Modulation of BK channel by MicroRNA-9 in neurons after exposure to HIV and methamphetamine

Abstract

MicroRNAs (miR) regulate phenotype and function of neurons by binding to miR-response elements (MRE) in the 3' untranslated regions (3'UTR) of various messenger RNAs to inhibit translation. MiR expression can be induced or inhibited by environmental factors like drug exposure and viral infection, leading to changes in cellular physiology. We hypothesized that the effects of methamphetamine (MA) and human immunodeficiency virus (HIV)-infection in the brain will induce changes in miR expression, and have downstream regulatory consequences in neurons. We first used a PCR-based array to screen for differential expression of 380 miRs in frontal cortex autopsy tissues of HIV-positive MA abusers and matched controls. These results showed significantly increased expression of the neuron-specific miR-9. In vitro, we used SH-SY5Y cells, an experimental system for dopaminergic studies, to determine miR expression by quantitative PCR after exposure to MA in the presence or absence of conditioned media from HIV-infected macrophages. Again, we found that miR-9 was significantly increased compared to controls. We also examined the inwardly rectifying potassium channel, KCNMA1, which has alternative splice variants that contain an MRE to miR-9. We identified alternate 3'UTRs of KCNMA1 both in vitro and in the autopsy specimens and found differential splice variant expression of KCNMA1, operating via the increased miR-9. Our results suggest that HIV and MA -induced elevated miR-9, leading to suppression of MRE-containing splice variants of KCNMA1, which may affect neurotransmitter release in dopaminergic neurons.

Figure 1. MiRs are significantly differentially expressed among Control, HIV+, and HIV+MA groups

Two-hundred-sixty miRs met significance criteria after Benjamini-Hochberg correction (A) by ANOVA. After post-hoc Tukey HSD testing for differential expression among all possible combinations, 464 tests met criteria for significance (B). Plotted are the sorted P values (circles) and the Benjamini-Hochberg significance threshold (line). The −ΔΔCT values were used to sort the miRs by hierarchical clustering of correlation with one another and is illustrated by a cell plot (C) with color coding for differential expression. Any miRs that were undetected in > 1 sample were excluded. White color corresponds to the mean expression, magenta corresponds to lower, and green corresponds to higher; with color intensity proportional to magnitude of the difference. This plot demonstrates a clear effect of HIV in the miR expression in the CNS. The y-axis labels indicate the subject ID and are sorted by Group. The majority of miRs that were significantly different by ANOVA were decreased expression, with some notable exceptions like miR-9, miR-124, and let-7d, which were elevated.

Figure 2. MiR-9, miR-124, and let-7d are significantly increased in SH-SY5Y cells after exposure to 10% conditioned media from HIV-infected MDM

Terminally differentiated SH-SY5Y cells were exposed to high p24 (1,700 pg/mL) (black bars) and low p24 (900 pg/mL) (grey bars) conditioned media from HIV-1 infected MDM, and RNA isolated and quantified by Taqman qPCR for miR-9, miR-124, and let-7d; with U6 and RNU44 as endogenous controls and calibrated to the average time 0 values. The Log (RQ) values are plotted. *(P < 0.05 compared to time 0), error bars indicate standard deviation of four independent samples.

Figure 3. MiR-9, miR-124, and let-7d are significantly increased in SH-SY5Y cells after exposure to MA

Terminally differentiated SH-SY5Y cells were exposed to 10 μM MA for three hr repeated daily and RNA isolated in 24 hr increments. MiR-9, miR-124, and let-7d were quantified by Taqman qPCR; with U6 and RNU44 as endogenous controls and calibrated to the average time 0 values. The Log (RQ) values are plotted. *(P < 0.05 compared to time 0), error bars indicate standard deviation of four independent samples.

Figure 4. Response elements to miR-9 in alternative splice variants of the BK channel 3′UTR

(A) Polyacrylamide gel electrophoresis of BK channel 3′RACE and sequence analysis reveals alternate 3′UTR of the KCNMA1 mRNA. (B) Schematic constructed based on our results and reference (66). Human BK gene information lists options of exonal assembly of the 3′ ends of the coding sequence and 3′UTR. Numbers in boxes represent exons. Letters represent the N-terminal amino acid sequences and (*) stop codon. Note that KCNMA1 alternative splices variants 1 and 5 (v.1 and v.5) are different lengths because v.5 contains an additional exon 33a. Variants 2, 3, 4, and 6 contain an alternate variant of exon 33 (33b), which contains a stop codon, and so they lack exon 34 and have a distinct, shorter 3′UTR from v.1 and v.5. The KCNMA1 v.1 and v.5 have a 8,270 bp 3′UTR (a) and an miR-9 MRE positioned only 85 bp from the stop codon, in contrast to the 3′UTR (b) for the other variants, where the MRE is positioned 1,111 bp from the stop codon. The black arrowhead indicates the position of the forward 3′RACE primer in exon 31. Black arrowheads illustrated in each 3′UTR denote the positions of two distinct primer pairs to quantify the two different 3′UTRs. A black bar denotes relative position of the in situ hybridization probe used to visualize the distinct KCNMA1 3′UTR variants in the human brain, shown in Figure 5. (C) Juxtaposition of the two 3′UTR variants against miR-9 at the miR-9 MRE illustrating stronger binding for v.1 and v.5 mRNA. Bases paired by Watson-Crick bond are depicted by a black line and G:U pairs depicted by two dots. The entire seed sequence, denoted by the boxes nucleotides, is aligned to the mRNA for 3′UTR (a), but not for 3′UTR (b). See Supplementary Materials for primer, probe, and MRE positions; and see also Materials and Methods for primer sequence information.

Figure 5. In situ hybridization depicting qualitative expression levels and location of the two BK channel 3′UTR variants in the human brain

Samples were obtained from a subset from those included in the initial screening procedure (Table 1). KCNMA1 3′UTR(a) is more abundant in Control frontal cortex (A) than KCNMA1 3′UTR(b) (D). In both the HIV+ (B) and HIV+MA (C) brains, KCNMA1 3′UTR(a) is less abundant. In the HIV+MA brain, KCNMA1 3′UTR(b) (F) is lower than in both the Control (D) and HIV+ (E) brains. As expected, hybridization for both variants of KCNMA1 is neuronal.

Figure 6. MiR-9 mediates downregulation of KCNMA1 3′UTR v.1-5 but not 3′UTR for v.2-3-4-6 in SH-SY5Y cells exposed to HIV and MA

Incubation of differentiated SH-SY5Y cells with HIV and MA caused a decrease in expression of KCNMA variants containing the 3′UTR(a) (A), and for 3′UTR(b) (B). Pre-incubation with Mercury LNA inhibitor specific to miR-9 stopped KCNMA1 mRNA downregulation. Black bars - cells pre-incubated with scrambled LNA sequence, grey bars - cells pre-incubated with anti-miR-9 LNA sequence. Comparisons noted in (A) P = 0.07; and for comparisons noted in (B) P < 0.01 by Student’s t test.