Complex Changes in the Efficiency of the Expression of Many Genes in Monogenic Diseases, Mucopolysaccharidoses, May Arise from Significant Disturbances in the Levels of Factors Involved in the Gene Expression Regulation Processes

Monogenic diseases are primarily caused by mutations in a single gene; thus, they are commonly recognized as genetic disorders with the simplest mechanisms. However, recent studies have indicated that the molecular mechanisms of monogenic diseases can be unexpectedly complicated, and their understanding requires complex studies at the molecular level. Previously, we have demonstrated that in mucopolysaccharidoses (MPS), a group of monogenic lysosomal storage diseases, several hundreds of genes reveal significant changes in the expression of various genes. Although the secondary effects of the primary biochemical defect and the inefficient degradation of glycosaminoglycans (GAGs) might be considered, the scale of the changes in the expression of a large fraction of genes cannot be explained by a block in one biochemical pathway. Here, we demonstrate that in cellular models of 11 types of MPS, the expression of genes coding for proteins involved in the regulation of the expression of many other genes at various stages (such as signal transduction, transcription, splicing, RNA degradation, translation, and others) is significantly disturbed relative to the control cells. This conclusion was based on transcriptomic studies, supported by biochemical analyses of levels of selected proteins encoded by genes revealing an especially high level of dysregulation in MPS (EXOSC9, SRSF10, RPL23, and NOTCH3 proteins were investigated). Interestingly, the reduction in GAGs levels, through the inhibition of their synthesis normalized the amounts of EXOSC9, RPL23, and NOTCH3 in some (but not all) MPS types, while the levels of SRSF10 could not be corrected in this way. These results indicate that different mechanisms are involved in the dysregulation of the expression of various genes in MPS, pointing to a potential explanation for the inability of some therapies (such as enzyme replacement therapy or substrate reduction therapy) to fully correct the physiology of MPS patients. We suggest that the disturbed expression of some genes, which appears as secondary or tertiary effects of GAG storage, might not be reversible, even after a reduction in the amounts of the storage material.


Introduction
Among about 20,000 known diseases caused by genetic defects, over 7000 are monogenic disorders, defined as diseases resulting from mutations in a single gene [1]. It is estimated that 1 in 50 humans suffer from such a disease [2]. Since the dysfunction of a single gene should result in the disturbance of the functions of just one protein or functional RNA species, one might predict that the mechanisms of monogenic diseases should be relatively simple. Following this path of thinking, the correction of this single gene or its product should result in the normalization of all functions of the affected organism. However, such predictions have turned out to be incorrect. It is not only the correction of

Transcriptomics
Transcriptomic analyses were performed exactly as described earlier [5]. Briefly, RNA was isolated and purified from 5 × 10 5 cells withdrawn from each culture (between the 4th and 15th passage). Four biological repeats of each isolation, purification, and further analyses were conducted and the results were used for statistical analyses. mRNA libraries were constructed and used for the preparation of cDNA libraries which were then sequenced using HiSeq4000 (Illumina, San Diego, CA, USA). For bioinformatic analyses, 4 × 10 7 raw reads from each sample were used which gave at least 12 Gb of raw data per sample. The GRCh38 human reference genome from the Ensembl database was used to map the raw readings. The BioMart interface was employed for the annotation and classification of transcripts. Raw data of the RNA-seq analysis are deposited at NCBI Sequence Read Archive (SRA), under accession no. PRJNA562649.

Statistical Analyses
For transcriptomic studies, statistical significance was determined using the results from 4 biological repeats of each RNA isolation procedure (n = 4) in the case of each cell line. A one-way analysis of variance (ANOVA) was used on log 2 (1 + x) values with normal continuous distribution. The Benjamini-Hochberg method was employed to calculate the false discovery rate (FDR). A post hoc Student's t-test with a Bonferroni correction was used for comparisons between two groups. These analyses were conducted with the R software v3.4.3, and the significance was assumed if p < 0.1, according to standards used in transcriptomic analyses [17][18][19][20][21][22][23][24][25]. In Western blotting experiments, mean values from 3 biological experiments (n = 3) ± standard deviation (SD) were used for statistical analyses, and the significance was assumed if p < 0.05.

Transcriptomic Analyses
Using the Gene Ontology database (http://geneontology.org/ (accessed on 3 March 2022), we assessed levels of transcripts of genes included in the term "gene expression" (GO:0010467 in the QuickGO; https://www.ebi.ac.uk/QuickGO accessed on 3 March 2022) in MPS cells relative to control fibroblasts (HDFa). We found that the expression of dozens of genes from this category significantly changed (either up-or down-regulated) in the cells derived from patients suffering from all MPS types in comparison to the control cells ( Figure 1). Among all tested MPS types, the lowest number of transcripts with altered levels occurred in MPS VII (33 transcripts) and the highest number occurred in MPS VI (99 transcripts). These results indicate that the expression of a significant number of genes coding for proteins involved in activities of other genes is affected by MPS, and this was valid for all tested MPS types.
Module (#DM-001, ProteinSimple) were employed. Levels of specific proteins were measured with the WES system, using the Total Protein Detection Module for Chemiluminescence (#DM-TP01, ProteinSimple) as a loading control.

Statistical Analyses
For transcriptomic studies, statistical significance was determined using the results from 4 biological repeats of each RNA isolation procedure (n = 4) in the case of each cell line. A one-way analysis of variance (ANOVA) was used on log2(1 + x) values with normal continuous distribution. The Benjamini-Hochberg method was employed to calculate the false discovery rate (FDR). A post hoc Student's t-test with a Bonferroni correction was used for comparisons between two groups. These analyses were conducted with the R software v3.4.3, and the significance was assumed if p < 0.1, according to standards used in transcriptomic analyses [17][18][19][20][21][22][23][24][25]. In Western blotting experiments, mean values from 3 biological experiments (n = 3) ± standard deviation (SD) were used for statistical analyses, and the significance was assumed if p < 0.05.

Transcriptomic Analyses
Using the Gene Ontology database (http://geneontology.org/ (accessed on 3 March 2022), we assessed levels of transcripts of genes included in the term "gene expression" (GO:0010467 in the QuickGO; https://www.ebi.ac.uk/QuickGO accessed on 3 March 2022) in MPS cells relative to control fibroblasts (HDFa). We found that the expression of dozens of genes from this category significantly changed (either up-or down-regulated) in the cells derived from patients suffering from all MPS types in comparison to the control cells ( Figure 1). Among all tested MPS types, the lowest number of transcripts with altered levels occurred in MPS VII (33 transcripts) and the highest number occurred in MPS VI (99 transcripts). These results indicate that the expression of a significant number of genes coding for proteins involved in activities of other genes is affected by MPS, and this was valid for all tested MPS types. We hypothesized that the previously described [5] changes in the levels of hundreds of transcripts in MPS cells might arise from an alteration in the expression of genes encoding factors involved in gene expression regulation at various stages. If such a hypothesis We hypothesized that the previously described [5] changes in the levels of hundreds of transcripts in MPS cells might arise from an alteration in the expression of genes encoding factors involved in gene expression regulation at various stages. If such a hypothesis is true, then changed amounts of regulatory factors could cause further alterations in activities of many (hundreds) different genes. Our analysis indicated that between 23 (in MPS VI) and 67 (in MPS IVB) transcripts from this sub-category (child term GO:0010468: regulation of gene expression) revealed changed levels in MPS cells relative to controls ( Figure 2). is true, then changed amounts of regulatory factors could cause further alterations in activities of many (hundreds) different genes. Our analysis indicated that between 23 (in MPS VI) and 67 (in MPS IVB) transcripts from this sub-category (child term GO:0010468: regulation of gene expression) revealed changed levels in MPS cells relative to controls ( Figure 2). We assumed that the most pronounced effects on the general activities of various genes should have specific genes coding for proteins involved in different stages of genetic information expression whose transcripts reveal especially high levels of changes. Therefore, transcripts with levels altered (down-or up-regulated) at least four times (i.e., log2FC > 2, where FC means 'fold change') in MPS cells relative to the HDFa control were identified among those included in GO:0010467 ("gene expression") ( Table 1). Although the 4-fold change was chosen arbitrarily, on the basis of previous studies one can indicate that this is a change level showing a significant and unambiguous alteration in gene expression which can affect cellular functions [18][19][20][21][22][23][24][25]. We assumed that the most pronounced effects on the general activities of various genes should have specific genes coding for proteins involved in different stages of genetic information expression whose transcripts reveal especially high levels of changes. Therefore, transcripts with levels altered (down-or up-regulated) at least four times (i.e., log 2 FC > 2, where FC means 'fold change') in MPS cells relative to the HDFa control were identified among those included in GO:0010467 ("gene expression") ( Table 1). Although the 4-fold change was chosen arbitrarily, on the basis of previous studies one can indicate that this is a change level showing a significant and unambiguous alteration in gene expression which can affect cellular functions [18][19][20][21][22][23][24][25].
Thirty-four transcripts fulfilled such a requirement, and they represented 28 genes (there were two or more kinds of transcripts of some genes) for factors involved in gene expression processes at different stages such as signal transduction (e.g., NOTCH3), transcription initiation (e.g., HOXC9), splicing (e.g., SRSF10), RNA degradation (e.g., EXOSC9), and translation (e.g., RPL23) (Table 1). Interestingly, in all transcripts but one (the COL4A2 first transcript), the direction of the change (down-or up-regulation) was the same for all MPS types which indicated the general tendency occurring in MPS irrespective of the specific type of the disease (Table 1). On the other hand, there were also clear differences between MPS types, as some genes represented by more than one transcript (MME, RPL10, COL4A2, and SPOCD1) revealed various levels of expression dysregulation in fibroblasts derived from patients with specific syndromes (Table 1). This might suggest that the kind(s) of stored GAG(s) might indirectly influence specific mechanisms regulating transcription or post-transcriptional modifications (including RNA degradation).   I  II  IIIA  IIIB  IIIC  IIID  IVA  IVB  VI  VII 1 If more than one transcript of a given gene could be identified in the RNA-seq analysis, they were indicated as transcript 1 (tr. 1), transcript 2 (tr. 2), and so on. Raw data for all transcripts are available at NCBI Sequence Read Archive (SRA), under accession no. PRJNA562649. 2 Down-regulated transcripts are marked with down-headed arrows (↓), and up-regulated transcripts are marked by up-headed arrows (↑).
For further analyses, we chose transcripts of four genes coding for proteins operating at various stages of the process of gene expression. They were characterized as having experienced significant changes in at least six MPS types. These genes were as follows: EXOSC9, coding for a component of the exoribonuclease complex which degrades RNA molecules [29], RPL23, coding for one of ribosomal subunits [30], SRSF10, encoding a protein involved in RNA splicing [31], and NOTCH3, encoding a receptor involved in the signal transduction process which regulates the activities of transcription factors [32]. The changes in the levels of these transcripts in fibroblasts representing all MPS types are summarized in Table 2. Despite the fact that the same tendency of changes in the levels of specific transcripts could be observed in all MPS types, the evident differences between the diseases might again suggest an influence of the kind(s) of stored GAG(s) on the gene expression regulation in the tested cells. Table 2. Genes coding for proteins involved in gene expression regulation whose transcripts revealed significantly changed levels in at least six MPS types relative to the control cells. The colored boxes indicate statistically significant differences (at FDR < 0.1; p < 0.1) between MPS and control cell lines: red boxes indicate down-regulation, and blue boxes indicate up-regulation relative to control.

Levels of Selected Proteins in MPS Cells
To investigate the expression of the selected genes (EXOSC9, RPL23, SRSF10, NOTCH3) in more detail, we measured the levels of their products in MPS cells relative to control cells. Amounts of protein were determined by Western blotting in all MPS types, and the results of these analyses are shown in  Table 2 and Figure 4). This indicated a relatively high compatibility of the results of these two kinds of experiments. However, in the case of the RPL23 gene and its product, we observed a

Effects of the Reduction in the GAG Levels in MPS Cells on the Abundance of the Selected Proteins
To test whether the elimination of the GAG storage in MPS cells can correct the levels of the selected proteins in the cells, we used genistein. This isoflavone has been demonstrated previously to significantly reduce the amounts of accumulated GAGs in MPS fibroblasts through the inhibition of their synthesis [28]. Therefore, MPS cells of types in which changed levels of EXOSC9, RPL23, SRSF10, and NOTCH3 were detected (see

Effects of the Reduction in the GAG Levels in MPS Cells on the Abundance of the Selected Proteins
To test whether the elimination of the GAG storage in MPS cells can correct the levels of the selected proteins in the cells, we used genistein. This isoflavone has been demonstrated previously to significantly reduce the amounts of accumulated GAGs in MPS fibroblasts through the inhibition of their synthesis [28]. Therefore, MPS cells of types in which changed levels of EXOSC9, RPL23, SRSF10, and NOTCH3 were detected (see Figure 4) were treated with 50 µM genistein for 48 h, and the effects on the abundance of these proteins were determined by Western blotting. The results of these experiments are demonstrated in Figure 5 (for measurement of EXOSC9), Figure 6 (for RPL23), Figure 7

Discussion
MPS are monogenic diseases caused by mutations in the genes involved in GAG metabolism, and it is commonly accepted that lysosomal GAG storage, resulting from the dysfunction of one of enzymes required for the degradation of these compounds, is the primary metabolic defect responsible for the severe symptoms found in patients [7]. However, recent studies have indicated that secondary and tertiary effects can considerably modulate the course of the disease due to various cellular dysfunctions arising not necessarily directly from the storage, but rather from the disruption of various processes in the cascade of changes in the regulatory mechanisms [17]. In this light, a failure to correct all symptoms in MPS patients treated with HSCT, ERT, SRT, or gene therapy, which theoretically should remove the primary metabolic defect [9][10][11][12][13][14][15][16], might be recognized from another point of view. The discovery of significant changes in the expression of~300-900 genes in MPS cells [5] indicated that the disrupted control of activities of many genes can contribute considerably to the disease pathomechanism. However, it was unclear how could such huge effects on the expression of hundreds of genes result from GAG storage.
In this report, we have demonstrated that the activities of many genes coding for proteins involved in different stages of expression of other genes are affected in MPS cells relative to control cells (Figures 1-3). Analyses of transcripts with especially high changes (FC > 2) in levels ( Table 1) showed that they include those coding for proteins required at various stages of gene expression, such as signal transduction, transcription, splicing, RNA degradation, translation, and others. Therefore, we hypothesized that different amounts of such proteins, controlling the expression of other genes, might strongly contribute to the dysregulation of hundreds of genes in MPS cells (as observed previously [5,[18][19][20][21][22][23][24][25][26][27]). For more detailed analyses, the following genes (and corresponding gene products) were chosen: EXOSC9 (coding for a structural component of the exosome which functions in the RNA degradation process [29]), RPL23 (coding for a large ribosomal subunit protein L23 [30]), SRSF10 (coding for a splicing regulatory protein [31]), and NOTCH3 (coding for a receptor protein involved in cellular signaling [32]). Apart from demonstrating especially high changes in the levels of transcripts of these genes in MPS cells, we also showed significant changes in the amounts of the corresponding proteins in these cells. Interestingly, these changes were not always compatible, i.e., in some cases, the directions of changes (up-or down-regulation) in the levels of the transcripts were opposite to those in the levels of proteins (Table 2 and Figures 4-8). Thus, for these genes, the regulation of their expression proceeds at both the transcriptional and pos-transcriptional stages, perhaps also including the control of translation.
Interestingly, it has been demonstrated previously that mutations in genes which were analyzed in more detail in MPS and revealed high levels of changes in their expression (EXOSC9, RPL23, SRSF10, and NOTCH3) result in the development of various diseases whose symptoms resemble those observed in MPS. Namely, EXOSC9 dysfunction leads to a neurodegenerative disease pontocerebellar hypoplasia type 1b [29], and neurodegeneration occurs in most MPS types [8]. Thus, one might assume that the downregulation of EXOSC9 expression ( Figure 5) can contribute to neurodegenerative processes in MPS patients. Moreover, RLP23 has been reported as a factor involved in various cellular processes, such as cell proliferation, apoptosis, and cell cycle, and all of them were demonstrated to be affected in MPS cells [18,21,33]. Furthermore, the SRSF10 protein is important for neurological processes and responses to viral infections [31], and abnormalities of both these groups of processes are evident in MPS [8,24]. Finally, mutations in the NOTCH3 gene cause various abnormalities, including arteriopathy and leukoencephalopathy, which occur also (to various degrees) in MPS [8], thus suggesting that the dysregulation of expression of this gene may be connected to MPS symptoms.
From the point of view of the development of effective therapies for MPS, it was important to test if a reduction in GAG storage results in the normalization of expression of the tested genes, measured by the levels of their final products-specific proteins. GAG levels were reduced by using genistein, a compound described previously by different research groups as an effective molecule in impairing the efficiency of GAG synthesis and decreasing the levels in cells [28,[34][35][36][37]. Perhaps surprisingly, treatment with genistein corrected the levels of expression of EXOSC9, RPL23, SRSF10, and NOTCH3 genes only in some tested MPS lines but not in others (Figures 5-8). Therefore, changes in the expression of these genes could be more stable than expected, and they might not necessarily be normalized if the primary metabolic defect is reversed. As a result, under the "therapeutic" conditions, sub-optimal levels of corresponding proteins (EXOSC9, RPL23, SRSF10, and NOTCH3) could still affect the expression of other genes, preventing the disappearance of cellular defects, and further correcting tissue and organ dysfunctions. Levels of these proteins in animal MPS models or MPS patients subjected to various therapeutical procedures remain to be determined. However, if this hypothesis is true it can explain, at least partially, a failure to reverse all MPS symptoms by any therapeutical approaches used to date, despite the efficient reduction in GAG storage.
The main limitations of this study were the use of a single cell line from each MPS type and the use of fibroblasts as the only kind of cells. Obviously, to perform a comprehensive study on a given disease, biological material from many patients should be employed, including samples from individuals representing different courses of the disease (mild, intermediate, severe). Moreover, fibroblasts are only representative and are model cells, while MPS affects virtually all kinds of cells. On the other hand, one must note that MPS are rare diseases with a severe course; thus, obtaining biological material from many patients is extremely difficult, if not impossible. For example, to date, only four patients suffering from MPS IX have been reported worldwide [8], making studies on a higher number of patients objectively impossible. Due to the severity of the disease, a lack of cooperativity among the vast majority of patients, and the related ethical aspects (unnecessary medical interventions), obtaining any biological material from MPS patients is difficult and problematic. In this light, we chose to use one cell line per each MPS type; however, our analyses indicated that the directions of changes in levels of transcripts of most genes are the same in various MPS types (Tables 1 and 2) which indicates that the regulatory mechanisms might be similar in the whole group of the diseases. Moreover, each experiment with RNA isolation was repeated four times (four independent biological repeats) making the results of the performed studies reliable. One should also note that fibroblasts have previously been used as model cells in many studies on MPS which has led to the discovery of the general mechanisms operating in this disease [8]. Therefore, if the conclusions do not concern the specific symptoms of patients but are rather restricted to the molecular mechanisms of the disease, such cells appear to be appropriate research models.

Conclusions
Changes in the expression of hundreds of genes in MPS cells may result from the dysregulation of activities of a smaller group of genes whose products are involved in the expression of other genes at various stages of this process. Dysregulation of gene expression in MPS cells is not fully normalized by the reduction in GAG levels, pointing to possible reasons for the failure to correct all MPS symptoms in patients subjected to all types of therapies used to date.