|Home | About | Journals | Submit | Contact Us | Français|
MicroRNAs (miRNAs) are ~22 nucleotides long, noncoding RNAs that control cellular function by either degrading mRNAs or arresting their translation. To understand their functional significance in ischemic pathophysiology, we profiled miRNAs in adult rat brain as a function of reperfusion time after transient middle cerebral artery occlusion. Of the 238 miRNAs evaluated, 8 showed increased and 12 showed decreased expression at least at 4 out of 5 reperfusion time points studied between 3 h and 3 days compared with sham. Of those, 17 showed >5 fold change. Bioinformatics analysis indicated a correlation between miRNAs altered to several mRNAs known to mediate inflammation, transcription, neuroprotection, receptors function, and ionic homeostasis. Antagomir-mediated prevention of mir-145 expression led to an increased protein expression of its downstream target superoxide dismutase-2 in the postischemic brain. In silico analysis showed sequence complementarity of eight miRNAs induced after focal ischemia to 877 promoters indicating the possibility of noncoding RNA-induced activation of gene expression. The mRNA expression of the RNases Drosha and Dicer, cofactor Pasha, and the pre-miRNA transporter exportin-5, which modulate miRNA biogenesis, were not altered after transient middle cerebral artery occlusion. Thus, the present studies indicate a critical role of miRNAs in controlling mRNA transcription and translation in the postischemic brain.
The mRNA levels and function are known to be controlled by several factors including transcription factors, histones, DNA methylation, and RNases. In addition, recent studies indicated that mRNAs are finely controlled by a group of small noncoding, evolutionarily conserved RNAs of ~22 nucleotides (nt) long called microRNAs (miRNAs) (Fire et al, 1998; Caplen et al, 2001; Grishok et al, 2001; Bartel, 2004; Ambros, 2004). Although their mechanism of action is not completely elucidated, miRNAs are thought to modulate mRNA function by binding to an 8-base pair complementary seed region in the 3′UTR. The binding of miRNAs is thought to either degrade the mRNAs or inhibit their translation (Humphreys et al, 2005; Jing et al, 2005). In addition, recent studies suggested that specific miRNAs can bind to target regions of the promoters of certain genes to induce their expression (Place et al, 2008) as well as can modulate gene expression by binding to a protein-coding region (Duursma et al, 2008). Several miRNAs can bind to a specific mRNA, and a single miRNA can bind to several mRNAs. Structural studies estimated that most miRNAs can act on ~200 mRNAs (Lewis et al, 2003). Because of this redundancy in function, the small number of miRNAs is thought to effectively control the huge number of mRNAs either sequentially or simultaneously. In addition, certain mRNAs and/or their protein products can control the expression of specific miRNAs in a feed-forward mechanism (Fazi et al, 2005). Many miRNAs were shown to control transcription factors that in turn control thousands of downstream genes (Miska et al, 2004; Krichevsky et al, 2006; Rogaev, 2005). Thus, miRNAs can significantly influence the cellular homeostasis under normal physiologic conditions.
Recent studies showed that many pathologic conditions significantly alter cerebral miRNA profiles, which might have profound effect on the disease outcome. It was shown that miRNAs are involved in brain tumor growth (Nicoloso and Calin, 2008), Alzheimer disease (Hébert et al, 2008), Down syndrome (Kuhn et al, 2008), schizophrenia (Beveridge et al, 2008), and stroke (Jeyaseelan et al, 2008).
An understanding of the functional significance of miRNAs will complement the existing knowledge gained from the mRNA studies to provide better therapeutic targets to prevent the secondary neuronal death after stroke. Hence, we presently analyzed the miRNA profiles as a function of reperfusion time after transient middle cerebral artery occlusion (MCAO) in adult rats. Using bioinformatics tools, we evaluated the target mRNAs and gene promoters of the miRNAs altered in the ischemic brain. We further evaluated the effect of preventing mir-145 induction on its down-stream target superoxide dismuate-2 (SOD2) expression employing the antagomir strategy.
In the nucleus, the miRNAs are transcribed as hairpin clusters of primary miRNAs (pri-miRNAs; 5′-capped polyadenylated transcripts), which will be converted to 70-nt stem loop structures (pre-miRNAs) by Drosha (a type-III RNase) in association with a cofactor Pasha (aka DiGeorge syndrome critical region gene 8) (Lee et al, 2003). The pre-miRNAs are transported from nucleus to cytosol by exportin-5 and acted on by another type-III RNase known as Dicer that deletes the terminal loop of pre-miRNAs to form mature miRNAs (Boyd, 2008). We also evaluated the effect of focal ischemia on the mRNA expression of Drosha, Pasha, Exportin-5, and Dicer.
Transient MCAO was induced in adult male spontaneously hypertensive rats (280–300 g; Charles River, Wilmington, MA, USA) under halothane anesthesia by the intraluminal suture method as described earlier (Vemuganti et al, 2004; Dhodda et al, 2004; Satriotomo et al, 2006; Kapadia et al, 2006; Tureyen et al, 2007, 2008). All the surgical procedures were approved by the Research Animal Resources and Care Committee of the University of Wisconsin-Madison and the animals were cared for 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).
Groups of rats subjected to transient MCAO were killed at various reperfusion periods (3, 6, 12, 24, and 72 h; n = 6 at each time point). Six sham-operated rats served as control. Total RNA was extracted from the ipsilateral cortex of each rat using the mirVana miRNA Isolation Kit (Ambion, Austin, TX) as per the manufacturer’s protocol. From each sample, 5 µg total RNA was size fractionated on a centrifugal filter (YM-100 Microcon; Millipore, USA). To the small RNAs (<300 nt), poly-A tails were added at the 3′end mediated by poly(A) polymerase, and the nucleotide tags were ligated to the poly-A tails. Each sample was hybridized to a microarray (LC Sciences, Houston, TX) that contained 238 known rat miRNA probes (12 repeats of each probe) from the Sanger miRBase version 9.0 (http://microrna.sanger.ac.uk/sequences/). These micoarrays were made by the in situ oligonucleotide synthesis directly on the microchips with the photogenerated acid coupled with conventional dimethoxytrityl chemistry using the μParaFlo microfluidic technology. Each of the detection probes on the microarray contained a coding segment (complementary to target miRNA) and a long spacer. The coding segment is the nucleotide sequence with a chemical modification and the spacer is a nonnucleotide molecule that extends the detection probe away from substrate to reduce the surface effects and hence increases the binding between target (miRNA in the sample) and the probe (corresponding sequence on the chip). This method of making chip increased the probe quality and reproducibility enabling high-quality process control and 10 times higher spot density than spotted arrays. The melting temperature of the detection probes on a microarray was balanced by incorporation of varying number of modified nucleotides with increased binding affinities. The hybridization reactions were performed overnight at 34°C in saline sodium phosphate EDTA buffer (900 mmol/L NaCl, 60 mmol/L Na2HPO4, 6 mmol/L EDTA, pH 6.8) containing 25% formamide. After hybridization, the labeling reaction was performed using tag-specific Cy5 dye. Hybridization images were collected using a laser scanner (GenePix 4000B, Molecular Devices, Sunnyvale, CA, USA) and digitized using Array-Pro image analysis software (Media Cybernetics, Bethesda, MD, USA).
The miRNA hybridization data were corrected by subtracting the background (calculated from the median of 5%–25% of the lowest-intensity cells) and normalized to the statistical median of all detectable transcripts using the locally weighted regression method, which balances the intensities of Cy5 labeled transcripts so that the differential expression ratios can be properly calculated (Bolstad et al, 2003). For background subtracting, the background was defined on each array as the average signal of the BKG0 spots (chemical linkers without the probes). The hybridization intensities above exp(5) (~150) were considered as significant as described earlier (Vagin et al, 2006), and intensities below 30,000 were considered as nonsaturated, as established with titration of several synthetic 20-nt RNA oligos (external controls) spiked into each sample. In addition, on each array there were 16 sets of spatially distributed internal control probes. These include PUC2PM-20B and PUC2MM-20B, which are the perfect match and the single-base mismatch sequences, respectively. The stringency of the intensity ratio of the PUC2PM-20B and PUC2MM-20B is expected to be larger than 30 indicating proper hybridization in each case. For proper analysis of signal intensities on each chip, both the internal controls and the test miRNA probes were repeated 12 times. On a microarray, the hybridization signal was linearly obtained from 1 to ~66,000 units. An miRNA transcript was considered detectable if it met the following criteria: (a) Signal intensity higher than three times the maximal background signal, (b) spot CV <0.5 (CV was calculated as (standard deviation)/(signal intensity)), and (c) the signals from at least 50% of the 12 redundant repeating probes above detection level. To avoid false positives, any spot that deviated >50% from the average value of the 12 repeating spots and/or spots with CV >0.5 were eliminated.
The data from different groups were normalized and analyzed statistically using ANOVA. The normalized hybridization data of the 36 microarrays (n = 6 for each group) were subjected to hierarchical cluster analysis using the Euclidian distance function. To increase the validity of the data, we generated cross-comparison matrices. For example, the data from the 6 sham chips were cross-compared with the 6 MCAO chips (at each time point) to generate a matrix of 36 comparisons. An miRNA transcript was assumed altered if it showed a statistically significant change in at least 30 out of 36 cross-comparisons (83% positive). The standard deviations of the sham and focal ischemia groups for each of the miRNAs profiled were <15%.
Real-time polymerase chain reaction (PCR) was performed to confirm the microarray data. The reverse transcription was performed using the TaqMan MiRNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA, USA) at 16°C for 30 mins, 42°C for 30 mins, and 85°C for 5 mins in buffer containing 10 ng low-molecular weight enriched RNA, 100 mmol/L dNTPs (with dTTP), 50 U reverse transcriptase, 0.4 U RNase inhibitor, and reverse transcriptase (RT) primer. PCR reactions were performed using the TaqMan® MiRNA Assay Kit (Applied Biosystems). Briefly, each reaction contained 10 µL, TaqMan 2 × Universal PCR Master Mix, 1 µL 20 × TaqMan® MicroRNA Assay reagent and 1.33 µL of the RT reaction product in a total volume of 20 µL. Real-time PCR was conducted at 95°C for 10 mins, followed by 40 cycles of 95°C for 15 sec and 60°C for 1 min. The threshold cycle (Ct) method (http://pebiodocs.com/pebiodocs/04303859.pdf) was used to determine the relative quantities of each miRNA. The sequences (5′−3′) of the amplified miRNA transcripts are as follows: UUCACAGUGGCUAAGUUCCGC (rno-mir-27a); GUCCAGUUUUCCCAGGAAUCCCU (rno-mir-145), UAUUGCUUAAGAAUACGCGUAG (rno-mir-137); UGGAAUGUAAG-GAAGUGUGUGG (rno-mir-206); ACAGCAGGCACAGACAGGCAG (rno-mir-214); UUGUGCUUGAUCUAACCAUGU (rno-mir-218); CUCAAACUAUGGGGGCACUUUUU (rno-mir-290); and UGAGGUAGUAGGUUGUAUAGUU (rno-let-7a). The mRNA expression of Drosha, Pasha, Dicer, and Exportin-5 was evaluated at 3 and 24 h of reperfusion in comparison to sham with real-time PCR using the SYBR-Green method as described earlier (Satriotomo et al, 2006, Tureyen et al, 2007, 2008). GAPDH and 18 secs rRNA were used as internal controls. The following primer sequences (5′–3′) designed with Primer Express Software (Applied Biosystems) based on the GenBank numbers given in parenthesis were used for real-time PCR: Drosha (NM_001107655): GGTGGCCGTTTACTTCAAAGG and GTCCATTGCTGCTCCCATTT; Dicer (XM_001068155): AGACCCACGGCAGCATTCT and GCGACGCAAAGATGGTGTTAT; and Exportin-5 (NM_001108789): AGTTGGCCGAAAAAACACAAA and TCCACCGAAACTTGACAACGT.
A cohort of rats subjected to transient MCAO or sham surgery were killed at 24 h of reperfusion (n = 4/group) after intracardial paraformaldehyde perfusion. Coronal brain sections (10 µm thick) from each rat were postfixed and acetylated by incubating in acetic anhydride/triethnolamine solution followed by washes in 1 × PBS. The sections were incubated in hybridization solution (50% formamide, 5 × SSC, 200 µg/mL yeast tRNA, 500 µg/mL salmon sperm DNA, 0.4 g Roche blocking reagent, and 5 × Denhardt’s solution) at room temperature for 5 h. The sections were incubated overnight at 22°C in hybridization solution containing 3 pmol of digoxin (DIG)-labeled LNA MiRCURY probes (Exiqon Inc, Woburn, MA, USA) at below the predicted Tm value of the probe used. The sections were washed at 60°C for 1 h in 0.2 × SSC and for 10 mins in 0.1 M Tris-HCl buffer (pH 7.5) and incubated in the blocking solution (10% fetal calf serum in 0.1 M Tris-HCl buffer) for 1 h at room temperature followed by labeling with anti-DIG antibodies (dilution 1:2000; Roche, IN, USA) overnight at 4°C.
The miRNA targets were identified using the RegRNA (http://regrna.mbc.nctu.edu.tw) with a high-stringent score cutoff of ≥ 140 and minimum free energy cutoff of ≤ −18.0. The RegRNA algorithm predicts targets based on a development of the open-source miRanda algorithm (http://www.microrna.org/miranda.html) (John et al, 2004) that incorporates the current biologic knowledge of target rules based on the mature rat miRNA sequences downloaded from mirBASE and the 5′-UTR regions of mRNAs from the NCBI database. An in silico analysis of the putative sequences in the promoters of rat genome complementary to the miRNAs induced after focal ischemia was conducted by using miRanda or RegRNA (http://regrna.mbc.nctu.edu.tw) algorithms with a very high stringency (energy threshold of −16.0 to −24.0 kcal/mol and a score threshold of = 100–160) by modifying the search target to scan promoter sequences instead of mRNA sequences as described earlier (Place et al, 2008). For this analysis, promoter sequences (1 kb in front of transcription start site) for rat genes were retrieved from Ensembl database. Hit positions on promoter are counted from −1 kb relative to the transcription start site.
Mir-145 is one of the miRNAs upregulated significantly in the postischemic brain from 3 h to 3 days after transient MCAO. Bioinformatics search showed that SOD2 is a major target mRNA of mir-145. To prove this experimentally, we tested the effect of preventing postischemic mir-145 activity using an antagomir that targets mir-145. In brief, antagomir-145 (miRIDIAN Hairpin Inhibitor IH-320377-06-0010; Dharmacon, Chicago, IL, USA) or control antagomir (miRIDIAN Hairpin Inhibitor negative control IN-001005-01; Dharmacon) was dissolved in artificial CSF (aCSF; 119 mmol/L NaCl, 3.1 mmol/L KCl, 1.2 mmol/L CaCl2, 1 mmol/L MgSO4, 0.50 mmol/L KH2PO4, 25 mmol/L NaHCO3, 5 mmol/L D-glucose, 2.2 mmol/L urea, pH 7.4) at a concentration of 20 nmol/mL and continuously infused into the lateral ventricles of cohorts of rats (n = 8/group) as described earlier (Vemuganti et al, 2004; Satriotomo et al, 2006). In brief, the antagomirs were filled into osmotic minipums (Alzet model 1030D), which pump at a rate of 1 µL/h (Alza Co., Palo Alto, CA, USA). Each pump was connected to an Alzet brain infusion stainless steel cannula by peristaltic tubing and primed overnight at 37°C to ensure immediate delivery after implantation. The cannula was stereotaxically implanted into the lateral ventricle [bregma; 0.8 mm posterior, −4.8 mm dorsoventral, −1.5 mm lateral; on the basis of the rat brain atlas of Paxinos and Watson (1998)] and secured to the skull with dental cement. The pump was placed in the skin fold on the neck of the rat. The cannula and pump implantation was conducted under halothane anesthesia. After 2 days, rats were subjected to transient MCAO for 1 h. At 1 day of reperfusion, from each group, four rats were killed and the ipsilateral cortical tissue was dissected for Western blotting. The remaining four rats from each group were killed after intracardial paraformaldehyde perfusion for immunohistochemistry. Western blotting was conducted as described earlier (Satriotomo et al, 2006). In brief, cortical tissue was homogenized in ice-cold 25 mmol/L Tris-HCl buffer (pH 7.4) containing 2 mmol/L EDTA and protease inhibitor cocktail (4-(2-aminoethyl) benzenesulfonyl fluoride, aprotinin, leupeptin, bestatin, pepstatin-A, and transepoxysuccinyl-L-leucylamido(4-guanidino)butane; Sigma Chemical Co., St Louis, MO, USA). Proteins were solubilized by adding Lamelli electrophoresis sample buffer (5% sodium dodecyl sulfate, 20% glycerol, 10% 2-mercaptoethanol, 125 mmol/L Tris-HCl, pH 6.8, and 0.004% bromophenol blue; Sigma Chemical Co.) and denatured by heating at 94°C for 3 mins. Samples (20 µg protein equivalent) were electrophoresed on Bio-Rad Criterion precast gels (4–20% polyacrylamide gradient), transferred to polyvinylidene difluoride membranes, and probed with polyclonal SOD2 antibodies (1:2000; Upstate Technologies, Temecula, CA, USA) and polyclonal HSP70 antibodies (1:3000; StressGen Biotechnologies, Victoria, BC, Canada) followed by HRP-conjugated anti-rabbit IgG (1:4000). The blots were stripped and reprobed with monoclonal anti-β-actin antibodies (1:10,000; Sigma Chemicals Co.). The protein bands recognized by the antibodies were detected by enhanced chemiluminescence according to the manufacturer’s instructions (Pierce, Rockford, IL, USA). Immunohistochemical staining was conducted as described earlier (Satriotomo et al, 2006). In brief, brains from the rats killed after paraformaldehyde perfusion were sectioned (coronal, 40 µm thick), and the sections were rinsed in 0.1 M Tris-buffered saline with 0.1% Triton-X100 (TBS-T; 3 × 5 mins), incubated in 1% H2O2 for 30 mins and washed in TBS-T (3 × 5 mins). The sections were incubated in 5% normal goat serum for 1 h and overnight in polyclonal SOD2 antibodies (1:1000; Upstate), washed in TBS-T (3 × 5 mins), incubated in biotinylated goat anti-rabbit antibody (1:1000; Vector Laboratories, Burlingame, CA, USA) for 1 h. The sections were washed in TBS and incubated for 30 mins in normal goat serum. Conjugation with avidin–biotin complex (1:100; Vectatin Elite ABC kit, Vector Laboratories) was followed by visualization with 3,3′-diaminobenzidine-hydrogen peroxide (Vector Laboratories). The sections were dehydrated, cleared, and mounted in Permount. Sections incubated without primary or secondary antibodies served as negative controls.
A 1 h transient MCAO in adult rats resulted in significant neuronal damage and infarction in the ipsilateral cortex and striatum that progressed with the time of reperfusion from 3 h to 3 days (Supplementary Figure 1). Microscopic examination of the Cresyl violet-stained brain sections showed a progressive loss of pyramidal neurons in the ipsilateral cortex between 6 h and 1 day of reperfusion (Supplementary Figure 1). These observations were similar to our previous studies (Satriotomo et al, 2006).
Of the 238 miRNAs spotted on the microarrays, 146 (61%) were observed to be expressed in the normal rat cerebral cortex (Supplementary Table 1). The expression levels for individual miRNAs ranged from 1 to 52,000 units (Figure 1A, top panel). After transient MCAO, 24 miRNAs showed increased expression and 23 miRNAs showed decreased expression at one or more reperfusion time points compared with sham (Table 1). Correlation plots of the expression levels of the miRNAs between sham samples showed no differences (Figure 1A, top panel), but correlation of sham samples with 24 h MCAO samples showed several up- and down-regulated miRNAs (Figure 1A, middle panel). The miRNA expression altered rapidly after focal ischemia. Eleven miRNAs altered by 3 h of reperfusion (three increased and eight decreased), and this number gradually increased as the reperfusion time progressed (Figure 1A, bottom panel). A maximal number of 46 miRNA altered at 3 days reperfusion (24 increased and 22 decreased) (Figure 1A, bottom panel). Hierarchical clustering shows the expression levels of all the 238 miRNAs analyzed as a function of reperfusion time after transient MCAO in comparison to sham (Figure 1B). Eight miRNAs showed a significant change (three increased and five decreased) at all time points (3 h–3 days reperfusion) after focal ischemia (Table 1). Whereas, 11 miRNAs showed altered expression (5 increased and 6 decreased) at 4 of 5 time points evaluated (Table 1). Of the 47 miRNAs altered after focal ischemia, 16 showed >10 fold change (6 increased and 10 decreased), and 11 showed 5 to 9.9-fold change (4 increased and 7 decreased) (Table 1). For a single miRNA, rno-mir-290 showed a maximal increase of 63-fold at 3 days of reperfusion (Table 1), and rno-mir-153 showed a maximal decrease of 52-fold at 1 day reperfusion (Table 1).
Using in situ hybridization, increased expression of rno-mir-145 and decreased expression of rno-mir-137 was confirmed in the ischemic core at 1 day reperfusion after transient MCAO (Figure 2). Compared with sham control, neither the contralateral cortex nor the periinfarct area on the ipsilateral cortex showed any change in the rno-mir-145 or rno-mir-137 expression pattern (Figure 2).
Using real-time PCR analysis, we evaluated the expression levels of three miRNAs that showed increased expression (mir-290, mir-145, and mir-206), three miRNAs that showed decreased expression (mir-137, mir-27a, and mir-218), and one miRNA that did not show any change in expression (let-7a) at 24 h reperfusion after transient MCAO compared with sham control. The real-time PCR essentially showed the same pattern of expression changes as observed with the microarray analysis (Figure 3A).
The real-time PCR analysis showed no significant changes in the mRNA expression of the RNases Drosha and Dicer, the Drosha cofactor Pasha, and the pre-miRNA transporter exportin-5 at either 3 h or 24 h reperfusion after transient MCAO compared with sham controls (Figure 3B).
Using the bioinformatics webtool microRNA.org, we searched the targets of each miRNA altered after focal ischemia. The 20 miRNAs altered in the postischemic brain (at least at 4 of the 5 reperfusion time points) showed several common mRNA targets. Surprisingly, many of those are known to be ischemia-responsive genes. In particular, we could correlate 14 inflammatory transcripts including IL1β, IL6, MIP1α, MCP1, ICAM1, Complement C3, COX2, and iNOS (Supplementary Figure 2) and 14 transcription factors including NF-kB, HIF1, Egr1, C/EBPβ PPARγ, IRF1, STAT3, and ATF3 (Supplementary Figure 3) to multiple miRNAs altered after focal ischemia. In addition, bioinformatics search also correlated ischemia-sensitive miRNAs to several mRNAs that code for neuroprotective proteins (HSP70, HSP27, HO1, PTBR, MT-1/2, Mn-SOD, Catalase, IGF-BP3, and osteopontin) (Supplementary Figure 4), and ion channels and neurotransmitter receptors (VGAT, NCKX2, TWIK, Na Channel I and II, Na/K-ATPase, IP3 receptor, GluR-b, GluR-c, and adrenergic receptor A1d) (Supplementary Figure 5). All these mRNA transcripts were previously shown to be altered in the postischemic brain (Vemuganti et al, 2002).
In silico analysis of the putative sequences in the promoters of rat genome complementary to the eight miRNAs induced at >4 reperfusion time points after transient MCAO showed 877 promoter hits (Table 2). Of these, 848 promoters have 1 hit and 29 have 2 hits. An interesting observation is that rno-mir-331 has a signature 7-bp motif (5′-ccccugg-3′) at its 5′end that enables rno-mir-331 to target only the promoters that contain the complementary sequence (5′-CCAGGG-3). None of the other 237 rat miRNA sequences scanned contained this motif. However, each of the 7 other miRNAs (induced after focal ischemia) tested also have unique 7 to 9-bp signature motifs that target specific sets of gene promoters (Table 2). Of the 8 miRNAs evaluated, rno-mir-331 showed maximum number of 270 hits, which included the promoters of many genes known to be induced after focal ischemia including Il6R, IL6st, PPARα, IL1β, SOD2, TGFβ2, TGFβ-R1, IGF-bp1, AQP2, and EPO (Table 3).
Mir-145 is one of the miRNAs that showed significant upregulation from 3 h to 3 days of reperfusion after transient MCAO (Table 1). Bioinformatics analysis with RegRNA and Miranda showed that rat miR-145 sequence has a complementary 8-bp targeting site in the 3′-UTR of rat SOD2 (GenBank#NM_017051) (Figure 4), which showed a very high minimum free energy of 165 and a score of −20.3. To experimentally show that mir-145 controls SOD2 protein expression in postischemic rat brain, we used an antagomir strategy. The cortex of sham-operated rats showed a very faint cytosolic Immunostaining of SOD2 (Figure 4). In the control antagomir infused rats, the SOD2 immunoreactivity increased in the neurons in the periinfarct area, but not in the ischemic core in the ipsilateral cortex at 1 day of reperfusion after transient MCAO (Figure 4). Whereas, infusion of an antagomir-145 that targets mir-145, the area that showed SOD2 immunoreactivity was much bigger with more number of neurons showing intense SOD2 immunostaining (Figure 4). Antagomir-145 group also showed smaller cortical infarcts than the control antagomir group (by 24%; n = 4/group; cortical infarct size was 179 ± 27 mm3 in control antagomir group and 136 ± 27 mm3 in the antagomir-145 group). When the periinfarct area from the two groups of rats was subjected to western blotting, the antagomir-145 group showed more SOD2 immunoreactive protein levels than the control antagomir group (n = 4/group) (Figure 4). Whereas, the immunoreactive protein levels of HSP70, which is known to be upregulated after focal ischemia but not a target of mir-145, was not influenced by antagomir-145 treatment (Figure 4). The real-time PCR analysis showed that antagomir-145 treatment significantly prevented the postischemic induction of mir-145 (by 75%; P <0.05) compared with control antagomir treated group (n = 4/group) (Figure 4). However, antagomir-145 had no effect on the levels of mir-206 (upregulated after focal ischemia), let-7a (unaltered after focal ischemia), mir-137, and mir-218 (both downregulated after focal ischemia) compared with control antagomir treatment (Figure 4).
In brief, the results of the present study showed that transient focal ischemia induces an extensive temporal change in the cerebral miRNAome but has no effect on miRNA synthetic machinery. Bioinformatics analysis showed that many of the miRNAs altered in the postischemic brain can target several mRNAs previously demonstrated to be altered after stroke. Furthermore, the miRNAs altered after focal ischemia showed complementarity to several gene promoters.
The molecular mechanisms that promote neuronal death, and thus, neurologic dysfunction after stroke are not understood completely. We and others showed that focal ischemia leads to extensive changes in the cerebral mRNA expression in rodents (Soriano et al, 2000; Vemuganti et al, 2002; Lu et al, 2003; Sharp et al, 2006; Du et al, 2006; Tang et al, 2006; Kapadia et al, 2006; Yan et al, 2007). This altered expression of various families of genes plays an important role in promoting the postischemic pathologic mechanisms like inflammation, ionic imbalance, edema, and receptor dysfunction that precipitate neuronal death after focal ischemia. Many studies showed that pharmacological manipulations can induce neuroprotection after focal ischemia by affecting the gene expression (Wang et al, 2004; Xu et al, 2005; Chen et al, 2006; Tureyen et al, 2007).
In the adult mammalian brain, thousands of mRNAs continuously transcribe and translate to form various proteins. Transcription factors, DNA methylation, RNA polymerases, ribosomes, and other components of the transcriptional and translational machinery are known to control mRNA and protein expression under normal as well as pathologic states. In addition, recent studies showed that noncoding RNAs including miRNAs are potent regulators of gene and protein expression. In rats, 286 miRNAs are identified so far and these are thought to control ~30% of all the mRNAs (http://microrna.sanger.ac.uk). As the action of an miRNA is dependent on the presence of a ~8-bp signature sequence in the 3′-UTR regions of mRNAs, most miRNAs can bind to a huge number of mRNAs either sequentially or simultaneously. In addition, as the 3′-UTR of most mRNAs contains complementary binding sites for multiple miRNAs, a specific mRNA can be controlled by several miRNAs (structural studies estimated that most miRNAs can inhibit ~200 mRNAs). Because of this redundancy in function, the small number of miRNAs can effectively control the huge number of mRNAs.
We currently observed that several miRNAs rapidly respond to focal ischemia and their expression changes by a very high magnitude. Furthermore, the ischemia-induced miRNA changes sustain for at least up to 3 days of reperfusion. We observed that 20% of the miRNAs analyzed (47 of the 238) were altered at one or more reperfusion time points after focal ischemia. This is striking as the number of mRNAs that show altered expression after focal ischemia is <3% of those profiled (Soriano et al, 2000; Vemuganti et al, 2002; Lu et al, 2003; Sharp et al, 2006; Du et al, 2006; Tang et al, 2006; Kapadia et al, 2006; Yan et al, 2007). Thus, the effect of focal ischemia seems to be greater on the genes that transcribe miRNAs than those transcribe proteincoding RNAs. In vertebrates, many miRNA genes are located in the intragenetic regions and the introns of the coding regions (Wang et al, 2007). Although ischemia can independently influence the transcription of mRNAs and miRNAs, an altered miRNA can significantly influence the translation of its downstream mRNAs. A recent study showed altered miRNA expression at 1 and 2 days of reperfusion after transient MCAO in Sprague–Dawley rats (Jeyaseelan et al, 2008). Despite the fact that we used spontaneously hypertensive rats, many miRNAs observed to be up- or downregulated at 1 and 3 days of reperfusion in the present study were also reported to follow the same pattern of change in the Sprague–Dawley rats (Jeyaseelan et al, 2008). In addition, our studies also showed that many of these miRNAs change as early as 3 h and the altered expression persists up to 3 days. Another recent study showed altered miRNA levels in rat carotid arteries after angioplasty and demonstrated that preventing miR-21 upregulation had a significant negative effect on neointimal lesion formation (Ji et al, 1997).
Although, the mechanisms that regulate miRNA transcription after focal ischemia are not known, changes in the miRNA synthetic RNases (Dicer and Drosha) and 5-exportin is not responsible. Bioinformatics showed that many ischemic-responsive miRNAs target several mRNAs known to be altered after focal ischemia. This indicates a possible crosstalk between mRNAs and miRNAs at the level of transcriptional regulation. It is also known that certain mRNAs and/or their protein products can control the expression of specific miRNAs (Fazi et al, 2005). The complex nature of miRNA–mRNA interaction controls granulopoiesis by the reciprocal regulation of transcription factors NFIA and C/EBPα and miRNA mir-223 (Fazi et al, 2005). Although NFI-A maintains mir-223 at low levels, replacing NFI-A by C/EBPα upregulates mir-223 which in turn suppresses NFI-A promoting granulocyte differentiation (Fazi et al, 2005). If the levels of a specific mRNAs are increased, that mRNA can instruct the expression of an upstream miRNA to prevent the translation of the mRNA as a feed-forward mechanism (Taganov et al, 2006). In addition, recent studies showed that miRNAs can also bind to promoter regions of certain genes to induce gene expression by a mechanism known as RNA-induced gene activation (Place et al, 2008). Thus, miRNAs and mRNAs can control each other bidirectionally, and hence, the relationship between an miRNA and an mRNA is not a simple, inverse one. Thus, after focal ischemia, an increased miRNA could prevent the translation of a downstream mRNA (by binding to the 3′UTRs) or could increase the expression of the mRNA transcription (by binding to the gene promoter). However, if an miRNA is decreased in expression, it can allow a downstream mRNA to translate or fail to sustain a gene transcription (if the binding of that miRNA to the promoter is a prerequisite for the normal level of transcription of that mRNA). Hence, altered miRNA expression has several consequences on mRNA transcription as well as translation.
Bioinformatics analysis plays a significant role in miRNA target prediction. However, to develop an miRNA as a therapeutic target, it is essential to experimentally validate its relation to the expression of the target protein. Mir-145 was one of the miRNAs observed presently to be upregulated after transient MCAO. Bioinformatics analysis with RegRNA and Miranda showed SOD2 as a likely target of mir-145 with very high stringency. As SOD2 is known to play a role in the antioxidant defense, its induction might promote cell survival after ischemia. We observed that after transient MCAO, the neurons that survived in the periinfarct area of the ipsilateral cortex show SOD2 protein expression, but mir-145 induction might not be allowing SOD2 protein to the necessary levels increase to promote any neuroprotection. When the postischemic mir-145 was neutralized with antagomir-145, the number of neurons that showed SOD2 was much higher in the ipsilateral cortex as well as the periinfarct area was much larger. Although these studies indicate a relationship between mir-145 and its target SOD2 in the postischemic brain, the protection afforded by antagomir-145 was only marginal as other miRNAs and mRNAs might also play a role in the postischemic neuronal damage.
As multiple miRNAs acting on common targets is an effective way to alter mRNA function, we evaluated the targets of some miRNAs altered after focal ischemia. In particular, we tested the 20 miRNAs that showed persistent change (altered expression at 4–5 consecutive reperfusion time points between 3 h and 3 days). We used the high-stringency bioinformatics tool microRNA.org that allows finding targets common to multiple miRNAs. It is striking that 14 miRNAs in various combinations target mRNAs that code for inflammatory genes, transcription factors, HSPs, anti-oxidant enzymes, growth factors, ion channels, and neurotransmitter receptors that are known to be altered after focal ischemia (Vemuganti et al, 2002). This indicates that altered expression of specific miRNAs and mRNAs after focal ischemia is not completely random. However, it is essential to experimentally establish the role of miRNA changes in modulating mRNAs that control various pathologic mechanisms in the postischemic brain.
To understand the possibility of RNA-induced gene activation in the ischemic brain, we conducted an in silico analysis of the miRNAs upregulated after focal ischemia against the promoter sequences of the whole rat genome. This analysis showed that the 8 miRNAs tested can bind to the promoters of many genes that play an important role in normal and ischemic brain. In particular, we observed that three miRNAs (rno-mir-324-3p, rno-mir-324-5p, and rno-mir-331) in various combinations can bind to the promoters of IL1β, IL6 receptor, IL6 signal transducer, CXCL chemokines (12α, 12β and 12γ), PPARα, HIF-1α, TIMP1, MMP9, TNFβ, TGFβ2, TGFβ-R1, TGFβ-R2, IGF1, IGF1-bp1, IGF1-bp4, AQP-2, AQP-3, iNOS, TNF-R, cytochrome-C oxidase subunits, EPO, SOD1, and SOD2.
Thus, the results of the present study indicate that focal ischemia significantly alters the temporal expression of many miRNAs, which might be controlling the mRNA transcription and translation, and thus, the resulting stroke pathophysiology. Future studies will show if modulating specific miRNAs can be a therapeutic option to prevent postischemic pathophysiological events and/or to promote plasticity and regeneration.
These studies were funded by NIH grants NS049448 AND NS061071.
Supplementary Information accompanies the paper on the Journal of Cerebral Blood Flow & Metabolism website (http://www.nature.com/jcbfm)