Sodium Salicylate Induces the Expression of the ImmunophilinFKBP51 and Biglycan Genes and Inhibits p34 cdc2 mRNA Both in Vitro and in Vivo *

One of the mechanisms proposed to explain the anti-inflammatory activity of sodium salicylate (NaSal) is based, at least in part, on its ability to inhibit nuclear factor-κB activation and inhibition of nuclear factor-κB-dependent gene expression. On the other hand, little is known about the ability of NaSal to activate gene expression. By differential display reverse transcription polymerase chain reaction, we identified several genes that are modulated upon treatment of mouse fibroblasts with NaSal. From the various cDNA fragments recovered from autoradiograms, we found that NaSal can increase the levels of mRNA for biglycan, the mouse homologue of the human eIF-3 p47 unit, and immunophilin FKBP51. NaSal-induced expression of these genes was time- and dose-dependent. Moreover,FKBP51 gene expression was augmented in vivo, in mice treated orally or intraperitoneally with NaSal. We also found that treating cells with NaSal can inhibit the expression of the p34 cdc2 kinase. The impact this inhibition on cell cycle was evaluated by measuring the content of DNA during the cell cycle. Treatment of cells with NaSal led to a G2/M arrest. By investigating the signaling events that regulate the expression of these genes and their biological activities, we can contribute to the understanding of the mechanism of NaSal.

Aspirin and salicylic acid constitute some of the most widely prescribed nonsteroidal anti-inflammatory drugs in the world (1). Diseases in which NSAIDs 1 can mitigate symptoms and the inflammatory process include some joint diseases, such as rheumatoid arthritis and osteoarthritis. These drugs can also retard the progression of Alzheimer's disease (2,3).
Several mechanisms have been proposed in order to explain the ability of NSAIDs to act as anti-inflammatory regulators, especially those ascribed to aspirin and its metabolite, sodium salicylate (NaSal). Vane (4), in his classical work, proposed that the therapeutic effects of aspirin could be ascribed to its ability to inhibit prostaglandin biosynthesis, probably via a direct inhibition of cyclooxygenases. Although only aspirin, and not NaSal, effectively inhibits the biosynthesis of cyclooxygenase isoforms in a dose-dependent manner, both compounds are potent anti-inflammatory agents (5,6). It has also been proposed that some effects of aspirin and NaSal that are apparently independent of prostaglandin biosynthesis are most likely due to the capacity of these drugs to insert themselves into the lipid bilayer of plasma membranes, where they disrupt signaling events and protein-protein interactions (7). Salicylates can also uncouple oxidative phosphorylation leading to ATP catabolism, diminishing intracellular ATP concentrations and, consequently, releasing micromolar amounts of adenosine, an autacoid with potent anti-inflammatory properties, into extra cellular fluids (8). At the molecular level, a more well defined action of NaSal and aspirin is the inhibition of NF-B activation (9). NF-B is known as a transcription factor that mediates the expression of a variety of genes that regulate the inflammatory response, including several cytokines and adhesion molecules (10,11). A possible pathway responsible for the inhibition of NF-B activation by NaSal appears to be related to its ability to activate p38 mitogen-activated protein kinase. Activation of p38 mitogen-activated protein kinase will lead to the inhibition of IB degradation and consequent inhibition of NF-B activation (12). Also, inhibition of RSK2 kinase by Na-Sal leads to the inhibition of NF-B-dependent gene expression (13). Furthermore, the ability of NaSal to block NF-B activation can be attributed to the inhibition of the kinase activity of IB kinase-␤ and its consequences, inhibition of IB phosphorylation and degradation and subsequent translocation of active NF-B to the nucleus (14). So far, a number of genes of which the expression is affected by NaSal at suprapharmacological or pharmacological concentrations have been identified. Aspirin or NaSal can partially or completely inhibit the induced expression of a large number of genes, such as those for iNOS (15), adhesion molecules (16), TF (tissue factor) (17), some chemokines (18), apolipoprotein A (19), prostaglandin H synthase (20), and several cytokines (21,22). Interestingly, a review of the literature shows a few genes that are induced by NaSal. Cytochrome P4502E1 in rat livers (23) and hsp70 in L929 cells (24) are the only examples of genes of which the expression levels are up-regulated by NaSal. The precise signaling pathway(s) responsible for NaSal-induced gene expression is not so well understood. Taking advantage of differential display RT-PCR (25), we decided to search for new genes, the expression of which can be modulated by NaSal and/or TNF. Here, we demonstrate that genes such as Biglycan and FKBP51 have their expression augmented in cells treated with NaSal, whereas the expression of p34 cdc2 kinase is diminished in NaSal-treated cells. The function of the proteins encoded by these genes appears to be related to the anti-inflammatory action of NaSal, and understanding the mechanism by which NaSal modulates the expression of these genes will contribute to our understanding of its mechanism of action.

EXPERIMENTAL PROCEDURES
Cell Culture and Treatments-Immortalized MEFs derived from sv129 mouse embryos were a gift from Drs. Charles Weissmann and Yi-Li-Yang (formally at the Institute of Molecular Biology I, University of Zurich, Zurich, Switzerland). MEFs were cultured at 37°C with 5% CO 2 in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, gentamicin (40 g/ml), and nonessential amino acids. Cells were treated with NaSal (Sigma) (20 mM for 6 h) and/or TNF (R&D Systems) (20 ng/ml for 4 h).
Mice and Administration of NaSal-Mice (pure sv129 background) were maintained in a sterile atmosphere with autoclaved food and acidified water. Four-to 8-week-old male mice were treated orally or intraperitoneally with 400 mg/kg NaSal. Control mice were treated with an equivalent volume of the vehicle. Animals were anesthetized and sacrificed at different time points, and organs were collected and immediately frozen in liquid nitrogen until RNA extraction.
RNA Isolation and DD-RT-PCR-Total RNA was isolated from near confluent MEFs or from organs using TRIZOL (Life Technologies, Inc.) according to the manufacturer's instructions. Differential display was performed following the protocol of Liang and Pardee (25). After DNase I treatment, 200 ng of total RNA were mixed with one of the four anchored oligo(dT) primers at a final concentration of 2.5 M (T 11 VA, T 11 VC, T 11 VG, and T 11 VT, where V is 3-fold degenerated for A, C, or G). Total RNA and anchored oligo(dT) were heated to 70°C for 10 min followed by an ice incubation for 2 min. Reverse transcription was carried out by the addition of reverse transcriptase buffer, dNTP (20 M each), 25 units of RNase inhibitor RNasin® (Promega), and 200 units of SuperScriptII (Life Technologies, Inc.). First strand cDNA synthesis was performed at 42°C for 60 min, followed by reverse transcriptase inactivation at 70°C for 10 min. One-tenth (2.0 l) of the cDNA first strand reaction was used as template for PCR amplification. PCR mixture contained 1ϫ PCR buffer (10 mM Tris-HCl, 50 mM KCl, 2.5 MgCl 2 ), a 20 M concentration of the same anchored oligo(dT) primer used in the reverse transcriptase reaction, 20 M each dNTP, 2.5 units of sequencing grade Taq DNA polymerase (Promega), 10 Ci of Redivue [␣-32 P]dCTP (Amersham Pharmacia Biotech), and 0.5 M of one random primer. Three random primers were used throughout this work: P13 (5Ј-CTGATCCATG-3Ј), P14 (5Ј-CTGCTCTCAA-3Ј), and P15 (5Ј-CTT-GATTGCC-3Ј). PCR amplification was carried out for 40 cycles (94°C for 30 s, 40°C for 2 min, and 72°C for 30 s, followed by 10 min postextension at 72°C). The PCR products were fractionated through a denaturing 6% polyacrylamide, 8 M urea gel at 1500 V. The gel was covered with a PVC film and exposed to x-ray films for 4 -16 h at Ϫ70°C. Bands of interest were excised from the gel, eluted in deionized water, and reamplified by PCR using the same conditions described for differential display, except that [␣-32 P]dCTP was replaced by nonradioactive dCTP.
cDNA Cloning and Sequencing Analyses-Reamplified cDNA fragments were purified by Wizard PCR preps (Promega) and cloned into pUC18 (SureClone system, Amersham Pharmacia Biotech) or pGEM-5zf(ϩ) from pGEM-T system (Promega). From each ligation, at least five white colonies were screened by PCR, using M13 forward and reverse primers that anneal to sequences flanking the multiple cloning sites. Miniprep plasmids from at least two independent clones containing the insert of expected size were sequenced (ABI Prisma TM , PE Applied Biosystems). Analysis of sequence homology was performed using a program from the National Center for Biotechnology Information.
Northern Blot Analysis and cDNA Probes-Fifteen micrograms of total RNA were fractionated through a 1% denaturing agarose gel and transferred by capillary onto a Nylon filter (Amersham Pharmacia Biotech). Prehybridization, hybridization, and washes were performed as described by Church and Gilbert (26). The KC chemokine probe corresponds to the full-length KC cDNA (GenBank TM accession number J04596) subcloned into pUC18 vector. The FKBP12 cDNA probe corresponds to a fragment of 529 base pairs from positions 735-1264 of the mRNA (GenBank TM accession number gb/X60203), obtained by RT-PCR. All other cDNA probes used for Northern blot analysis correspond to the clones listed in Table I and were generated by digestion with appropriate restriction enzymes flanking the cloned insert. Clone 14T366 is a 275-base pair fragment of the mouse GAPDH mRNA (positions 952-1226, GenBank TM accession number gb/M32599) and was used for ensuring equal loading. Probes were labeled by random priming using Redivue [␣-32 P]dCTP and the Rediprime labeling system (Amersham Pharmacia Biotech). Nylon filters were exposed to Kodak Hyperfilm (Amersham Pharmacia Biotech) at -70°C, with intensifying screen.
Fluorescent-activated Cell Sorter (FACS) DNA-MEFs (5-9 ϫ 10 6 cells) were treated with NaSal (20 mM) as indicated in the Fig. 5 legend, harvested by trypsinization, washed twice with phosphate buffer saline, and frozen at -70°C in citrate buffer (5% Me 2 SO, 40 mM sodium citrate, and 250 mM sucrose). For flow cytometry analysis, 10 6 cells were treated with trypsin (30 mg/liter) for 10 min at room temperature, then with trypsin inhibitor (0.5 g/liter) plus RNase A (0.1 g/liter) for 10 min at room temperature, and finally with propidium iodide (416 mg/liter) for 15 min in an ice bath. Cells were filtered through a 20 m nylon filter, and the amount of incorporated propidium iodide was determined by FACS analysis (FACScalibur, Becton Dickinson).

RESULTS
Differential Display RT-PCR and the Identification of NaSal and/or TNF-modulated Genes-A total of 44 PCRs were carried out in the presence of [␣-32 P]dCTP by using one of the four anchored dT 11 VN primers in combination with one of three random 10-mers. On average, each lane in a differential display gel yielded about 80 -100 distinguishable bands (Fig. 1). Fifty-nine bands representing cDNAs differentially displayed were excised and cloned. At least two clones from each ligation were isolated and sequenced, yielding a total of 94 clones with different inserts that were used for reverse Northern blot analysis. The size of cloned cDNA fragments ranged from 147 to 613 base pairs. The vast majority of clones were found to contain both random and anchored primers at their 5Ј and 3Ј ends, respectively. In a few clones, we found either the anchored or the random primers at both ends. The nucleotide sequences of the cloned cDNA fragments were compared against the nonredundant and expressed sequence tag data bases available on , with TNF (20 ng/ml for 4 h) (t), or with NaSal plus TNF (nt). The cDNA first strand was obtained by reverse transcriptase reaction using DNA-free total RNA (200 ng) in the presence of anchored oligoT 11 VN. cDNA was used as template for PCR in the presence of the same anchored primer and one of the 10-nucleotide-long random primers (brackets labeled 13, 14 or 15). Radioactive PCR products were fractionated through a 6% denaturing polyacrylamide gel and exposed to x-ray film. Numbered arrowheads indicate some of the cloned fragments. network services. Of the 94 distinct clones that were generated, 40 were found to be identical to known mouse genes, 20 had significant similarity to rat or human genes, 15 were similar to expressed sequence tags (either mouse or human expressed sequence tags), and 19 showed nonsignificant homology with the sequences deposited in GenBank TM . A summary of the clones of which the differential expression was confirmed by Northern blot analysis is given in Table I.
Evaluation of Differential Gene Expression upon Stimulation with NaSal and/or TNF-As a positive control for the stimulation of gene expression by TNF, we investigated the expression of KC chemokine and TSG-14 genes. Both are known to be up-regulated by TNF with functional NF-B sites on their promoter region (27,28). The mRNA steady-state level of both KC (Fig. 2) and TSG-14 (not shown) was elevated after TNF treatment but not after NaSal treatment. Furthermore, the TNF-induced expression of KC and TSG-14 was not impaired when the cells were co-stimulated with NaSal.
We next determined the steady-state levels of mRNA for other cDNAs identified by the DD-RT-PCR. As shown in Fig. 2, when MEFs were stimulated with NaSal, we detected an increase in the steady-state level of mRNA for biglycan, eIF-3p47 homologue, FKBP51, and es64. The es64 gene was also induced by TNF, but the combined TNF/NaSal treatment appears to have a synergistic effect.
The clone 14VC221, corresponding to p34 cdc2 kinase, was chosen from the DD-RT-PCR because it appeared to be inhibited by NaSal. Indeed, when we performed Northern blot analysis using total RNA from untreated control cells or from cells treated with NaSal, TNF, or both together, we observed a weaker signal for the p34 cdc2 kinase mRNA in lanes corresponding to cells exposed to NaSal (Fig. 2).
We next decided to further characterize the modulatory effect of NaSal on the expression of the above-mentioned genes, and because very little is known about stimulation of gene expression by NaSal, we gave priority to those genes of which the mRNA levels were up-regulated in the presence of the drug. First, we tested whether the up-regulation of FKBP51 observed in NaSal-treated MEFs could be reproduced in vivo, by measuring its mRNA level in organs of NaSal-treated mice. Northern blot analysis of spleen-derived total RNA from mice treated orally (Fig. 3A) or via intraperitoneal injection (Fig. 3B) using the 15VC462 clone as probe revealed that, indeed, FKBP51 mRNA is augmented in NaSal-treated mice. In orally treated mice, the peak of FKBP51 mRNA occurs between 4 and 6 h after treatment, whereas in mice injected intraperitoneally, the peak of induction occurs at 3 h postinjection (Fig. 3). Neither oral nor intraperitoneal treatment with NaSal led to the augmented expression of the FKBP12 gene.
We also determined the kinetics of biglycan and eIF-3 p47 homologue induction in MEFs treated with NaSal. As shown in Fig. 4, whereas no basal expression of biglycan could be de-tected, untreated cells had a measurable level of eIF-3 p47 mRNA, and the peak of mRNA steady-state levels for both biglycan and eIF-3 p47 homologue were obtained after 9 h of FIG. 2. Northern blot analysis for detection of differentially expressed genes. MEFs were treated as described in Fig. 1. Twelve micrograms of total RNA were fractionated through a 1% denaturing agarose gel and transferred to nylon filters. Cloned cDNA fragments isolated from the DD-RT-PCR were used as radioactive probes to detect mRNA levels of the corresponding genes. Blots were also reprobed with the 32 P-labeled clone 14T366 (corresponding to the mouse GAPDH mRNA at positions 952-1226) as control for RNA loading. a Corresponding to the bands assigned in Fig. 1. b Only significant stretches of matching nucleotides between the query and subject sequences were analyzed. stimulation (Fig. 4A). Co-stimulation of MEFs with NaSal and TNF for either 4 or 8 h did not alter the mRNA levels of biglycan or eIF-3 p47 homologue (Fig. 4B). Reduced Levels of the p34 cdc2 Kinase mRNA in NaSal-treated Cells Led to Cell Cycle Arrest at the G 2 /M Transition-In order to evaluate the physiology of the inhibitory effect of NaSal on p34 cdc2 kinase gene expression, we compared, by FACS analysis, the percentage of cells during the various phases of the cell cycle in cultures that were treated with NaSal and their control counterparts. We first determined, by Northern blot analysis, that maximum reduction in p34 cdc2 kinase mRNA level occurred in cells treated with 20 mM NaSal for 9 -12 h (Fig. 5A). However, at these time points, a significant amount of "dying" cells could be observed, as also pointed out by the reduced levels of GAPDH as compared with earlier points, regardless of our effort in accurately loading the gels. Therefore, we decided to perform cell cycle analysis in cells treated with NaSal for 6 h. Fig. 5B shows one representative of two experiments in which we observed that, in NaSal-treated cultures, 7.68% of cells were at the G 2 /M transition phase, as compared with 3.69% of cells at the same transition phase in control cultures of cells. In Fig. 5C, we show the average of a duplicate experiment that confirms the increased percentage of cells at the G 2 /M transition phase in NaSal-treated cultures. DISCUSSION Among the NSAIDs, acetylsalicylic acid, better known as aspirin, and other aspirin-like drugs are the most widely prescribed (1) and have been used clinically for more than a century. Nevertheless, only recently have we gained some information on the molecular aspects of the mechanism by which these drugs can modulate the expression of inflammation related genes. From the seminal work of Vane (4), suggesting that inhibition of prostaglandin biosynthesis is the major antiinflammatory action of these drugs, we now know that NaSal can also inhibit the activity the activation of p38 kinase, a key event in the inhibition of NF-B activation (12). What is less clear is how NaSal and other aspirin-like compounds can activate gene expression. Thus far, only a few genes, such as the hepatic cytochrome p4502E1 (23,29), are known to be induced by NaSal. It has also been suggested that some NSAIDs can induce the expression of the HSP70 gene in human monocytes (30) and in L929 cells (24). However, in HeLa cells, NaSal did not activate HSP70 gene transcription. In these cells, it appears that the transient acquisition of DNA binding activity by FIG. 3. NaSal induces the expression of the FKBP51 gene in vivo. Mice (sv129) were treated orally (A) or intraperitoneally (B) with NaSal (400 mg/kg) as indicated. At several different time points, animals were sacrificed, and organs were collected and immediately frozen on liquid N 2 . Fifteen micrograms of total RNA from spleen were fractionated through a 1% denaturing agarose gel and transferred to nylon filters. Filters were hybridized with [␣-32 P]dCTP-labeled probes for FKBP51 (clone 15VC462, described in Fig. 2) and for FKBP12 (fragment of the cDNA, cloned into pUC as described under "Experimental Procedures"). Filters were also hybridized with a radioactively labeled cDNA probe for the mouse GAPDH gene to control for RNA loading. the transcription factor HSF1 is not sufficient to drive gene expression (31).
Using the differential display technique (25), we identified a series of genes of which the expression is modulated by NaSal on mouse embryonic fibroblast.
Among the genes of which the mRNA level was found to be diminished in NaSal-treated cells, we identified the cycle-dependent p34 cdc2 kinase (32). We observed that upon treatment of mouse embryonic fibroblasts with NaSal, p34 cdc2 kinase mRNA decreased in a dose-and time-dependent manner, leading to G 2 /M arrest in mouse cells. It was previously reported that aspirin or indomethacin reduces the levels of p34 cdc2 kinase protein in HT-29 colon adenocarcinoma cells, reducing the proportion in the S phase (33). Our results corroborate these findings and are in agreement with the notion that NaSal can inhibit the proliferation of tumor cell lines by blocking cells from entering into S phase and also by enhancing apoptosis (34,35). From the technical point of view, the fact that we cloned three independent cDNA fragments corresponding to p34 cdc2 kinase, all of them with a similar pattern in the gels, provides an excellent internal control for the reliability and reproduc-ibility of the differential display, showing a correlation of this finding with the data already reported in the literature.
The identification of at least four genes up-regulated by NaSal opens a new field of investigation of NaSal mode of molecular action. These genes, which have not yet been identified as being responsive to NaSal, encode a small chondroitin/ dermatan sulfate proteoglycan biglycan (PG-I) (36,37), FK506binding protein p51 (FKBP51) (38), es64 (homologous to the human StAR protein) (39), and a putative eukaryotic translation initiation factor 3 (eIF-3) p47 subunit (40).
Of these genes, only the Biglycan gene has had its promoter structure identified and functionally characterized. In humans, its induction appears to be controlled via protein kinase A and interaction of SP3-and SP1-like factors (41). More recently, a binding site for the transcription factor cKrox, originally found as regulator of type I collagen gene expression (42,43), was also found at position -248 to -230 of the human biglycan gene (44). In the mouse, the expression of the biglycan gene appears to be regulated by SP-1, AP-1, and AP-2 transcription factors with low levels of expression in skin (37).
Based on these observations, and principally on the functional characterization of the promoter region of the other NaSal-stimulated genes described herein, it will be possible to investigate the molecular events in NaSal-triggered signal transduction. Further characterization of cis-elements and trans-acting factors that regulate the expression of other Na-Sal-stimulated genes will greatly contribute to our understanding of the molecular aspects of the mode of action of NSAIDs.
The characterization of the induced expression of the FKBP51 gene by NaSal and other NSAIDs, such as aspirin or indomethacin, could be of great importance for the understanding of the mode of action of aspirin. The FKBP51 gene was initially identified as a gene regulated by glucocorticoids in murine thymocytes (45) and subsequently characterized as member of family of intracellular receptors for immune suppressant drugs, such as FK506 and rapamycin (38). Moreover, FKBP51 was also identified by DD-RT-PCR as being probably involved in the adipocyte differentiation (46). Interestingly, NaSal was a potent inducer of FKBP51 both in vitro and in vivo, but it failed to induce the expression of the FKBP12 gene, the intracellular receptor for cyclosporin.
This differential induction of the two genes provides an important piece of evidence for a specific signaling cascade triggered by NaSal; a comparison of the promoter structure of these genes will help in the elucidation of NaSal mode action and could have an important impact in the administration of immune suppressant drugs. Fifteen micrograms of total RNA were fractionated through a 1% denaturing agarose gel and transferred nylon filters. Cloned cDNA fragment 14VC1 was used as radioactive probe to detect the levels of the p34 cdc2 kinase mRNA. Filters were also hybridized with a mouse GAPDH cDNA probe to control for RNA loading. B and C, 5-9 ϫ 10 6 cells were treated as indicated and the harvested by trypsinization and centrifugation. Propidium iodide was incorporated via trypsinization and dye uptake was measured by FACS analyzer (FACScalibur, Becton Dickinson). On average, 12,000 cells were scanned per time point. The averages plus S.D. of duplicates are shown. In B is shown a representative experiment, and in C is shown the average of two independent experiments.