Down-Regulation of miR-23a-3p Mediates Irradiation-Induced Neuronal Apoptosis

Abstract

Radiation-induced central nervous system toxicity is a significant risk factor for patients receiving cancer radiotherapy. Surprisingly, the mechanisms responsible for the DNA damage-triggered neuronal cell death following irradiation have yet to be deciphered. Using primary cortical neuronal cultures in vitro, we demonstrated that X-ray exposure induces the mitochondrial pathway of intrinsic apoptosis and that miR-23a-3p plays a significant role in the regulation of this process. Primary cortical neurons exposed to irradiation show the activation of DNA-damage response pathways, including the sequential phosphorylation of ATM kinase, histone H2AX, and p53. This is followed by the p53-dependent up-regulation of the pro-apoptotic Bcl2 family molecules, including the BH3-only molecules PUMA, Noxa, and Bim, leading to mitochondrial outer membrane permeabilization (MOMP) and the release of cytochrome c, which activates caspase-dependent apoptosis. miR-23a-3p, a negative regulator of specific pro-apoptotic Bcl-2 family molecules, is rapidly decreased after neuronal irradiation. By increasing the degradation of PUMA and Noxa mRNAs in the RNA-induced silencing complex (RISC), the administration of the miR-23a-3p mimic inhibits the irradiation-induced up-regulation of Noxa and Puma. These changes result in an attenuation of apoptotic processes such as MOMP, the release of cytochrome c and caspases activation, and a reduction in neuronal cell death. The neuroprotective effects of miR-23a-3p administration may not only involve the direct inhibition of pro-apoptotic Bcl-2 molecules downstream of p53 but also include the attenuation of secondary DNA damage upstream of p53. Importantly, we demonstrated that brain irradiation in vivo results in the down-regulation of miR-23a-3p and the elevation of pro-apoptotic Bcl2-family molecules PUMA, Noxa, and Bax, not only broadly in the cortex and hippocampus, except for Bax, which was up-regulated only in the hippocampus but also selectively in isolated neuronal populations from the irradiated brain. Overall, our data suggest that miR-23a-3p down-regulation contributes to irradiation-induced intrinsic pathways of neuronal apoptosis. These regulated pathways of neurodegeneration may be the target of effective neuroprotective strategies using miR-23a-3p mimics to block their development and increase neuronal survival after irradiation.

Keywords: Bim; MOMP; Noxa; microRNA (miR), Puma; neuronal apoptosis; radiation.

Figure 1

The expression of pro-apoptotic members of the Bcl-2 family is upregulated, and miR-23a-3p is downregulated in the cortex and hippocampus of irradiated mice. Tissues and neurons were collected at 30 min, 6 h, 24 h, and 7 days after 10 Gy whole-brain irradiation. Total RNA was used for qPCR analysis. qPCR quantification of Puma, Noxa, Bax mRNA levels (A); and miR-23a-3p (B) in cortex, hippocampus, and ex vivo neurons, n = 6/group for brain tissues, n = 5/group for ex vivo neurons, with two technical replicates. Data represent the mean ± SD of one-way ANOVA and Tukey post-hoc analysis, * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001 vs. control animals.

Figure 2

Irradiation induces the activation of DNA damage and p53 pathways in primary cortical neurons. Neurons were collected 30 min, 6 h, and 24 h after 8 Gy irradiation. Image of a representative experiment. Whole-cell lysates were separated by SDS-polyacrylamide gel and immunoblotted with antibodies against Ph-ATM (Ser1981), Ph-ATR (Ser428) (approximately 300kDa), γ-H2A.X (Ser139), and Ph-p53 (Ser15) (representative images). Protein levels were quantified by densitometry; then, they were normalized to β-actin and to the parent protein. Data are presented as fold change compared with untreated control levels. The experiment was repeated 3 times with similar results, n = 3/group in each experiment. Data represent the mean ± SD of one-way ANOVA and Tukey post-hoc analysis, * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001 vs. control RCN.

Figure 3

IR up-regulated select pro-apoptotic members of the Bcl-2 family, mitochondrial membrane permeabilization, and release of pro-apoptotic molecules. Neurons were collected at 30 min, 6 h, and 24 h after 8 Gy irradiation. Total RNA was used for qPCR analysis. qPCR quantification of Puma, Noxa, Bim, and Bax (A); Bid, Bok, and Bak1 (B), and Bcl-xL and Bcl-2 (C) mRNA levels in a representative experiment. The experiment was repeated three times. N = 5/group in each experiment, with two technical replicates. Data represent the mean ± SD of one-way ANOVA and Tukey post-hoc analysis, * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001 vs. control RCN. RCNs were collected at 24 h after 8 Gy irradiation. Cytosolic fractions were separated by SDS-polyacrylamide gel and immunoblotted with antibodies against cytochrome c and GAPDH. Image of the representative experiment (D). Protein levels were quantified by densitometry, normalized to approximately a 37 kDa band of GAPDH, and presented as a fold change compared to control levels. The experiment was repeated 3 times with similar results, n = 3/group in each experiment. Data represent the mean ± SD. Statistical significance assigned by one-tailed t-test, ***p < 0.001 versus control. Irradiation induces neuronal cell death. Neurons were irradiated with 8 Gy. Twenty-four hours later, LDH release was measured. Data are expressed as a percentage of control untreated neurons as well as completely permeabilized cells (100% cell death) (E). The experiment was repeated six times with similar results, n = 6/group. Data represent the mean ± SD. Statistical significance was assigned by the one-tailed t-test, p < 0.0001 vs. control. Rat cortical neurons (RCNs) were collected at 30 min, 6 h, and 24 h after ionizing radiation (IR). Whole-cell lysates were separated by SDS-polyacrylamide gel and immunoblotted with antibodies against Puma, cleaved caspase-3, poly (ADP-ribose) polymerase family, member 1 (PARP), cleaved α-fodrin, and apoptotic peptidase activating factor 1 (Apaf-1). Image of a representative experiment (F). Protein levels were quantified by densitometry, normalized to β-actin, and presented as a fold change compared with untreated control levels (G). The experiment was repeated three times. N = 3/group in each experiment. Data represent mean ± SD of one-way ANOVA and Tukey post-hoc analysis, * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001 vs. control RCN.

Figure 4

Irradiation decreased levels of cellular miR-23a-3p, and miR-23a-3p, PUMA, and Noxa mRNAs within the RNA-induced silencing complex. Neurons collected at 30 min, 6 h, and 24 h after 2, 8, and 32 Gy irradiation. Total RNA was used for qPCR analysis. qPCR quantification of miR-23a-3p (A) and miR-27a-3p (B) levels in a representative experiment. The experiment was repeated three times. N = 3/group in each experiment for miR-23a-3p, n = 5/group in each experiment for miR-27a-3p, with 2 technical replicates. Data represent the mean ± SD of one-way ANOVA and Tukey post-hoc analysis * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001 vs. control RCN. Neurons were collected 3 h after 8 Gy treatment, subjected to RIP with Ago2 antibodies, and samples used for qPCR analysis. qPCR quantification of miR-23a-3p (C), Puma (D) and Noxa (E) levels in precipitates after RIP in a representative experiment. The experiment was repeated two times, n = 3/group in each experiment with 2 technical replicates. Data represent the mean ± SD. Statistical significance assigned by one-tailed t-test, *** p < 0.001 versus control, n = 3/group.

Figure 5

miR-23a-3p mimic reverses the decrease of endogenous miR-23a-3p in irradiated primary cortical neurons and attenuates neuronal cell death. RCNs were transfected with miR-23a-3p mimics and negative control mimics before irradiation. At 3 h after exposure to 8 Gy, neurons were harvested for RNA isolation. qPCR quantification of miR-23a-3p in a representative experiment (A). The experiment was repeated two times. N = 3/group in each experiment with 2 technical replicates. Data represent mean ± SD. Statistical significance assigned by one-way ANOVA and Tukey post hoc analysis; * p < 0.05, *** p < 0.001 vs. control RCN; ^ p < 0.05 vs. irradiated negative control mimics. LDH release was measured at 24 h after irradiation as a percentage of control untreated neurons as well as completely permeabilized cells (100% cell death). Analysis of a representative experiment (B). The experiment was repeated four times with similar results, n = 6/group in each experiment. Data represent mean ± SD. Statistical significance assigned by one-way ANOVA and Tukey post hoc analysis, * p < 0.05, *** p < 0.001, vs. control RCN. ^ p < 0.05, ^^^^ p < 0.0001 vs. negative control mimic transfected cells (-ve mim) RCN. RCNs were treated as described above and harvested 3 h after 8 Gy irradiation. Lysates from each sample were subjected to RNA-binding protein immunoprecipitation (RIP) using Ago2 antibodies, followed by qPCR analysis for levels of miR-23a-3p, Puma, and Noxa in the RISC. Analysis of a representative experiment (C) n = 3/group in each experiment with 2 technical replicates. Data represent mean ± SD. Significance assigned by one-way ANOVA and Tukey post hoc analysis, miR-23a-3p *** p = 0.0001 vs. control RCNs; ^^ p < 0.0039 vs. 8 Gy + miR-ve mimic; Puma *** p = 0.0004 vs. control RCNs; ^ p < 0.0471 vs. 8 Gy + miR-ve mimic; Noxa *** p = 0.0006 vs. control RCNs; ^ p < 0.0228 vs. 8 Gy + miR-ve mimic. IR causes the down-regulation of miR-23a-3p at the transcription stage. qPCR quantification of pri-miR-23a-3p. Analysis of a representative experiment (D). Data represent the mean ± SD of one-way ANOVA and Tukey post-hoc analysis, n = 6 for the control group, n = 3 for all other groups with 2 technical replicates in each experiment; * p < 0.05, ****p < 0.0001 vs. control RCN. miR-23a-3p mimic attenuates the irradiation-induced elevation of Puma, Noxa, and Bim in primary cortical neurons. RCNs were transfected with miR-23a-3p mimics and negative control mimics before irradiation. Neurons were harvested for RNA isolation at 1 h, 3 h, 6 h, and 24 h after exposure to 8 Gy. qPCR quantification of Puma, Noxa, Bim, and Bax and p21. Analysis of a representative experiment (E). The experiment was repeated three times, n = 6 for control, n = 3 for all other groups in each experiment with 2 technical replicates. Data represent mean ± SD. Statistical significance assigned by one-way ANOVA and Tukey post hoc analysis, ** p < 0.01, *** p < 0.001, **** p < 0.0001 vs. control RCN, ^^^ p < 0.001, ^^^^ p < 0.0001 vs. negative control mimic transfected cells (-ve mim) RCN.

Figure 6

miR-23a-3p attenuates DNA damage response and p53 activation in primary cortical neurons following irradiation. RCNs were transfected with miR-23a-3p mimics, and negative control mimics before irradiation. Neurons were collected at 30 min, 6 h, and 24 h after 8 Gy irradiation. Whole-cell lysates were separated by SDS-polyacrylamide gel and immunoblotted with antibodies against Ph-ATM (Ser1981), Ph-ATR (Ser428) (approximately 300 kDa), γ-H2A.X (Ser139), H2A.X, Ph-p53 (Ser15), p53, p21 and β-actin (representative image). Image of a representative experiment (A). Protein levels were quantified by densitometry, normalized to β-actin and the parent proteins for γ-H2A.X and Ph-p53, and presented as fold change compared with untreated control levels (B–F). The experiment was repeated 3 times with similar results, n = 3/group in each experiment. Data represent mean ± SD. Statistical significance assigned by one-way ANOVA and Tukey post-hoc analysis, * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001 vs. control RCN, ^^ p < 0.01, ^^^^ p < 0.0001 vs. negative control mimic transfected cells (-ve mim) RCN.

Figure 7

miR-23a-3p attenuates DNA damage response in primary cortical neurons following irradiation. Representative microscopy images from 30 min, 6 h, and 24 h of RCNs stained for γ-H2A.X (red), and 4′,6-diamidino-2-phenylindole (DAPI) (blue) (A). Data were calculated and plotted for all fields together as a cumulative frequency distribution without binning for γ-H2A.X foci count (B) and nuclear staining intensity (C). Eight fields were acquired per treatment as detailed above, and an average of 117 nuclei were identified per field.

Figure 7

miR-23a-3p attenuates DNA damage response in primary cortical neurons following irradiation. Representative microscopy images from 30 min, 6 h, and 24 h of RCNs stained for γ-H2A.X (red), and 4′,6-diamidino-2-phenylindole (DAPI) (blue) (A). Data were calculated and plotted for all fields together as a cumulative frequency distribution without binning for γ-H2A.X foci count (B) and nuclear staining intensity (C). Eight fields were acquired per treatment as detailed above, and an average of 117 nuclei were identified per field.

Figure 8

miR-23a-3p attenuates caspase-dependent neuronal apoptosis after irradiation. Neurons were transfected with miR-23a-3p mimics and miR-ve mimics and collected at 6 h after 8 Gy irradiation. Cytosolic fractions were separated by SDS-polyacrylamide gel and immunoblotted with antibodies against cytochrome c and GAPDH. Image of a representative experiment (A). Protein levels were quantified by densitometry, normalized to an approximately 37 kDa band of GAPDH, and presented as a fold change compared to control levels, n = 4 for control, n = 5 for other groups in each experiment. The experiment was repeated three times. Data represent mean ± SD. Significance assigned by one-way ANOVA and Tukey post-hoc analysis, **** p < 0.0001 vs. control RCNs; ^^^^ p < 0.0001 vs. 8 Gy + miR-ve mimic. RCNs were transfected with miR-23a-3p mimics, and negative control mimics before irradiation. Neurons were collected at 30 min, 6 h, and 24 h after 8 Gy irradiation. Whole-cell lysates were separated by SDS-polyacrylamide gel and immunoblotted with antibodies against Puma, α-fodrin, and cleaved caspase-3. Image of a representative experiment. (B). Protein levels were quantified by densitometry, normalized to β-actin, and presented as fold change compared with untreated control levels (C). The experiment was repeated 3 times with similar results, n = 3/group in each experiment. Data represent mean ± SD of one-way ANOVA and Tukey post hoc analysis, *** p < 0.001, **** p < 0.0001 vs. control RCN, ^^^^ p < 0.0001 vs. negative control mimic transfected cells (-ve mim) RCN.

Figure 9

The schematic illustration of the role of miR-23a on the IR-induced neuronal outcome. Pro-apoptotic events are shown in red, anti-apoptotic events are shown in green. Confirmed changes are underlined and shown in solid arrows.