Nonsteroidal anti-inflammatory drugs modulate cellular glycosaminoglycan synthesis by affecting EGFR and PI3K signaling pathways

In this report, selected non-steroidal anti-inflammatory drugs (NSAIDs), indomethacin and nimesulide, and analgesics acetaminophen, alone, as well as in combination with isoflavone genistein as potential glycosaminoglycan (GAG) metabolism modulators were considered for the treatment of mucopolysaccharidoses (MPSs) with neurological symptoms due to the effective blood-brain barrier (BBB) penetration properties of these compounds. We found that indomethacin and nimesulide, but not acetaminophen, inhibited GAG synthesis in fibroblasts significantly, while the most pronounced impairment of glycosaminoglycan production was observed after exposure to the mixture of nimesulide and genistein. Phosphorylation of the EGF receptor (EGFR) was inhibited even more effective in the presence of indomethacin and nimesulide than in the presence of genistein. When examined the activity of phosphatidylinositol-3-kinase (PI3K) production, we observed its most significant decrease in the case of fibroblast exposition to nimesulide, and afterwards to indomethacin and genistein mix, rather than indomethacin used alone. Some effects on expression of individual GAG metabolism-related and lysosomal function genes, and significant activity modulation of a number of genes involved in intracellular signal transduction pathways and metabolism of DNA and proteins were detected. This study documents that NSAIDs, and their mixtures with genistein modulate cellular glycosaminoglycan synthesis by affecting EGFR and PI3K signaling pathways.


Measurement of kinetics of GAG synthesis.
Glycosaminoglycan synthesis was estimated by measurement of 35 S or 3 H incorporation into GAG chains. Cells were plated in a number of 2 × 10 4 per well in 48-well plate and incubated overnight to allow the attachment. Next, cells were treated for 48 hours in standard DMEM supplemented with appropriate amounts of particular NSAIDs, genistein, mixtures of selected NSAIDs with genistein or DMSO only (Ctrl). After 48 h of incubation, cells were labelled for 24 hours with 20 mCi/ml of H 2 [ 35 S]O 4 or 10 mCi/ml of 3 H-GlcN (Hartmann Analytic GmbH, Braunschweig, Germany) in growth medium without inorganic sulfates (Minimum Essential Medium Eagle, Joklik's modified, Sigma-Aldrich Co. LLC., St. Louis, USA) or growth medium without glucose (Thermo Fisher Scientific Inc., Paisley, UK) mixed with standard DMEM medium (1:1) supplemented with tested compounds or DMSO. Next, cells were washed six times with Dulbecco's phosphate-buffered saline (DPBS, Thermo Fisher Scientific Inc., Paisley, UK) and digested for 3 hours at 65 °C with 0.5% papain (Sigma-Aldrich Co. LLC., St. Louis, USA) (prepared in 200 mM phosphate buffer [Na 2 HPO 4 -NaH 2 PO 4 ], pH 6.4, containing 100 mM sodium acetate, 10 mM Na 2 EDTA and 5 mM L-cysteine). Incorporation of radioactive precursors was measured in a MicroBeta 2 scintillation counter (Perkin Elmer Inc., MA, USA) and Quant-iT ™ PicoGreen ® dsDNA Reagent (Thermo Fisher Scientific Inc., Paisley, UK) was used to determine DNA concentration in papain digested samples. Incorporation of radioactive precursors was calculated per DNA amount (cpm/ng of DNA) and expressed as relative to control cultures (treated with 0.05% DMSO only).
To study the effects of a combination of tested compounds in order to provide evidence of a significant superiority compared to the single agents the Combination Index (CI) with the use of the Bliss Independence model was determined 19,20 . It is based on the principle that drug effects are outcomes of probabilistic processes and assumes that drugs act independently in such a manner that neither of them interferes with the other (different sites of action), but each contributes to a common result. CI is recognized as the standard measure of combination effect that indicated a greater (CI < 1), lesser (CI > 1) or similar (CI = 1) effect than the expected additive effect.
EGFR and PI3K phosphorylation analysis. Efficiency of phosphorylation inhibition of the epidermal growth factor receptor (EGFR) or phosphoinositide 3-kinase (PI3K) pathway was assessed using MUSE ® Cell Analyzer (Merck Millipore, Germany) using a Millipore's Muse ™ Activation Dual Detection kits (Merck Millipore, Germany). HDFa cells were seeded in 6-well plates at a density of 2 × 10 5 per well. Following an overnight incubation, standard growth medium was substituted with the medium containing 0.05% DMSO only, supplemented with 100 ng/ml EGF in 0.05% DMSO, or with 100 ng/ml EGF in 0.05% DMSO and particular tested compounds at the indicated final concentrations 21 . After 24 hours EGFR pathway activation was determined by Muse ® EGFR-RTK Activation Dual Detection kit (Merck Millipore, Germany) by using a phospho-specific anti-phospho-EGFR (Tyr1173)-Alexa Fluor ® 555 and an anti-EGFR-PECy5 according to the manufacturer's protocol. Similarly, PI3 kinase pathway activation was assessed using an anti-phospho-Akt (Ser473), Alexa Fluor ® 555 and an anti-Akt, PECy5 provided in the PI3K Activation Dual Detection kit (Merck Millipore, Germany). Additionally, Jurkat T cells treated with 0.5 mM wortmannin were used as a control of PI3K activation status determination.
Total RNA extraction. For transcriptomic analyses 1 × 10 5 cells were seeded and grown for 24 h. Then, the medium was substituted with either NSAIDs-free one, containing DMSO (Ctrl), or the one supplemented with appropriate amounts of tested drug, genistein or mixture of drug and genistein. Total RNA was extracted from cells using the High Pure RNA Isolation Kit (Roche Applied Science, IN, USA) and quantified with the Quant-it TM RiboGreen ® assay kit (Thermo Fisher Scientific Inc., Paisley, UK) following the manufacturer's instructions. In addition, the quality of each RNA sample was assessed using the RNA 6000 Nano Assay on the Agilent 2100 Bioanalyzer (Agilent Technologies Inc., CA, USA).

DNA microarray processing and real-time quantitative RT-PCR assays for mRNA analysis.
Whole genome microarray analysis of three biological replicates was performed using Illumina's Human HT-12v4 Expression BeadChips, targeting more than 25,000 genes with more than 48,000 probes (Illumina Inc., CA, USA), for all tested conditions. BeadChips were scanned using an Illumina BeadArray Reader and the Bead Scan Software (Illumina Inc., CA, USA). The quality of microarray data was controlled by examining raw and adjusted intensity histograms. The assay performance and data extraction was done according to Moskot et al. 10 All gene expression data have been deposited in the NCBI Gene Expression Omnibus (GEO series accession number GSE63239), according to the MIAME (minimum information about a microarray experiment) standards. An overview of experiment performance was gained by clustering samples using a correlation metric (Illumina ® BeadStudio Data Analysis software). The Pearson correlation coefficient method was used to calculate "expression distance values" across experiments and to group samples that have similar expression patterns. The values ranging between 0.98 and 0.99 for biological replicates indicate a high degree of reproducibility and strong correspondences between expression profiles.
Quantitative real-time Reverse Transcription PCR (real-time qRT-PCR) was performed to measure the mRNA levels of the studied genes. using the Light Cycler 480 II detection system (Roche Applied Science, IN, USA). Expression values were normalized against three control genes ACTB, SDHA and YWHAZ of constant expression level (2 −ΔΔct method). Determination of reference genes for real-time qRT-PCR was assessed using commercially available RealTime ready Human Reference Gene Panel (cat no. 05339545001, Roche Applied Science, IN, USA). Moreover, statistical analyses of the normalized gene expression data were performed in Prism (GraphPad). For both DNA microarray and real-time qRT-PCR analyses, a fold change (FC) greater or equal to 1.3 and below and equal 0.7 was considered as a relevant criterion for genes being significantly differentially expressed, with a p value of <0.05.
Gene Ontology and Gene Set Enrichment studies. The transcriptomic analysis was conducted through combining gene set enrichment analysis (GSEA) 22 , with the use of GOrilla 23 and REViGO (visual interaction gene network analysis) 24 tools for reconstructing crucial molecular networks and target genes identification. For GSEA study goals Reactome was used as reference database for ontology distribution with FDR q-value below 0.1 overall as a limitation for ontology selection. Tests for enrichment of genes with shared Biological Process Gene Ontology (GO) terms were performed using the online tool GOrilla. REViGO was used to eliminate redundant GO terms, and multiple-test correction for significant GO. Genes exhibiting fold change ≥1.3 and below and equal 0.7 with a p value less than 0.05, defined as differential expressed genes (DEGs), were examined.

Results
Selection of analgesic and nonsteroidal anti-inflammatory drugs and study of their cytotoxic and antiproliferative activity. It has been previously demonstrated that acetaminophen, indomethacin and nimesulide are among analgesics and NSAIDs which inhibit GAG synthesis 13,14,25 . Moreover, mixtures of those compounds with genistein, flavonoid responsible for impairment of production and enhancement of degradation of glycosaminoglycans 6,10,11 , were included in this work in order to see whether they might result in even more effective reduction of GAG accumulation than the use of any of these compounds alone. The concentrations of acetaminophen (5-500 mM), indomethacin (1.5-50 mM), and nimesulide (0.3-100 mM) used are close to clinically relevant molar concentrations, while genistein of 100 mM as previously applied 6,10,11 .
We first tested selected drugs in order to establish their cytotoxicity and antiproliferative activity in cultured fibroblasts. For this, we measured metabolic activity of HDFa, MPS IIIA, IIIB and VI cells treated with different concentrations of acetaminophen (5, 25, 100, and 500 mM), indomethacin (1.5, 5, 15, and 50 mM), or nimesulide (0.3, 2, 10, and 100 mM) for different periods of time (24, 48 h, and 7 days). This study demonstrated that 24 and 48 h treatment with clinically achievable low concentrations of tested compounds has little cytotoxic effect ( Fig. 1). Additionally, no remarkable inhibitory effect of particular compounds on fibroblasts growth after 7 days was visible, while clear antiproliferative activity was observed for mixtures used, i.e. both indomethacin or nimesulide with genistein ( Fig. 1). Tested compounds were added individually at the final concentration of 20, 100, and 500 mM for acetaminophen, 0.4, 2, and 10 mM for indomethacin and 1, 5, and 25 mM for nimesulide, 100 mM for genistein, or applied as mix of 10 mM indomethacin and 100 mM genistein, or 25 mM nimesulide and 100 mM genistein to cell culture. We found that two of all tested compounds, i.e. indomethacin and nimesulide, but not acetaminophen, revealed ability to inhibit significantly GAG synthesis in HDFa, MPS IIIA and VI (Fig. 2). The most pronounced impairment of glycosaminoglycan production measured by estimation of the amount of incorporated 35 S was observed after 72-hour exposure to nimesulide and genistein mixture in HDFa (48% decrease) and MPS IIIA (40% reduction). Furthermore, statistically significant differences in GAG synthesis relative to untreated cells were documented by estimation of the amount of incorporated 35 S with some exceptions as follows: (i) at the level of around 30% reduction for HDFa treated for 72 hours with 25 mM nimesulide (when both radioactive precursors applied), HDFa and MPS IIIA treated for 72 hours with 100 mM genistein, and MPS VI in the presence of 10 mM indomethacin; (ii) at the level of 25% decrease for HDFa treated for 72 hours with 10 mM indomethacin and for 7 days with 25 mM nimesulide, as well as for MPS VI exposed to 25 mM nimesulide for 72 hours; and (iii) at the level of 15-20% decline for HDFa treated for 72 hours with 1 mM nimesulide (glucosamine, D-[1-3 H] incorporation), and for 7 days with 5 mM nimesulide, as well as for MPS IIIA exposed to 25 mM nimesulide for 72 h. Interestingly, acetaminophen stimulated rather than inhibited GAG synthesis (Fig. 2). In addition, the Combination Index (CI) with the use of the Bliss Independence model was determined for the action of tested mixtures in HDFa and MPS IIIA fibroblasts. For both cell lines the CI value was 1.9 in the case of indomethacin and genistein combination indicating an antagonism. When the mixture of nimesulide and genistein was used, the CI was equal to 1.0, resulting in an additive effect.

Inhibition of phosphorylation of EGF receptor in the presence of indomethacin, nimesulide and genistein.
To test whether compounds used in this study impair GAG synthesis by the same mechanism as that proposed for genistein, we measured efficiency of phosphorylation of EGFR in the presence and absence of investigated NSAIDs, i.e. indomethacin and nimesulide, effectively inhibiting GAG synthesis in fibroblasts. In addition, tests with the use of mixture of indomethacin and genistein, and also nimesulide and genistein were performed. The extent of EGFR pathway activation by measuring EGFR phosphorylation relative to the total EGFR expression in given cell population was investigated. By doing such, the levels of both the total and phosphorylated protein were measured simultaneously in the same cell, resulting in a normalized and accurate measurement of EGFR activation Sections E1 and E2 stand for viability of HDFa and MPS IIIA cells treated with mixtures of indomethacin (I) and genistein (G), and also nimesulide (N) and genistein (G). Statistically significant differences in cell viability relative to control cells (treated with 0.05% DMSO only) are indicated with * for p < 0.05, ** for p < 0.01. Statistical analysis was performed using ANOVA with Tukey's HSD Post Hoc test. after drug stimulation. We found that both NSAIDs tested separately revealed statistically significant inhibition of EGFR phosphorylation, even more effective than that of genistein, or compounds' combinations ( Fig. 3).
PI3K activity decrease in fibroblasts treated with NSAIDs and genistein. In this part of the work we studied further how the inhibition of EGFR phosphorylation by investigated NSAIDs influences the activity of phosphatidylinositol-3-kinase, a downstream effector of EGF receptor, in fibroblasts treated with particular indomethacin and nimesulide alone or with their mixture with genistein. When examined the activity of PI3K produced we observed its most significant decrease in the case of cell exhibition to nimesulide. Indomethacin in combination with genistein resulted in more effective inhibition of PI3K activity than used alone, however still less pronounced when compared to nimesulide (Fig. 4).  In general, the microarray studies revealed that the fibroblasts responded to this treatment rather with minor changes in global expression profiles of only few dozen of genes, slightly modulating transcript level of genes involved in GAG metabolism pathway and lysosomal function, affecting however a number of genes related to different signal transduction pathways and metabolism of DNA and proteins, substantially when high concentrations of the NSAIDs were applied.

Activity of genes involved in GAG and
We used a real-time quantitative RT-PCR approach to examine in more detail the expression patterns of genes involved in GAG metabolism, as well as in lysosome biogenesis and function, whose activities were modulated in fibroblasts at tested conditions, as observed in microarray studies. Because significant changes in the activity of these genes were generally visible to the cells treated with 25 mM nimesulide, 100 mM genistein, and the mix of these two compounds, we implemented such conditions for real-time qRT-PCR analyses. Among 45 genes covered by Real Time ready Custom Panel, 7 of GAG metabolism pathways (5 of GAG synthesis and 2 of GAG degradation) and 7 sequences coding for lysosomal proteins (except for these 2 of GAG degradation), all revealing modulated activity in transcriptomic analyses, were detected (Table 1) Moreover, no transcriptomic data regarding modulation of expression of prostaglandin synthesis were available due to the problem with low level analysis being related to elimination of the results for these transcripts after background adjustment. This could have resulted from the use of uninflamed, wild type HDFa fibroblasts, i.e. not induced by inflammatory stimuli. Though, we used a real-time quantitative RT-PCR approach to examine in more detail the expression patterns of two prostaglandin synthase genes, PTGS1 and PTGS2, in fibroblasts exposed not only to 10 mM indomethacin or 25 mM nimesulide as it was in the case of microarray analysis, but also to their mixtures with 100 mM genistein. We found that indomethacin, but not nimesulide induced a 1.3-fold decrease of PTGS1 mRNA level, while PTGS2 expression was down-regulated by nimesulide, not however by indomethacin, when compared with the non-treated counterpart by using both the ACTB and YWHAZ reference control genes. Peculiarly, the analysis showed that both genes PTGS1 and PTGS2 were however significantly enriched in their transcript level in response to the compound combinations' treatment of cells. In addition, our studies revealed that selected NSAIDs and/or their combination with genistein can significantly modulate mRNA level of genes coding for products involved in regulation of cellular signaling pathways, and DNA and protein metabolism ( Fig. 5 and Table 2). Gene Ontology analyses performed with the GOrilla tool of microarray data revealed that at the top of the list of NSAID-regulated genes, were those mainly involved in metabolic processes (for both 10 mM indomethacin and 25 mM nimesulide 24 h treatment) in the cell (Fig. 5). Accordingly, GSEA conducted on the Reactome pathways showed a significant enrichment of signal transduction cascades genes (involved mainly in the Wnt, TCR, SCF/c-kit associated with PI3K signaling), and among DNA and protein metabolism genes, regulated by 10 mM indomethacin, and also 25 mM nimesulide, after 24 h exposition ( Table 2).

Discussion
This work exploits and expands on recent breakthroughs in the understanding of cellular cross-talk to develop novel therapeutic approaches based on the use of analgesic nonsteroidal anti-inflammatory drugs to treat rare inherited lysosomal diseases, such as mucopolysaccharidoses caused by altered glycosaminoglycan metabolism. The encouraging results of research on the molecular mechanism of isoflavone genistein-mediated impairment of GAG production prompted us to continue the study on the implementation of the non-enzymatic SRT using GAG synthesis inhibitors to treat MPS, especially those with neurological symptoms 7,10,11 . Furthermore, the latest investigations exhibited that among different analgesics and NSAID compounds acetaminophen, indomethacin and nimesulide are substances remarkably inhibiting GAG synthesis 13,14 . It has been shown that the reason for this glycosaminoglycan level reduction is a deficiency of cellular sulphates, a key factor balancing the production and degradation of GAGs, resulting from NSAID medication with side-effect in form of sulphate depletion, which in turn takes place in the treatment of arthritis [25][26][27] . There is different matter in the case of patients suffering from mucopolysaccharidosis. The excess of GAGs, due to lack of or significantly reduced activity of one of the enzymes responsible for degradation of these compounds, affects cells, tissues, organs, and consequently the whole body. Thus, the performance of GAG synthesis reduction through the use of agents causing the sulphate depletion in the human body may yield in a related effect to that observed in the case of genistein. It is reported in the literature that the main mechanism of action of indomethacin and nimesulide is the inhibition of prostaglandin synthesis by blocking the activity of the enzyme cyclooxygenase (COX) isoforms, officially known as prostaglandin-endoperoxide synthase (PTGS) isoforms, COX-1 (PTGS1; non-selectively blocked by indomethacin), COX-2 (PTGS2; non-selectively blocked by indomethacin and selectively blocked by nimesulide) and/or COX-3 (selectively blocked by acetaminophen), which in turn leads to reduced signs of inflammation 28 . Another work, describing stimulation of angiogenesis, cell proliferation and differentiation in carcinogenesis by prostaglandins, indicated their additional function as activators of intracellular signaling pathway through EGF receptor 26 . Based on this, a hypothesis can be considered that the reduction of prostaglandin level, resulting from the inhibition of their synthesis in the cell exposed to the above substances responsible for the COX activity decrease, might lead to the inhibition of EGF receptor activity, thereby blocking the cellular signal transduction process essential for the expression of genes encoding the relevant enzymes involved in the metabolism of GAGs. However, this phenomenon has remained unclear. Therefore, we intended to validate the extent of action of not only flavonoids, but also selected drugs, and to learn interactions at the molecular level in terms of modulation of cellular responses by these substances.
As acetaminophen, indomethacin and nimesulide belong to drugs routinely used in medication, their cytotoxicity and anti-proliferation features are already widely documented and are consistent with those obtained in our work (Fig. 1). Moreover, the properties concerning the impairment of GAG synthesis via all three selected for this work compounds were verified, when used alone and in combination with genistein (Fig. 2). Interestingly, most significant inhibition of glycosaminoglycan production was observed in cells treated with a mix of nimesulide and genistein, resulted in an additive effect of these two compounds. The question remains what is the molecular mechanism by which the tested compounds reduce efficiency of GAG synthesis. It is known from the literature that analgesic and nonsteroidal anti-inflammatory drugs inhibit synthesis of prostaglandins 28 , and that prostaglandins function in cells as activators of intracellular signaling pathway through EGF receptor 29 . On the other hand, results of our previous studies suggested that genistein impairs GAG synthesis due to its inhibitory effect on 3 for at least one of the conditions tested, n ≥ 3, with the p-value < 0.05), and denote differences for samples treated with tested compound or their mixture, against untreated samples, with respect to reference mRNA expression at a constant level. n.d. stands for no data available, as no transcriptomic data regarding modulation of the expression of these genes were available due to the problem with low level analysis being related to elimination of the results for these transcripts after background adjustment. * Stands for result based on raw data from Moskot et al. 11 . the kinase activity of the epidermal growth factor receptor (EGFR) 7 . At this time, we showed that indomethacin and nimesulide can inhibit EGFR and PI3K signaling cascades (Figs 3 and 4). It appears that agents that target the EGFR/PI3K axes and signaling cascades show a considerable promise for treatment of MPS patients. Our recent findings provided information on genistein targetome responsible for modulation of expression of GAG metabolism and lysosomal biogenesis and function genes 10,11 . In the work reported here, in turn, we observed that indomethacin and nimesulide can downregulate glycosaminoglycan production by blocking phosphorylation of EGF receptor (Figs 2 and 3), thus possibly impairing a signal transduction pathway necessary for activation of genes coding for enzymes involved in this anabolic process. The reduction in EGFR activity by tested NSAIDs, even more effective than by genistein (Fig. 3), may also be associated with the modulation of expression of genes which products act as activators of various intracellular signaling pathways. It was therefore interesting to study the effect of selected NSAIDs on cell transcriptome, with particular attention paid to GAG and lysosome-related genes' activities, as well as to genes involved in synthesis of prostaglandins and various intracellular signaling pathways. We noticed some changes in expression of particular GAG metabolism-related and lysosomal function genes, with significant variation in the activity of several transcripts linked to intracellular signal transduction pathways and metabolism of DNA and proteins (Fig. 5, Tables 1 and 2). However, it should be noted at this point that essentially a rather subtle effect of the tested substances on the global activity of genome of human dermal fibroblasts was detected. These observations seem to be valuable particularly in the context of the safety in use of these compounds for humans.
The deregulation of GAGs may concern either all types of GAGs or some of them, depending on the agent being used. Our studies showed that the tested compounds, i.e. indomethacin, nimesulide and genistein exhibit  deregulation of all types of glycosaminoglycans due to EGF receptor signaling cascades' inhibition. In addition, in the case of cell treatment with NSAIDs, substances responsible for depletion of cellular sulphates, only sulphated GAGs (i.e. heparan sulfate, dermatan sulfate, chondroitin sulfate and keratan sulfate) seemed to be deregulated. Sulphates are basic and indispensable building blocks of GAGs. With the exception of hyaluronate, all glycosaminoglycans contain sulphate groups in ester linkages with the hydroxyl groups of the amino sugar residues. Nevertheless, in future study detailed quantification of every GAG individually will be of interest. Summing up, the results presented in this work suggest that some of the non-steroidal anti-inflammatory drugs, and their combinations with isoflavone genistein, can be considered as a method for improvement of efficiency of therapy for mucopolysaccharidoses, especially their neurological forms. It is believed that our investigations can provide new information regarding the molecular mechanism of action of substances decreasing the synthesis of compounds being pathologically accumulated in the cells of people suffering from various types of genetically determined lysosomal storage diseases. On the other hand, proper choice of conditions of the drug applications are not trivial, and extensive studies are necessary for optimization of each particular combined treatment. Previous studies documented preliminary experiments with genistein-based SRT called GET-IT combined with ERT on MPS I culture cells as an effective approach 30 . Currently, the results of our studies on this kind of approach on MPS mouse models, conducted within the framework of another large research grant, are very promising. The same can be presumed on the development of a high efficacy of NSAID and ERT or HSCT combined therapies for MPSs, representing a promising method in indications of unmet medical needs. Understanding of these issues may result in serious progress in therapeutic approaches.