Synthetic CpG oligonucleotides induce a genetic profile ameliorating murine myocardial I/R injury

Abstract We previously demonstrated that pre‐conditioning with CpG oligonucleotide (ODN) 1668 induces quick up‐regulation of gene expression 3 hours post‐murine myocardial ischaemia/reperfusion (I/R) injury, terminating inflammatory processes that sustain I/R injury. Now, performing comprehensive microarray and biocomputational analyses, we sought to further enlighten the “black box” beyond these first 3 hours. C57BL/6 mice were pretreated with either CpG 1668 or with control ODN 1612, respectively. Sixteen hours later, myocardial ischaemia was induced for 1 hour in a closed‐chest model, followed by reperfusion for 24 hours. RNA was extracted from hearts, and labelled cDNA was hybridized to gene microarrays. Data analysis was performed with BRB ArrayTools and Ingenuity Pathway Analysis. Functional groups mediating restoration of cellular integrity were among the top up‐regulated categories. Genes known to influence cardiomyocyte survival were strongly induced 24 hours post‐I/R. In contrast, proinflammatory pathways were down‐regulated. Interleukin‐10, an upstream regulator, suppressed specifically selected proinflammatory target genes at 24 hours compared to 3 hours post‐I/R. The IL1 complex is supposed to be one regulator of a network increasing cardiovascular angiogenesis. The up‐regulation of numerous protective pathways and the suppression of proinflammatory activity are supposed to be the genetic correlate of the cardioprotective effects of CpG 1668 pre‐conditioning.

proposed over the years. For a detailed overview, we refer to some excellent recent review publications. [2][3][4][5] One concept for reducing I/R damage is the application of physical (repetitive hypoxic stimulation, hypo-/hyperthermia), pharmacological (eg anaesthetics or opioids) or immunological effectors, such as ligands for Toll-like receptors (TLRs).
Both pre-and post-conditioning with TLR ligands can trigger a myocardial inflammatory response, which may potentially diminish the damage induced by I/R. 5,6 Successful implementation of this approach in a murine I/R model has been demonstrated by our group using ligands for TLR 2, 4 and 9. [7][8][9] In particular, synthetic TLR9 ligands, such as CpG oligodeoxynucleotides (CpG ODNs), prompt a strong innate immune response with only moderate side effects in the human organism. 10 Hence, they are suitable for use as immunostimulatory agents as well as vaccine adjuvants in clinical trials. Their potential to stimulate the innate immune system depends on the cytosine and guanine (CG) triphosphate deoxynucleotide motif as well as on the degree of methylation. 11 In a previous study, we could demonstrate that the intraperitoneal (ip) injection of C57BL/6 wild-type mice with the CpG ODN 1668 thioate significantly boosted levels of inflammatory serum mediators and concomitantly raised the number of immune cells in blood and the spleen. 12 In addition, this intervention up-regulated gene expression of myocardial pattern recognition receptors. After myocardial ischaemia and reperfusion induced in a closed-chest model 16 hours after pre-conditioning, infarct size as well as rates of failure of cardiac function was reduced by up to 75%, which did not occur in the control groups. Protein and RT-PCR analyses (performed 3 hours after start of reperfusion) strongly suggested that an excessive CpG-dependent up-regulation of interleukin-10 (IL-10) was one key factor in reducing infarct size and improving cardiac function. This was further supported by the fact that blocking IL-10 prevented the cardioprotection afforded by CpG pre-conditioning. Interestingly, genome profiling revealed alterations in the level of activity of IL-10-dependent genes as well, considered to elicit suppressive networks that terminate the inflammatory processes that sustain I/R injury. Furthermore, CpG pre-conditioning together with I/R induced massive escalation of gene expression in general, highly exceeding what was observed after CpG priming or I/R alone. Bioinformatic analysis revealed that genes mediating the inflammatory response or immune cell trafficking were among the most prominently up-regulated functional groups.
Despite these findings, such striking changes in gene regulation have been described incompletely, as pathway and network analyses have only been performed over a short period of time, that is up to 3 hours, following I/R. It has been demonstrated that, on the one hand, both harmful and protective mechanisms are distinctly altered at 24 hours after I/R, suggesting that this vulnerable time-point may be a critical period for any interventional therapeutic strategy. [13][14][15][16] On the other hand, the protective time span of immunologic preconditioning strategies to minimize I/R injury has been shown to be most pronounced in the 24-72 hours following administration of TLR ligands. 17,18 This prompted us to extend the observational period and perform comprehensive microarray and biocomputational analyses at a later time-point to illuminate the protective mechanisms of CpG pre-conditioning during this temporal "black box." We wondered whether long-term effects, such as the observed restoration of cardiac function mediated by CpG, might have a preceding correlate at the level of gene expression. Our results may help further elucidate the protective molecular mechanisms of CpG ODN-mediated amelioration of myocardial I/R injury.

| MATERIALS AND METHODS
All mice were handled in accordance with the Guide for Use and Care of Laboratory Animals (NIH publication No. 85-23, revised 1996). Protocols were approved by the local animal authority (protocol no. AZ 8.87-51.04.20.09.392; LANUV, Recklinghausen, Germany).

| Surgical techniques, experimental groups and protocols
Ten-to 12-week-old male C57BL/6 mice (Charles River Laboratories, Sulzfeld, Germany) were included. To prevent an inflammatory reaction due to surgical trauma, a chronic closed-chest model of myocardial I/R with left parasternal incision was applied. Surgical procedures have been described and depicted in detail in a previous publication. 19 In brief, in mice anesthetized with isoflurane (2.5 Vol%; Forene, Abbott GmbH, Wiesbaden, Germany) and mechanically ventilated (respiratory rate 105*min À1 , 200 lL tidal volume; Minivent, Hugo Sachs Elektronik-Harvard Apparatus GmbH, March-Hugstetten, Germany), a left parasternal incision was performed and an 8-0-prolene suture was passed underneath the left anterior descending artery (LAD). Both ends of the suture were threaded through a polyethylene tube and were tightened to confirm correct position by emerging paleness of the distal myocardium. Afterwards, both ends of the suture were released again, formed into a knot, placed subcutaneously, and the chest was closed with 6-0 sutures. Analgesia was obtained by application of buprenorphine (0.1 mg/kg s.c.), and mice were allowed to recover from anaesthesia. After 5-7 days, all mice were treated with an ip dose of D-Galactosamine N (DGalN, 0.2 mg/g body weight (BW), Roth, Karlsruhe, Germany) to slow down the hepatic degradation of CpG ODNs (Figure 1). 20 In control experiments, DGalN alone did not induce an inflammatory response (data not shown). Animals were randomized to ip received either a stimulatory CpG ODN (1668 thioate; 5 0 -TCCAT-GACGTTCCTGATGCT; TIB Molbiol, Berlin, Germany; 0.25 nmol/g BW), a non-CG containing control ODN (1612 thioate; 5 0 -GCTA-GATGTTAGCGT; TIB Molbiol; 0.25 nmol/g BW) or PBS as control.
In previous experiments, the dosage of CpG ODN 1668 had been tested. Pre-conditioning was performed 16 hours before I/R, and this was chosen because our previous experiments showed that preconditioning with the TLR4 ligand lipopolysaccharide results in a reduction in infarct size after this time span, and that priming with CpG ODN applied under the same conditions (interval and concentration) as used here attenuated cardiac hypertrophy induced by transverse aortic constriction. 8,21 Myocardial infarction was then induced for 1 hour (for time course see Figure 1). Mice were sedated with propofol (1% in 0.2 mL; 2 mg ip). Electrocardiogram (ECG; PowerLab, ADInstruments GmbH, Spechbach, Germany) was recorded. The original skin incision was reopened, and the subcutaneously placed thread was manipulated without opening the chest.
Thereafter, tension was carefully applied on the loop to achieve an ischaemia, operationalized here as a significant ST-elevation. One hour after ischaemia, the loop was released and reperfusion induced the recession of the ST-segment elevation. Finally, the skin was closed, followed by 24 hours of reperfusion.

Data from 4 independent experiments and 3 untreated controls
were used for all statistical analyses. Expression analyses were performed using BRB ArrayTools. Data were background-corrected, flagged values were removed, spots in which both signals were <100 were filtered out, ratios were log base 2-transformed, and Lowess intensity-dependent normalization was applied to adjust for differences in labelling intensities of the Cy3 and Cy5 dyes. 22 Analysis was restricted to genes present on >50% of the arrays after filtering.
The gene expression profile of all treatment groups was compared with that of the untreated control groups. A P value cut-off of 0.0001 was used to identify genes whose expression was significantly up-regulated after CpG ODN stimulation when compared with controls. Data were evaluated using Ingenuity Pathway Analysis

| Statistical analysis
Genes that were expressed differentially in the treatment groups were identified using a random-variance t test. The random-variance t test is an improvement over the standard, separate t test, as it permits sharing information among genes about within-class variation without assuming that all genes have the same variance. 24 The behaviour of specific gene subsets and functional groups was analysed with the IPA Comparison Tool. Ingenuity Pathway Analysis does provide a denominator for each functional group. Based on the entity of genes known for a specific function and the entity of genes significantly regulated in a data set, the probability that the regulation of a specific functional group is different from a random distribution is calculated. Differences between regulators in terms of number of genes regulated were established by chi-square analysis and Fisher's exact test. Differences in fold changes between groups of genes were established by t test analysis (2-sided).
To identify biological functions expected to increase or decrease according to the observed gene expression changes in the data set, we used the IPA Downstream Effects analytic. For each biological function, a statistical quantity was computed, called the activation zscore. The activation z-score is used to infer likely activation states of biological functions based on comparison with a model that assigns random regulation directions (for more information see http://pages.ingenuity.com/rs/ingenuity/images/0812%20downstrea m_effects_analysis_whitepaper.pdf).

| Temporal pattern and functional groups of gene regulation 24 hours following myocardial I/R in CpG ODN 1668-treated mice
We previously demonstrated that I/R following pre-treatment with CpG ODN 1668 induced early pronounced myocardial gene activation (30 minutes to 3 hours following I/R) with selected induction of specific functional groups, compared to pre-conditioning with control ODN. 12 In this present work, we focused on genetic profile changes at a later time-point, 24 hours after I/R. Both the PBS and the control ODN groups showed only a moderate increase in differential gene expression (165 and 126 genes, respectively), whereas in the sham group (surgical procedure without subsequent pre-conditioning or I/R), 89 genes were differentially regulated, allowing us to determine the effects on gene expression elicited by induction of I/R alone. Differential gene expression in the CpG ODN group, showed a reduced number of regulated genes compared to the 3 hours timepoint, yet they still remained at a high activation level (434 genes, P < .0001 compared to control ODN group) and reflected the effect of CpG ODN 1668. Gene groups mediating "cellular movement" or "cellular function and maintenance" were among the top scoring functional categories, depicted in detail in Table 1. These groups primarily focus on the restoration of cellular integrity.
In order to validate our murine cardiac I/R model, we evaluated toxicity functions as a quality control. Toxicity functions in IPA relate to organ dysfunction and mark symptomatic pathological endpoints. The gene expression pattern identified "cardiac infarction" (P < 2.26E-11), "cardiac dysfunction" (P < 1.18E-6) and "cardiac inflammation" (P < 1.96E-5) to be highly significantly induced (supplementary data, Table S1).  pathway activation, which began to decline after 24 hours (Figure 2).

|
In contrast, CpG pre-treatment prompted strong activation of several defined pathways. This stimulation even increased activation along these pathways (ie "cAMP-mediated signalling") or in other cases at least sustained activation at high levels over the prolonged observation period of 24 hours after I/R (eg "dendritic cell maturation", "IL-8 signalling" and "acute-phase response signalling"). Some pathways ("Gai signalling") instead showed a delayed activation pattern following pre-conditioning with CpG, not yet detectable 3 hours after I/R.
Lastly, CpG-mediated activation status of some of the pathways began to decline after 24 hours (eg "role of pattern recognition", "TREM1 signalling" and "production of nitric oxide"). Of note, none of the pathways were altered in the sham group.
In contrast to these pathway activations, several pathways, described as mediating important proinflammatory responses, were instead suppressed at 24 hours, despite having initially been upregulated at 3 hours following I/R in the CpG ODN 1668 group.

| Genes predicted to increase survival of
cardiomyocytes are strongly up-regulated by CpG ODN 1668 pre-conditioning at 24 hours post-I/R The extent of irreversible ischaemic damage in the early phase after acute MI is clearly and independently associated with adverse ventricular remodelling and worse patient outcome. 25 According to IPA analysis, several genes known to influence cell survival were strongly induced at 24 hours post-I/R in the CpG pre-treatment group, compared to control. The P value in IPA for this functional group was 8E-09, implying that the difference in regulation between these groups group is far away from a random distribution. Of these significantly up-regulated genes, 75% have been described as increasing cell survival, for example, BCLA1, MT2, CTGF, Saa3 and Usp17la.
Another subgroup of these highly up-regulated genes includes those known to decrease the rates of death of cardiomyocytes, for example, TIMP1, HSPB1, NAD+ and APOD (Table 3).

| Genes mediating cardiovascular angiogenesis
are strongly up-regulated by CpG ODN 1668 preconditioning at 24 hours post-I/R Angiogenesis is a key factor in long-term restoration of a sufficient blood supply to the myocardium. Twenty-four hours post-I/R, genes known to regulate cardiovascular angiogenesis were markedly upregulated in the CpG ODN 1668 pre-conditioning group compared to control. Of these up-regulated genes, 82.4% have been described to increase cardiovascular angiogenesis. A complete overview is given in Figure 3A (see also Table S2).
Specific upstream regulators induce these aforementioned genes.
On this study, we identified the IL1 complex as one regulator of the network acting to increase cardiovascular angiogenesis at 24 hours in the CpG pre-conditioning group ( Figure 3B), including all of the most significantly activated genes from Figure 3A.   While I/R following pre-conditioning with the control ODN displayed a very weak influence on promoting potentially protective pathways, CpG pre-treatment resulted in early and sustained activation of safeguarding signal transduction mechanisms. Signalling activity involving cAMP and protein kinase A (PKA) was increased in CpG-pretreated mice and exerted protective effects in myocardial I/R. 37 Dendritic cells, whose maturation was boosted by TLR9 pre-conditioning, infiltrated the infarcted heart and bolstered healing and restorative processes by controlling recruitment of monocytes and macrophages. 38 CpG pre-treatment exclusively induced the up-regulation of a number of genes 24 hours after I/R that are implicated in lower rates of myocardial cell death. Specifically, we identified TIMP1, APOD and NAD+, among others, to have been highly propagated. [39][40][41] It is known that TIMP1 inhibits metalloproteinases and it was shown to directly mediate protection from I/R injury in a rat model of myocardial infarction. 39 Other genes, such as BCLA1, MT2, Saa3 and Usp17la, have likewise been demonstrated to exercise positive influences on cell survival. [42][43][44][45][46] The activation of angiogenesis represents another requisite aspect of restoration initiated after myocardial ischaemic injury.
Increasing angiogenesis in infracted tissue was demonstrated to occur after pre-conditioning using CpG ODNs, restoring myocardial function after I/R in a very recent study by Zhou et al. 47 In line with this, our study identified a set of genes with specific implications for the fostering of angiogenesis to be significantly up-regulated in TLR9-pretreated mice. Among the leading up-regulated genes we discovered were S100A9, S100A8, ARG1, CFB and CCL2, all of which have been demonstrated to exert specific positive influences on cardiovascular angiogenesis. [48][49][50][51] Moreover, using IPA, we identified the IL1 gene complex as one regulator of increased cardiovascular angiogenic activity 24 hours following I/R in the CpG-pretreated animals, which influenced the target genes mentioned above. The IL1 complex contains genes coding for both anti-and proinflammatory cytokines, including IL-1A, IL-1B and IL-1RA. 52 Indeed, gain-of-function experiments could demonstrate that overexpression of IL-1RA mediates protection after myocardial I/ R. 53 As our previous analyses revealed, positive clinical effects of CpG pre-conditioning are accompanied by an heightened up-regulation of IL-10 activity, suggesting that IL-10 is a key regulator of TLR9-driven cardioprotection. 12 Several other publications have confirmed this basic finding, demonstrating it not only after the use of CpG, but also following ischaemic, vagal or pharmacologic (eg using morphine) pre-or post-conditioning. 9  Obviously, our study has limitations. The IL1 complex prompting of increased cardiovascular angiogenic activity 24 hours following I/R in CpG-pretreated animals was not verified using blocking or knockdown/knockout approaches in this study, and, hence, remains speculative. Moreover, although we had previously found a cardioprotective impact of CpG-mediated IL-10 up-regulation by blocking IL-10R1, we did not repeat these experiments in this study. However, taken together, our data reveal an early, yet prolonged induction of differential gene expression by pre-conditioning with CpG ODN. The bolstering of numerous protective pathways and functions and the, presumably mainly IL-10-mediated, suppression of proinflammatory activity may be the genetic correlate of the clinically observed cardioprotective effects of CpG preconditioning.