MicroRNAs (miRNAs) are small, single-stranded, non coding RNA molecules that regulate gene expression of their complementary messenger RNA targets [1,2]. Taking into account, the significance of these regulatory molecules in gene regulation, discovery of dysregulated miRNAs could benefit to clarify the molecular differences of miRNAs serum profiles between the autism spectrum disorder (ASD) subjects and typically developing children (TDC). The prevalence of ASD has increased considerably (from 2-5/10,000 to 1:68 children) during the last two decades [3,4]. Increased attention and changes made to the diagnostic criteria by the medical community have necessarily contributed to this tendency [5], as well as several environmental factors, acting particularly over early postnatal neurodevelopment and prenatal development [6]. In addition, advanced parental age at time of conception has also been shown to pose a risk of ASD [7]. Despite major advances in the understanding of ASD pathogenesis, inter individual heterogeneity and the levels of complexity has largely complicated the delivery of extensive scientific learning into efficient clinical practices. Currently, the golden standard to diagnose ASD is built on information from several sources: family caregivers, teachers, standardized behavioral scales and clinical observations. Also, pervasive development of the patients is only monitored by clinical scales with uncertain reliability, especially in early age. Specific and sensitive quantitative biomarkers, evaluated through experimental methods, electrophysiological techniques and brain imaging, would greatly support clinicians in providing more timely referrals to behavioral intervention programs’ earlier diagnoses and evidence-based prognosis [8]. Quantitative polymerase chain reaction (qPCR) technologies offer a accurate tools to search for specific miRNAs profiles applicable as potential biomarkers for neurodevelopmental disorders. The choice of miR-328-3p and miR-3135a as potential biomarkers in ASD is based mainly on our preliminary data showing differential expression changes of these miRNAs from small RNA sequencing. In addition, there is no scientific data on the role of these miRNAs in ASD, although the importance of many miRNA molecules has been shown in the etiopathogenesis of several neurological disorders. The current study was performed in order to shed light on the probable participation of miRNAs in ASD. The main aim of the current research was to analyze differential expression of miR-328-3p and miR-3135a in serum of ASD subjects and their relevance as potential blood-based biomarkers.
The control group representing TDC was assigned with the aim to match by sex and age to the ASD group. Inspection of all children in the healthy group for absence of autistic features was done by clinical examination and CARS. Children with known infectious, oncological, metabolic or genetic conditions were excluded from the study. No children were receiving any drug therapy when they were recruited.
MicroRNA-specific quantitative reverse transcription polymerase chain reaction primer sets.
MiRNAs | Primer Sequences (5’>3’) |
---|---|
miR-197-5p SL | CTC AAC TGG TGT CGT GGA GTC GGC AAT TCA GTT GAG CCT CCC AC |
miR-197-5p F | ACA CTC CAG CTG GGC GGG TAG AGA GGG CAG T |
miR-500a-5p SL | CTC AAC TGG TGT CGT GGA GTC GGC AAT TCA GTT GAG TCT ACC CC |
miR-500a-5p F | ACA CTC CAG CTG GGT AAT CCT TGC TAC CTG G |
miR-424-5p SL | CTC AAC TGG TGT CGT GGA GTC GGC AAT TCA GTT GAG TTC AAA AC |
miR-424-5p F | ACA CTC CAG CTG GGC AGC AGC AAT TCA TGT |
Cel-miR-39 Spike-in | UCA CCG GG GUA AAU UA |
Cel-miR-39 SL | CTC AAC TGG TGT CGT GGA GTC GGC AAT TCA GTT GAG CAA GCT GA |
Cel-mi-39 F | ACA CTC CAG CTG GGT CAC CGG GTG TAA ATC |
Universal R | GTC GGC AAT TCA GTT GAG |
MiRNAs: microRNAs; SL: stem-loop. F: forward; Spike-in: exogenous RNA control; R: reverse.
The qPCR data showing a DNA melt profile result for amplification of the specific single product in qRT-PCR analysis. Specific single products corresponding to exogeneous spike-in control cel-miR-39 (panel A); miR-328-3p (panel B) and miR-3135a in ASD and healthy control patients, were confirmed by monitoring the dissociation curve (melting curve analysis). The melting temperatures of miR-3135a amplicons were 76 ± 1 °C (panel B), whereas spiked-in cel-miR-39 control had a melting temperature of 77 ± 1 °C (panel A), and melting temperatures of miR-328-3p amplicons were 79 ± 1 °C (panel C), respectively.
Differential expression of serum miRNAs in ASD patients. Quantitative RT-PCR analysis of miR-3135a and miR-328-3p levels. The circulating serum miRNAs signatures were identified by miRNA-specific stem-loop qRT-PCR analysis in the ASD and control groups. Expression levels of the analyzed miRNAs were normalized to spiked-in cel-miR-39 control and expressed in relation to controls.
Receiver operating characteristic curve analysis using differentially expressed serum miRNAs. The ROC for miR-3135a, and miR-328-3p signature in patients with ASD was performed to evaluate the prediction accuracy of selected biomarkers. The dotted diagonal line represents random classification accuracy (AUC 0.5). The ROC curves were drawn for miR-3135a, and miR-328-3p, which yielded 0.828 and 0.858 as AUC values, respectively. Combined ROC curve describe the logistic regression (LOGREGR) of the differentially expressed miRNA members (miR-3135a and miR-328-3p). Diagnostic sensitivity of combined classifiers was 78.9% with the corresponding specificity of 88.9%. The combination of the miRNAs showed a correspondence to that using only miR-328-3p as a biomarker.
Fold change difference of two down-regulated serum miRNAs between the ASD and TDC groups. Data are expressed as fold change of mean 2–ΔΔCt for each miRNA after being normalized with spike-in cel-miR-39 control.
The obtained data was subsequently used to assess the diagnostic specificity and sensitivity of analyzed serum miRNAs. Diagnostic sensitivities of miR-3135a and miR-328-3p for ASD were 76.3 and 78.9%, respectively. The corresponding specificities were 88.9 and 88.9%, and AUC 95% confidence interval (95% CI) were 0.828 (0.715-0.911%), and 0.858 (0.749-0.932%), respectively. In addition, a combined ROC analysis was done. The combined ROC curve analysis showed a better diagnostic value than individual miRNAs in ASD. Area under the ROC AUC (95% CI) 0.858. Confidence interval (0.749-0.932). Together, these results indicate that the identified serum miRNAs, alone or in combination, can discriminate between ASD cases and TDC with high accuracy.
The KEGG pathway enrichment analysis for the targets of the identified serum miR-3135a.
The KEGG pathway enrichment analysis for the targets of the identified serum miR-328-3p.
There are many reports in the literature for the potential use of miRNAs as biomarkers for neural disorders, such as Parkinson’s and Alzheimer’s disease [10,11]. However, only few studies have been investigating miRNA in serum samples and their importance as biomarkers is still not fully understood. Moreover, the expression profiles of miRNAs in ASD have been examined, including studies from lymphoblastoid cell cultures [12, 13, 14].
Exploring the miRNA expression patterns as potential serum-based biomarkers for ASD diagnosis is still in its infancy. The question of how exactly miRNAs regulate their target genes through fine molecular mechanisms in the context of ASD pathogenesis are not fully understood. It is currently known that individual miRNAs can have several target genes, and thus, have an impact on more than one pathway. It was found that miRNAs can regulate translation of a wide range of proteins in neurons [15] including proteins involved in neuronal migration [16], channels [17] and neuronal morphology [18]. A functioning miRNA system is obligatory in astrocytes, with loss of miRNA biogenesis that could lead to seizures and neuro-degeneration [19]. In order to examine the potential role of the two differentially expressed miRNAs, we obtained a list of their validated target genes and constructed a custom script that used the KEGG database for pathways in which the validated target genes participate. Some studies revealed that miR-3135a and miR-328-3p were involved in cancer [20,21]. Moreover, miRNA-328 dysregulation has also been associated with several complex neurological conditions, such as аmyotrophic lateral sclerosis (ALS) [22], Alzheimer’s [23] and prion diseases [24]. At present, there is no scientific data on the role of miR-3135a and miR-328-3p in ASD. Further research is necessary to explore how miR-3135a and miR-328-3p function in ASD.
The results of the current study present evidence that addition of circulating biomarker investigations has the potential to improve the specificity of screening and lower the age of diagnosis. Thus, we suggest that an ideal biomarker should be:
The most remarkable findings of our study were that serum miR-3135a and miR-328-3p could discriminate ASD patients from healthy controls. However, the specific pattern of these miRNAs and their appearance in the medical tests as biomarkers requires subsequent confirmation.
The discovery of new miRNAs biomarkers for ASD requires the integration of experiments from different fields including:
Some promising candidates found so far are miR-132, miR-7 and miR-195. Deregulation of miR-132 was reported in lymphoblastoid cell lines (LCLs) and postmortem cerebellar cortices from ASD patients [26, 27, 28]. The biological functions of miR-132 and its targets have been validated by many animal studies [13,29,30]. MiR-7 has been shown to be up-regulated in the saliva and postmortem anterior prefrontal cortices. This miRNA is located in the ASD-associated copy number variation (CNV) locus [31, 32, 33] and its functional relevance has been validated using animal studies [34,35]. Deregulation of miR-195 is found in LCLs and serum samples, and it is also disrupted by an ASD-associated CNV locus [25,13].
The role of miRNAs in psychiatric disorders and ASD will be further elaborated using continuously improved relevant approaches. In addition, meta-analysis of miRNAs, covering genetic variation, expression and biological function will provide valuable information for the potential role of miRNA in ASD, and this could help the diagnosis and prognosis of ASD and psychiatric disorders Moreover, miRNA biomarkers could be very useful in distinguishing of different subtypes of psychiatric disorder. Finally, our results contribute to the new course of miRNA research in ASD biology but it is only a small part of the long validation process of miRNA dysregulation in ASD patients.