Significance of NF-kappaB signaling and PARP 1 activity in the TNF-induced inhibition of PHEX gene expression in human osteoblasts *

Although loss of bone mineral density is a common symptom of chronic inflammatory diseases, its mechanisms are still poorly understood. The PHEX gene encodes a Zn-endopeptidase expressed in osteoblasts and contributes to bone mineralization. Data derived from rodents has indicated co-repression of the PHEX gene by the NF-κB pathway and poly(ADP-ribose) polymerase 1 (PARP1). The aim of this study was to determine the molecular mechanism involved in TNF-mediated downregulation of PHEX expression in human osteoblasts and human osteosarcoma cell line. We observed that activation of the NF-κB pathway by TNF was manifested as a nuclear increase in RELA and NFKB1 heterodimer. We found that TNF reduced PHEX expression and the proteasome inhibitor reversed this effect in osteosarcoma cell line. Contrary to the effects seen in rodents, inhibition of PARP1 enzymatic activity did not significantly reverse the effect of TNF on the human PHEX gene expression. EMSA studies showed that the number of adenines in the PHEX proximal promoter is crucial for the transcription factors’ interactions within that region. The obtained results support the hypothesis indicating the existence of a molecular mechanism of gene repression that involves a poly adenine-rich region of the proximal gene promoters and PARP1 transcriptional activity.


INTRODUCTION
Decreased bone mineral density (BMD) is common in patients with chronic inflammatory conditions associated with increased levels of inflammatory mediators (Viswanathan & Sylvester, 2008).Osteoporosis, characterized by progressive loss of BMD, leads to an increased risk of fracture and is one of the most common bone diseases (Sponholtz et al., 2014).The lifetime risk of an osteoporotic fracture in the United States is estimated at almost 40% for women and approximately 13% for men (Johnell & Kanis, 2005).Moreover, osteoporosis-related fractures are associated with considerable morbidity and mortality in older people (Johnell & Kanis, 2004;Johnell et al., 2004).Both in vivo and in vitro studies have demonstrated that pro-inflammatory cytokines, such as interleukin-6 (IL-6), interleukin-1 beta (IL-1β), and tumor necrosis factor (TNF), influence the age-and estrogen-related decrease in BMD (Sponholtz et al., 2014).These immune factors have been shown to promote the proliferation, activity, and survival of osteoclasts while inhibiting the survival of osteoblasts (Clowes, Riggs, & Khosla, 2005).
The changes in the bone matrix are associated with the activity of three major cell types: bone forming osteoblasts (OBs) and osteocytes, and bone resorbing osteoclasts (OCs) (Teti & Zallone, 2009).The balance between the two opposite processes is the basis for maintaining bone homeostasis and may be easily upset, leading to a decrease in BMD.This also occurs during chronic autoinflammatory conditions, such as inflammatory bowel diseases (IBD).IBD patients have increased levels of proinflammatory cytokines, including TNF, and altered bone homeostasis (Viswanathan & Sylvester, 2008).The pathogenesis of osteopenia and osteoporosis in IBD has been suggested to result from altered rates of bone formation and resorption secondary to a multifactorial mechanism (Ghishan & Kiela, 2011), which includes two major categories: (1) poor nutritional status and malabsorption and (2) inflammation with IBD-associated inflammatory cytokines affecting the bone more directly.Studies conducted on IBD animal models indicated that malabsorption and inflammation are not two independent contributors to IBD-associated BMD loss, and that both inflammation and inflammatory cytokines actively contribute to the abnormalities in intestinal and renal mineral and vitamin absorption (Ghishan & Kiela, 2011).
The renal-skeletal-gut axis (Kiela & Ghishan, 2009) is one of the main bone homeostatic mechanisms.This axis regulates systemic levels of calcium (Ca 2+ ) and inorganic phosphate (Pi), which are important for systemic homeostasis, including bone formation.PHEX (phosphate-regulating gene with homologies to endopeptidases on the X chromosome), which encodes a Zn-metalloen-dopeptidase that is expressed primarily in OBs, osteocytes, and odontoblasts, plays a key role in this process (Kiela & Ghishan, 2009).A pivotal role of this gene in Pi homeostasis was established in vitamin D-resistant, X-linked hypophosphatemic rickets (XLH) (Du et al., 1996;F. Francis, 1995).Inactivating mutations in the PHEX gene in this familial disorder results in hypophosphatemia, aberrant vitamin D levels, high level of serum alkaline phosphatase and osteomalacia.PHEX gene mutation is associated with decreased expression and activity of the Na + /P i cotransporter NaPi-IIa (NPT2; SLC34A1) responsible for phosphate re-absorption in the renal proximal tubules (Collins & Ghishan, 1994;Dixon et al., 1998;Drezner, 2000;Hernando, Biber, Forster, & Murer, 2005;Murer, Hernando, Forster, & Biber, 2003), which are also pivotal elements of the renal-skeletal-gut axis.PHEX affects renal Pi re-absorption via an indirect action of phosphaturic factors termed phosphatonins.The fibroblast growth factor 23 (FGF-23) is one of the most extensively researched phosphatonin known to inhibit renal Pi handling and bone mineralization (Kiela & Ghishan, 2009), and it has been hypothesized that PHEX inactivates FGF-23 via proteolytic degradation (Bowe et al., 2001).
The nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) signaling pathway plays a key role in inflammation and immune responses.In OCs NF-κB is activated during inflammation leading to increased bone resorption.While the actions of the NF-κB signaling pathway in osteoclasts are well understood (Novack, 2011), less is known about this pathway in OBs, and about its contribution to the BMD regulation.To examine the effects of NF-κB on OBs, transgenic mice expressing a dominant negative form of the inhibitor of nuclear factor kappa B kinase subunit gamma (IKK-DN) were generated (Chang et al., 2009).Young transgenic mice expressing an IKK-DN in OBs had normal numbers of OBs, but increased trabecular bone mass, bone formation rates, and expression of OB marker genes, although this effect did not persist into adulthood (Chang et al., 2009).Young IKK-DN transgenic mice showed enhanced c-Jun N-terminal kinases (JNK) activation and increased expression of fos-related antigen 1 (Fra1), a protein that was required for the stimulatory effects of NF-κB blockade in OBs.
During chronic inflammation such as IBD, circulating and/or infiltrating lymphocytes and other mononuclear cells produce cytokines that may influence bone metabolism by altering the balance of bone mineral deposition and resorption.Decreased BMD is a common outcome of IBD: 31 to 59% of adult IBD patients are classified as osteopenic, whereas 5 to 41% are actually diagnosed with osteoporosis, although the rates of up to 70% of adult and pediatric IBD patients with low BMD have been reported (Rodriguez-Bores, Barahona-Garrido, & Yamamoto-Furusho, 2007).In murine experimental colitis, bone PHEX expression was inhibited by TNF and this effect correlated with reduced OBs mineralization in vitro (Uno et al., 2006).TNF treatment and chemically induced colitis decreased murine Phex mRNA expression via a transcriptional mechanism, and the adenine-reach region (PolyA) of the promoter located -116 to -110 nt upstream from the transcriptional start site (TSS) was necessary for the TNF-mediated inhibition (Uno et al., 2006).Further work revealed that murine Phex promoter constitutively binds poly(ADP-Ribose) polymerase 1 (PARP1) in the PolyA region and after TNF treatment RELA proto-oncogene NF-κB subunit (RELA, synonym: p65), a component of the NF-κB pathway, is poly(AD-P)-ribosylated (PARylated) by PARP1 to increase its activity as a trans-repressor (Majewski, Thurston, Ramalingam, Kiela, & Ghishan, 2010).Moreover, PARP1 activity was indispensable for the effects of TNF, as demonstrated by a blunted response to the cytokine in the presence of a PARP1 inhibitor or overexpressed poly(ADP-ribose) glycohydrolase (PARG), an enzyme responsible for de-PARylation, and by a complete abrogation of the response to TNF in PARP1 knockout mice (Majewski et al., 2010).These discoveries indicated a novel cooperative mechanism of TNF mediated gene regulation in OBs involving NF-κB and PARP1.
PARP1 protein is involved in numerous important processes at the cellular level, e.g.continuous monitoring of DNA integrity to initiate its repair.Therefore, PARP1 frequently interacts with DNA, and an increase in its positioning close to transcription start sites has also been documented (Krishnakumar et al., 2008).Indeed, recent microarray analysis of the colonic tissue showed extensive transcriptional reprogramming (Larmonier et al., 2016), thus confirming important role of PARP1 as a transcriptional modulator.Both, PARP1 and NF-κB pathway have been suggested to play a crucial role in inflammatory disorders and other reports have also pointed at PARP1 as a coactivator of the NF-κB pathway (Hassa & Hottiger, 2002).
Considering both conserved and divergent elements in mouse and human PHEX gene promoters (human PHEX proximal promoter region does not possess a long PolyA region characteristic for mice Phex; Fig. 1), we aimed to verify the cooperativity of PARP1 and NF-κB pathway in osteoblasts isolated from adult human bones (HOb) as well as in cells derived from human osteosarcoma (MG-63).
Real-Time RT-PCR analysis.One mg of total RNA isolated with TRIzol was reverse-transcribed using the Maxima First Strand cDNA Synthesis Kit for RT-qP-CR and 25 ng or 100 ng of cDNA, regardless of experiments, were used for real time PCR analysis using commercially available primers and TaqMan probes (Hs 02758991_g1 for GAPDH and Hs 01011692_m1 for PHEX), Luminaris Color Probe qPCR Master Mix and the StepOne Real-Time PCR System.The resulting data were analyzed using the comparative cycle threshold (Ct) method as means of relative quantification of gene expression, normalized to an endogenous reference glyceraldehyde-3-phospate dehydrogenase (GAPDH) and relative to a calibrator (normalized Ct value obtained from control osteoblasts), and expressed as 2 -ΔΔ Ct (Applied Biosystems User Bulletin number 2: Rev B "Relative Quantitation of Gene Expression").
Transient transfections.HOb and MG-63 cells were seeded in 6 or 24-well plates (0.5×10 6 or 0.5×10 5 cells/well, respectively), and cells were transfected with pNiFty2-Luc reporter vector (1 μg/ml of medium) using the TransIT-LT1 transfection reagent (3 μg/ml).pNiF-ty2-Luc activity was estimated using Dual-Glo Luciferase Assay System and expressed as luciferase activity (RLU, relative light units) per 10 μl of cell lysate.In some experiments, the cells were transfected with the expression vectors pFLAG-CMV2-RELA or/and pFLAG-CMV2-NFKB1 (0.5 μg/ml of medium), or the respective control pCMV2-null.After 15-24 h post transfection, the cells were used for total RNA isolation, and PHEX mRNA level was estimated as described in the "Real-Time RT-PCR Analysis" section.
Electrophoretic Mobility Shift Assay (EMSA).Nuclear protein (NP) for EMSA was prepared from the MG-63 cells as previously described (Slomiany, Kelly, & Kurtz, 2000).Double-stranded, synthetic oligonucleotides were end-labeled with [γ 32 P]dATP and cleaned using Mini Quick Spin DNA Column, according to the manufacturers' protocols.For each reaction, 30 000 cpm of the probe were incubated at room temperature for 20 min with 5-10 μg of NP, 4 μl of Gel Shift Binding 5× Buffer, 1 μg of poly(d(I-C)), and H 2 O to a final volume of 20 μl.For competition studies, a 100 excess of unlabeled probe was added to the reaction.Antibodies for supershift assays (NFKB1, NFKB2, RELA and REL) were purchased from Santa Cruz Biotechnology (CAT: sc-7178, sc-298X, sc-372X and sc-272X, respectively).Binding reactions were loaded onto a 6% DNA retardation gel and were separated at 250 volts in 0.5 TBE.Gels were dried and then exposed to an X-ray film.The oligonucleotides used for the EMSA are depicted in Table 1.
Statistical analysis.The results were presented as mean values ± S.E. for each parameter.Unless otherwise indicated, there were six observations in each group.Influence of TNF on mean levels of relative NF-κB activation were compared using Welch two sample t-test for HOb and MG-63 cells separately.Two-way ANOVA was used to estimate effects of TNF and cLβL, TNF and 3-AB, RELA and nuclear factor kappa B subunit 1 (NFKB1, synonym: p50) on relative expression of PHEX.Marginal tests were performed (type III sum of squares) and interaction between main factors was assessed.In all the tests we used significance level α=0.05.Assumptions of normality and variance equality were checked using Shapiro-Wilk and Bartlett tests, respectively.The calculations were performed in R ver.3.3.1 (R Core Team, 2016).

TNF increase the NF-κB activity in human osteoblasts
To confirm the activation of NF-κB pathway under our experimental conditions, we transfected cells with NF-κB-Luc luciferase reporter vector, pNiFty2-Luc.Cells were transfected in 24 well plates upon reaching ~50% confluency, and the media were changed 15-24 h post-transfection to one without or with TNF (20 ng/ ml).24 hours later, the cells were lysed and analyzed for luciferase activity.The t-test analyses indicated significant increase of NF-κB activity in TNF-treated HOb (Fig. 2A) and MG-63 osteoblasts (Fig. 2B) in comparison to untreated cells (CTRL) (p<0.001;95% confidence intervals 1769-2541 and 2821-4079, respectively).After TNF-treatment of HOb and MG-63 cells we found 4to 10-fold increase in the activity of NF-κB pathway, respectively.
To identify the key components of the NF-κB complex activated by TNF, we performed electromobility shift assays (EMSA).MG-63 cells at ~80% confluency were treated with TNF (20 ng/ml) and incubated for 24 hours.Nuclear protein (NP) was extracted from control and TNF treated osteoblasts as previously described (Slomiany et al., 2000).Double-stranded, commercially available NF-κB consensus oligonucleotides were end-labeled with [γ 32 P] dATP and incubated with NP together with (Super shift) or without (EMSA) the antibodies recognizing individual NF-κB effector proteins.TNF induced a formation of protein/DNA complexes as shown in the picture (Fig. 3A).Incubation of NP isolated from TNF-treated osteoblasts with antibodies (Fig. 3B) demonstrated that complexes binding with NF-κB consensus oligo consisted mainly of RELA (synonym: p65) and NFKB1 (synonym: p50).Supplementation of EMSA reaction with antibodies targeting nuclear factor kappa B subunit 2 (NFKB2, synonym: p52), REL proto-oncogene NF-κB subunit (REL, synonym: c-REL), or control IgG did not affect protein/DNA complex formation, thus indicating that in response to TNF stimulation, human osteoblasts activate primarily the classical NFKB1/ RELA-dependent pathway.

TNF-mediated PHEX inhibition is NF-κB-dependent
The aim of this experiment was to study the mechanism of TNF influence on the PHEX gene expression.To this end, cells that achieved approximately 80% confluency in the cell culture were treated with TNF (20 ng/ml) and incubated for 24 hours.To test whether NF-κB activation is required for the inhibitory effects on the PHEX mRNA level, cells were pre-treated with clasto-lactacystin β-lactone (cLβL, 10 μM), an irreversible proteasome inhibitor.Proteasome inhibition stabilizes IκBα (nuclear factor of kappa light polypeptide gene enhancer in B-cells inhibitor alpha; NFKBIA), a natural inhibitor of nuclear translocation of the NF-κB complex.For both HOb (Fig. 4A) and MG-63 (Fig. 4B), the average level of PHEX mRNA was significantly different from zero (p<0.001) in the control groups.TNF significantly decreased the PHEX mRNA level compared to control for both HOb (p<0.05) and MG-63 (p<0.001).However, the effect of cLβL was cell line dependent: the level of PHEX mRNA increased significantly in HOb (p<0.001) upon stimulation but remained the same in MG-63 cells.Yet, in MG-63 cells, we observed a strong interaction between TNF and cLβL (p<0.001),i.e. pretreatment with cLβL reversed the inhibitory effect of TNF on PHEX mRNA level, which increased by 46% in the cLβL and TNF pre-treated cells, compared to 87% in the cells pre-treated with TNF only.
To test the effects of RELA and NFKB1 proteins on PHEX expression, HOb and MG-63 cells were seeded in 6-well plates and transfected with pFLAG-CMV2-  RELA or/and pFLAG-CMV2-NFKB1, or the control vector pCMV2-null (CTRL).24 h after transfection, total RNA was isolated and PHEX mRNA expression was estimated with RT-qPCR.The statistical analysis did not find any differences in PHEX mRNA level between control and either HOb (Fig. 5A) or MG-63 (Fig. 5B) cells, regardless of RELA, NFKB1 or NFKB1/RELA overexpression.

PARylation inhibitor 3-AB does not reverse the effects of TNF on PHEX expression
To test if PARP1 enzymatic activity modulates the effect of TNF (20 ng/ml) on the PHEX mRNA level, osteoblasts at ~80% confluency were pre-treated with PARP inhibitor 3-aminobenzamide (3-AB, 1 mM).Both in HOb (Fig. 6A) and in MG-63 (Fig. 6B) cells, mRNA level of PHEX was significantly different from zero (p<0.001) in the control groups.TNF treatment significantly decreased PHEX mRNA level, as compared to control, in both HOb (p<0.01) and MG-63 (p<0.001).Pre-treatment with 3-AB neither changed the level of PHEX mRNA when applied alone, nor reversed the inhibitory effect of TNF in both cell lines.

Differential protein/DNA complex formation at the proximal mouse, rat, and human Phex/PHEX promoter after TNF treatment
The aim of this experiment was to study the potential of species-specific Phex/PHEX proximal promoter segments in binding transcription factors using electromobility shift assay (EMSA).To this end, MG-63 cells that achieved approximately 80% confluency were treated with TNF (20 ng/ml) and incubated for 24 hours.The nuclear proteins (NP) was isolated and used for a   test with double-stranded oligonucleotides that were complementary for analogous proximal promoter regions of mice (mPhex), rat (rPhex) and human (hPHEX), containing (mouse, rat) or not (human) adenine-rich regions (PolyA), as well as oligonucleotides complementary to different lengths of human PHEX promoter: hPHEX(1) and hPHEX( 2).The sequences are presented in Table 1.Double-stranded oligonucleotides were end-labeled with [γ 32 P]dATP, incubated with NP (10 μg), and divided on the retardation gel.We found that the length of the PolyA region in the promoters determined the level of migrating of the protein-DNA complexes (Fig. 7A).The EMSA signal was the highest for the mPhex promoter, lower for the rPhex promoter, and the hPhex promoter did not bind to any proteins (Fig. 7A).Comparison of the signal detected in EMSA coming from mPhex and the two different hPHEX parts of the proximal promoter, additionally confirmed the capability of the PolyA region of binding NP (Fig. 7B).The EMSA signal was detected for the mPhex promoter, but both different hPHEX(1) and hPHEX(2) parts of the promoters did not bind any proteins in the same experimental conditions.

DISCUSSION
We found that TNF reduces the expression of the human PHEX gen in both HOb and MG-63 osteoblasts, and these results are in concordance with those obtained after TNF treatment of a rat osteosarcoma cell line (UMR-106) (Majewski et al., 2010).Additionally, other published data coming from the rodent cells confirm this observation.Recently, it was shown that in differentiated mouse osteocyte-like IDG-SW3 cells, Phex mRNA levels decreased after treatment with the TNF, IL-1β or LPS (Ito et al., 2015).It should be noted that in the preliminary dose-response experiments (not shown), we had determined that it was necessary to use 2 to 4 higher concentration of TNF (20 ng/ml) to elicit the same effect on the PHEX expression in human osteoblasts, compared to the rat UMR-106 (10 ng/ml) or mouse IDG-SW3 (5 ng/ml) cells, respectively.Thus, we subsequently used this higher concentration of TNF in all the experiments conducted on HOb and MG-63 cells.The difference between the required dose in the tested cell systems may be related to species-specific expression of the TNF-receptors, and nonsignaling (decoy) TNF re-ceptors, or may be simply a unique characteristic of each cell line.It was previously confirmed that human osteoblasts derived from mesenchymal stem cells and MG-63 human osteosarcoma cells, used in our study, have similar complement of receptors for TNF (Bu et al., 2003).
Our data demonstrated that the NF-κB signaling pathway is involved in TNF-induced repression of human PHEX gene.Using the pNiFty2-Luc reporter vector with five NF-κΒ repeated transcription factor binding sites, we confirmed the increase in the NF-κB signaling pathway activity in TNF-treated OBs.Moreover, we found that the inhibition of proteasome activity by pretreatment of OBs with cLβL reversed the effects of TNF on PHEX mRNA expression.These results strongly correspond with data derived from rat UMR-106 cells (Majewski et al., 2010).Interestingly, our data not only indicate that inhibition of NF-κB signaling pathway activity reversed the effect of TNF, but also indicate that human PHEX can be under constitutive repression of this pathway.In HOb cells PHEX expression increased significantly after treatment with cLβL in both control and TNF treated groups, when compared to the other groups.These effects were not visible in MG-63 cells.It is important to note that MG-63 cells are osteosarcoma cells derived from malignant bone tumors.They share some osteoblastic features (Declercq et al., 2004), but their chromosomal modifications lead to unusual cellular and molecular functions (Al-Romaih et al., 2003).While osteosarcoma-derived cells are generally used as osteoblastic models, their molecular malfunctions are not yet fully studied.Ultimately, the expression of extracellular matrix proteins contributing to the osteosarcoma-osteoid is changed, hence the differences compared to normal OBs (Benayahu, Shur, Marom, Meller, & Issakov, 2001).In EMSA and supershift analyses, we identified NF-κB pathway components, which levels increased in OBs nuclei following the TNF treatment.EMSA from TNF-treated MG-63 cells with NF-κB consensus probe showed an increased formation of DNA/protein complexes, which consisted predominantly of NFKB1 and RELA, highlighting the dominant role of its heterodimer in the response to TNF in human OBs.These data are consistent with our previous data derived from experiments conducted on rat UMR-106 cells (Majewski et al., 2010).However, in experiments with overexpression of NFKB1 and RELA in human OBs, we did not observe a suppressive effect on PHEX mRNA expression.This contrasts with the rat Phex expression, which was inhibited by the overexpression of both RELA alone and RELA with NFKB1 (Majewski et al., 2010).
Considering the previously reported importance of PARP1 in regulating Phex expression in rodents (Majewski et al., 2010), we investigated whether inhibition of PARP1 enzymatic activity by 3-AB would alter the response to TNF.Interestingly, 3-AB did not abrogate the effects of TNF on the expression of PHEX mRNA in the human osteoblasts in previously effective dose (1 mM).It suggested that the role of PARP1 in the regulation of human PHEX promoter activity may be less important than in rodents.
Although PARP1 is an abundant and ubiquitous nuclear enzyme that was originally identified as a key factor in the DNA repair pathway, it has now been shown to positively and negatively affect gene transcription and chromatin structure under both basal and signal-activated conditions (Nguewa et al., 2005).Investigations of gene expression profiles in PARP1-deficient embryonic stem cells and in liver cells from PARP-1 KO mice demonstrated that 3.5% of the transcriptome was regulated by  1. Effects of TNF on the expression of the PHEX in human osteoblasts PARP1, as 2.4 % were positively regulated (Ogino et al., 2007).Moreover, PARP1 has been described as one of the major molecules involved in the propagation of inflammatory stimuli and has been proposed as a target for anti-inflammatory treatment (Peralta-Leal et al., 2009).PARP1 has been shown to affect gene transcription in several ways: (1) it may act as an enhancer-binding factor similar to classical sequence-specific DNA-binding activators or repressors; (2) it may work as a transcriptional co-regulator or (3) it may act as a modifying enzyme that catalyzes the NAD-dependent addition of PARylation to numerous nuclear proteins.In the experiments conducted on mice and in UMR-106 cells, it was demonstrated that in TNF-treated OBs RELA is translocated to the nucleus, where it is PARylated by PARP1 (Majewski et al., 2010).This chemical modification is critical for TNF-induced inhibition of rodent Phex expression, and this response does not occur in PARP1-deficient mice and is reversed by the PARP1 inhibitor 3-AB or by overexpression of PARG (Majewski et al., 2010).According to this study, TNF treatment of UMR-106 resulted in binding of the NFKB1/RELA complex to the two sites located between the transcription start site and the PolyA region of the Phex gene promoter (-76/-57 and -35/-16).The same region identified as the binding region for the NFKB1/RELA heterodimer was found in the human PHEX promoter; thus, it is possible that the human NFKB1/RELA heterodimer acts in the same way as in the rodent Phex promoter.However, additional investigations are necessary to confirm this.Using both molecular and proteomic methods, it was shown that PARP1 enzyme binds to the PolyA region in rodents and, under the influence of TNF, PARylates RELA subunits, thereby increasing their affinity for the proximal Phex promoter, and consequently inhibiting RNA polymerase II complex (Majewski et al., 2010).The human PHEX proximal promoter region does not possess a long PolyA region characteristic for mice Phex (Fig. 1).This difference in the promoter structures provides the opportunity to verify the hypothesis of the increased importance of PARP1, which has been demonstrated to act as a regular transcription factor that depends on proximal promoter organization.Current EMSA studies have shown that the number of adenines in the PolyA regions of promoters from three examined species is crucial for the level of protein interactions with this region of the gene.It seems to be very interesting whether other genes containing PolyA regions that are located close to the TSS region may be regulated by the described mechanism.
In conclusion, TNF inhibits PHEX gene expression in human OBs via a mechanism only partially consistent with rodent Phex gene.It probably requires a NFKB1/ RELA NF-κB heterodimer, but it does not rely on PARP1 activity, as we previously reported for rat and mouse.Differences in the organization of the PHEX gene promoters and different requirements for the NF-κB complex likely explain these divergent mechanisms.

Figure 1 .
Figure 1.In silico analysis of the promoter regions of human PHEX (hPHEX), mouse Phex (mPhex) and rat Phex (rPhex).The analysis included 133 nucleotides upstream the transcription start site (+1).Nucleotide sequences in the brackets indicate the binding sites for the transcription factors of the NF-κB pathway proteins; the AP1 binding site and poly adenine-rich region (PolyA) are marked.

Figure 2 .
Figure 2. The activity of the NF-κB signaling pathway in the control (CTRL) and TNF treated HOb (A) or MG-63 (B) osteoblasts.The charts show mean (n=6) ± standard error.Statistical significance is indicated in the results section.

Figure 3 .
Figure 3. Electrophoretic mobility shift assay (EMSA) of the nuclear proteins isolated from MG-63 cells with (TNF) or without (CTRL) 24 h TNF treatment (A).Supershift/blocking EMSA analysis of the NF-κB/DNA complexes with extra anti-NF-κB pathway antibody (B).The probe was NF-κB consensus (Promega, sequence: AGTT-GAGGGGACTTTCCCAGG). (A) The picture shows specific (NF-κB) and nonspecific (ns) migrating protein/DNA complexes.A longer electrophoresis and application of the anti-NF-κB antibody (B) enabled for identification of the main effectors of NF-κB pathway (RELA and NFKB1) migrating to the nuclei of the osteoblasts after TNF treatment.

Figure 4 .
Figure 4.The relative level of Phex mRNA in (A) HOb and (B) MG-63 cells shown as mean (n=5) ± standard error.The level of mRNA in the CTRL group was normalized to 1.Both control (CTRL) and TNF treated osteoblasts (TNF) were additionally pre-treated with the proteasome inhibitor clasto-Lactacystin β-lactone (cLβL).Statistical significance is indicated in the results section.

Figure 5 .
Figure 5.The relative level of Phex mRNA in (A) HOb and (B) MG-63 cells shown as mean (n=6) ± standard error.The level of mRNA in the CTRL and cells overexpressing RELA, NFKB1 or both RELA and NFKB1 (group NFKB1/RELA).Overexpression was obtained by transfection of cells with expression plasmids pFLAG-CMV-RELA and pFLAG-CMV2-NFKB1.The level of mRNA in the control group (transfection with control plasmid pC-MV2-null) was normalized to 1. Statistical significance is indicated in the results section.

Figure 6 .
Figure 6.The relative level of Phex mRNA in (A) HOb and (B) MG-63 cells shown as mean (n=6) ± standard error.The level of mRNA in the CTRL group was normalized to 1.Control (CTRL) and TNF treated osteoblasts (TNF) were additionally pre-treated with the inhibitor of poly(ADP) ribosylation (PARylation) 3-aminobenzamide (3-AB).Statistical significance is indicated in the results section.

Figure 7 .
Figure 7. Electrophoretic mobility shift assay (EMSA) of the nuclear proteins isolated from TNF-treated MG-63 cells after 24 h.(A) EMSA analysis of protein-DNA complexes with part of human (hPHEX), rat (rPhex) and mouse (mPhex) proximal promoter.(B) EMSA analysis of the protein-DNA complexes with two human PHEX(1), PHEX(2) promoters differing in length, and a mouse (mPhex) promoter.Probe sequences used in the EMSA are presented in Table1.