Research Article For reprint orders, please contact:
[email protected]
Genome-wide miRNA expression profiling of human lymphoblastoid cell lines identifies tentative SSRI antidepressant response biomarkers Aim: Over 30% of patients with major depression do not respond well to first-line treatment with selective serotonin reuptake inhibitors (SSRIs). Using genome-wide expression profiling of human lymphoblastoid cell lines (LCLs) CHL1 was identified as a tentative SSRI sensitivity biomarker. This study reports on miRNAs implicated in SSRI sensitivity of LCLs. Methods: Eighty LCLs were screened from healthy adult female individuals for growth inhibition by paroxetine. Eight LCLs exhibiting high or low sensitivities to paroxetine were chosen for genome-wide expression profiling with miRNA microarrays. Results: The miRNA miR-1513p had 6.7-fold higher basal expression in paroxetine-sensitive LCLs. This corresponds with lower expression of CHL1, a target of miR-151-3p. The additional miRNAs miR-212, miR-132, miR-30b*, let-7b and let-7c also differed by >1.5-fold (p 1.8; and 260/230 nm >2.0. Microarray experiments & data analysis Affymetrix (CA, USA) GeneChip® miRNA 2.0 arrays, which contain the entire repertoire of currently known small ncRNA transcripts involved in gene expression (15,644 probe sets for 131 organisms, including 1105 human miRNAs, 1105 human pre-miRNAs and 2302 human small nucleolar RNAs) were used for genome-wide miRNA-expression analysis according to the supplier instruction manual, as described in the Affymetrix website [103] . Microarray analysis was performed on CEL files using Partek Genomics SuiteTM (Partek Inc., MO, USA) [104] . Data were normalized and summarized with the robust multiaverage method [15] . Batch effect removal was applied to the different samples, to remove individual variations, followed by one-way analysis of variance (ANOVA). Human miRNAs that were differentially expressed in the two phenotypic edges of the LCLs (fold difference >1.5; p < 0.05) were obtained ( Supplementary Table 1 ; see www.futuremedicine.com/doi/suppl/10.2217/ pgs.12.93). Clustering analysis was done by future science group
Tentative miRNAs SSRI response biomarkers
Research Article
Partek Genomics Suite software with Pearson’s dissimilarity correlation and average linkage methods.
Table 1. Mature sequences of selected miRNAs. miRNA
Mature miRNA sequence
Applied Biosystems assay ID
Real-time PCR Real-time PCR assays were performed for validating the differential microarray expression patterns of selected miRNAs using the same groups of LCLs according to their paroxetine sensitivity phenotypes. Real-time PCR assays were performed using TaqMan® MicroRNA Assays (Applied Biosystems, CA, USA) containing primer and probe sets (as detailed in Table 1) designed to detect and quantify mature miRNAs as described [9] . Following the extraction of total RNA from LCLs growing in optimal conditions and with no drug exposure, RNA was converted into cDNA with specific RT miRNA primers supplied with the TaqMan miRNA assay, using TaqMan® MicroRNA transcription kit (Applied Biosystems; catalog number: PN 4366596, containing 10× RT buffer, dNTP mix, RNase inhibitor and Multiscribe RT enzyme). The reverse transcription reaction was preformed in a thermal cycler using the following program: 16°C for 30 min, 42°C for 30 min; 85°C for 5 min; and 4°C thereafter. Realtime PCR experiments were conducted with mixtures (20 µl final volume) containing 1.5 µl of cDNA template, 10 µl 2× TaqMan® Universal PCR Master Mix (Applied Biosystems, catalog number: PN 4324018), and 1 µl of specific TaqMan MicroRNA Assay probe. U6 snRNA was employed as control [18–20] . Real-time PCR reactions were performed using a standard program of Applied Biosystems PRISM 7900 sequence detection system (Applied Biosystems). Comparative critical threshold (Ct) values, obtained by real-time PCR analysis, were used for relative quantification of miRNA expression. In this study, each ∆Ct value means Ct value of the target transcript relative to the control transcript:
miR-151-3p
CUAGACUGAAGCUCCUUGAGG
2254
miR-212
UAACAGUCUCCAGUCACGGCC
515
miR-132
UAACAGUCUACAGCCAUGGUCG
457
let-7b
UGAGGUAGUAGGUUGUGUGGUU
2619
let-7c
UGAGGUAGUAGGUUGUAUGGUU
379
miR-30b* (miR-30b-3p)
CUGGGAGGUGGAUGUUUACUUC
2129
Ct of target transcript - Ct of control transcript
The ∆∆Ct value represents ∆Ct value of lines with high paroxetine sensitivity compared with lines with low paroxetine sensitivity. DDCt = (Ct of target transcript - Ct of U6) highsensitivity (Ct of target transcript - Ct of U6) low sensitivity
Fold differences were obtained by using the formula 2-∆∆Ct. All PCR reactions were conducted in triplicate. future science group
Results Human LCLs from healthy adult female donors (n = 80) from the National Laboratory for the Genetics of Israeli Populations biobank (see Methods section [102] ) were phenotyped for their sensitivity to growth inhibition by the SSRI drug paroxetine during 3 day incubations, as described previously [9] . From the 80 cell lines screened, eight ‘edge’ cell lines, four from each extreme phenotype for paroxetine-mediated growth inhibition sensitivity, showing mean growth inhibition of 54 ± 2.2% and 24 ± 6.9% respectively (p = 1.7 × 10-4 ; Figure 1) were selected for comparative miRNA-expression analysis by microarray experiments. miRNA was prepared from the selected LCLs while growing under optimal conditions with no drug exposure, converted to cDNA, and subjected to whole-genome expression analysis with Affymetrix GeneChip miRNA 2.0 arrays (see Methods). The array experiments have identified 42 human miRNAs whose basal expression levels differed by >1.5-fold and p 1.5-fold differential expression and a p-value of
