2.2. Overexpression of miRNAs in NPCs of IID Patients

A total of 766 miRNAs were analyzed in all patients and controls. Only 245 miRNAs were expressed in both IID patients and controls (Figure 2a). The HTqPCR, gplots, and limma bioconductor statistical analyses showed that just 20 miRNAs were significantly different (p ≤ 0.05) in IID patients compared to controls (Figure 2b), which were miR-25, miR-15b, let-7c, miR-32, miR-9, miR-30c, miR-27a, miR-19b, miR-93, miR-27b, miR-17, miR-138-1*, miR-324-5p, let-7d, miR-597, miR-28-3p, miR-24, miR-29a, miR-31 and miR-135a*. Of these, miR-25, miR-15b, let-7c, miR-32, miR-9, miR-30c, and miR-27a showed levels of expression 50 times more in IID patients than controls (Figure 2b). miR-25 presented the highest expression; therefore, it was used to analyze its possible targets and effects on IID.

Open in a separate window

Figure 2

Expression analysis of the neurosphere miRNAs of IID patients and controls. (a) miRNA expression graph obtained for 754 miRNAs that were analyzed in each sample using TaqMan^® Array MicroRNA Cards (cards A and B). (b). List of the 20 miRNAs that were statistically significantly (p ≤ 0.05) overexpressed in IID patients compared to controls.

2.3. mRNA Expression Profile and Selection of Targets for miR-25

Of the 28,869 mRNAs that were analyzed, 21,190 showed changes in their expression when IID patients and controls were compared (Figure 3). In IID patients, 9829 were overexpressed (1.19 to 1.99 times higher) and 11,361 were under-expressed (−1.0 to −3.0 times lower) (Figure 3a). The statistical analysis of the mRNAs showed that only 85 mRNAs that were under-expressed and 56 that were overexpressed were significantly different (p ≤ 0.05; Figure 3b).

Open in a separate window

Figure 3

mRNAs expression profile from the neurospheres of four IID patients and four controls, representative of all samples analyzed. (a) The image shows the unsupervised hierarchical data pooling of 28,869 mRNAs for each sample. Colors represent expression levels of each mRNA; the red regions are the highly expressed genes (9829), the green regions are the weakly-expressed or nonexistent ones (11,361), and the black regions represent the genes without change (7679). (b) The volcano plot shows the expression of mRNAs from the 25 IID patients and 25 controls. In the scatter diagram, cut-off points were made in the values of fold-change (log2) at −0.26 and 0.26 to identify those genes that have changes in their expression that are valid by qRT-PCR and statistically significant (p ≤ 0.05 or 1 in −Log10). The most significant mRNAs are found at the top and appear symmetrically placed about the vertical line (Fch = 0). The 56 overexpressed genes are found in the upper-right corner and the 85 under-expressed genes appear in the upper-left corner. The yellow boxes in the heat map and the blue circle in the volcano plot show the location of some of the under-expressed mRNAs (including ROBO2) mentioned in this paper.

In this work, we exclusively analyzed the under-expressed mRNAs found in all patients compared to the controls that have reported implications in neurodevelopment.

In silico analysis with the DAVID database [20] provided the functional annotation, classification, and grouping of each of the 85 genes encoding mRNAs that were under-expressed. Of these 85 genes, those that have been reported to have an implication in neurodevelopment are ROBO2, PTN, PAX7, TIMP1, ARHGEF5, MMP1, CDH13, SERPINB5, GPM6B, SNAI2, CTTNBP2, CCDC23, ERBB4, TP63, VSNL1, DSG3, SLC4A4, SEMA3E (Table 1).

The prediction of the target mRNAs of miR-25 was carried out by aligning the ROBO2, PTN, PAX7, TIMP1, ARHGEF5, MMP1, CDH13, SERPINB5, GPM6B, SNAI2, CTTNBP2, CCDC23, ERBB4, TP63, VSNL1, DSG3, SLC4A4, SEMA3E with the miR-25 sequences using the TargetScan [21], miRanda [22] and DIANA-microT 2023 [23] databases. Considering a score of mirSVR of −0.1 to −1.0, a value of 0.5–0.6 for the conserved sites (PhastCons), and the seed region of six to eight consecutive nucleotides, the best target mRNAs for miR-25 were ROBO2, which had the greatest statistical significance and the greatest number of alignments (in four regions: two with the seed sequence and another two regions with 12 and 15 nucleotides that included the complete seed); ERBB4, which aligned only once with 13 nucleotides; SLC4A4, which aligned in two positions with 6 and 11 nucleotides; SNAI2 and TIMP1 genes which aligned at two positions with 10 and 16 nucleotides, respectively; and CDH13, CTTNBP2, DSG3, TP63, and SEMA3E mRNAs, which only aligned with the seed region (six nucleotides) of miR-25. PTN, PAX7, ARHGEF5, MMP1, SERPINB5, GPM6B, CCDC23, and VSNL1 did not present alignments to the seed region of miR-25. Due to this data, it was determined that the main target of miR-25 was the mRNA of ROBO2 (Table 2), which was therefore studied to show that it is negatively regulated by miR-25.

2.4. KEGG and GO Functional Annotation and Enrichment Analysis

The Kyoto Encyclopedia of Genes and Genomes (KEGG) [24] and Gene Ontology (GO) [25,26] analyses of biological processes showed that these 13 gene targets for miR-25 are involved to an extent in axon guidance, adherent junctions, focal adhesion, development of the central nervous system, and synaptic transmission (Table 1 and Figure 4). The KEGG analysis showed that other cellular pathways are involved to a lesser extent, including those of adhesion molecules, focal adhesion, and neurotrophic signaling pathways, the regulation of actin in the cytoskeleton, and some associated diseases, such as Alzheimer’s, Huntington’s, and Prion disease; GO analysis showed biological processes such as neural differentiation, neural migration, cytoskeleton organization, cell–cell junction organization, etc. All of these are related and associated with each other. It should be noted that ROBO2 plays a part in the axon guidance pathway, cell adhesion, and central nervous system development (Figure 4).

Open in a separate window

Figure 4

Analysis of pathways (KEGG) and biological processes (GO) of selected under-expressed genes. It was observed that axon guidance was the pathway and biological process in which most of the selected genes participate and, therefore, they all are associated with each other.

2.5. Validation of Expression Obtained in the Microarray of the ROBO2 mRNA by qRT-PCR

The ROBO2 expression obtained from the microarray was validated by qRT-PCR in eight IID patients and eight controls, randomly selected. The amplification of mRNAs of IID patients and controls showed that the expression of ROBO2 (green curves, Figure 5a) is much lower than in the controls (yellow curves, Figure 5a). When calculating the value of the expression of the patients compared to the controls, it was found that the expression in the IID patients is at least six times lower and the difference in the expression of ROBO2 was significant with p ≤ 0.05 (Figure 5b). The relative expression of the controls was taken from the baseline, so it was assigned a value of 1 (Figure 5b). This analysis corroborated the down-regulation of ROBO2 in IID patients observed in the microarray.

Open in a separate window

Figure 5

Validation of ROBO2 expression in 25 IID patients and 25 controls by qPCR. (a) Graph of expression of ROBO2 in patients and controls; green curves show the expression in patients and in yellow, the controls. (b) Relative expression of ROBO2; the expression value in controls was taken as the basal value, so the value of 1 was assigned to it. * p ≤ 0.05.

4.2. Sampling and Human Progenitor Neural Cell (NPC) Culture Characterization

The samples and NPC cultures (neurospheres) were obtained and characterized according to the methods reported by Jimenez-Vaca et al. [18]. The following primary antibodies were used for the characterization of the NPCs: anti-Nestin, anti-Sox2, anti-βIII-tubulin, and 4′,6′-diamidino-2-phenylindole dihydrochloride (DAPI) to stain the nuclei. All the antibodies used were manufactured by Zymed Invitrogen Corporation (Camarillo, CA, USA).

4.3. MicroRNA Expression Profile

MicroRNA was extracted from each patient and control sample neurospheres using the mirVana™ Kit (miRNA Isolation Kit, Invotrogen, Camarillo, CA, USA) according to the manufacturer’s instructions. The total miRNA extracted was quantified in a NanoDrop (Thermo Fisher Scientific, Waltham, MA, USA). To determine the expression profile of the miRNAs from patients and controls, a TaqMan^® Array Human MicroRNA Cards (cards A and B) specific to human miRNAs (Applied Biosystems, Waltham, MA, USA) was used. This assay allows accurate quantitation of 756 human microRNAs. Four endogenous controls (MammU6) on each card aid in data normalization. Also, Megaplex™ RT Primers, Human v2.0 (Applied Biosystems, Waltham, MA, USA), primers were used as necessary for the reaction. The procedure is based on an RT-PCR reaction, cDNA preamplification with TaqMan^® enzyme, and a dilution according to manufacturer.

4.4. mRNA Expression Profile

Total mRNA was obtained from 2.0 × 10⁶ neurospheres from each patient and each control using the Trizol technique. First, 1 mL of trizol was added to the cells, they were then lysed and incubated for 5 min at room temperature (RT). Next, 40 μL of chloroform was added, mixed vigorously for 15 s, and incubated for 5 min at RT. It was centrifuged at 12,000× rpm at 4 °C for 10 min. The aqueous phase was transferred to a new tube and 50 μL of isopropyl alcohol was added, mixed, and incubated at 4 °C for 10 min. It was centrifuged at 7500× rpm for 5 min and the pellet obtained was washed twice with 500 μL of 75% ethanol. Finally, the mRNA was resuspended in 30 μL of RNase-free water. The expression analysis was performed using the GeneChip^® Human Gene 1.0 ST Array (Affymetrix, Santa Clara, CA, USA). This involves an RT-PCR–based method and biotin labeling to obtain cRNA. The cRNA is fragmented for subsequent hybridization microarray control. An aliquot was taken to determine the size of the fragments, which should be 35 to 200 bases. For hybridization, labeled cRNA fragments are mixed with B2 oligonucleotide control (3 nM), 20× eukaryotic hybridization controls, 2× hybridization buffer, and DMSO, in a 300 μL final volume. The hybridization cocktail was heated to 99 °C for 5 min. Hybridization was maintained for 16 h and subsequently washed, stained, and scanned through the Fluidics Station 400 or 450/250 workstation. Data analysis was performed with DataAssist v1.0 software (the intensities for each probe in GeneChip Expression Analysis Fundamentals data).

4.5. Selection of Differentially-Expressed Genes with Function in the CNS and Neurodevelopment

The search for any relationship between genes with differential expression and some function in the CNS was carried out using the DAVID database [20], which allows annotation, classification, and functional grouping of each gene, in addition to determining the signaling and regulation pathways in which they are involved.

4.6. In Silico Prediction of Target Genes for miR-25

To identify which genes are targets for miR-25, an in silico alignment of each mRNA sequence with significant under-expression against the mature miR-25 was conducted. For the selection of a possible target of miR-25, the TargetScan [21], miRanda [22], and DIANA-microT 2023 [23] databases were used. The mRNAs that presented the best alignments in their 3’ UTR region with miR-25 were selected.

4.7. Functional Annotation (KEGG) and GO Enrichment Analysis

The KEGG (Kyoto Encyclopedia of Genes and Genomes) [24] and the Gene Ontology (GO) [25,26] databases were used to analyze the under-expressed genes previously selected as targets of miR-25. These analyses allowed us to determine the pathways of molecular interactions as well as the processes, functions, and locations in which these genes participate within the cell.

4.8. Validation of ROBO2 Expression by qRT-PCR

To corroborate the expression data of the microarray of the selected target gene (ROBO2), which is the target best aligned to miR-25, validation was carried out by qRT-PCR with TaqMan^® probes specific for the mRNA of ROBO2 in samples from 8 patients and 8 controls previously used in the microarray.

The cDNA synthesis was performed in two steps. In the first step, 2 ug of sample, 250 ng of random primers, 1 mL of 10 mM dNTPs mix, and RNase-free water were incubated at 65 °C for 5 min to obtain a total reaction volume of 12 mL after adding all the reagents. Subsequently, the solution was cooled on ice. In the second step, the previous solution was incubated at 37 °C for 2 min after adding 4 mL of 5× buffer, 2 mL of 0.1 M DTT, and 1 mL of RNaseOUT™ (Invitrogen, CA, USA) Recombinant Ribonuclease Inhibitor (40 units/mL). A total of 1μL of the enzyme M-MLV RT (200 units/mL Invitrogen, CA, USA) was added, and the solution was incubated at 25 °C for 10 min. It was then incubated at 37 °C for 50 min. Reverse transcription samples were stored at −20 °C until use. For real-time qPCR, the StepOne™ Real-Time System model from Applied Biosystems was used with the corresponding software and the expression assays from the same company, which are previously validated, optimized and exist for practically all the human genes. Reactions were carried out according to the manufacturer’s protocol. Each reaction (20 mL) contained 10 mL TaqMan Universal PCR Master Mix NoAmpErase^® UNG (Applied Biosystems, CA, USA) (2×), 1 mL TaqMan Gene Expression Assay Mix (20×), 8 mL RNase-free water, and 1 mL of sample ten-fold–diluted reverse transcription product. For each assay, a negative control was performed with RNase-free water, and the RPB1 subunit of RNA polymerase II (POLR2A) was used as an endogenous control. The quantification of the relative expression of the genes was carried out by the method of the second derivative 2^−ΔΔCt. The TaqMan^® Gene Expression Assay IDs for ROBO2 and internal controls were ROBO2 (Hs00326067_m1) and POLR2A (Hs00172187_m1).

We thank Ana Luisa Jimenez-Vaca for her support in carrying out this work.