Time-Dependent Profiles of MicroRNA Expression Induced by Ischemic Preconditioning in the Gerbil Hippocampus

Experimental Model

Experiments were performed with male Mongolian gerbils (Japan SLC Inc., Shizuoka, Japan). The 10- to 12-week-old gerbils weighing about 80 g were maintained in a temperature-regulated room (23–25°C) at a 12-h light/dark cycle for at least 7 days. The gerbils were fasted, but were allowed free access to water overnight before surgery. In total, 18 gerbils were randomized into three groups. When the animals began to regain consciousness after anesthesia, both common carotid arteries (CCAs) of the first and second groups (n = 8 each) were occluded with surgical microclips (RS-5424; Roboz Surgical Instrument Co., Gaithersburg, MD, USA) for 2 min of tCCAO. For the second group (n = 8), secondary occlusion of both CCAs for 2 min was induced 2 days later, and then in the same way for tertiary occlusion, with a total of three rounds of tCCAO at 2-day intervals. Gerbils of the sham control (SC) group were treated only once with a neck incision without any artery occlusion. The animals were allowed to recover for 1 day, 7 days, 1 month, or 6 months at ambient temperature (23–25°C), then were deeply anesthetized with pentobarbital (IP, 40 mg/kg; Kyoritsu Seiyaku Corporation, Saitama, Japan) and decapitated. SC animals were sacrificed in the same way just after the sham operation.

tCCAO and Reperfusion

Transient global forebrain ischemia was induced into gerbils by occluding bilateral CCA, as per our previous method (3). Briefly, gerbils were lightly anesthetized after inhaling a nitrous oxide/oxygen/isoflurane (69%:30%:1%; Air Water Inc., Okayama, Japan) mixture. Body temperature was monitored in all animals, and was maintained at 37 ± 0.3°C by using a heating pad (model BMT-100; Bio Research Center, Nagoya, Japan) during surgery. A midline neck incision was made, and both CCAs were exposed. Animals began to regain consciousness after the end of anesthesia, and both CCAs were occluded with surgical microclips. After 2 min of tCCAO, cerebral blood flow was restored by removing the microclips. At 1 day, 7 days, 1 month, and 6 months after reperfusion, the animals were euthanized as described above (n = 2) using a previously defined method (28). Then, all brains were dissected, quickly frozen in powdered dry ice, and stored at −80°C until subsequent preparation.

All animal experiments were approved by the Animal Committee of the Graduate School of Medicine and Dentistry, Okayama University. All institutional and national guidelines for the care and use of laboratory animals were followed.

Total RNA (Including miRNA) Isolation From Gerbil Brain Samples

The bilateral hippocampal CA1 region of the frozen brain was dissected and combined in a cryostat (HM 500M; Microm Inc., Mound, MN, USA) at −18°C. Total RNA containing miRNA was isolated from about 5 μg of the hippocampal CA1 region suspended in RNA-Later ICE (Ambion Inc., Austin, TX, USA) using a mirVana^™ miRNA Isolation Kit (Ambion). The concentration, quality, and integrity of total RNA were determined using a NanoDrop Spectrophotometer (ND-1000; NanoDrop Tech, Rockland, DE, USA). The ratio of OD₂₆₀/OD₂₈₀, which was >1.8 for all RNA samples, was used to ascertain the quality of RNA. RNA samples were stored at −80°C before miRNA array studies.

Gerbil miRNA Sequencing

For each SC or ischemic brain sample, an equal amount of total RNA was pooled from three biological replicates to generate enough RNA (approximately 1,000 ng) for deep sequencing (Table 1). Briefly, the isolated total RNAs of each sample were resolved on a denaturing 15% polyacrylamide gel (Bio-Rad Laboratories, Hercules, CA, USA) for size selection, and these small RNAs (≤30 bases) were ligated to a pair of Solexa adaptors at the 5′ and 3′ ends using T4 RNA ligase. After ligation and purification, adapter-ligated small RNAs were reverse transcribed and amplified for 15 cycles (cDNA RT-PCR) with a pair of adapter complementary primers in order to produce sequencing libraries. PCR products were then purified and directly sequenced using an Illumina HiSeq2000 instrument (Illumina Inc., San Diego, USA).

OPEN IN VIEWER

Table 1.

MicroRNAs Targeted by the Transient Ischemia-Induced MicroRNA

miRNA	Number of Targets	Target Sequence (5′ to 3′)
mmu-miR-15a-5p	337	UAGCAGCACAUAAUGGUUUGUG
sha-miR-24 L+3R+1	1,986	AUCUGGCUCAGUUCAGCAGGAAC
oan-let-7d-3p_R+2	8	CUAUAUAGCCUACUGCCUUCCU
mmu-miR-125b-5p	354	UCCCUGAGACCCUAACUUGUGA
mmu-miR-132–5p	371	AACCGUGGCUUUCGAUUGUUAC
mmu-miR-181c-5p	448	AACAUUCAACCUGUCGGUGAGU
mmu-miR-378a-5p	1,074	CUCCUGACUCCAGGUCCUGUGU

mmu-miRNA, Mus musculus (mouse) microRNA; sha-miR, Sarcophilus harrisii (Tasmanian devil) microRNA; oan-let-7d: Ornithorhynchus anatinus (platypus) let-7d microRNA.

miRNA Microarray Analysis

Groups of gerbils subjected to tCCAO were sacrificed at various reperfusion periods (1 day, 7 days, 1 month, and 6 months; n = 2 at each time point). To perform the miRNA array experiment, total RNA (2–5 μg) was size fractionated on a Microcon centrifugal filter (YM-100; Millipore, Billerica, MA, USA). To the small RNAs (<300 nucleotides), poly-A tails were added at the 3′ end mediated by poly(A) polymerase, and nucleotide tags were ligated to the poly-A tails. Each sample was hybridized to a microarray (LC Sciences, Houston, TX, USA) that contained gerbil miRNA probes. Thereafter, an oligonucleotide was ligated to the poly-A tail for later staining with fluorescent dye Cy3. Hybridization was performed overnight using μParaFlo^® microfluidic technology. Hybridization images were collected using a laser scanner (GenePix 4000B; Molecular Device, Sunnyvale, CA, USA) and digitized using Array-Pro image analysis software (Media Cybernetics, Bethesda, MD, USA). Correlations between fluorescence signal intensities from the laser scanner and expression levels of different miRNA sequences were generally proportional. Since fluorescence intensity was the read-out of the laser scanner, fluorescence intensity data were expressed with no absolute units in the figures of this study.

Statistical Analysis

All data in this study are presented as means ± SD. Data from different groups were normalized and statistically analyzed using one-way ANOVA, followed by Tukey's post hoc comparison. The normalized hybridization data of the microarrays were subjected to hierarchical cluster analysis using the Euclidian distance function. To increase the validity of the data, cross-comparison matrices were generated. For example, data from the two sham chips were crosscompared with the two tCCAO chips (at each time point) to generate a matrix of four comparisons. Statistical analyses were performed using SPSS (ver. 13.0; IBM, Armonk, NY, USA). All tests were considered statistically significant at a value of p < 0.05, as per our previous report (12).

miRNA Expression Profile in Gerbil Hippocampus

A total of 1,067 miRNAs (LC Sciences, Houston, TX; USA) were screened in samples of the CA1 region from gerbils subjected to tCCAO. A total of 251 miRNAs were detected in samples of one or three doses of tCCAO. Since no miRNA data for gerbils are available, the sequences of mouse and some other animal species' miRNAs were used for this study (Fig. 1A, Table 1). Each column in Figure 1B–E represents a single sample, and each row represents a single miRNA. Black squares represent the median level of miRNA, and green or red squares represent lower or higher levels than the median level of miRNA expression, respectively.

OPEN IN VIEWER

Figure 1.

Time-dependent profiles of microRNA expression in gerbil hippocampus. Hierarchical clustering of miRNAs detected in the hippocampal CA1 region showed four distinct clusters for neuroprotection and protein expression. Each column represents a single sample, and each row represents a single miRNA. Black squares represent median level of miRNA, and green or red squares represent lower or higher levels than the median level of miRNA expression, respectively (A). Clusters B, C, D, and E contained miRNAs related to neurogenesis, 43 kDa transactivation response DNA-binding protein (TDP43), fused in sarcoma/translocated in liposarcoma (FUS/TLS), and heat shock protein 70 (HSP70), respectively. tCCAO, transient common carotid artery occlusion; mmu-miR, Mus musculus (mouse) microRNA; oan-let-7b, Ornithorhynchus anatinus (platypus) let-7d microRNA; sha-miR-24, Sarcophilus harrisii (Tasmanian devil) microRNA 24.

Hierarchical clustering of the miRNAs detected in the brain showed four distinct clusters for neurogenesis and protein expression (Fig. 1B–E). Cluster B contained Mus musculus (mouse)-let-7b (mmu-let-7b), Macaca mulatta (rhesus monkey; mml)-miR-9, mmu-miR-10b, mmu-miR-15a, mmu-miR-15b, Equus caballus (horse; eca)-miR-16, mmu-miR-16, mmu-miR-20 (a, b), mmu-miR-28 (a, b), mmu-miR-29 (a, b, c), mmu-miR-34 (a, b, c), mmu-miR-124, mmu-miR-128, oan-miR-128, mmu-miR-125b, Ornithorhynchus anatinus (platypus; oan)-miR-125, mmu-miR-133 (a, b), mmu-miR-146b, Gorilla gorilla gorilla (gorilla; ggo)-miR-146b and rno-miR-146b, related to neurogenesis (18,27,29); cluster C contained mmu-let-7 (b, d, i), oan-let-7b, Canis familiaris (dog; cfa)-mir-24, mmu-miR-24, Sarcophilus harrisii (Tasmanian devil; sha)-miR-24, mmu-miR-33, mmu-miR-132, mmu-miR-143, mmu-miR-330, and mmu-miR-574, related to 43 kDa trans-activation response DNA-binding protein (TDP43) (9,10, 24); cluster D contained mmu-miR-9, oan-miR-9, mmu-miR-12, mmu-miR-132, mmu-miR-192, mmu-miR-199 (a, b), and sha-miR-199a, related to fused in sarcoma/translocated in liposarcoma (FUS/TLS) (38); cluster E contained mmu-miR-181 (c, d), mmu-miR-200 (a, b, c), and mmu-miR-378 (a, b, c, d), related to heat shock protein 70 (HSP70) (41,49).

Target Analysis of Differently Expressed miRNAs

Among the above four clusters, miRNA expression profiles for mmu-miR-15a-5p, sha-miR-24_L+3R+1, oan-let-7b-3p_R+2, mmu-miR-125b-5p, mmu-miR-132–5p, mmu-miR-181c-5p, and mmu-miR-378a-5p were compared in the CA1 subfield of gerbils subjected to one or three doses of 2-min tCCAO. Heat maps showed the pattern of miRNA expression among the different experimental groups. These expressed miRNAs were used for further analyses (Figs. 2, 3, 4).

OPEN IN VIEWER

Figure 2.

mmu-miR-15a-5p expression during neuroprotection. From the signal strength of 820 ± 22 in SC brains (gray circle), the signal strength slowly, but progressively, increased in the CA1 region after 2-min tCCAO for one dose (open diamonds) and three doses (filled squares). Note the higher strength after three doses versus one dose (*p < 0.05 vs. SC, **p < 0.01 vs. SC).

OPEN IN VIEWER

Figure 3.

MicroRNA changes for FUS/TLS and TDP43. MicroRNA changes from baseline SC brains (gray circles) for FUS/TLS (A, C) and TDP43 (B, D). The signal strength of mmu-miR-125b-5p and mmu-miR-132–5p showed a similar increasing pattern in both groups of 2-min tCCAO for one dose (open diamonds) and three doses (filled squares) (*p < 0.05 vs. SC) (A, C). In contrast, the peak time of sha-miR-24 and oan-let-7b-3p accelerated to 1 day from three doses (filled squares) to one dose (open diamonds), then gradually returned until 6 months (*p < 0.05 vs. SC, **p < 0.01 vs. SC) (B, D).

OPEN IN VIEWER

Figure 4.

MicroRNA changes for HSP70. Note the peak time acceleration to 1 day and 7 days from one dose (open diamonds) to three doses (filled squares), with a gradual return to 6 months in mmu-miR-181c-5p and mmu-miR-378a-5p (*p < 0.05 vs. SC, **p < 0.01 vs. SC.

mmu-miR-15a-5p Expression for Neurogenesis

The signal strength of mmu-miR-15a-5p for the hippocampal CA1 region is shown in Figure 2. Signal strength was 820 ± 22 in the SC brains (Fig. 2, circle). After one dose of 2-min tCCAO, the signal strength was slowly, but progressively, induced in the CA1 region. From the baseline of the SC level, the strength of miRNA increased to 943 ± 153, 1,125 ± 196, 1,163 ± 37, and 1,373 ± 244 at 1 day, 7 days, 1 month, and 6 months, respectively (Fig. 2, open diamonds). In the group of three doses of 2-min tCCAO, the signal strength showed a similar pattern to that of the one dose, but the strength was much higher: 1,301 ± 186, 1,438 ± 311, 1,525 ± 356, and 1,386 ± 166 at 1 day, 7 days, 1 month, and 6 months, respectively (*p < 0.05 vs. SC, **p < 0.01 vs. SC, filled squares) (Fig. 2).

sha-miR-24 and oan-let-7b-3p Expression for TDP43

The signal strength of sha-miR-24_L+3R+1 was 7,098 ± 920 in the hippocampal CA1 region of the SC brains (Fig. 3B, circle). From the baseline level of the SC brain, the strength of miRNA increased to 9,327 ± 2,889 and 12,134 ± 240 at 1 day and 7 days, peaked at 1 month (13,790 ± 2,885) then slightly returned at 6 months (11,543 ± 1,776) after one dose of 2-min tCCAO (Fig. 3B, open diamonds). In the group of three doses of 2-min tCCAO, the peak time accelerated to 1 day (13,122 ± 781), then gradually returned until 6 months. The strength of miRNA was 12,421 ± 2,319, 12,452 ± 1,159, 10,586 ± 167 at 7 days, 1 month, and 6 months, respectively (*p < 0.05 vs. SC, **p < 0.01 vs. SC, filled squares) (Fig. 3B).

The signal strength of oan-let-7b-3p_R+2 was 18 ± 2 in the hippocampal CA1 region of the SC brains (Fig. 3D, circle). From the baseline level of the SC brain, the strength of miRNA increased to 17 ± 1, 25 ± 16, 30 ± 18, and 46 ± 11 at 1 day, 7 days, 1 month, and 6 months, respectively, after one dose of 2-min tCCAO (Fig. 3D, open diamonds). In the group of three doses of 2-min tCCAO, the signal strength peaked at 1 day (63 ± 18), then returned from 7 days to 6 months. The strength of miRNA was 46 ± 6, 35 ± 1, and 32 ± 6 at 7 days, 1 month, and 6 months, respectively (*p < 0.05 vs. SC, **p < 0.01 vs. SC, filled squares) (Fig. 3D).

mmu-miR-125b-5p and mmu-miR-132–5p Expression for FUS/TLS

The signal strength of mmu-miR-125b-5p and mmu-miR-132–5p for the hippocampal CA1 region is shown in Figure 3A and 3C. The signal strength of mmu-miR-125b-5p (Fig. 3A) showed a similar increasing pattern in both groups of 2-min tCCAO at both doses. From the baseline level of the SC brain (17,508 ± 2,915) (Fig. 3A, circle), the strength of mmu-miR-125b-5p was 18,392 ± 2,680, 25,851 ± 715, 27,273 ± 166, 22,717 ± 3,889 at 1 day, 7 days, 1 months, and 6 months after one dose of 2-min tCCAO (Fig. 3A, open diamonds), and 24,879 ± 456, 26,344 ± 152, 24,574 ± 1,301, 26,728 ± 1,236 at 1 day, 7 days, 1 months, and 6 months after three doses (*p < 0.05 vs. SC, filled squares) (Fig. 3A).

The signal of mmu-miR-132–5p increased and peaked at 1 day then was downregulated from 7 days to 6 months in both groups (Fig. 3C). From the baseline level of the SC brain (705 ± 30, Fig. 3C, circle), the strength of mmu-miR-132–5p was 769 ± 209, 710 ± 57, 601 ± 37, 541 ± 46 at 1 day, 7 days, 1 month, and 6 months after one dose of 2-min tCCAO (Fig. 3C, open diamonds), and 1,009 ± 152, 637 ± 64, 574 ± 71, 678 ± 176 at 1 day, 7 days, 1 month, and 6 months after three doses (*p < 0.05 vs. SC, filled squares) (Fig. 3C).

mmu-miR-181c-5p and mmu-miR-378a-5p Expression for HSP70

The signal strength of mmu-miR-181c-5p was 530 ± 45 in the hippocampal CA1 region of the SC brains (Fig. 4A, circle). From the baseline level of the SC brain, the signal strength of miRNA increased to 2,056 ± 2,091, 2,587 ± 1,144, and 3,094 ± 144 at 1 day, 7 days, and 1 month, peak at 1 month, then returned slightly at 6 months (2,479 ± 892) after one dose of 2-min tCCAO (Fig. 4A, open diamonds). In the group of three doses of 2-min tCCAO, the peak time accelerated to 1 day (3,330 ± 224), then gradually returned to 3,278 ± 772, 2,675 ± 226, and 1,870 ± 264 at 7 days, 1 month, and 6 months, respectively (*p < 0.05 vs. SC, **p < 0.01 vs. SC, filled squares) (Fig. 4A).

The signal strength of mmu-miR-378a-5p was 56 ± 8 in the hippocampal CA1 region of the SC brains (Fig. 4B, circle). From the baseline of the SC level, the strength of miRNA decreased to 56 ± 2, 48 ± 7, 42 ± 1, and 52 ± 8 at 1 day, 7 days, 1 month, and 6 months, respectively, after one dose of 2-min tCCAO (Fig. 4B, open diamonds). In the group of three doses of 2-min tCCAO, signal strength was induced in the CA1 region with a peak time at 7 days (75 ± 23), then returned from 1 month to 6 months. The strength of miRNA was 58 ± 1, 68 ± 3, and 55 ± 0 at 1 day, 1 month, and 6 months, respectively (*p < 0.05 vs. SC, **p < 0.01 vs. SC, filled squares) (Fig. 4B).