Gain of survival signaling by down-regulation of three key miRNAs in brain of calorie-restricted mice

MiRNA expression targeting Bcl-2 in calorie-restricted mouse brain

The comparative expression of the miRNAs of interest was screened by miRNA microarrays (MMChips) in brain samples of CR- and ad lib-fed mice. The brain tissues were obtained from the National Institute on Aging tissue bank; C57/black 6 mice were calorie- restricted to 60% of normal ad lib caloric intake, according to the standardized regimen, starting at 4 months of age. By 12 months of age, the CR impact is fully established; therefore this age group was selected along with two older groups, 24 and 28 months (Supplementary Figure 1A). For each group, total RNA from brain tissues of three mice was used, to control for inter-animal variations. In addition, three MMChips were used for each mouse brain RNA sample, to generate statistically significant results of 9 datasets per age group. The Significance Analysis of Microarray(SAM) software listed miRNAs showing significant change (Figure 1A). MicroRNAs exhibit variations in expression in CR-fed mouse brain tissues, compared to their ad lib-fed counterparts, with inverse expression of age-dependent patterns (Figure 1A).

Figure 1. (A) Based on microRNA microarray chip data, a heat map indicating the expression levels of miRNAs with respect to age and mouse diet category (Ad libitum, AL and Calorie restricted, CR) in brain tissues was generated. Intensity of color represents relative expression, with blue representing decreased expression, and red representing increased expression. The comparison is drawn on the basis of expression among the mouse categories. (B) Graphical representation of miRNA (mmu-miR-34a, mmu-miR-30e and mmu-miR-181a-1*) expression in brain tissues, which exhibit the greatest fold changes among 12, 24 and 28 months in CR-fed and Ad lib mice. Samples are represented by age in months (12, 24, or 28). AL and CR for Ad lib and calorie restricted respectively. (C) Western blot analysis of Bcl-2, with alpha-tubulin as normalization control. The histogram shows average densitometry measurements of Bcl-2 after normalization. (D) Schematic diagram showing miRNA binding sites on the 3'UTR region of the Bcl-2 gene. (**p<0.01, **p<0.0001 in panels B and C; all histograms represent average ± std.dev.)

Among these, three microRNAs, mmu-mir-181a-1*, mmu-mir-34a and mmu-mir-30e, exhibit the most significant down-regulation in brains of CR-fed mice, in an age-dependent manner, compared to ad lib-fed mice (Figure 1B). The expression profiles of these three microRNAs were further confirmed by quantitative real-time PCR (qRT-PCR) (Figure 1B). These three miRNAs were selected because they manifest correlated fold changes at 24 and 28 months between CR- and ad lib-fed mice. All three microRNAs display significantly decreased expression in CR-fed mice at 28 months of age, compared to ad lib-fed mice. A disparity exists between MMChip results and qPCR in miR-30e expression, when 24 month old CR-fed mice are compared to 24 month old ad lib-fed control mice; the array results show decreased expression, whereas qPCR results show increased expression in 24-month CR-fed mice (Figure 1B). The qPCR data also show a significant difference in miR-34a expression between CR- and ad lib-fed mice at 28 months of age, whereas array data point to almost similar expression levels at that age. These anomalies seem to be based on the fact that microarray screening is less sensitive, and the analyses are based on signal strength from hundreds of loci on the arrays. Even after normalization with negative and positive controls, the assay is less sensitive than qPCR assay, which quantifies a single miRNA at a given time. Interestingly, a common denominator among all these three miRNAs is the age-dependent decrease in expression in CR-fed compared to ad lib-fed mice.

The miRBase targets version 5 database (http://microrna.sanger.ac.uk/) reported a set of nine genes which are targeted by all three miRNAs (Figure 1B); three other databases were employed to further confirm the predicted relationship between these miRNAs and these target genes. Bcl-2 was a consistent result from all the databases (Target Scan, Pictar and RNA 22). This analysis pointed to the fact that even human Bcl-2 may be a target for these miRNAs, as the seed sequences of these three miRNAs are well conserved between mice and humans (data not shown). In ad lib-fed mouse brain tissues Bcl-2 expression levels show an age- dependent decline (Figure 1C). In CR brain tissues, Bcl-2 expression is not only maintained, but rather increases as the mice reach the advance age of 28 months. This persistent increase in expression levels with age is inversely related to the reduced expression of miRNAs at the age of 28 months (Figure 1C).

We further analyzed the number of target sites for these miRNAs on the 3' UTR regions of the predicted target genes. Based on these exercises of algorithmic prediction, the 3'UTR of Bcl-2 and binding sites of the three miRNAs are depicted in the schematic diagram (Figure 1D).

In situ detection of miRNAs and Bcl-2 in calorie-restricted mouse brain tissues

MicroRNA -181a-1*,-34a and -30e expression patterns at different ages in both CR- and ad lib-fed mice were determined by in situ hybridization (ISH), using LNA (locked nucleic acid) probes (Figure 2A), in coronal serial sections of 10 μm thickness from 12, 24 and 28 month old brains. The densitometry values from the acquired images were normalized with the values from a hybridization control where a scrambled LNA probe was used. The expression profiles were consistent with our array screening and qPCR validation results. All three miRNAs showed reduced staining intensity in CR-fed mouse brains, compared to ad lib-fed mice; this difference was seen not only in cortical, but also in hippocampal, regions (Figure 2A).

Figure 2. Expression of three key MicroRNAs and Bcl-2 in CR mouse brain tissues. (A) In situ hybridization (ISH) detection of miRNAs (mmu-miR-34a, mmu-miR-30e and mmu-miR-181a-1*)in mouse brain tissues from both CR- and Ad lib-fed mice. Note increased hybridization signal of these miRNAs in the Ad lib group across all age groups, compared to CR. Graphical representations of densitometric analysis of miRNA (mmu-miR-34a, mmu-miR-30e and mmu-miR-181a-1*) hybridization in cortex and hippocampal regions. (B) Increased immunohistochemical staining of Bcl-2 in cortex and hippocampal region of CR, compared to Ad lib mice brain. Graphical representations of densitometric analysis of miRNA (mmu-miR-34a, mmu-miR-30e and mmu-miR-181a-1*) staining in cortex and hippocampal regions. (**p<0.01, **p<0.0001; all histograms represent average ± std.dev.)

The predicted target for these miRNAs, Bcl-2 expression, was detected by immunohistochemistry, as described in Methods. Increased staining of Bcl-2 was observed in CR-fed, compared to ad lib-fed, mice, in an age-dependent manner (Figure 2B). Densitometry values from the acquired images were normalized with values from a blank control (picture not shown). Cortical and hippocampal regions evenly show the elevated expression of Bcl-2 in CR- fed mice, in an age-dependent manner. Taken together, the correlated inverse expression of these miRNAs and Bcl-2 was validated by in situ hybridization and immunohistochemistry study.

MiRNA-dependent down-regulation of Bcl-2 in human and mouse cells

To evaluate the repression of Bcl2 by any of the three microRNAs, we used transfection to over-express miR-34a, -30e, or -181a in human embryonic kidney cells (HEK-293 cells) and mouse fibroblast cells (NIH/3T3 cells). The fact that miRNA binding sites on the 3'UTR region of Bcl-2 are conserved between mice and men gave us the opportunity to test this relationship in a mouse-independent system. Our results show that indeed Bcl-2 expression is repressed in all three miRNA-transfected cultures, by individual cell immnocytochemistry of transfected cultures with anti-body to this protein (Figure 3A-B). The densitometry values from the acquired images provide quantitative indices for this repression effect, plotted as histograms of miRNA expression in relationship to the repressed level of Bcl2 in these transfected cells.

Figure 3. Effect of over-expression of microRNAs on Bcl-2 expression in HEK-293 and NIH/3T3 cells. Immunocytochemistry analysis of miRNA (mmu-miR-34a, mmu-miR-30e and mmu-miR-181a-1*) suppression of endogenous Bcl-2 in both 293 (A) and NIH-3T3 (B) cell strains. The cells transfected with plasmid showing fluorescence and Bcl-2 immunostaining are from the respective identical fields. In each set (HEK-293 and NIH-3T3 cells), miRNA-transfected cells have significant Bcl-2 staining reduction, compared to Bcl-2 signal in neighboring cells lacking miRNA transfection. This pattern does not pertain in scrambled plasmid-transfected 293 or NIH-3T3 cells; scrambled plasmid-transfected cells show Bcl-2 staining intensity identical to un-transfected cells. The histograms show the quantification of fluorescence for plasmid transfection and the corresponding Bcl-2 expression in the cells. (**p<0.01, **p<0.0001; all histograms represent average ± std.dev.)

MiRNAs effect Bcl-2/Bax ratio favoring pro-survival signaling

Since Bcl-2 is an anti-apoptotic, and Bax a pro-apoptotic, gene, their ratio is a profound indicator of cell survival; this survival ratio can be altered either by increased Bcl-2 expression or decreased Bax expression. In CR-fed mice the Bcl-2/Bax ratio showed an increased trend in an age-dependent manner, compared to ad lib-fed mice (Figure 4A). To validate this antagonistic pattern between Bcl2 and Bax protein levels, we examined their ratios in microRNA-transfected HEK-293 and NIH/3T3 cultures. Total cellular protein fractions were isolated, as described in Methods. Decreased expression of Bcl-2 in miRNA-transfected cultures indeed led to a decline in Bcl-2/Bax ratio (Figure 4 B-C); densitometry values from the acquired Western blot images were normalized with values from alpha-tubulin blots as the loading control. Taken together, each of these three miRNAs is able to repress Bcl-2 in transfected cells, with subsequent increase of Bax proteins, thus resulting in a reduced Bcl2/Bax ratio. Finally, these results provide validation that Bcl-2 is a shared target for silencing by all three miRNAs at the post-transcriptional level.

Figure 4. Bcl-2/Bax ratios in calorie-restricted mouse brain tissue and over-expressed miRNA- transfected cell lines. Western blot analysis of Bcl-2, Bax and alpha-tubulin as normalization control. (A) In CR-fed mouse brain tissue, increased Bcl-2 is observed compared to Ad Lib (AL)-fed mice, while Bax expression stays constant. Increased Bcl-2/Bax ratio is observed in CR-fed mice, compared to Ad Lib-fed mice. Histogram represents ratios of Bcl-2 and Bax expression. (B) In miRNA- (mmu-miR-34a, mmu-miR-30e and mmu-miR-181a-1*) transfected HEK-293 cells, Bcl-2/Bax ratio declines, as Bcl-2 expression is considerably diminished in miRNA-transfected cells, compared to scrambled and vector controls. Histogram represents ratios of Bcl-2 and Bax expression. (C) In line with results in HEK-293 transfected cells, miRNA-transfected NIH/3T3 cells also exhibit a decline in Bcl-2 expression, resulting in a decline of the Bcl-2/Bax ratio. Histogram represents ratios of Bcl-2 and Bax expression. (**p<0.01, **p<0.0001; all histograms represent average ± std.dev.)

Altered expression of apoptosome-specific proteins Caspase-9 and Caspase-3

To validate the effect of Bcl-2 on apoptosome-mediated apoptosis, we examined the expression of Caspase-9 and Caspase-3. In line with Bcl-2 expression, the expression of cleaved Caspase-9 and Caspase-3 is significantly reduced in age-dependent manner from young to old age groups in CR-fed, compared to ad lib-fed, mouse brain samples (Figure 5A). Western blot assays were done using brain tissue samples from three different animals from CR- and ad lib- fed mice from each age group. For Western blot, in situ and qPCR assays, the specimens used were from the same mice, to maintain homogeneity and avoid inter-animal variation. The level of α-tubulin expression was used as the internal control for normalization of values of these two proteins.

Figure 5. Altered expression of active apoptosome-specific proteins in calorie-restricted mouse brain tissue and in over-expressed miRNA cell lines. Cleaved Caspase-9 (C-Caspase-9) and cleaved Caspase-3 (C-Caspase-3) are important components of the apoptosome complex. (A) In calorie-restricted (CR)-fed mouse brain tissues, levels of c-Caspase-9 and -3 decline in an age-dependent manner, compared to Ad Lib (AL)-fed mice. (B) Over-expression of three miRNAs in HEK-293 and NIH/3t3 cells produces a significant increase in cleaved form as compared to scrambled and vector alone controls. Increase in these apoptosome-specific proteins indicates the pro-apoptotic effect of these miRNAs. (**p<0.01, **p<0.0001; all histograms represent average ± std.dev.)

MiRNA-mediated repression of Bcl-2 3'UTR and elevated apoptosis in cell lines

To address the cascade effect from over-expression of the three microRNAs to repress Bcl-2 expression, with subsequent increase of Bax and downstream corresponding increase in Caspase activities, the miRNA-transfected HEK-293 and NIH/3T3 cultures were further examined by Western blot analysis with antibodies to Caspases, whose autocatalytic cleavage is essential to execute apoptotic death. Our results show that in transfected cells, the decreased Bcl-2/Bax ratio is accompanied by a significant increase in cleaved Caspase-9 and Caspase-3 (Figure 5B). Further validation of the three microRNAs’ ability to repress Bcl-2, and induce the chain reaction of an increase in Bax and Caspase cleavage, was achieved by direct repression assays with co-transfection experiments, where miR-34a, -30e, or 181a-1* expression plasmids were used to co-transfect along with cloned 3'-UTR reporter construct in HEK-293 cells. At 72 hours, the green fluorescence corresponding to the miRNA constructs can be observed, with fading red fluorescence of the reporter expression of 3'UTR constructs, compared to co-transfection experiments using control vector carrying a scrambled sequence. Image analysis of the intensities between the miR-plasmid and reporter constructs was performed; repression of the Bcl-2 3'UTR reporter intensities was significant in the miR-plasmid co-transfected cells, but not in their control counterparts.

Further functional impact of these miRNAs’ repression of bcl2 expression was performed by examining putative increases of apoptotic death in the transfected cells. MiRNA- transfected HEK-293 and NIH/3T3 cells were subjected to apoptosis inducers such as staurosporine, a general protein kinase inhibitor isolated from the bacterium Streptomyces staurosporeus. Our results show that indeed transfected cells show dosage-dependent decrease in LD₅₀. Since the transfection assays rarely acquire 100% efficiency, cell death percentages were calculated based on the positively transfected cell number, rather than the total cells in the cultures; the increase of apoptosis in cells with transduced microRNA expression indeed corresponds to the decrease of LD₅₀ indices.

Taken together, our work has shown that: during aging, miR-34a, -30e, and -181* are not up-regulated in CR brain, while pro-survival gene Bcl-2 is maintained at the same level in older mice as their younger counterparts, accompanied by a decrease of pro-apoptosome proteins such as Bax and Caspase 9 cleaved products. Functional validation of the three microRNAs’ silencing effect on Bcl-2, with subsequent induction of pro-cell death signaling, was documented by transfection studies, and suggests the role of these miRNAs in apoptosis. Thus, during aging, down-regulation of the three microRNAs may indeed reflect a gain of survival signaling in neurons of older CR mice.

MicroRNA profiling

Small RNA samples (1 μg) were labeled with digoxigenin (DIG) using the DIG Oligonucleotide Tailing Kit, 2nd Generation (Roche Diagnostics, U.S.A.) as described earlier [34]. Our own mouse miRNA DNA nitrocellulose membrane microchips (MMChips) were printed bearing 367 anti-sense mature miRNA sequences from miRBAse (http://microrna.sanger.ac.uk/Version 8.2) as previously described [35]. The membranes were scanned by an Expression 1680 scanner (Epson, U.S.A.), and data was acquired using Array-Pro Analyzer 4.5 software (Media Cybernetics, MD, U.S.A). Net intensity was derived from whole cell area measurement, and corrected using mean intensity of ring background of surrounding spots. Microarray data analyses were performed with SAM software, version 3.02 (Significance Analysis of Microarrays, Stanford University, CA), including pairwise comparison (T statistics) between age groups. Kolmogorov-Smirnov statistics were generated by Gene Set Enrichment Analysis (GSEA) software [36].

Mouse tissue samples

Tissue samples were obtained from the National Institute of Aging (http://www.nia.nih.gov/ResearchInformation/ScientificResources/AgedRodentTissueBankHandbook/RodentTissueBankResourcesInfo.htm) from CR- and AL-fed mice, at age groups of 4, 12, 24, and 28 months. The tissue samples were rapidly frozen at −80°C for analysis. Tissue samples were processed for protein, RNA extraction and in situ experiments as described below. Animal tissue related processing and experiments were done in accordance to the approved institutional (University of Louisville) biosafety board protocol #05-001.

Total and small RNA extractions

For total RNA extraction, the frozen tissue blocks were homogenized with Trizol according to the manufacturer's instructions. Small RNA enrichment was done using incubation with PEG 8000, as described previously [37].

Protein extraction

The frozen tissues were diced, then homogenized in two volumes (g mL–1) of RIPA buffer (150 mm NaCl, 10 mm Tris, pH 7.2, 0.1% SDS, 1.0% Triton X-100, 1% deoxycholate, 5 mm EDTA, pH 8.0) containing 1× protease inhibitor (Calbiochem, San Diego, CA) and centrifuged at 10,000 × g for 10 min. The supernatant containing the proteins was harvested. The Bradford method was used to quantify total protein concentration with BioRad reagents (BioRad, Hercules, CA).

QRT-PCR validation

Quantitative real-time PCR was carried out using Taqman miRNA assay reagents to verify miRNA expression profiles. Two hundred nanograms of total RNA were reverse transcribed to cDNA; the real-time PCR reactions were performed on a 7500 Fast System Real-Time PCR cycler (Applied Biosystems, Foster City, CA), according to the manufacturer's instructions. MicroRNA expression fold changes between ages across the groups were calculated using the delta Ct method, using RNU44 for normalization.

Western Blotting

The brain tissues (12, 24, 28 month ad lib- and CR-fed mice, n=3 each) were homogenized as described previously [34]. Following this procedure, membranes were incubated overnight at 4°C with primary antibodies: rabbit anti-Bcl-2, rabbit anti-Bax, rabbit anti-Caspase-3 and rabbit anti-Caspase-8 antibodies (sc-783, sc-6236, sc-7148 and sc-7890; Santa Cruz Biotechnology Inc, Santa Cruz, CA) and rabbit anti-Caspase-9 antibody (ab32539; Abcam Inc., Cambridge, MA). Blots were developed as described before [34].

Tissue processing for in situ hybridization and immunohistochemistry

The frozen brain tissue blocks were prepared by placing them in a cutting mold on dry ice, covered with Optimal Cutting Temperature (O.C.T.) compound (Sakura, Torrance, CA). The blocks were kept at −80 °C overnight; then 10 μm sections were collected in the coronal plane by cryostat (Leica, Solms, Germany), and mounted onto Superfrost^®Plus microscopic slides (Fisher Scientific, Pittsburgh, PA). Slides were stored at −80°C until further use.

MicroRNAs locked nucleic acid (LNA) in situ hybridization

The frozen sections were fixed with 4 % paraformaldehyde (PFA) for 30 min at room temperature, washed in 0.1 M phosphate-buffered saline (PBS) for 20 min, and treated with acetylation solution (92.8 % diethylpyrocarbonate (DEPC)-treated water, 1.3 % triethanolamine, 0.2 % HCl, 5.7 % acetic acid) for 20 min. The in situ hybridization was performed for miR-34a, miR-30e and 181a-1* as described previously [34].

Immunohistochemistry

The sections were then blocked with 10 % goat serum for 30 minutes (Invitrogen, Carlsbad, CA), and incubated with rabbit anti-Bcl-2 antibody (sc-783; Santa Cruz Biotechnology Inc, Santa Cruz, CA) overnight at 4°C. After incubation, the sections were washed thrice with PBS for 5 minutes, and incubated with Alexa fluor 555 goat anti-rabbit IgG (#4413; Cell Signaling Technology, Inc., Danvers, MA). For signal detection and analysis the procedure was followed as described earlier [34].

Construction of Bcl-2 3'-UTR Reporter Construct and miRNA expression constructs

Mouse Bcl-2 full length 3'-UTR was cloned into ptdTomato-C1 vector from Clontech (Mountain View, CA), following the gap repair method [38]. Briefly, upstream and downstream miniarms were amplified from mouse C57BL/6J genomic DNA using the following primers. The retrieve construct was made by cloning these two miniarms into EcoRI plus BamHI linearized ptdTomato-C1, following the Circular Polymerase Extension Cloning method [39]. After sequencing verification, the retrieve construct was linearized with EcoRV and electroporated into recombineering-competent bacterial artificial chromosome (BAC) clone containing the Bcl-2 3'-UTR genomic sequence. The BAC clone was made recombineering-competent by transformation with pKD46 [40]. Bcl-2 3'-UTR in ptdTomato-C1 was selected by kanamycin resistance, and verified by full coverage sequencing.

Expression constructs of mouse miR-34a, -30e, -181a, and miR-93 over were made by cloning miRNA precursor sequences with a flanking genomic sequence amplified from mouse genomic DNA (C57BL/6J) into pCDH-CMV-MCS-EF1-copGFP vector (System Biosciences, Mountain View, CA). The primers used for miRNA-specific primer sequences were mmu-miR-34a-Fwd: TGCTGTACCCTGCTGCTT, mmu-miR-34a-Rev: ACTCCAGACCCAGTGCCTC, mmu-miR-93-Fwd: TTGAATCCATGAGGGCTAGG, mmu-miR-93-Rev: TGTTCAGCTGTCCTGTGAGG, mmu-miR-30e-Fwd: GCAAGCCCTGGGTCACCTCC, mmu-miR-30e-Rev: AGCAATGCCTGACTGCATCACA, mmu-miR-181a-Fwd: GCCAAGCTCTGGGCTGGAGG and mmu-miR-181a-Rev: TCGGGTCCTGCTTTGCGG TT. The miRNA expression constructs were grown in Stbl3 strain (Invitrogen, Carlsbad, CA).

Cell transfection/co-transfection studies

Human embryonic kidney (HEK) 293 cells and NIH/3T3 cells (ATCC, Manassas, VA) were used for transfection/co-transfection experiments with the cell line nucleofector kit V by Nucleofector System (Lonza Walkersville Inc., Walkersville, MD), according to the manufacturer's instructions. In brief, 1 × 10⁶ cells were suspended in 100 μL nucleofector solution, and transfected with 5 μg of plasmids (vector control, scrambled control, mmu-miR-34a, mmu-miR-30e or mmu-miR-181a-1*). For co-transfection, the Bcl-2 3'UTR construct was co-transfected with vector alone, scrambled control, miR-34a, miR-30e and miR-181a-1* (1:1 ratio). The procedure followed has been described previously [34].

Apoptosis Assay

Human embryonic kidney (HEK) 293 cells and NIH/3T3 cells (ATCC, Manassas, VA) were used for transfection/co-transfection experiments with the cell line nucleofector kit V from Nucleofector System (Lonza Walkersville Inc. Walkersville, MD) according to the manufacturer's instructions. MiRNA expression plasmids were transfected, and transfection efficiency was calculated. Various doses of Staurosporine (2nM to 10nM) were used, and LD₅₀ was calculated based on the cell death ratios in miRNA-transfected and their respective control transfected cells. LD₅₀doses for miRNA transfected cells were calculated to determine the effect of miRNA on cell survival. Cells were stained with Tryphan blue as per the manufacturer's instructions http://cellgro.com/.

Statistical analysis

All tests were performed using SPSS Professional Edition software (IBM). Descriptive statistics including mean and s.e.m. along with one-way ANOVAs followed by multiple comparison tests and two tailed T test were used to determine significant differences. P < 0.05 was considered significant.