Resveratrol preconditioning induces cerebral ischemic tolerance but has minimal effect on cerebral microRNA profiles

Abstract

The health benefits of the plant-derived polyphenol resveratrol were established in multiple disease systems. Notably, pre-treatment with resveratrol was shown to be neuroprotective in several models of cerebral ischemia. Mechanisms of resveratrol-mediated neuroprotection have been explored in the context of canonical resveratrol targets, but epigenetic and non-coding RNA processes have not yet been evaluated. Resveratrol was shown to alter microRNAs in cancer and cardiac ischemia. Previous studies also showed that ischemic preconditioning that induces ischemic tolerance significantly alters cerebral microRNA levels, particularly those that target neuroprotective pathways. Therefore, we tested if resveratrol-mediated ischemic tolerance also alters microRNA expression with a goal to identify microRNAs that are amenable to manipulation to induce neuroprotection after cerebral ischemia. Hence, we tested the microRNA profiles in mouse brain following intraperitoneal administration of resveratrol that induced significant tolerance against transient focal ischemia. We analyzed microRNA profiles using microarrays from both Affymetrix and LC Sciences that contain probes for all known mouse microRNAs. The results show that there is no consistent change in any of the microRNAs tested between resveratrol and vehicle groups indicating that microRNAs play a minimal role in resveratrol-mediated cerebral ischemic tolerance.

Introduction

The polyphenol resveratrol is known to induce antioxidant enzymes, such as superoxide dismutase and glutathione peroxidase, and thus promotes free radical scavenging.^1,2 Oxidative stress is a known promoter of secondary brain damage, and resveratrol administration was shown to be neuroprotective following stroke,^3,4 subarachnoid hemorrhage (SAH),^5,6 traumatic brain injury (TBI),^7,8 and spinal cord injury (SCI).^9,10 Interestingly, resveratrol treatment was also shown to induce ischemic tolerance similar to other preconditioning (PC) paradigms.^11,12 A recent study of resveratrol PC demonstrated long-term tolerance against cerebral ischemia via silent mating type information regulation 2 homolog 1 (SIRT1)-mediated upregulation of brain-derived neurotrophic factor (BDNF).¹³ Despite these advances in knowledge, the molecular mechanisms that contribute to PC-induced ischemic tolerance are not completely understood. However, PC induced by a brief ischemic event was reported to alter cerebral microRNA (miRNA) profiles.^14,15 Resveratrol administration was also shown to alter miRNA levels in cancer tissue and the ischemic heart.^16–18 Hence, using microarray analysis, we currently evaluated if resveratrol-induced ischemic tolerance changes the miRNA profiles in rodent brain.

Materials and methods

Adult, male C57BL6/J mice (11 to 12 weeks; Jackson Laboratories, Bar Harbor, ME) were used for all experiments. All mice were in good health and weighed 25–28 g. The adult mouse is an appropriate animal for studying stroke, as the methods for inducing focal ischemia and subsequent mechanistic evaluation are well developed and routinely used in mice. Additionally, the rodent brain is similar to the human brain with regard to its RNA expression profiles and the way that it responds to a neurological insult. Animals were housed in a level 1 facility on a 12-h day/night cycle with ad libitum access to chow and water. All surgical procedures and drug administration took place during the day in a designated area in the level 1 facility. Animal protocols were approved by the “University of Wisconsin-Madison Research Animal Resources and Care Committee,” and animals were cared in accordance with the “Guide for the Care and Use of Laboratory Animals, U.S. Department of Health and Human Services Publication Number. 86-23 (revised).” Studies were conducted and reported in accordance to the ARRIVE guidelines.

Focal ischemia

Mice were randomly assigned to treatment groups and injected intraperitoneally with either resveratrol (10 mg/kg; Cayman, Ann Arbor, MI) or vehicle (7:3 normal saline: Kolliphor EL; BASF, Wyandotte, USA). At three days after the drug administration, cohorts of mice were subjected to a 1-h transient middle cerebral artery occlusion (MCAO) under isoflurane anesthesia as described earlier.¹⁹ Mice were euthanized at three days of reperfusion; brain sections were stained 2,3,5-triphenyltetrazolium chloride (TTC), and the relative infarct volume (calculated as a percentage of the hemisphere) was quantified by a blinded investigator using ImageJ software as described earlier.¹⁹ Mice that failed to show an infarct and those that showed a hemorrhage were excluded. No animals were excluded as a result of MCAO-related mortality.

Microarray and real-time PCR analysis

Cohorts of mice injected with either resveratrol or vehicle were euthanized at three days after drug administration, and the total RNA was isolated from the cerebral cortex using miRVana RNA isolation kit (Ambion, USA). The RNA samples were subjected to miRNA microarray analysis as described earlier²⁰ using GeneChip miRNA 4.0 array with probes validated by miRBase release 20 (Affymetrix USA) and miRNA microarray with probes validated by miRBase release 21 (LC sciences USA). We used an n = 3/group for Affymetrix analysis and n = 5/group for LC Sciences analysis. The fold change (FC) was calculated by the respective microarray data analysis software. In case of LC Sciences data, FC was calculated as 2 (avg. group2/avg. group1) where each value was calculated using an internal control (n = 5/group). In case of Affymetrix data, FC was calculated using bi-weighted average signal of each group (n = 3/group). In both cases, data were analyzed by an investigator blinded to the experimental details. Sample sizes were chosen based on previous studies in our lab that showed that n = 3 to 5/group will yield statistically significant data in microarray analysis. Real-time PCR was conducted by SYBR-green method using let-7a-5p and 18s rRNA as housekeeping controls for miRNAs and mRNAs, respectively.²⁰ LNA primers (Exiqon, Vedbaek, Denmark) and oligonucleotide primers (Integrated DNA Technologies, Coralville, IOWA) were used for amplifying miRNAs and mRNAs, respectively. The FC was calculated using the ddCt method. The sequences and genome location of miRNAs were identified with the NCBI BLAST database (http://blast.ncbi.nlm.nih.gov/Blast.cgi), the BLAT database (https://genome.ucsc.edu), and the miRBase database (miRbase.org).

Results

Resveratrol PC-induced ischemic tolerance

At three days of reperfusion, mice treated with resveratrol three days prior to transient MCAO showed a significantly smaller infarct than vehicle-treated group (by ∼46%; p < 0.05; n = 6 to 7/group) (Figure 1). Resveratrol treatment also induced the expression of BDNF, a known resveratrol-inducible gene,^13,21–23 by 3.11- and 2.41-fold over the vehicle control at one and three days, respectively (p < 0.05; n = 3/group) (Figure 1). However, resveratrol treatment had no significant effect on the levels of miR-124-5p and miR-21a-5p (miRNAs previously reported to be altered with resveratrol treatment in other disease systems) in the mouse cerebral cortex (Figure 1, panels d and e).

Figure 1.

Resveratrol preconditioning induces ischemic tolerance. Three days following resveratrol (10 mg/kg; i.p.) or vehicle administration, mice were subjected to a 1-h transient MCAO and were euthanized at three days of reperfusion. Panel A shows serial brain sections stained with TTC from representative mice of the two groups. The percentage of infarcted hemisphere was used to quantify infarct volumes. Panel B shows infarct volume quantitated from the TTC-stained serial brain sections. Panel C shows BDNF mRNA expression in vehicle and resveratrol-treated groups. Panels D (mmu-miR-124-5p) and E (mmu-miR-21a-5p) show expression levels of murine miRNAs in vehicle and resveratrol-treated groups. Bars are mean ± SD of n = 7 for vehicle and 6 for resveratrol groups in B and n = 3/group in C, D, and E. *p < 0.05 by unpaired Student’s t-test.

Resveratrol PC had no consistent effect on cerebral miRNA profiles

We analyzed the effect of resveratrol on miRNA profiles using two different microarray platforms from LC Sciences and Affymetrix. While the LC Sciences arrays are spotted only for probes-specific for mouse miRNAs, the Affymetrix arrays are spotted with probes for miRNAs from 203 different species. The LC Sciences arrays showed that eight mouse-specific miRNAs were altered in the resveratrol-treated group compared to vehicle control (two up- and six down-regulated; > 2 FC; p < 0.05; n = 5/group). Whereas, Affymetrix arrays showed altered levels (>2-fold; p < 0.05; n = 3/group) of 17 mouse-specific miRNAs at one and/or three days after resveratrol treatment (four miRNAs altered exclusively at one day; 11 miRNAs altered exclusively at three days; two miRNAs altered at both one and three days) (Table 1). However, none of these miRNAs were observed to be altered in both analyses (LC Sciences and Affymetrix), demonstrating lack of consistent changes between platforms. Although statistically significant, most of these changes (19 out of 25 miRNAs altered) were low-abundance changes (two to threefold). This mild alteration in FC in addition to the lack of consistency between miRNA profiles suggests that the 25 miRNAs that were seen to be altered need further validation to indicate them as specific or positive changes.

Table 1.

Resveratrol-induced changes in the levels of murine miRNAs.

LC sciences microarrays		Affymetrix microarrays
miRNA	Δ Fold	miRNA	Δ Fold at one day	Δ Fold at three days
mmu-miR-335-5p	2.30	mmu-miR-3962	−2.06	−2.80
mmu-miR-7a-1-3p	2.16	mmu-miR-467g	−2.19	−3.07
mmu-miR-149-5p	−2.17	mmu-miR-327	2.31	NC
mmu-miR-764-5p	−3.18	mmu-miR-6973b-5p	2.28	NC
mmu-miR-381-5p	−3.35	mmu-miR-3082-5p	2.23	NC
mmu-miR-7081-5p	−3.79	mmu-miR-7053-3p	−2.23	NC
mmu-miR-1298-5p	−8.17	mmu-miR-497-5p	NC	2.06
mmu-miR-7239-3p	−13.30	mmu-miR-7005-5p	NC	−2.07
		mmu-miR-148a-3p	NC	−2.11
		mmu-miR-7020-5p	NC	−2.20
		mmu-miR-1247-3p	NC	−2.23
		mmu-miR-1188-5p	NC	−2.32
		mmu-miR-539-5p	NC	−2.38
		mmu-miR-7052-5p	NC	−2.39
		mmu-miR-7052-5p	NC	−2.39
		mmu-miR-34b-5p	NC	−2.86
		mmu-miR-7b-5p	NC	−3.73

All the fold changes shown were mean values (with < 15% SD) and were significantly different from the vehicle group (p < 0.05; Affymetrix: one-way ANOVA; LC Sciences: unpaired Student’s t-test). Samples were analyzed at one and three days after resveratrol or vehicle administration (n = 5/group for LC Sciences arrays and n = 3/group for the Affymetrix arrays). Probes for all these transcripts are present on both arrays, but none showed parallel change in both arrays. NC: no change.

When miRNAs were analyzed using the Affymetrix arrays, 97 non-murine miRNAs were observed to be altered (>2 FC; p < 0.05; n = 3/group) at one and/or three days after resveratrol treatment (47 miRNAs altered exclusively at one day; 44 miRNAs altered exclusively at three days; six miRNAs altered at both one and three days) (Supplementary Tables S1 to S3). Of those, 13 showed partial sequence similarity to the murine orthologous miRNAs (Table S1), and the remaining 84 were not coded by mouse genome as analyzed by BLAST and BLAT (Tables S2 and S3). As the probes on the Affymetrix arrays are for full-length miRNAs, they are not supposed to bind to the miRNA sequences with partial conservation. Hence, we consider all these also as non-specific and/or false-positive changes.

Using real-time PCR, we analyzed the levels of miR-21a-5p, miR-124-5p, miR-214-5p, miR-20b-5p, and miR-328-5p that were shown to be altered after resveratrol treatment in different experimental systems.^18,24–28 We did not detect any changes in the levels of any of these miRNAs between the resveratrol and vehicle groups. We presented the data for miR-21a-5p and miR-124-5p (Figure 1). However, data for miR-20b-5p, miR-328-5p, and miR-214-5p were not presented as the levels of these miRNAs were below the level of detection of real-time PCR in the mouse cerebral cortex in all groups.

Discussion

The plant-derived polyphenol resveratrol is a well-characterized antioxidant whose benefits have been demonstrated in models of atherosclerosis, stroke,^3,4 SAH,^5,6 TBI,^7,8 SCI,^9,10 chronic neurodegeneration, and cancer.^29,30 Resveratrol treatment in vitro in cancer cell lines as well as in vivo in rodent models of tumorigenesis showed distinctive alterations in miRNA expression, including the pro-survival oncomiRs miR-21 and miR-663.^16,25,26 Resveratrol-responsive miRNAs were also shown to promote survival in the ischemic heart.^18,31 Interestingly, these studies of resveratrol treatment show that pro-survival miRNA levels decrease in cancers, but increase in the ischemic heart, suggesting that its effects are organ- or disease-specific.

Recent studies also showed that resveratrol treatment provides robust ischemic tolerance similar to ischemic PC.^11–13 In rat hippocampal slices subjected to oxygen–glucose deprivation, SIRT1 inhibitor sirtinol reversed the ischemic tolerance induced by resveratrol PC as well as ischemic PC.¹¹ Further studies showed that the transcription factor nuclear factor (erythroid-derived 2)-like 2 (NRF2) and the neurotrophic factor BDNF also mediate resveratrol PC-induced ischemic tolerance.^13,32 Interestingly, ischemic PC was shown to significantly alter miRNA profiles in both rats and mice.^14,15 Many of the miRNAs altered after PC are putative candidates for inducing ischemic tolerance and neuroprotection. Furthermore, manipulation of miRNAs was shown to protect the post-ischemic brain, notably by preventing inflammation,³³ apoptotic/necrotic neuronal death,^34–36 and oxidative stress.^37,38

The present study shows that the resveratrol-induced cerebral ischemic tolerance might not involve miRNAs. To identify robust changes, we analyzed the effect of resveratrol on miRNA profiles with two different miRNA platforms (LC Sciences and Affymetrix). Both contain full-length probes for all the known mouse miRNAs. Although each array analysis showed changes in the miRNA profiles in resveratrol-treated group over vehicle control, there was no consistency between the two arrays. None of the miRNAs showed altered levels in both arrays. This suggests that any perceived alterations after one-dose intraperitoneal resveratrol administration (that induces ischemic tolerance) in the expression of some miRNAs might be either non-specific or false positives. However, some mouse-specific miRNAs observed to be altered after resveratrol treatment by each array platform might need further validation in order to rule them out as non-specific or false positives.

Previous studies showed that resveratrol treatment alters miRNA levels in various experimental systems. The miRNA miR-124 was shown to mediate the resveratrol-induced improvement in learning and memory in aging mice,²⁴ and miR-20b was shown to mediate the resveratrol-induced anti-angiogenic actions in rat ischemic myocardium.¹⁸ In vitro resveratrol treatment suppressed miR-21 expression in human glioblastoma cells²⁵ and human prostate cancer cells²⁶ and induced miR-328 expression in human osteosarcoma cells.²⁸ In a Parkinson’s disease mouse model, resveratrol treatment for 21 days increased miR-214 levels in mid brain.²⁷ However, we failed to find any change in any of these miRNAs in mouse brain after resveratrol treatment. This discrepancy might be due to the long-term resveratrol administration protocols and much higher doses of resveratrol used in many of the above studies. Differences in experimental systems (such as in vitro cultured cancer cells) and organs evaluated (heart versus brain) might also be responsible for the lack of changes in these miRNAs in resveratrol-treated groups in our experiments. These studies suggest that the mechanism of action of resveratrol might be organ and/or dose-specific.

There are limitations in miRNA analysis by different techniques. Consistency of results between different miRNA microarray platforms is not perfect.³⁹ Furthermore, the sensitivity of the detection and analysis between microarrays and real-time PCR is also not perfect in most cases, particularly for low-abundant changes. However, we detected many stably expressed miRNAs in parallel studies between the Affymetrix and LC Sciences microarray platforms suggesting optimization to be internally consistent. Nevertheless, consistent alterations in miRNA levels due to resveratrol PC were not detected even after accounting for limitations in the microarray and real-time PCR techniques.

To date, few studies evaluated the effect of resveratrol on miRNAs in non-cancerous brain cells. Two studies showed that treatment with resveratrol analog CAY10512 had no effect on miRNA profiles in neurons treated with H₂O₂ or astrocytes treated with metal sulfate.^40,41 Bigagli et al. showed that long-term resveratrol treatment in neurons prevented the downregulation of select miRNA clusters with age, but failed to counter the senescence-induced cell death.⁴² Intracerebroventricular administration of a daily dose of resveratrol for one week was shown to improve memory formation and long-term potentiation by altering the levels of miRNA-124, miRNA-134, cAMP response element binding protein, and BDNF in eight- to nine-month-old mice.²⁴ Whereas, the present study shows that a single intraperitoneal injection of resveratrol that induces robust ischemic tolerance had no effect on miRNA levels. It is noteworthy that this study does not explore the effect of long-term resveratrol treatment or increased dosage on cerebral miRNA profiles. Although our results are negative, they do not argue against the benefits of resveratrol-induced PC, which might be effective by a plethora of mechanisms, including induction of neuroprotective pathways, reactive oxygen/nitrogen species scavenging, suppression of inflammation, and upregulation of neurotrophic factors like BDNF.⁴³

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Mary Lopez received a fellowship from the UW Cellular & Molecular Pathology Graduate Training Program supported by the National Institutes of Health Award T32GM081061. This work was partially supported by the Department of Neurological Surgery, University of Wisconsin-Madison.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Authors’ contributions

All persons who meet authorship criteria are listed as authors, and all authors certify that they have participated sufficiently in the work to take public responsibility for the content, including participation in the concept, design, analysis, writing, or revision of the manuscript. Furthermore, each author certifies that this material or similar material has not been and will not be submitted to or published in any other publication before its appearance in the Journal of Cerebral Blood Flow and Metabolism. ML conducted the experiments and analyzed the data. RV designed the studies. RJD provided the guidance for the conceptual design. ML and RV wrote the manuscript.

Supplementary material

Supplementary material for this paper can be found at http://jcbfm.sagepub.com/content/by/supplemental-data