Expression of Long Noncoding RNAs in Fibroblasts from Mucopolysaccharidosis Patients

In this report, changes in the levels of various long non-coding RNAs (lncRNAs) were demonstrated for the first time in fibroblasts derived from patients suffering from 11 types/subtypes of mucopolysaccharidosis (MPS). Some kinds of lncRNA (SNHG5, LINC01705, LINC00856, CYTOR, MEG3, and GAS5) were present at especially elevated levels (an over six-fold change relative to the control cells) in several types of MPS. Some potential target genes for these lncRNAs were identified, and correlations between changed levels of specific lncRNAs and modulations in the abundance of mRNA transcripts of these genes (HNRNPC, FXR1, TP53, TARDBP, and MATR3) were found. Interestingly, the affected genes code for proteins involved in various regulatory processes, especially gene expression control through interactions with DNA or RNA regions. In conclusion, the results presented in this report suggest that changes in the levels of lncRNAs can considerably influence the pathomechanism of MPS through the dysregulation of the expression of certain genes, especially those involved in the control of the activities of other genes.


Introduction
Lysosomal storage disease (LSD) constitutes a group of disorders caused by a lack of the activity of specific lysosomal enzymes (such as acid hydrolases), membrane proteins, and deficiencies in lysosomal transporters or signals required for delivery of an enzyme to the lysosome. This leads to a deficiency in the lysosomal degradation of various macromolecules, and, in turn, to their accumulation in lysosomes [1,2].
Mucopolysaccharidoses (MPS) are a group of LSDs caused by a significant deficiency in the activity of lysosomal enzymes involved in the degradation of glycosaminoglycans (GAGs) [3]. GAGs are long, non-branched polysaccharides that are degraded due to their cleavage by endohexosidases or exoglucuronidases, followed by the removal of their shorter fragments by the subsequent actions of other enzymes [3]. The actions of all GAG-degrading enzymes are correlated and sequential, i.e., the next enzyme can only act after the action of the previous one. Thus, if only one enzyme is deficient, the whole degradation pathway is highly impaired.
There are 13 types/subtypes of MPS occurring in humans (plus one that, to date, has only been identified in an animal model), which are classified according to a lack or deficiency of a specific enzyme and the kind of stored GAG(s) [4,5]. These diseases are inherited in an autosomal recessive manner, except MPS II, which is a sex-linked disease. MPS is a rare disease, but its specific prevalence depends on the particular type/subtype. For example, MPS I occurs at a frequency of 1 per 88,000 live births, while there are only 4 cases of MPS IX described in the literature [6]. There are some MPS symptoms occurring in all or most types, such as coarse facial features, organomegaly, and changes in the cardiovascular system and bones. However, there are also symptoms specific for particular types, such as mental deterioration in some clinical forms of MPS I and II and in all patients with MPS III [4,5,7]. hypothesis that ncRNAs may be potential targets for novel therapeutics to be used against inherited diseases [49].
There is only very limited knowledge regarding the role of ncRNAs in MPS. The only studies conducted in this field and published recently concerned a mouse model of MPS I in which only one miRNA was investigated. It was demonstrated that the modulation of the level of miR-17 is responsible for the impaired expression of the Neu1 gene [50]. This gene codes for neuraminidase 1, which is involved in the degradation of gangliosides [51], which are compounds that act as secondary storage material in MPS [6].
It is worth mentioning that studies on ncRNA in the pathogenesis of LSD-published to date-have focused solely on miRNA molecules. The role of lcnRNAs remains unknown, and it is definitely worth investigation as the influence of such RNAs on pathomechanisms has been demonstrated for various neurodegenerative diseases (see above). Thus, the aim of this work was to assess the potential role of lncRNAs in MPS. In this first (to our knowledge) study on lncRNAs in MPS, we employed a transcriptomic approach that should facilitate the demonstration of a global picture of the changes in the levels of these RNA species as well as the estimation of their potential roles in the pathomechanisms of these complex diseases. Moreover, it was assumed that studies on lncRNAs in MPS might help to explain the incomplete efficiency of therapies based solely on reducing GAG levels, and perhaps indicate new methods for the development of combined therapies that could involve the modulation of the levels of lncRNAs and/or miRNA, thus improving the control of the expression of specific genes, which is crucial for the effective abolition of all cellular defects occurring in MPS.

Lines of Human Fibroblasts and Cell Cultures
Patient-derived fibroblasts were used in this study. The cell lines were purchased from the NIGMS Human Genetic Cell Repository at the Coriell Institute for Medical Research (this Institute incorporates all required forms of bio-ethical approval). MPS cell lines were characterized by the presence of specific mutations or dysfunction of the corresponding enzyme (when mutations were not determined). Their features are presented in Table 1. HDFa cell line (healthy fibroblasts) was used as a control. Cells were cultured in the DMEM medium, which was supplemented with 10% fetal bovine serum. Standard mixture of antibiotics was added to this medium. Cultures of fibroblasts were incubated at 37 • C, with 5% CO 2 saturation, and at 95% humidity.

Transcriptomic Analyses
Transcriptomic analyses were performed according to the previously described procedure (for statistical analyses, results from 4 independent biological experiments were used) [52]. Briefly, 5 × 10 5 cells were withdrawn from cell cultures (passages between 4 and 15) and homogenized using the QIAshredder columns in the presence of guani-dine isothiocyanate and β-mercaptoethanol. RNA was extracted with the RNeasy Mini kit (Qiagen, Hilden, Germany) and treated with Turbo DNase (Life Technologies, Life Technologies, Carlsbad, CA, USA). RNA quality was tested using RNA Nano Chips in the Agilent 2100 Bioanalyzer System (Agilent Technologies, Santa Clara, CA, USA). cDNA libraries were constructed based on mRNA libraries using the Illumina TruSeq Stranded mRNA Library Prep Kit (Illumina, San Diego, CA, USA). Sequencing was performed by employing HiSeq4000 (Illumina, San Diego, CA, USA). Raw reads (4 × 10 7 from every single experiment) were analyzed, yielding over 12 Gb of raw data per sample. In order to map the raw readings, the GRCh38 human reference genome, taken from the Ensembl database (https://www.ensembl.org/, last accessed on 1 December 2022), was employed. For annotation and classification of transcripts, the BioMart interface was used. Raw data of the RNA-seq analysis have been deposited in the NCBI Sequence Read Archive (SRA) (accession no. PRJNA562649).

Statistical Analyses
In the transcriptomic analyses, statistical significance was calculated based on results obtained from 4 independent biological experiments (n = 4). For the primary assessment of values with normal continuous distribution, one-way ANOVA was used with log 2 (1 + x). To calculate the false discovery rate (FDR), the Benjamini-Hochberg method was used. To compare two groups, post hoc Student's t-test was used with Bonferroni correction. These calculations were performed using the R software v3.4.3 (https://cran.r-project.org/bin/ windows/base/old/3.4.3/, last accessed on 1 December 2022 last accessed on 5 December 2022). The differences were considered significant when p < 0.1, which is used as a standard in transcriptomic analyses [15][16][17][18][19][20][21][22][23].

Results
To investigate changes in the levels of lncRNAs in the MPS cells relative to the control fibroblasts, we conducted transcriptomic analyses using cell lines derived from patients suffering from most MPS types (I, II, IIIA, IIIB, IIIC, IIID, IVA, IVB, VI, VII, and IX). The RNA-seq studies were performed as described in Section 2. Although only one cell line per MPS type was used, every experiment was performed independently four times; thus, there were four biological repeats (n = 4). Since our goal was to investigate potential changes in the levels of lncRNAs in MPS patients for the first time, we also considered all MPS types as a group of diseases and assessed whether any changes were common for most or all types. This made our transcriptomic analyses reliable, as indicated in previously published articles [15][16][17][18][19][20][21][22][23]. For analyses concerning long RNA molecules, the standard RNA isolation procedure allowed for the purification of both mRNA and lncRNA species, which enabled us to compare the abundance of specific kinds of these nucleic acids in a single experiment.
Our global analysis of the abundance of RNA molecules indicated that there are changes in the levels of various lncRNAs in every MPS type. Between 2 (in MPS II) and 13 (in MPS IX) lncRNA transcripts were significantly changed (either up-or down-regulated) relative to the control fibroblasts ( Figure 1). These results indicated that there can be potential changes in the lncRNA-mediated regulation of the expression of various genes in MPS cells.
To assess which kinds of lncRNA are commonly up-or down-regulated in MPS cells, we assessed specific long non-coding transcripts that revealed either increased or decreased abundance relative to the control cells in fibroblasts of at least four MPS types. Five of such lncRNAs were identified, and are specified in Table 2. Importantly, in every case, a specific lncRNA was either up-or down-regulated in all MPS types wherein significant differences relative to the control cells occurred (no cases of up-regulation in some and down-regulation in other MPS types were identified). Therefore, it appears that similar changes in the levels of specific lncRNA occur in all/most of the MPS types. To assess which kinds of lncRNA are commonly up-or down-regulated in MPS cells, we assessed specific long non-coding transcripts that revealed either increased or decreased abundance relative to the control cells in fibroblasts of at least four MPS types. Five of such lncRNAs were identified, and are specified in Table 2. Importantly, in every case, a specific lncRNA was either up-or down-regulated in all MPS types wherein significant differences relative to the control cells occurred (no cases of up-regulation in some and down-regulation in other MPS types were identified). Therefore, it appears that similar changes in the levels of specific lncRNA occur in all/most of the MPS types.
In the next step, we analyzed how many and which lncRNAs occurred at especially increased or decreased levels in MPS cells. For this kind of analysis, we grouped lncRNAs into cohorts, revealing specific log2FC values (where FC states for fold change). These calculations indicated that the levels of most affected lncRNAs changed a few times; however, there were long non-coding transcripts that had been up-or down-regulated over ~6 times (log2FC>2.5 or <-2.5) (Figure 2).
Specific genes encoding lncRNAs occurring at especially elevated or decreased levels (over ~6-fold, i.e., log2FC>2.5 or <-2.5) in the MPS cells have been identified, and they are presented in Table 3. Again, the direction of the changes was always the same for each specific lncRNA type, i.e., every lncRNA was either up-or down-regulated in all MPS types. This confirms the uniform character of the changes in the levels of lncRNAs in MPS, irrespective of the disease type.  Table 2. Genes encoding lncRNAs whose expression is significantly changed in at least four MPS types relative to the control cells. Changes in expression (transcripts' levels) of particular genes in MPS types are depicted, indicating log 2 fold change (FC) relative to control cells; up-regulation is marked in green while down-regulation in marked in blue; non-statistically significant differences are not colored. I  II  IIIA  IIIB  IIIC  IIID  IVA  IVB  VI  VII  IX  GAS5 0 In the next step, we analyzed how many and which lncRNAs occurred at especially increased or decreased levels in MPS cells. For this kind of analysis, we grouped lncRNAs into cohorts, revealing specific log 2 FC values (where FC states for fold change). These calculations indicated that the levels of most affected lncRNAs changed a few times; however, there were long non-coding transcripts that had been up-or down-regulated over 6 times (log 2 FC > 2.5 or <−2.5) (Figure 2).

Transcript log 2 FC of Selected Transcripts' Levels in at Least 4 MPS Types vs. HDFa Line
Specific genes encoding lncRNAs occurring at especially elevated or decreased levels (over~6-fold, i.e., log 2 FC > 2.5 or <−2.5) in the MPS cells have been identified, and they are presented in Table 3. Again, the direction of the changes was always the same for each specific lncRNA type, i.e., every lncRNA was either up-or down-regulated in all MPS types. This confirms the uniform character of the changes in the levels of lncRNAs in MPS, irrespective of the disease type.   I  II  IIIA  IIIB  IIIC  IIID  IVA  IVB  VI  VII I  II  IIIA  IIIB  IIIC  IIID  IVA  IVB  VI  VII

Transcript log 2 FC > 2.5 or <−2.5 of Selected Transcripts' Levels in Particular MPS Type vs. HDFa Line
To identify potential target genes for specific lncRNA-mediated regulations, we employed the NPInter v4.0 data base (http://bigdata.ibp.ac.cn/npinter4/#, last accessed on 5 December 2022). Then, among all the genes potentially regulated by a given lncRNA, we selected those which revealed significant changes in at least one MPS type, as indicated by the results of our transcriptomic (RNA-seq) analyses for mRNAs (NCBI Sequence Read Archive (SRA) accession no. PRJNA562649). The results of these analyses are presented in Figure 3. They revealed that increased levels of the GAS5 lncRNA correlate with the upregulation of the HNRNPC gene in MPS IIIC, the down-regulation of the FXR1 gene in MPS IIIC, and the down-regulation of the MATR3 gene in MPS IIIB ( Figure 3A). Higher abundance of the MEG3 lncRNA occurred simultaneously with increased levels of TP53 mRNA in MPS IIIC ( Figure 3B) and decreased levels of TARDBP mRNA in MPS IIIB ( Figure 3C).
The down-regulation of LINC00856 lncRNA correlated with the down-regulation of the MATR3 gene in MPS IIIB ( Figure 3D). by the results of our transcriptomic (RNA-seq) analyses for mRNAs (NCBI Sequence Read Archive (SRA) accession no. PRJNA562649). The results of these analyses are presented in Figure 3. They revealed that increased levels of the GAS5 lncRNA correlate with the upregulation of the HNRNPC gene in MPS IIIC, the down-regulation of the FXR1 gene in MPS IIIC, and the down-regulation of the MATR3 gene in MPS IIIB ( Figure 3A). Higher abundance of the MEG3 lncRNA occurred simultaneously with increased levels of TP53 mRNA in MPS IIIC ( Figure 3B) and decreased levels of TARDBP mRNA in MPS IIIB (Figure 3C). The down-regulation of LINC00856 lncRNA correlated with the down-regulation of the MATR3 gene in MPS IIIB ( Figure 3D).

Discussion
Although MPS is a group of monogenic diseases characterized by the accumulation of GAGs in lysosomes, recent studies clearly indicated that primary storage is not the only cause of significant changes in cellular physiology. Among various secondary and tertiary defects that significantly contribute to the development of specific MPS symptoms in patients, the dysregulation of the expression of a battery of genes has been demonstrated to significantly influence the structures and functions of cellular organelles as well as different cellular processes [15][16][17][18][19][20][21][22][23]. Recent findings have suggested that changes in the expression of genes coding for regulators of activities of other genes might be responsible for triggering chains of processes that considerably contribute to dysfunctions of cells, tissues, organs, and entire organisms [16]. Since non-coding RNA molecules can modulate the expression of hundreds or thousands of genes [24][25][26][27], we investigated whether the levels of

Discussion
Although MPS is a group of monogenic diseases characterized by the accumulation of GAGs in lysosomes, recent studies clearly indicated that primary storage is not the only cause of significant changes in cellular physiology. Among various secondary and tertiary defects that significantly contribute to the development of specific MPS symptoms in patients, the dysregulation of the expression of a battery of genes has been demonstrated to significantly influence the structures and functions of cellular organelles as well as different cellular processes [15][16][17][18][19][20][21][22][23]. Recent findings have suggested that changes in the expression of genes coding for regulators of activities of other genes might be responsible for triggering chains of processes that considerably contribute to dysfunctions of cells, tissues, organs, and entire organisms [16]. Since non-coding RNA molecules can modulate the ex-pression of hundreds or thousands of genes [24][25][26][27], we investigated whether the levels of these kinds of regulatory RNAs are changed in MPS cells. In this study, we have focused on lncRNAs, as this kind of non-coding RNAs was not investigated previously in MPS cells.
In this report, RNA-seq analyses of the biological material of fibroblasts derived from patients suffering from most MPS types indicated, for the first time, that the levels of various lncRNAs are significantly changed relative to control cells ( Figure 1, Table 2). Some lncRNAs were up-or down-regulated especially significantly (over 6-times; Figure 2, Table 3), and using the NPInter v4.0 data base, we identified potential target genes whose expression could be regulated by the actions of these non-coding transcripts. Then, by analyzing the levels of mRNAs derived from the same cells, genes that could be affected by specific lncRNAs and whose expression was up-or down-regulated were indicated. Therefore, we assume that the expression of these genes is likely modulated by changes in the levels of specific lncRNAs in MPS cells. The putative pairs of such lncRNAs and proteinencoding genes are as follows: GAS5 lncRNA-HNRNPC, FXR1, and MATR3; MEG3 lncRNA-TP53 and TARDBP; and LINC00856 lncRNA -MATR3 ( Figure 3). Interestingly, this putative lncRNA-mediated regulation was especially pronounced in different subtypes of Sanfilippo disease (MPS III).
One may ask: what are the roles of the genetic elements involved in the abovedescribed regulations? It has been demonstrated that the GAS5 lncRNA controls cancer development [53,54]. Its potential target genes, HNRNPC and FXR1, whose expression is changed in MPS cells, encode heterogeneous nuclear ribonucleoprotein C (a protein involved in the control of splicing and translation) [55] and RNA-binding protein fragile-X mental retardation autosomal 1 (a factor that participates in the mRNA transport from the nucleus and is bound to polysomes) [56,57], respectively. The MATR3 gene product, matrin 3, is a protein that interacts with nucleic acids and is able to bind to both DNA and RNA fragments, though its exact functions are still poorly understood [58,59]. Like GAS5, the MEG3 lncRNA is a tumor-growth inhibitor [60,61]. Potential MEG3-regulated genes, whose changed expression correlated with increased levels of this lncRNA in MPS cells, include TP53 and TARDBP. These genes encode the major tumor-suppressor protein p53 [62] and TAR DNA-binding protein 43 (or TDP-43) (an RNA/DNA-binding protein that is involved in RNA transactions) [63], respectively. Previously, the LINC00856 lncRNA was found to be involved in the response to acute lymphoblastic leukemia [64], and the down-regulation of MATR3 correlated with decreased levels of this lncRNA in MPS cells.
Interestingly, as indicated in the preceding paragraph, genes strongly affected by lncRNAs in MPS cells are involved in the regulation of the expression of other genes, which is mainly performed through interactions with nucleic acids. It is, therefore, tempting to speculate that the dysregulation of the expression of regulatory genes might significantly contribute to the development of cellular dysfunctions in MPS cells. If such a dysregulation is strong enough, the cellular changes might become irreversible or hardly reversible. In such a case, the pathogenic cascade of cellular misfunctions would not be corrected by the simple removal of the primary storage material (GAGs). This might explain why various therapies for MPS fail to treat all the symptoms of patients, despite the normalization of GAG levels in body fluids [9][10][11]. In this light, specific lncRNAs might be considered as potential targets for novel auxiliary drugs against MPS. Namely, the primary therapeutic intervention (ERT, hematopoietic stem cell transplantation, gene therapy, or others) would cause the normalization of GAG levels, while the accompanying, lcnRNA-specific drug might modulate the expression of specific genes, leading to the corrections of otherwise hardly improved cellular processes.
The observed correlations between the changes in the levels of lncRNAs in MPS cells and the expression efficiency of the target genes were either direct or reverse (Figure 3). Various lncRNAs can regulate gene expression by either sponging miRNAs (alleviating negative regulation) or interacting directly with mRNA molecules (modulating their abundance and translational efficiency) [33,34,65]. Therefore, when the directions of the changes in the levels of specific lncRNA and an mRNA of its target gene are the same, the lncRNA acts, most probably, through the interaction with miRNA(s) and the prevention of mRNA degradation. On the other hand, when the directions of changes are opposite, direct lncRNA-mRNA interactions are more likely.
This study, which demonstrates changes in levels of lncRNAs in MPS cells for the first time, also has some limitations. One of these was the use of only one cell line per MPS type. However, the biological experiments were repeated four times, which made the results of the RNA-seq analyses reliable [15][16][17][18][19][20][21][22][23]. The presented conclusions can also be corroborated by the fact that the directions of the changes in the levels of specific lncRNAs were always the same in all MPS types. Another limitation was that our study only focused on transcriptomic analyses. Indeed, further studies with different investigatory methods of lncRNA functions are necessary to understand the details of their roles in MPS. Nevertheless, this is the first report signaling the importance of changes in the levels of lncRNAs in MPS; thus, it can open a new field of research into understanding the molecular mechanisms of the pathogenesis of this disease.

Conclusions
For the first time, changed levels of various lncRNAs were demonstrated in 11 types/subtypes of MPS. Increased or decreased levels of specific lncRNAs correlated with the up-or down-regulation of the expression of genes that are potential targets for lncRNA-mediated