWLEDGEMENTS

References