MicroRNAs miR-17 and miR-20a Inhibit T Cell Activation Genes and Are Under-Expressed in MS Whole Blood

It is well established that Multiple Sclerosis (MS) is an immune mediated disease. Little is known about what drives the differential control of the immune system in MS patients compared to unaffected individuals. MicroRNAs (miRNAs) are small non-coding nucleic acids that are involved in the control of gene expression. Their potential role in T cell activation and neurodegenerative disease has recently been recognised and they are therefore excellent candidates for further studies in MS. We investigated the transcriptome of currently known miRNAs using miRNA microarray analysis in peripheral blood samples of 59 treatment naïve MS patients and 37 controls. Of these 59, 18 had a primary progressive, 17 a secondary progressive and 24 a relapsing remitting disease course. In all MS subtypes miR-17 and miR-20a were significantly under-expressed in MS, confirmed by RT-PCR. We demonstrate that these miRNAs modulate T cell activation genes in a knock-in and knock-down T cell model. The same T cell activation genes are also up-regulated in MS whole blood mRNA, suggesting these miRNAs or their analogues may provide useful targets for new therapeutic approaches.


Introduction
Multiple Sclerosis (MS) is a disease which affects mainly young people that often leaves them disabled in their most productive years.It is a relatively common disease with an incidence somewhere between 1-2 per 1000 and the rate appears to be increasing [1].It is assumed that a T cell dysregulation results in demyelination and axonal loss throughout the central nervous system [2].Most patients have a relapsing remitting course (RRMS), which is unpredictable and is observed as episodes of acute inflammation that results in neurological dysfunction, which in the majority of cases responds to immunomodulatory steroid treatment.Relapsing remitting disease is characterized by some level of myelin repair, whereas in the progressive form myelin repair seems to be insufficient or ineffectual resulting in progressive disability without any observable signs of recovery.
In recent years genome wide association studies have identified that not only is there an association with haplotype in the Human Leukocyte Antigen (HLA) region, but also in both the IL-2 and the IL-7 receptor genes, CD56, CD226 and CLEC16A, which together are considered to contribute to a predisposition to develop the disease [3][4][5].
In the largest single genome wide association study searching for genetic risk factors for MS, polymorphisms associated with the disease confirmed the importance of immune dysfunction [5].These identified polymorphisms alone or in combination do not explain the significant differences in immune function associated with this disease.
MicroRNAs (miRNAs) have recently been discovered to be regulatory modulators of gene expression [6].A striking feature of these mRNA regulators is their evolutionary conservation, indicating their level of importance in the control of gene expression [7].MiRNAs bind to the 39 UTR of target mRNA through base pairing, resulting in target mRNA cleavage or translation inhibition [8].On average each miRNA regulates about 200 genes, and the outcome of regulation is cell state and type specific [9].Their dysregulation has been associated with many diseases, and the potential for modulating their action by therapeutic intervention has excited much interest [10].This is particularly so in immune-related diseases.The miRNA transcriptome of immune cell subsets are distinct, suggesting that naı ¨ve, effector and central memory T cell [11] and regulatory T cell [12] function is dependent on the miRNA regulation.Gross changes to mouse miRNA regulation by deletion of the genes which mediate it, Dicer and Drosher, results in T cell abnormalities and autoimmune disease [13].Autoimmune diseases such as MS are ameliorated by drugs modulating T cell function like interferon beta and glatiramer acetate, and with monoclonal antibodies specific to T cell surface markers [14].The genes which are associated with this, and other autoimmune diseases, are predominately expressed in antigen presenting cells and T cells, and there is a marked T cell activation gene expression pattern in MS whole blood [15][16][17].One of the difficulties of studying multiple sclerosis is the acquisition of samples unaffected by the influence of immunomodulatory treatment.In this report we have investigated miRNA expression profiles of a series of effectively treatment naı ¨ve MS patients (i.e.all patients were free of disease modifying therapy for at least 3 months) compared to a healthy age-matched control group.The findings of this study demonstrate the differences between the immune function in MS without the influence of any treatment regiment.

Results
We measured the known miRNA transcriptome in whole blood from 59 MS patients and 37 healthy controls using the Illumina sentrix array matrix.Of these 59 patients, 18 had a primary progressive (PPMS), 17 a secondary progressive (SPMS) and 24 a relapsing remitting course.The patient demographics are shown in table S1.To ensure we were not measuring effects due to diurnal variation in immune function, all samples were collected between the hours of 9am-1pm for both the MS and healthy control populations.All of our patients were of Caucasian origin, and had not received any immunomodulatory therapy within the previous 3 months.The female to male ratio was 2:1 similar to the disease incidence in the general MS population.The age range for the entire group was 32 to 81 years with a mean age of 54 years, mean time since diagnosis was 20 years, ranging from 0 to 58 years, mean expanded disability status scale (EDSS) was 4.5 ranging from 0 to 6.There was no significant difference in the demographics between the group assessed by microarray and the one assessed by RT-PCR.As expected the disease duration and the EDSS was higher in the SP and PP group compared to the RR group.
MiRNA expression analysis revealed that out of 733 miRNAs assessed, 26 were down-regulated and 1 up-regulated in MS whole blood, based on a false discovery rate of less than 1% (Table 1).The fold change ranged between 1.36 and 21.59.The downregulation of miRNAs was across all the subtypes of MS.Out of the 27 we focused on miR-17 and miR-20a as the differential expression analysis revealed that these miRNAs were the most significantly different in the MS group compared to the controls.Both miRNAs are known to be involved in immune function and their importance was confirmed by RT-PCR performed between 3 and 6 times.Both miRNAs were differently expressed across all MS subtypes (Figure 1).Most significantly different was miR-17 with a p-value of 7.61 e-05 (see Table 2).
To better define the effects of miR-17 and miR-20a on gene expression we conducted knock-in (with synthetic miRNA) and knock-down (with LNA modified anitmiRs) experiments using Jurkat cells (performed in triplicate).Genes down-regulated by the miR-17 transformants, and up-regulated by the transformants of small interfering RNAs (siRNAs) corresponding to it were identified using Illumina HT12 microarrays and differential expression analysis (table S2).Similarly, genes potentially regulat-ed by miR-20a were identified (table S3).Putative targets for miRNAs can be identified from their sequences, and we identified these using target prediction metadata collated in the miRecords database (http://mirecords.umn.edu/miRecords).
Genes that were differentially expressed as a result of knock-in or knock-down were compared against the predicted targets of the respective miRNA using in silico analysis.The resulting lists of genes were considered more likely to be directly affected by the changes in miRNA.
Gene expression data was used to identify gene pathways overrepresented in the set of dysregulated genes in the transformed Jurkat cells using GeneGo maps module of Metacore (Genego, MI) (tables S4, S5).Translation (P,10 24 for miR-20a), cholesterol biosynthesis (P,10 25 for miR-20a, 10 24 for miR-17) and immune response pathways (P,10 24 for miR-17) were over-represented in the list of genes dysregulated by changes in miRNA expression.
We then tested if the genes regulated by the miRNAs miR-17 and 20a in Jurkat cells were also over-represented in MS whole blood.We previously reported, from a comparison of whole blood transcriptomes of 150 MS patients and controls, that T cell activation genes were markedly up-regulated in MS (P,10 214 ) [15].We compared the gene expression profiles from transformed Jurkat cells against non-transformed cells.Putative miR-17 or miR-20a target genes were then compared with gene expression profiles from the MS whole blood samples to identify genes likely to be affected by these miRNAs.
From just the mRNA and transformed Jurkat cell list, the pathways represented were very similar to the pathways overrepresented in mRNA alone (Figure 2).Genes which were on all three lists were the most likely to be regulated by the miRNA in MS.From all 3 lists, transcription (P,10 26 ) and immune response pathways (P,10 25 ) were over-represented for miR-20a; and the Th17 immune response pathway (P,10 23 ) for miR-17.

Discussion
Dysregulation of the immune system in MS is considered to be a fundamental aspect of disease initiation and development in this neurodegenerative disease.In this report we have investigated the miRNA transcriptome in a series of MS patients who were not on current immunomodulatory treatment.59 clinically definite MS patients (mean age of 54.2 years), enriched for PP subtype (n = 18), with 24 RR and 17 SP were included in the analysis compared to 37 controls (mean age 48.0 years).The mean disease duration even for the relapsing remitting group was over 15 years.The selection of treatment naı ¨ve patients suggests that they were not a group with high disease activity.It could be argued that the significance of the findings of miR-17 and miR-20a being down regulated is underestimated compared to a more active northern European MS population.From the miRNA transcriptome comprising 733 miRNA species only one was up-regulated whereas 26 miRNAs were down-regulated with a false discovery rate less than one.From this data we focused on two miRNAs, miR-17 and miR-20a, as these were all under-expressed in all subtypes of MS and are known to be involved in the control of immune function [18].
Although miR-17 and miR-20a are encoded in the same cistron (miR-17-92), other miRNAs from this cistron were not underexpressed, and these two miRNAs were not identically expressed.There appears to be post-transcriptional regulation of the cistron miRNAs, allowing them to regulate different processes [19].The translation pathway of lymphocytes is orchestrated by miR-17 and miR-20a [20].Elevated expression of miR-17 and related miRNAs has been found to be associated with lymphoma and other cancers, and is probably important in maintaining an undifferentiated phenotype and resisting apoptosis [21,22].A reduction of miR-17-92 miRNAs observed in the present study, has also been associated with differentiation [23] and could potentially be involved in MS-associated T cell activation.We have previously demonstrated that mRNAs from these pathways were up-regulated in all clinical subtypes of MS [15].In order to test whether these down-regulated miRNAs, miR-17 and 20a inhibit the expression of these T cell mRNAs we used knock in (with the miRNAs) and knock-down (with the antisense or siRNAs) assays of Jurkat cells, a model for T cells.
Jurkat cells are derived from a T cell lymphoma patient, and have traits of a partially differentiated T cell line [24,25].The transformed Jurkat cells were not stimulated, whereas the T cells in whole blood represent a myriad of subsets in various states of stimulation.However, as these miRNAs under-expressed in MS were known to increase transcription of genes involved with translation and immune response pathways, we expected they may also contribute to the pattern of gene dysregulation seen in MS whole blood mRNA.The set of genes down-regulated by miR-17 and miR-20a in transformed Jurkat cells overlap with the genes and the pathways they are predicted to control, suggesting these miRNAs contribute to the gene dysregulation observed in whole blood.To our knowledge there have been only five other publications investigating the role of miRNA in MS [26][27][28][29][30], four of which focus on the immune system in MS and the other on active and inactive central nervous system lesions.Otagui et al. [29] focused more on up-regulation during relapse.As they only examined small number of patients, their qPCR did not reach statistical significance.Keller et al. [28] examined 20 RRMS patients on treatment and identified a number of miRNAs that were up or down regulated.The report by Du et al. [26] identified miR-326 as a major determinant of disease in Chinese MS but not neuromyelitis optica (an inflammatory demyelinating disorder thought to be mediated by antibodies targeting aquaporin 4 [31]).In our study we did not identify any statistically significant change in miR-326 between MS patients and controls.Furthermore, subgroup analysis confirmed this finding.Since we do not know why this miRNA was chosen for analysis by Du et al. [26] we are unable to put forward any functional explanation to describe our discrepant results apart from differences observed in MS patients from Asian or Caucasian origin.
The results of this investigation indicate that miR-17 and miR-20a are implicated in the development of MS.Of special interest is miR-17 which we found to be down regulated in peripheral blood.Interestingly, Lindberg et al. [30] also identified miR-17 as being associated with MS in CD4 + cells, but the relationship was in the opposite direction to what we observed.This is likely to be a result of a type 1 error given that only 8 patients and 10 controls were examined by Lindberg et al. in their initial miRNA expression analysis and only 15 patients and 10 controls in their confirmatory study.This in combination with our study design that included the three major subgroups of MS whereas the patients participating in the Lindberg report were diagnosed with relapsing remitting disease only could potentially explain the differences observed in miR-17 expression.In our analysis we have examined 59 cases against 37 controls thereby providing a more robust statistical analysis.Another possibility for the discrepancy is that we examined expression in whole blood collected in PAXgene RNA tubes, which stabilizes RNA on collection, and includes a high proportion of neutrophil RNA.One possibility is that miR-17 is very labile, and if it degraded in the 90 minutes or so longer it took Lindberg et al [30] to purify their miRNA -it could have degraded preferentially in the healthy control cells because of the different regulatory environment.Also, it may be that miR-17 is downregulated in MS neutrophils cf controls, so that the net miR-17 we measured was less, even if it was more abundant in MS CD3 cells.
Even if the miRNAs under-expressed in MS were not directly contributing to the immune cell signature observed in MS whole blood, the excessive T cell activation signature seen in MS [15][16][17]; and other autoimmune diseases [32] suggest agents which can  S4 and S5).Most significant pathways are represented in the form of pie-charts where each slice represents 2log 10 of the p-value of that pathway as a proportion of the sum of the 2log10 of the p-values of the over-represented pathways (P,0.05 for miR-17, p,0.001 for miR-20a and mRNA (MS vs. HC)).The pvalues were determined by Metacore Pathway analysis based on the chi squared value for the expected compared to observed number of genes identified from that pathway in the list of dysregulated genes.(DE differential expression).doi:10.1371/journal.pone.0012132.g002algorithms.The predicted target genes found to be modulated in response to specific changes miRNA expression in electroporated cells were then examined to determine their potential for biological implications in the context of MS.

Figure 1 .
Figure 1.Relative expression of target miRNA.Relative expression of target miRNA compared to endogenous control miRNA in MS subtypes and healthy controls, validation by Q-RTPCR.The two miRNAs were significantly under-expressed in MS patients compared to healthy controls (*P,0.05,**P,0.01),error bars 6 SEM.doi:10.1371/journal.pone.0012132.g001

PP 3 .
20E-02 4.68E-02 P-values of the relative expression as determined by Q-RTPCR of miR-17 and miR-20a in MS patients as a whole and their respective subgroups compared against a healthy aged-matched population.doi:10.1371/journal.pone.0012132.t002

Figure 2 .
Figure 2. Gene networks dysregulated in MS are regulated by miR-17 and miR-20a miRNAs.The pathways enriched with genes both dysregulated in MS-whole blood and modulated in miR-17/20a transformed Jurkat cells were identified using Metacore (TablesS4 and S5).Most significant pathways are represented in the form of pie-charts where each slice represents 2log 10 of the p-value of that pathway as a proportion of the sum of the 2log10 of the p-values of the over-represented pathways (P,0.05 for miR-17, p,0.001 for miR-20a and mRNA (MS vs. HC)).The pvalues were determined by Metacore Pathway analysis based on the chi squared value for the expected compared to observed number of genes identified from that pathway in the list of dysregulated genes.(DE differential expression).doi:10.1371/journal.pone.0012132.g002

Table 1 .
miRNAs dysregulated in MS whole blood.

Table 2 .
P values of relative expression of target miRNAs.