Pentoxifylline-induced protein expression change in RAW 264.7 cells as determined by immunoprecipitation-based high performance liquid chromatography

Although pentoxifylline (PTX) was identified as a competitive non-selective phosphodiesterase inhibitor, its pharmacological effect has not been clearly elucidated. The present study explored the effect of low dose 10 μg/mL PTX (therapeutic dose) compared to high dose 300 μg/mL PTX (experimental dose) in RAW 264.7 cells through immunoprecipitation-based high performance liquid chromatography (IP-HPLC), immunohistochemistry, and western blot. 10 μg/mL PTX increased the expression of proliferation (Ki-67, PCNA, cyclin D2, cdc25A), epigenetic modification (KDM4D, PCAF, HMGB1), protein translation (DOHH, DHPS, eIF5A1), RAS signaling (KRAS, pAKT1/2/3, PI3K), NFkB signaling (NFkB, GADD45, p38), protection (HSP70, SOD1, GSTO1/2), survival (pAKT1/2/3, SP1, sirtuin 6), neuromuscular differentiation (NSEγ, myosin-1a, desmin), osteoblastic differentiation (BMP2, RUNX2, osterix), acute inflammation (TNFα, IL-1, CXCR4), innate immunity (β-defensin 1, lactoferrin, TLR-3, -4), cell-mediated immunity (CD4, CD8, CD80), while decreased the expression of ER stress (eIF2α, eIF2AK3, ATF6α), fibrosis (FGF2, CTGF, collagen 3A1), and chronic inflammation (CD68, MMP-2, -3, COX2) versus the untreated controls. The activation of proliferation by 10 μg/mL PTX was also supported by the increase of cMyc-MAX heterodimer and β-catenin-TCF1 complex in double IP-HPLC. 10 μg/mL PTX enhanced FAS-mediated apoptosis but diminished p53-mediated apoptosis, and downregulated many angiogenesis proteins (angiogenin, VEGF-A, and FLT4), but upregulated HIF1α, VEGFR2, and CMG2 reactively. Whereas, 300 μg/mL PTX consistently decreased proliferation, epigenetic modification, RAS and NFkB signaling, neuromuscular and osteoblastic differentiation, but increased apoptosis, ER stress, and fibrosis compared to 10 μg/mL PTX. These data suggest PTX has different biological effect on RWA 264.7 cells depending on the concentration of 10 μg/mL and 300 μg/mL PTX. The low dose 10 μg/mL PTX enhanced RAS/NFkB signaling, proliferation, differentiation, and inflammation, particularly, it stimulated neuromuscular and osteoblastic differentiation, innate immunity, and cell-mediated immunity, but attenuated ER stress, fibrosis, angiogenesis, and chronic inflammation, while the high dose 300 μg/mL PTX was found to alleviate the 10 μg/mL PTX-induced biological effects, resulted in the suppression of RAS/NFkB signaling, proliferation, neuromuscular and osteoblastic differentiation, and inflammation.

Introduction Pentoxifylline (PTX), a xanthine derivative, is primarily used as an antiproteolytic agent to treat muscle pain in people with peripheral artery disease [1,2] by activating cAMP/EPAC/ AKT signaling [3,4]. It has been frequently reported that PTX remarkably suppressed the secretions of pro-inflammatory cytokines and the nuclear factor-kappa B (NFkB) activation [5][6][7], and reduced chronic inflammation [5,8]. PTX appeared to have anti-fibrotic effect on radiation-induced lung fibrosis by modulation of PKA and PAI-1 expression as possible antifibrotic mechanisms [9], and PTX therapy with vitamin E showed prevention of radiationinduced fibrosis in breast cancer patients [10]. PTX is suggested as an oral osteogenic drug for the treatment of post-menopausal osteoporosis [11]. As PTX given before tooth extraction is prophylactic, it might affect healing in a positive way by optimizing the inflammatory response [12]. Many authors suggest that PTX may increase the anticancer potential of anticancer drugs such as cisplatin or doxorubicin as well as reduce side effects of these drugs [13][14][15][16].
As RAW 264.7 cells are immortalized macrophages which are mainly involved with wound healing and tumor progression, the present study utilized RAW 264.7 cells for in vitro protein expression experiment. Although PTX has short half-life (0.39-0.84 h for the various doses and 0.96-1.61 h for the metabolites), its therapeutic dose for adult human is usually 400 mg (Trental), three times a day [17,18]. Therefore, in this study, RAW 264.7 cells were primarily treated with 10 μg/mL PTX, which is similar to human therapeutic dose (6.7 mg/kg, Trental). However, in the pilot study to know the trends of protein expressions by PTX, 10 μg/mL PTX increased the expression of some proliferation-related proteins, RAS and NFkB signaling proteins, and even some inflammatory proteins in RAW 264.7 cells. These results were contrary to many reports insisting the anti-proliferative and anti-inflammatory effect of PTX. However, it was found that many experiments for PTX-induced effects on cells and animals were frequently performed by using higher dose of PTX, 100-500 μg/mL [5,8,[19][20][21][22], than therapeutic dose of PTX, about 10 μg/mL. In order to elucidate the different pharmacological effect depending on the dose of PTX, the present study was performed to compare 10 μg/mL PTXinduced protein expressions with 300 μg/mL PTX-induced protein expression in RAW 264.7 cells.
As the essential protein signalings are intimately correlated and cross-talked with each other to maintain cellular homeostasis during proliferation, differentiation, inflammation, apoptosis, and senescence, it is necessary to know the global protein expression involving multiple signaling pathways in order to explain or predict the fate of cells involved with diseases or drug therapy. The present study examined PTX-induced global protein expression changes in RAW 264.7 cells through IP-HPLC, immunocytochemistry (ICC), and western blot. Particularly, IP-HPLC analysis is available to determine protein expression levels in different biological fluids, such as blood plasma, urine, saliva [23,24], inflammatory exudates [25][26][27], cancer tissues [28,29], cell culture extract [30][31][32][33][34], and blood plasma [32,35]. Contrary to enzymelinked immunosorbent assay (ELISA). IP-HPLC uses protein A/G agarose beads in chaotic buffer solution and micro-sensitive UV spectroscopy to determine protein expression level

Immunocytochemical staining analysis
When approximately 70% confluent RAW 264.7 cells were spread over the surfaces of twowell culture slide dishes, the cells were treated with 10 μg/mL PTX for 12, 24, or 48 h, while the control cells were treated with 100 μL of normal saline. The cells on the culture slides were fixed with 4% paraformaldehyde solution for 20 min, permeabilized with cooled methanol for 10 min at -20˚C, and applied for immunohistochemistry using selected antisera (the same ones used in IP-HPLC, Table 1); Ki-67 for cellular proliferation, KMD4D and PCAF for epigenetic modification, TNFα, IL-6, TLR3, and TLR4 for inflammation, GSTO1/2, LC3, and GADD153 (CHOP) for endoplasmic reticulum stress, PARP-1 and caspase 3 for apoptosis, NSEγ for neural differentiation, MYH2 for muscular differentiation, TGF-β1, RUNX2, OPG, and BMPR2 for osteoblastic differentiation.
The cells treated with 10 μg/mL and 300 μg/mL PTX for 0, 12, 24, and 48 h were collected with phosphate-buffered saline (PBS) separately, treated with trypsin-ethylene-diamine-tetraacetic acid (trypsin-EDTA) for one minute, and washed with PBS, and followed by cell lysis with ice-cold RIPA buffer (Sigma Aldrich, USA). The lysates were centrifuged at 12,000 g for 20 min at 4˚C. The protein concentration of the supernatant was quantified using a Bradford assay (BioRad, USA). Equal amounts (30 μg/lane) of the sample proteins were separated by 8, 10, 15, or 20% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) in Tris-glycine SDS running buffer (25 mM Tris, 0.1% SDS, and 0.2M glycine), and transferred to a nitrocellulose membrane. The membranes were blocked with 5% nonfat dry milk in TBST buffer (25 mM Tris-HCl, 150 mM NaCl, 0.1% Tween 20, pH 7.5) for 1 h. After washing three times with TBST buffer, the membrane was incubated with each primary antibody (dilution ratio = 1:1000, the same antibody used in IP-HPLC) and horseradish peroxidase-conjugated secondary antibody for 1 h separately. The protein bands were then detected using an enhanced chemiluminescence system (Amersham Pharmacia Biotech, Piscataway, NJ, USA) and digitally imaged using a ChemiDoc XRS system (Bio-Rad Laboratories, Hercules, CA, USA). The level of β-actin expression was used as an internal control to normalize the expression of the target proteins. The size and intensity of protein bands from the cells treated with 10 μg/mL PTX for 0, 12, 24, and 48 h were demonstrated, and the protein bands from 10 μg/ mL PTX-treated cells were compared with those from 300 μg/mL PTX-treated cells.
The supernatant of the antibody-incubated column was removed, and followed by immunoprecipitation-based IP-HPLC. Briefly, each protein sample was mixed with 5 mL of binding buffer (150mM NaCl, 10mM Tris pH 7.4, 1mM EDTA, 1mM EGTA, 0.2mM sodium vanadate, 0.2mM PMSF and 0.5% NP-40) and incubated in the antibody-bound protein A/G agarose bead column on a rotating stirrer at room temperature for 1 h. After multiple washing of the columns with Tris-NaCl buffer, pH 7.5, in a graded NaCl concentration (0.15-0.3M), the target proteins were eluted with 300μL of IgG elution buffer (Pierce, USA). The immunoprecipitated proteins were analyzed using a precision HPLC unit (1100 series, Agilent, Santa Clara, CA, USA) equipped with a reverse-phase column and a micro-analytical UV detector system (SG Highteco, Hanam, Korea). Column elution was performed using 0.15M NaCl/20% acetonitrile solution at 0.5 mL/min for 15 min, 40˚C, and the proteins were detected using a UV spectrometer at 280 nm. The control and experimental samples were run sequentially to allow comparisons. For IP-HPLC, the whole protein peak areas (mAU � s) were obtained and calculated mathematically using an analytical algorithm (see S1 Fig) by subtracting the negative control antibody peak areas, and protein expression levels were compared and normalized using the square roots of protein peak areas. The ratios of the protein levels between the experimental and control groups were plotted into line and star graphs. Protein expressional changes of less than ±5%, ±5-10%, ±10-20%, or over ±20% changes were described as minimal, slight, significant, or marked, respectively [30][31][32][33]40]. The housekeeping proteins including β-actin, α-tubulin, and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) were simultaneously used as internal controls.
In the previous study, the IP-HPLC results were compared with western blot data using cytoplasmic housekeeping protein (β-actin), the former showed minute error ranges less than ± 5% which were appropriate for statistical analysis, while the latter showed a large error range of more than 20% which were impossible to be analyzed statistically [40] (see S2 Fig). Therefore, the present study mainly performed IP-HPLC, and its results were compared to representative findings of ICC and western blot performed with some selected antisera, even though ICC and western blot are usually involved with great error range (�20%).

Double IP-HPLC
The double IP-HPLC was designed to detect a protein complex or a binding body contained two different proteins. The first IP-HPLC was performed using the first antibody against one protein of complex as the above procedures to get 300 μL protein elute, which was applied as a protein sample in the second IP-HPLC. Subsequently, the second IP-HPLC was performed using the second column containing protein A/G beads bound with the second antibody against the other protein of complex. The protein elute sample was incubated with the protein A/G beads in the second column, and followed the same procedures of IP-HPLC described above.

Global protein expression indexes
As the overexpression and under-expression of essential 150 proteins observed in this study showed characteristics of 21 cellular functions affected by 10 μg/mL PTX. The maximum expression value (%) of upregulated proteins and the minimum protein expression values (%) of downregulated proteins at 12, 24, 48 h after 10 μg/mL PTX treatment were selected and plotted into a star graph.

Statistical analysis
Proportional data (%) of the experimental and control groups were plotted on line graphs and star plots. Their analyses were repeated two to six times until the standard deviations reached �±5%. The line graphs revealed time-dependent expression changes between the relevant proteins, and the star plots revealed the different expression levels of all proteins examined. The results were analyzed by measuring the standard error (s ¼ � ffi ffi ffiffi ). The expression of the control housekeeping proteins, i.e., β-actin, α-tubulin, and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was non-responsive (�5%) to 12, 24, or 48 h of PTX treatment [40,41] (see S1 File).

Proliferation by cytological cell counting assay
10 μg/mL PTX-treated RAW 264.7 cells were evenly spread on two-well culture slide dishes and cultured for 48 h. Their monotonous small round nuclei were stained with hematoxylin, and then well distinguishable under microscope. They were increased in cell number depending on time, at 0, 12, 24, and 48 h (Fig 1A-1D).

Immunocytochemical observation
The characteristic protein expressions were observed in RAW 264.7 cells through immunocytochemical (ICC) staining, that is, Ki-67, a marker of proliferation was strongly positive in 10 μg/mL PTX-treated cells at 12, 24, and 48 h compared to the untreated controls, and KDM4D and PCAF, markers of histone demethylation and acetylation, respectively, were strongly positive at 12 and 24 h. For the immunoreaction of inflammatory proteins, 10 μg/mL PTX-treated cells showed stronger positivity of TNFα, TLR3, and TLR4 at 24 and 48 h compared to the untreated controls, and increased positivity of IL-6 at 12, 24, and 48 h (Fig 2). 10 μg/mL PTX-treated cells showed slight increase of immunoreaction for GSTO1/2 (a marker of antioxidant and cellular stress) and LC3β (a marker of autophagosome biogenesis), compared to the untreated controls. Caspase 3, a marker of apoptosis executioner was markedly positive in 10 μg/mL PTX-treated cells at 12, 24, and 48 h, while the immunoreaction of GADD153 (CHOP, a marker of endoplasmic reticulum stress) and PARP-1 (a marker of DNA damage) was almost similar in the experimental and control cells (Fig 3).
10 μg/mL PTX-treated cells showed slight increase of NSEγ immunoreaction at 24 and 48 h compared to the untreated controls, and marked increase of MYH2 (a marker of muscular differentiation) immunoreaction at 12, 24, and 48 h. Regarding the markers of osteoblastic differentiation, RUNX2, OPG, BMPR2, and TGF-β1 were markedly positive in 10 μg/mL PTXtreated cells at 12, 24, and 48 h (Fig 4). These results indicate 10 μg/mL PTX affect RAW 264.7 cells to have a potential for neuro-muscular and osteogenic differentiation.

PLOS ONE
cadherin) at 12, 24, and 48 h than the untreated controls. Among the cMyc/MAX/MAD network proteins, the bands of cMyc and MAX were gradually attenuated at 12, 24, and 48 h, while the bands of MAD1 were increased. The proteins relevant to epigenetic modification, KDM4D and HDAC10 were increased in 10 μg/mL PTX-treated cells at 12, 24, and 48 h versus the untreated controls. RAS signaling proteins, KRAS, HRAS, NRAS, ERK1, and p-ERK1 were also increased at 12, 24, and 48 h (Fig 5).
TNFα, an inflammatory cytokine, TLR2 and TLR4, markers of innate immunity were increased in 10 μg/mL PTX-treated cells at 12, 24, and 48 h compared to the untreated controls. ER stress proteins, eIF2AK3 and p-eIF2AK3, a marker for autophagy formation, LC3β, and an apoptosis executing protein, caspase 3 were coincidently increased at 24 and 48 h. 10 μg/mL PTX-treated cell showed stronger bands of osteoblastic differentiation proteins, TGF-β1, BMP2, RUNX2, and ATF4 than the untreated controls, and slightly strong bands of nerve differentiation proteins, NSEγ and NF1, and muscle differentiation proteins, MYH2 and desmin at 12, 24, and 48 h compared to the untreated controls ( Fig 6).
The expressions of inflammatory proteins, TNFα and TLR2 were rarely affected by 300 μg/ mL PTX compared to the untreated controls, while increased by 10 μg/mL PTX at 12, 24, and 48 h. And the expressions of osteogenesis proteins, BMP2, RUNX2, and ATF4 were slightly decreased by 300 μg/mL PTX, while increased by 10 μg/mL PTX at 12, 24, and 48 h. On the other hand, the TGF-β1 expression was rarely affected by 300 μg/mL PTX, while slightly increased by 10 μg/mL PTX at 12 and 24 h. And the expression of house-keeping protein, βactin was almost not affected by 10 μg/mL and 300 μg/mL PTX at 12, 24, and 48 h (Fig 7).

Immunoprecipitation-based high performance liquid chromatography (IP-HPLC) analysis
10 μg/mL PTX-treated RAW 264.7 cells were extensively explored for different protein expression by IP-HPLC using 409 antisera, and 300 μg/mL PTX-treated cells were simply done using
Effects of 10 μg/mL PTX on the expression of cMyc/MAX/MAD network proteins. The expression of cMyc (regulator genes and proto-oncogenes that code for transcription factors) was increased by 8.2% at 12h after 10 μg/mL PTX treatment but gradually decreased to the untreated control level at 48 h, the expression of MAX (bHLH-Zip protein forming heterodimer with cMyc) was decreased by 3% at 48 h, while the expression of MAD1 (bHLH-Zip protein forming heterodimer with MAX which can oppose functions of Myc-MAX heterodimers) was increased by 4.8% at 48 h versus the untreated controls (Fig 8C and 8D). Whereas the double IP-HPLC using first antibody of cMyc and second antibody of MAX or MAD1 showed that the heterodimers of cMyc and MAX were increased by 11% at 24 h and 11.9% at 48 h, while the heterodimer of cMyc and MAD1 was decreased by 4.2% at 12 h and 2.1% at 24 h compared to the untreated controls (Fig 8E and 8F). On the other hand, the expressions of cMyc/MAX/MAD network interacting proteins, CDK4 and cyclin D2 were increased by 4.2% at 12 h and 10.1% at 12 h, respectively, but the expression of p27 (cyclin-dependent kinase inhibitor 1B) was minimally affected by PTX (�5%) (Fig 8C and 8D).
In the double IP-HPLC using antisera of cMyc/MAX and cMyc/MAD1, the cMyc-MAX heterodimer was increased by 11% at 24 h and 11.9% at 48 H, while the cMyc-MAD heterodimer was decreased by 4.2% at 12 h and 2.1% at 24 h compared to the untreated controls. On the other hand, CDK4-p27 complex was consistently reduced by 4.6%, at 12 h, 4.1% at 24 h, and 4.7% at 48h in the double IP-HPLC using CDK4 and p27 antisera (Fig 8E and 8F).
Effects of 10 μg/mL PTX on the expression of p53/Rb/E2F signaling proteins. 10 μg/ mL PTX decreased the expression of tumor suppressor proteins, that is, p53 by 17.5% at 24 h, Rb1 by 9.8% at 48 h, and phosphorylated Rb1 (p-Rb1) at 5.5% in RAW 264.7 cells versus the untreated controls, while the expression of objective transcription factor, E2F1 was increased by 10.6% at 12 h. And the Ser/Thr-kinase components of cyclin D2 and CDK4 were upregulated by 10.1% and 4.2% at 12 h, respectively, and a cyclin-dependent kinase inhibitor protein, p21CIP1 was also upregulated by 5.2% at 48 h (Fig 9A and 9B).
Effects of 10 μg/mL PTX on the expression of Wnt/β-catenin signaling proteins. 10 μg/ mL PTX increased the protein expressions of Wnt1 (by 13% at 24 h), β-catenin (by 18.5% at 24 h), adenomatous polyposis coli (APC, by 13.8% at 48 h), snail (by 13.1% at 48 h), and T-cell factor 1 (TCF1, a transcription factor, by 13.4% at 12 h) versus the untreated controls, while decreased the protein expression of snail (a transcription factor for the repression of adhesion molecule E-cadherin) by 8% at 12 h and AXIN2 (a regulator of β-catenin stability) by 8.7% at  24 h. On the other hand, E-cadherin (a type-I transmembrane protein stabilized by β-catenin) and VE-cadherin (vascular endothelial cadherin) were increased by 17.4% at 48 h and 21.9% at 12 h, respectively, while the expressions of Wnt signaling cofactors, MMP9 and vimentin which are necessary in the process of epithelial-mesenchymal transition, were decreased by 10.4% and 12.9% at 24 h, respectively (Fig 9E and 9F).
These results indicate Wnt/β-catenin signaling was enhanced with concomitant upregulation of Wnt1, β-catenin, APC, and TCF1 by 10 μg/mL PTX, but the activation of Wnt/β-catenin signaling was not followed by upregulation of snail, AXIN2, MMP9, and vimentin, but led to the overexpression of E-cadherin and VE-cadherin as shown in the results of double IP-HPLC (Fig 9G and 9H).
Effects of 10 μg/mL PTX on the expression of protein translation proteins. RAW 264.7 cells treated with 10 μg/mL PTX showed increase in the expression of protein translation protein: deoxyhypusine hydroxylase (DOHH, by 7.6% at 12 h), deoxyhypusine protein synthase (DHPS, 17.1% at 24 h), eukaryotic translation initiation factor 5A1 (eIF5A1, 6.5% at 12 h), and eIF5A2 (7.5% at 24 h) versus the untreated controls. On the other hand, the essential factor for protein synthesis to form a ternary complex (TC) with GTP and the initiator Met-tRNA, that is, eIF2α and p-eIF2α were decreased by 8.7% at 12 h and 5.7% at 48 h, respectively, and eukaryotic translation initiation factor 2-α kinase 3 (eIF2AK3, a protein kinase R (PKR)-like endoplasmic reticulum kinase (PERK)) was decreased by 9.3% at 12 h, but p-eIF2AK3 was reactively increased by 6.7% at 48 h (Fig 10C and 10D).
Effects of 10 μg/mL PTX on the expression of NFkB signaling proteins. 10 μg/mL PTX had different effects on the expression of NFkB signaling proteins in RAW 264.7 cells. PTX markedly upregulated nuclear factor kappa-light-chain-enhancer of activated B cells (NFkB) by 20.7% at 24 h but slightly downregulated ikappa B kinase (IKK) by 6.8% at 12 h versus the untreated controls, and subsequently increased the expression of downstream effector proteins of NFkB signaling, that is, p38 mitogen-activated protein kinase (p38) by 13.1% at 48 h, phosphorylated p38 (p-p38) by 12.3% at 48 h, growth arrest and DNA damage 45 (GADD45) by 18.7% at 48 h, multiple drug resistance (MDR) by 12.5% at 48 h, protein kinase C (PKC) by 8.6% at 24 h, p-PKC1α by 11.5% at 24 h, steroid receptor co-activator-1 (SRC1) by 12.3% at 48 h, and A-kinase anchoring proteins (AKAP13) by 16.4% at 24 h.
Inflammatory proteins upregulated by 10 μg/mL PTX. Among the inflammatory proteins, some proteins were upregulated by 10 μg/mL PTX in RAW 264.7 cells as follows, that is, tumor necrosis factor α (TNFα) by 10.6% at 24 h, interleukin-1 (IL-1) by 5 On the other hand, the expression of MMP12, TLR2, TLR7, and C-reactive protein (CRP) showed a trend of increase but were only minimally affected by PTX (� 5%) (Fig 11E and 11F).    19.5% at 24 h. On the other hand, the expression of CD40 and programmed cell death protein 1/1 (PD-1, CD279) showed a trend of decrease but were only minimally affected by PTX (� 5%) (Fig 12A and 12B).
Effects of 10 μg/mL PTX on the expression of FAS-mediated apoptosis proteins. RAW 264.7 cells treated with 10 μg/mL PTX showed increases in the expression of FAS-mediated apoptosis proteins, that is, FAS ligand (FASL) by 20 (Fig 12E and 12F).
Effects of 10 μg/mL PTX on the expression of survival and aging-related proteins. 10 μg/mL PTX-treated RAW 264.7 cells showed increases in the expression of survival and aging-related proteins, that is, pAKT1/2/3 by 6.5% at 24h, PKC by 8.6% at 24h, pPKC1α by 11.5% at 24h, SP-1 by 8.3% at 48h, a stress responsive deacetylase sirtuin 6 by 5.8% at 24 h, a master regulator of the regulatory pathway in the development and function of regulatory T cells FOXP3 by 7.4% at 12h, a type-I membrane protein related to β-glucuronidase Klotho by 7.1% at 12h, and PLCβ2 by 13.1% at 48h versus the untreated controls. Whereas PTX decreased the expression of telomere reverse transcriptase (TERT) by 10.4% at 12h, SP3 by 16.4% at 48h, sirtuin 1 by 9% at 24h, and FOXP4 by 11% at 48h. The expressions of sirtuin 7, FOXO1 (a transcription factor negatively regulating adipogenesis), and FOXO3 were only minimally affected by PTX (� 5%) (Fig 13C and 13D).
Effects of 10 μg/mL PTX on the expression of endoplasmic reticulum stress proteins. 10 μg/mL PTX had different effects on the expression of endoplasmic reticulum stress proteins in RAW 264.7 cells. PTX downregulated the proteins contributing to ER stress signaling; eIF2AK3, (PERK, which functions as an ER kinase) by 9.3% at 12 h, eIF2α and p- eIF2α (essential factors for protein synthesis also responsible for ER stresses) by 8.7% at 12 h and 5.7% at 48 h, respectively, activating transcription factor 6α (ATF6α) by 9% at 12 h, binding immunoglobulin protein (BIP, a HSP70 molecular chaperone) by 8.4% at 12 h, serine/threonine-protein kinase/endoribonuclease inositol-requiring enzyme 1 α (IRE1α) by 8.2% at 24 h, PGC-1α by 12.1% at 24 h, caveolin-1 by 8.3% at 48 h, and AIF by 10.9% at 48 h compared to the untreated controls. On the other hand, PTX upregulated the proteins contributing to ERstress environment in cells; HSP70 by 17.2% at 12 h, p-eIF2AK3 by 6.7% at 48 h, ATF4 (cAMP-response element binding protein 2) by 8% at 24 h, AP1M1 (the medium chain of the trans-Golgi network clathrin-associated protein complex AP-1) by 12.9% at 48 h, calnexin (a chaperone for the protein folding in the membrane of the ER) by 8.5% at 12 h, LC3β (microtubule-associated proteins 1A/1B light chain 3B contributing to autophagosome biogenesis) by 11.6% at 48 h, and endothelin-1 (inducing Ca 2+ release from ER) by 17.8% at 24 h. The expressions of GADD153 (C/EBP homologous protein (CHOP)) and p-GADD153 were only minimally affected by PTX (� 5%) (Fig 13E and 13F).
Effects of 10 μg/mL PTX on the expression of SHH/PTCH/GLI and Notch/Jagged signaling proteins. 10 μg/mL PTX was found to influence the expression of SHH/PTCH/GLI signaling proteins positively or negatively in RAW 264.7 cells. PTX upregulated the upstream proteins of SHH/PTCH/GLI signaling; sonic hedgehog (SHH) by 14.6% at 24 h, patched homolog 1 (PTCH1, the receptor for sonic hedgehog) by 8.8% at 48 h, and CD44 (HCAM, the activator of SHH signaling) by 9.8% at 48 h versus the untreated controls, while downregulated the downstream proteins of SHH/PTCH/GLI signaling; GLI1(Glioma-associated oncogene, the effectors of SHH signaling) by 12.9% at 24 h, EpCAM (epithelial cell adhesion molecule, involved in SHH signaling) by 3.1% at 48 h, and BCL2 (GLI binding site in BCL2 promoter, upregulated by GLI1) by 4.6% at 12 h.
Effects of 10 μg/mL PTX on the expression of oncogenesis proteins. 10 μg/mL PTX was found to influence the expression of oncogenesis proteins positively or negatively in RAW 264.7 cells. PTX decreased the expression of tumor suppressor proteins; breast cancer type 1 susceptibility protein (BRCA1) by 8.4% at 12 h, breast cancer type 2 susceptibility protein (BRCA2) by 18.6% at 24 h, neurofibromin 1 (NF1, a GTPase-activating protein that negatively regulates RAS/MAPK pathway activity) by 6.3% at 12 h, ataxia telangiectasia caused by mutations (ATM, a serine/threonine protein kinase recruited and activated by DNA double-strand breaks) by 13.9% at 24 h, maspin (a mammary serine protease inhibitor, serpin superfamily) by 9.5% at 24 h, deleted in malignant brain tumors 1 protein (DMBT1, a glycoprotein that interacts between tumor cells and the immune system) by 18% at 24 h, methyl-CpG-binding domain protein 4 (MBD4, a DNA repair enzyme that removes mismatched U or T) by 6.7% at 12 h, p53 by 17.5% at 24 h, retinoblastoma protein (Rb1) by 9.8% at 48 h, but increased the expression of PTCH1 (Protein patched homolog 1, a suppressor of smoothened release, which signals cell proliferation) by 10.1% at 24 h. On the other hand, PTX increased the expression of oncogenic proteins; carcinoembryonic antigen (CEA) by 9.6% at 12 h, 14-3-3 θ proteins (a phosphoserine binding protein that regulates Cdc25C) by 10.9% at 48 h, survivin (a negative regulator of apoptosis) by 11.9% at 48 h, mucin 4 (an anti-adhesive glycoprotein that contributes to tumor development and metastasis) by 15.6% at 24 h, Yes-associated protein 1 (YAP1, a potent oncogene that binds to 14-3-3) by 10.3% at 12 h, but decreased the expression of mucin 1 (a glycoprotein with extensive O-linked glycosylation of its extracellular domain, oncogenic epithelial membrane antigen) by 21.2% at 48 h. the expression of phosphatase and tensin homolog (PTEN, tumor suppressor protein) and PIM1 (proto-oncogene serine/threonine-protein kinase) were only minimally affected by PTX (� 5%) (Fig 15A and 15B).
Effects of 10 μg/mL PTX on the expression of angiogenesis proteins. 10 μg/mL PTX reduced the expression of major angiogenesis proteins in RAW 264.

Discussion
The pharmacological effect of PTX was frequently investigated in different cell types including RAW 264.7 cells [19][20][21][22] and animals [50][51][52][53] by using higher dose PTX, 100-500 μg/mL, rather than the therapeutic dose in human (about 10 μg/mL). It was also reported that the low dose PTX, 10, 25, 50 mg/kg, led to an increase in the expression of caspase 3 and TNFα in the rat hippocampus following lipopolysaccharide (LPS)-induced inflammation [54]. And the TNFα production by lipopolysaccharide (LPS)-stimulated human alveolar macrophages was significantly suppressed in the presence of PTX at concentration of 2 mM and 1 mM (278.3 μg/mL), but not at 0.5 mM, 0.1mM (27.8 μg/mL), and 0.01 mM, while production of IL-1β, IL-6, and GM-CSF remained unaffected. These data indicate PTX showed anti-inflammatory effect selectively depending on its concentration [55].
In the present study, RAW 264.7 cells, which are originally murine monocytes, were explored for PTX-induced protein expression changes by administrating with two different doses, 10 μg/mL PTX similar to the therapeutic dose in human, and 300 μg/mL PTX which was frequently used in cell and animal experiments. First of all, the 10 μg/mL PTX-induced effect was compared with the 300 μg/mL PTX-induced effect. However, in clinical application, even the therapeutic dose PTX, about 10 μg/mL, produces diverse side effects including belching, bloating, stomach discomfort or upset, nausea, vomiting, indigestion, dizziness, flushing, angina, palpitations, hypersensitivity, itchiness, rash, hives, bleeding, hallucinations, arrhythmias, and aseptic meningitis [56,57]. Therefore, it is thought that the low dose 10 μg/mL PTXinduced protein expression may be more informative to know the real pharmacological effect of PTX in human than the high dose 300 μg/mL PTX-induced protein expression in cell culture. Therefore, in the present study, 10 μg/mL PTX-induced effect on cells was more extensively investigated than 300 μg/mL PTX-induced effect.
In the global protein expression of RAW 264.7 cells by 10 μg/mL PTX, a competitive nonselective phosphodiesterase inhibitor which is known to raise intracellular cAMP and activate PKA, actually enhanced RAS signaling by upregulating AKAP13, pAKT1/2/3, PKC, p-PKC1α, KRAS, and HRAS, and subsequently activated histone/DNA demethylation and acetylation by upregulating KDMD4, HMGB1, and PCAF, and downregulating HDAC10, MBD4, DMAP1, DNMT1 and EZH2, and subsequently induced cellular proliferation by upregulating cMyc/ MAX/MAD network proteins, Wnt/β-catenin signaling objective protein TCF1, proliferation activating proteins Ki-67, PCNA, PLK4, cyclin D2, and cdc25A in this study. On the other hand, the expression of CDK inhibitors, p14, p15/16, and p21 were compensatory upregulated. The double IP-HPLC to assess the amount of protein complex containing two different target proteins showed the increase of cMyc-MAX heterodimer (Fig 8E and 8F) and β-catenin-TCF1 complex (Fig 9G and 9H) which led to cell cycle progression [58], and the decrease of cMyc-MAD1 heterodimer and CDK4-p27 complex concomitantly. On the other hand, the double IP-HPLC also revealed the increase of E2F1-Rb1 and CDK4-p21 complexes (Fig 9C and 9D), which mitigated cell cycle progression [59].
Although the expression of cMyc and MAX were decreased by 10 μg/mL PTX at 12, 24, and 48 h, the expression of MAD1 was increased in western blot and IP-HPLC, and the double IP-HPLC showed dominant increase of cMyc-MAX heterodimer and decrease of cMyc-MAD heterodimer versus the untreated controls (Fig 8E and 8F). Therefore, it is suggested that cMyc and MAX are competitively utilized to form cMyc-MAX heterodimer against cMyc-MAD heterodimer at 12, 24, and 48 h after 10 μg/mL PTX treatment, and then the unbound The 10 μg/mL PTX-treated RAW 264.7 cells showed crosstalk between activated RAS and NFkB signalings, and were progressed to cytodifferentiation by upregulating some growth factors and SHH/PTCH signaling. Particularly, it is evident that the 10 μg/mL PTX-treated RAW 264.7 cells have potentials of neuromuscular and osteoblastic differentiation, which are able to influence on objective adjacent cells [63][64][65]. The active neuromuscular and osteoblastic differentiations were also observed in ICC and western blot in this study. On the other hand, regarding the expression of osteogenesis proteins, 300 μg/mL PTX downregulated the osteoblastic differentiation proteins, BMP2, RUNX2, osterix, osteocalcin, osteopontin, and osteonectin, but upregulated the osteoclastic differentiation protein, RANKL, and compensatory increased the expression of OPG compared to 10 μg/mL PTX.
In the present study, 300 μg/mL PTX consistently decreased the expression of proliferation-, RAS/NFkB signaling-, inflammation-, and osteogenesis-related proteins but apoptosis proteins compared to 10 μg/mL PTX, therefore, it is thought that the high dose 300 μg/mL PTX may somehow disturb the protein expressions and give a harmful effect on RAW 264.7 cells, and resulted in the increase of FAS-mediated apoptosis, whereas the low dose 10 μg/mL PTX showed characteristic protein expression of competitive non-selective phosphodiesterase inhibitor, which may be helpful for the investigation of PTX pharmacological effect in human.
In addition, 10 μg/mL PTX was found to increase the potential of oncogenesis by upregulating the oncogenic proteins, CEA, 14-3-3θ, survivin, mucin 4, and YAP1, and downregulating the tumor suppressor proteins, P53, Rb1, BRCA1, BRCA2, NF1, ATM, maspin, and DMBT1. Nevertheless, 10 μg/mL PTX did not increase the expression of DNA repair enzymes, MBD4 and PARP-1, and exogenous stress responsible proteins, JNK1 and JAK2, but showed the overexpression of antioxidant proteins, SOD1 and GSTO1/2, and cell protection proteins, HSP70, sirtuin 6, and leptin, and resulted in the attenuation of ER stress. Therefore, it is suggested that 10 μg/mL PTX does not exert oncogenesis in RAW 264.7 cells, but maintains the cellular homeostasis, even though there appears slight elevation of some oncogenic protein expression.
The 10 μg/mL PTX-induced protein expression changes of different signaling pathways were summarized in a diagram of Fig 19. The panoramic protein signaling diagram illustrated main axes of protein signaling pathways in cells based on IP-HPLC data obtained in this study, therefore, it may indicate the real status of pharmacological effect of PTX in RAW 264.7 cells. We thought this PTX-induced protein expression of different signaling pathways should be corrected or added by further precise protein expression investigation using different cells and animals.
https://doi.org/10.1371/journal.pone.0261797.g019 PLOS ONE neuromuscular and osteoblastic differentiation, and stimulated acute inflammation, innate immunity, and cell mediated immunity, while reduced chronic inflammation, ER stress, and fibrosis but reactively increased FAS-mediated apoptosis and oncogenic potential in RAW 264.7 cells. On the other hand, 300 μg/mL PTX was found to decrease RAS/NFkB signaling compared to 10 μg/mL PTX, and subsequently attenuated proliferation, epigenetic activation, inflammation, neuromuscular and osteoblastic differentiation but increased apoptosis and fibrosis.