Pharmacologic Inhibition of Tpl2 Blocks Inflammatory Responses in Primary Human Monocytes, Synoviocytes, and Blood*

Tumor necrosis factor α (TNFα) is a pro-inflammatory cytokine that controls the initiation and progression of inflammatory diseases such as rheumatoid arthritis. Tpl2 is a MAPKKK in the MAPK (i.e. ERK) pathway, and the Tpl2-MEK-ERK signaling pathway is activated by the pro-inflammatory mediators TNFα, interleukin (IL)-1β, and bacterial endotoxin (lipopolysaccharide (LPS)). Moreover, Tpl2 is required for TNFα expression. Thus, pharmacologic inhibition of Tpl2 should be a valid approach to therapeutic intervention in the pathogenesis of rheumatoid arthritis and other inflammatory diseases in humans. We have developed a series of highly selective and potent Tpl2 inhibitors, and in the present study we have used these inhibitors to demonstrate that the catalytic activity of Tpl2 is required for the LPS-induced activation of MEK and ERK in primary human monocytes. These inhibitors selectively target Tpl2 in these cells, and they block LPS- and IL-1β-induced TNFα production in both primary human monocytes and human blood. In rheumatoid arthritis fibroblast-like synoviocytes these inhibitors block ERK activation, cyclooxygenase-2 expression, and the production of IL-6, IL-8, and prostaglandin E2, and the matrix metalloproteinases MMP-1 and MMP-3. Taken together, our results show that inhibition of Tpl2 in primary human cell types can decrease the production of TNFα and other pro-inflammatory mediators during inflammatory events, and they further support the notion that Tpl2 is an appropriate therapeutic target for rheumatoid arthritis and other human inflammatory diseases.

Tpl2 is a member of the MAPKKK family of serine/threonine kinases. It resides upstream of the MAPKKs MEK1 and MEK2, and, in turn, the MAPKs ERK1 and ERK2. This Tpl2-MEK-ERK signaling module is activated in response to various pro-inflammatory stimuli, and it regulates the expression of several proinflammatory mediators (1)(2)(3)(4)(5)(6)(7)(8). Notably, Tpl2 regulates the expression of TNF␣ protein, which is critical to the initiation and progression of many inflammatory disorders including rheumatoid arthritis (9 -13). Rheumatoid arthritis occurs in nearly 1% of the adult population in most western countries, with an annual incidence of ϳ0.04%, and it can develop into a chronic, debilitating condition characterized by joint pain/ swelling/stiffness, restricted mobility, and the erosion of bone and cartilage in affected joints (14). Protein therapeutics such as ENBREL (etanercept) (soluble TNFRII-Fc), and REMICADE (infliximab) and HUMIRA (adalimumab) (anti-tumor necrosis factor ␣ antibodies) (TNF␣) 3 that bind and neutralize TNF␣ have proven clinically efficacious, and they have provided tremendous medical benefit. The success of these drugs has validated the use of anti-TNF␣ therapies for treating arthritic and inflammatory diseases, and they have necessitated the search for small molecule inhibitors with similar or related mechanisms of action. Due to its physiological roles in cytokine signaling networks and its key role in the production of TNF␣, a highly selective, small molecule inhibitor of Tpl2 should constitute an effective therapy for TNF␣-driven inflammatory disorders such as rheumatoid arthritis.
Genetic studies with tpl2 Ϫ/Ϫ mice have shown that Tpl2 is required for the expression of TNF␣ in circulating plasma following the administration of LPS in vivo, and cultured macrophages from tpl2 Ϫ/Ϫ mice exhibited markedly reduced TNF␣ production following LPS stimulation (2,8). In addition, LPSstimulated tpl2 Ϫ/Ϫ macrophages exhibited a restricted defect in ERK activation, with signaling through the JNK and p38 MAPK pathways largely unaffected (2,8). Tpl2-deficient mice exhibited a strongly attenuated progression of disease in a model of TNF␣-dependent inflammatory bowel disease, and TNF␣-induced ERK activation was completely ablated in tpl2 Ϫ/Ϫ macrophages (1,4,5). These data implicate the Tpl2-MEK-ERK module as a fundamental signaling component in cells of the innate immune system. However, Tpl2 signaling also regulates adaptive immunity. MEK/ERK activation in response to CD40 stimulation is ablated in tpl2 Ϫ/Ϫ B cells, and a study of tpl2 Ϫ/Ϫ dendritic cells demonstrated that Tpl2 is a negative regulator of IL-12 production in that cell type (4,8). Dendritic cell IL-12 is known to induce Th1-type T cell differentiation, and Th1-skewed immune responses in tpl2 Ϫ/Ϫ mice were observed following ovalbumin immunization or Leishmania major infection in vivo.
Although the foregoing studies indicate that Tpl2 is an important regulator of both innate and adaptive immunity, all of those studies were performed with Tpl2-deficient mice and cells derived from those mice. Currently, there are almost no reports of the targeted down-regulation of Tpl2 in human cell types (for one such report, see Ref. 6). These studies are necessary to further validate Tpl2 as a therapeutic target for human inflammatory disorders.
We have used highly selective and potent small molecule inhibitors of Tpl2 activity to validate several functions of Tpl2 in human cells. First, we demonstrate that Tpl2 is required for the LPS-induced activation of MEK and ERK in primary human monocytes. Second, we show that inhibition of Tpl2 blocks LPS-and IL-1␤-induced TNF␣ production in primary human monocytes and human blood. Third, our Tpl2 inhibitors block ERK activation, COX-2 expression, and the production of various soluble pro-inflammatory mediators in RA-FLS. Taken together, our data confirm physiological roles for Tpl2 in human inflammatory cell types, and they provide essential insights into the value of Tpl2 as a target for therapeutic intervention in human inflammatory diseases such as rheumatoid arthritis.
Immunoblot Antibodies-All immunoblot antibodies were from Cell Signaling Technology, except for anti-Cot/Tpl2 and anti-COX-2 from Santa Cruz Biotechnology (Santa Cruz, CA). All secondary/detection antibodies were from GE Healthcare.
HeLa Assays-HeLa cells were obtained from the American Type Culture Collection (Manassas, VA) and passaged in DMEM supplemented with 10% FBS (Sigma). For transfection assays (Figs. 1 and 2), human TPL2 and TPL2-K167R cDNAs (16) were subcloned into a replication-defective Type 5 adenoviral vector ("pAdori," Genetics Institute, Andover, MA) (17) under the control of the CMV promoter and all constructs were confirmed by sequencing. Cells were transfected for ϳ4 h using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions, and were then washed with DMEM and allowed to incubate in DMEM supplemented with 0.5% FBS, or, for inhibitor studies, in DMEM, 0.5% FBS containing Tpl2 inhibitors or vehicle (Me 2 SO). Final Me 2 SO concentrations were 0.25%. In some experiments HeLa transfectants were trypsinized and re-plated before the addition of compounds. 16 -24 h later media were harvested and IL-8 levels were measured by ELISA (BioSource International, Camarillo, CA). Compound toxicity was assessed using the WST-1 Cell Proliferation Reagent as described by the manufacturer (Roche Applied Science). Cells were lysed with the addition of 1ϫ LDS sample buffer (Invitrogen), briefly sonicated and boiled, and analyzed by standard immunoblot analysis. Immunoreactive band intensities were quantitated on an Image Station 2000MM (Eastman Kodak Co.).
For inhibitor studies with TNF␣-stimulated, non-transfected HeLa cells (Fig. 7), the cells were plated in DMEM, 0.5% FBS 1 day prior to testing. Inhibitors were added 30 -45 min before stimulation with 100 ng/ml rhTNF␣ (R&D Systems). Cell lysates were then prepared in 1ϫ LDS buffer and immunoblotted. Media aliquots from 4-h stimulations were analyzed for IL-8 by electrochemiluminescence detection on a Sector6000 plate reader according to the manufacturer's instructions (Meso Scale Discovery, Gaithersburg, MD). Toxicity was assessed using the WST-1 Reagent.
Primary Human Monocytes-Human blood buffy coats were purchased from the Blood Transfusion Service at Massachusetts General Hospital (Boston, MA). Monocytes were prepared by negative selection. For each preparation, EDTA was added to the buffy coat to a final concentration of 1 mM, and the buffy coat was incubated with the RosetteSep Monocyte Enrichment antibody mixture from Stem Cell Technologies (Vancouver, Canada) for 20 min (6 ml of mixture per 50 ml of  NOVEMBER 16, 2007 • VOLUME 282 • NUMBER 46 buffy coat). The buffy coat was then diluted with an equal volume of phosphate-buffered saline supplemented with 2% FBS and 1 mM EDTA, layered on an equal volume of Ficoll-Paque (Sigma), and centrifuged at 500 ϫ g for 20 min (brake on low or off). Cells located within ϳ1 cm above the interface were removed, layered on an equal volume of Ficoll-Paque, and centrifuged as above. All cells above this second interface were collected, pelleted at 500 ϫ g for 5 min, and washed twice with phosphate-buffered saline, 2% FBS, 1 mM EDTA (brake on high). The cell pellet was suspended to 2 ϫ 10 6 cells/ml in RPMI 1640 (pH 7.3) supplemented with 0.5% FBS. Cellular yields were routinely 65-80% CD14-positive as judged by flow cytometry. Monocytes were plated in 48-well formats at 400 l (ϳ800,000 cells) per well, and incubated at 37°C, 5% CO 2 for Ն30 min. Tpl2 inhibitor or vehicle (Me 2 SO) was added for 30 -45 min. Final Me 2 SO concentrations were 0.25%. Salmonella typhimurium LPS (Sigma) was used at 10 ng/ml, and rhIL-1␤ (BioSource International) was used at 100 ng/ml. Cell lysates were prepared in 1ϫ LDS buffer and immunoblotted. For cytokine determinations, monocytes were stimulated for 3-4 h and media aliquots were analyzed by electrochemiluminescence detection. Toxicity was assessed using the WST-1 Reagent.

Tpl2 Inhibitors Block Inflammation in Human Cell Types
Whole Blood-Cytokine levels in human blood were analyzed as described previously (18). Blood samples were stimulated for 3-4 h with LPS or rhIL-1␤ at 10 or 100 ng/ml, respectively.
RA-FLS-Cryopreserved RA-FLS (Cell Applications, San Diego, CA) were thawed and cultured in Synoviocyte Growth Medium according to the supplier's instructions. In all experi-ments, media was changed to RPMI 1640, 0.5% FBS, and Tpl2 inhibitor or vehicle (Me 2 SO) was added 30 -45 min before stimulation with 100 ng/ml rhIL-1␤ or rhTNF␣. Cell lysates were prepared in 1ϫ LDS buffer and immunoblotted. Media aliquots from 4-h stimulations with rhIL-1␤ were analyzed for PGE 2 by ELISA (Assay Designs, Ann Arbor, MI), and for IL-6 and IL-8 by electrochemiluminescence detection. Media aliquots from 24-h stimulations with rhIL-1␤ were analyzed for MMPs by Luminex (R&D Systems). Media aliquots from 24-h stimulations with rhTNF␣ were analyzed for PGE 2 by ELISA. Toxicity was assessed using the WST-1 Reagent.
Peritoneal Macrophages-All animal procedures were approved by the Institutional Animal Care and Use Committee (IACUC). Mice were provided food and water ad libitum. tpl2 Ϫ/Ϫ mice (backcrossed onto C57BL/6 for Ͼ10 generations) were obtained from Dr. Philip Tsichlis (Thomas Jefferson University). These mice were crossed with 129S5 mice to produce F1 heterozygotes. The F1 mice were inter-crossed to produce F2s. Female TPL2 ϩ/ϩ , TPL2 ϩ/Ϫ , and tpl2 Ϫ/Ϫ F2 littermates (8 -12 weeks of age) were injected intra-peritoneally with 1 ml of sterile 4% Brewer thioglycollate broth (BD Biosciences), and 72 h later peritoneal exudate cells were isolated by lavage with RPMI 1640. The cells were washed and plated in RPMI 1640, 5% FBS, 100 units/ml penicillin, 100 g/ml streptomycin at ϳ3 ϫ 10 6 cells per 35-mm dish. After ϳ5 h, non-adherent cells were removed by repeated washing, and the adherent macrophages were subjected to various treatments. Total macrophage protein levels were quantitated by Bradford assay (Bio-Rad). S. typhimurium LPS was from Sigma, and rmIL-1␤ and rmTNF␣ were from BioSource International. Cell lysates were prepared in 1ϫ LDS buffer and immunoblotted as above. Cytokine levels in media supernatants were measured by electrochemiluminescence detection.

RESULTS
A Chemical Series of Selective Tpl2 Inhibitors-We have developed a series of compounds that inhibit Tpl2, and the synthesis and structure-activity relationships of the early prototypes in this series have been described (18,19). Subsequent, more selective, and more potent entries in this series are described herein and elsewhere (20). Compounds 1, 2, and 3 in this article (structures shown in Figs. 2A and 3A) are representative of our chemical series of Tpl2 inhibitors inasmuch as they are highly selective for Tpl2 inhibition when tested against a panel of 13 kinases. This panel includes kinases known to be involved in TNF␣ production (e.g. p38 and MK2) plus other serine/threonine and tyrosine kinases. As an example of the highly selective nature of our Tpl2 inhibitors, Table 1 shows the cell-free IC 50 values in all 13 kinase assays for the C8-chloronaphthyridine-3-carbonitrile shown in Fig. 2A and named here as Compound 1. Among these 13 kinases, Compound 1 was least selective against CaMKII, but it was 2000-fold less potent against CaMKII than Tpl2 (Tpl2 IC 50 ϭ 12 nM). Importantly, relative to the unsubstituted parent molecule (compound 31 in Ref. 19), the C8-chloro substitution in Compound 1 significantly reduced potency against EGFR (20).

Tpl2 Inhibitors Block Inflammation in Human Cell Types
ing Tpl2 Inhibitors-To develop a Tpl2-dependent cell-based assay, we utilized HeLa cells transiently transfected with a plasmid that encodes full-length human Tpl2 driven by the CMV promoter. The overexpression of Tpl2 from this plasmid augmented the steady-state phosphorylation levels of endogenous MEK and ERK and the production and secretion of the proinflammatory cytokine IL-8 ( Fig. 1). This TPL2-driven production of IL-8 is consistent with our published findings (21). These responses required the catalytic activity of overexpressed Tpl2 because they did not occur in the presence of the kinaseinactive Tpl2-K167R. Thus, our HeLa transfectants provided MEK/ERK activation and IL-8 production as cellular responses that are induced by Tpl2 kinase activity. Indeed, Compound 1 inhibited IL-8 production in a dose-dependent manner with an IC 50 of 1 M (Fig. 2B). Compound 1 also inhibited MEK and ERK phosphorylation in these cells (data not shown).

Tpl2 Catalytic Activity Is Required for the LPS-induced Activation of MEK and ERK in Primary Human
Monocytes-Next, we screened our Tpl2 inhibitors by measuring their inhibition of LPS-induced MEK phosphorylation in freshly isolated primary human monocytes. To ensure that our compounds were inhibiting Tpl2 in these cells, we sought a correlation between potency in the cell-free Tpl2 enzyme assay and potency in the monocyte phospho-MEK assay. Of 140 compounds with cellfree Tpl2 IC 50 values Յ0. Selective Inhibition of Tpl2 in Primary Human Monocytes by Compound 1-The high degree of selectivity of our compounds for Tpl2 inhibition (see Table 1) was also observed in LPSstimulated primary human monocytes. As shown in Fig. 4A, Compound 1 potently inhibited LPS-induced MEK and ERK phosphorylation in human monocytes with IC 50 values of 0.5 M for LPS-induced MEK phosphorylation and 0.1 M for ERK phosphorylation. However, Compound 1 did not inhibit phosphorylation of p38, MK2, c-Jun, or IB␣ at concentrations as a high as 5 M. These data demonstrate that Compound 1 selectively inhibits Tpl2 in LPS-stimulated primary human monocytes, and they are consistent with the restricted defect in ERK activation in LPS-stimulated peritoneal macrophages from tpl2 Ϫ/Ϫ mice (2,8). The failure to block NF-B signaling in monocytes was corroborated with assays that employed an NF-B-luciferase reporter plasmid stably integrated into HEK293 cells: Compound 1 did not inhibit TNF␣or IL-1␤induced luciferase activity in those cells at concentrations as high as 10 M (data not shown).
Another MAPKKK known to activate MEK is the kinase Raf (see Ref. 22, and references therein). To rule out the inhibition of Raf by our compounds in monocytes, we preincubated monocytes with increasing concentrations of Compound 1 and then stimulated with PMA, which selectively activates the Raf-MEK-ERK cascade. As shown in Fig. 4B, Compound 1 did not inhibit PMA-induced MEK and ERK phosphorylation. These data indicate that Compound 1 does not inhibit Raf activity in primary human monocytes. This is consistent with the  fact that Compound 1 did not inhibit Raf-1 in the cell-free assay (Table 1). After establishing that Compound 1 is selective for Tpl2 inhibition in monocytes, we tested its inhibition of LPS-induced TNF␣ production. Compound 1 potently inhibited monocyte LPS-induced TNF␣ with an IC 50 of 0.6 M ( Table 2), which is consistent with the dramatic reduction in LPS-induced TNF␣ production in tpl2 Ϫ/Ϫ mouse macrophages (2,8). To exclude the possibility that reductions in LPS-induced TNF␣ were caused by prolonged exposure of the cells to inhibitors with cytotoxic properties, we routinely measured cell viability at the conclusion of the assay by adding the tetrazolium salt WST-1 (a 5-tetrazolino-1,3-benzene disulfonate; Roche Diagnostics) to the culture media, and then measuring spectrophotometrically its rapid cleavage to formazan by mitochondrial dehydrogenases, the activity of which is directly related to the number of viable cells in the culture. A decrease in the amount of formazan dye produced relative to control cultures was interpreted as evidence of a toxic result. In 35 of 41 tests of primary monocytes, 5 M Compound 1 caused either no cytotoxic effects or less than a 50% reduction in the appearance of the formazan dye (relative to controls). In the remaining 6 tests, 50% cytotoxicity was observed at 4.3 Ϯ 0.14 M (mean Ϯ S.E.). These data indicate that for almost all human monocyte preparations tested, cytotoxic effects of Compound 1 only manifested at a concentration that was at least 1 order of magnitude higher than the IC 50 for LPS-induced TNF␣ (ϳ0.6 M), and that, therefore, its potency for TNF␣ inhibition in these cells does not result from cytotoxic effects.
Given that LPS-induced IL-12p40 is not decreased in tpl2 Ϫ/Ϫ mouse macrophages (8), we measured IL-12p40 levels in culture supernatants from LPS-treated primary human monocytes that had been pre-treated with Compound 1. Compound 1 inhibited LPS-induced IL-12p40 in primary monocytes at least 10-fold less potently than it inhibited LPS-induced TNF␣ ( Table 2). We conclude that the pharmacologic inhibition of Tpl2 is sufficient for both a dramatic and selective reduction in LPS-induced TNF␣ production in primary human monocytes.
Rodriguez et al. (6) used Tpl2-specific small interfering RNA to knock down endogenous Tpl2 in HeLa cells. Using this technique these authors demonstrated that Tpl2 is required for IL-1␤-induced ERK activation in this cell type. Our results showing that the pharmacologic inhibition of Tpl2 in human monocytes blocks IL-1␤-induced MEK/ERK activation and TNF␣ production prompted us to use a genetic approach to confirm that Tpl2 mediates IL-1␤-induced ERK activation and TNF␣ production in cells from the monocyte-macrophage

Inhibition of LPS-and IL-1␤-induced TNF␣ Production in Human
Blood by Compound 1-Although circulating blood cells are not precisely the same as the cells found in arthritic joints, it is reasonable to assume that circulating immune cells do migrate to arthritic joints and participate in the disease process. Therefore, as one predictor of the in vivo efficacies of potential anti-arthritic drugs, we tested our Tpl2 inhibitors for inhibition of TNF␣ production in an in vitro human blood TNF␣ assay. Consistent with its ability to inhibit TNF␣ production in isolated human monocytes, Compound 1 inhibited LPS-and IL-1␤-induced TNF␣ production in human blood with IC 50 values of 5.4 and 7.2 M, respectively ( Table 2). The higher IC 50 values in the blood TNF␣ assay as compared with the monocyte TNF␣ assay may be due to a high degree of plasma protein binding for Compound 1. Extensive investigations into the structure-activity relationships of our Tpl2 inhibitors in human blood are described elsewhere (18 -20).
Inhibition of Inflammatory Responses in RA-FLS by Compound 1-Having established Compound 1 as a candidate antiinflammatory and anti-arthritic compound, we tested whether it could inhibit inflammatory responses in RA-FLS. As shown in  . ERK phosphorylation is ablated and TNF␣ production is reduced in IL-1␤-stimulated tpl2 ؊/؊ macrophages. A, immunoblot analysis of peritoneal macrophages from TPL2 ϩ/ϩ , TPL2 ϩ/Ϫ , and tpl2 Ϫ/Ϫ mice stimulated with 10 ng/ml LPS, 100 ng/ml mIL-1␤, or 100 ng/ml mTNF␣ for 45 min. Results are representative of five independent experiments. Arrowheads indicate the phosphorylated and unphosphorylated forms of ERK1/2, and the long and short forms of Tpl2. B and C, peritoneal macrophages were stimulated with 10 ng/ml LPS, 100 ng/ml mIL-1␤, or 100 ng/ml mTNF␣ for the indicated times. Cytokine levels in the culture supernatants were normalized to levels of macrophage monolayer proteins. Units of y axes: picgrams of cytokine per g of macrophage monolayer. In B and C, n ϭ 3 animals for TPL2 ϩ/ϩ , n ϭ 4 animals for TPL2 ϩ/Ϫ , and n ϭ 4 animals for tpl2 Ϫ/Ϫ . Values represent the mean Ϯ S.D.
and Table 3). In all RA-FLS isolates, incubation with Compound 1 for Ն4 h, in either the presence or absence of IL-1␤, caused no cytotoxic effects at all concentrations tested (Յ10 M). The production of COX-2 and PGE 2 are ablated in LPSstimulated tpl2 Ϫ/Ϫ mouse macrophages (3), so our results suggest that Tpl2 is required for IL-1␤-induced COX-2 expression and PGE 2 production in RA-FLS. IL-1␤ also stimulated RA-FLS to produce IL-6, IL-8, and the matrix metalloproteinases MMP-1 and MMP-3. The physiological characteristics of RA-FLS explants can vary from one isolate to the next; however, Compound 1 consistently and reproducibly inhibited IL-1␤-induced IL-6, IL-8, and MMP-3 in the RA-FLS isolates used here ( Table 3). The inhibition of IL-1␤-induced MMP-1 was somewhat less reproducible: in three of five donors the IC 50 was 4.0 Ϯ 0.8 M, but in two of these five donors there was little or no inhibition. Compound 1 caused no cytotoxicity under these conditions. Thus, barring only a few exceptions, Compound 1 inhibited IL-1␤-induced IL-6, IL-8, MMP-1, and MMP-3 in RA-FLS.
As Tpl2 is required for cellular responses to TNF␣ in murine macrophages and fibroblasts (1, 4), we tested whether Compound 1 could inhibit TNF␣-stimulated responses in human monocytes, blood, RA-FLS, and the human fibroblastic HeLa cell line. In our hands, TNF␣ treatment did not induce the production of cytokines and other inflammatory mediators in human monocytes and blood, nor did it induce MEK/ERK phosphorylation in monocytes. In four RA-FLS isolates PGE 2 production was induced 5-30-fold by stimulation with TNF␣ for 24 h, and this induction was potently inhibited by Compound 1 with an IC 50 of 0.7 Ϯ 0.1 M (Table 3). In HeLa cells IL-8 production was induced ϳ8-fold upon stimulation of HeLa cells for 4 h with TNF␣, and Compound 1 inhibited this induction with an IC 50 of 2.5 M (Fig. 7A). Moreover, TNF␣induced ERK phosphorylation in HeLa cells was potently inhibited by Compound 1 with an IC 50 of 0.2 M (Fig. 7B). In these TNF␣ induction experiments with RA-FLS and HeLa cells no cytotoxicity was observed under the conditions employed. Thus, we conclude that Tpl2 mediates cellular responses to TNF␣ stimulation in these human fibroblastic cell types.

DISCUSSION
Recent clinical successes with protein therapeutics that bind and neutralize TNF␣ have confirmed a role for TNF␣ in the progression of rheumatoid arthritis, inflammatory bowel disease, psoriasis, psoriatic arthritis, and ankylosing spondylitis (13). Thus, a reasonable approach to the development of therapies for these diseases is to identify drugs that block the signaling pathways that induce TNF␣ production, and for this reason we have created a series of inhibitors of the Tpl2 MAPKKK. Many of our 8-chloro-quinoline-3-carbonitrile (e.g. Compound 1) and 1,7-naphthyridine-3-carbonitrile (e.g. Compounds 2 and 3) Tpl2 inhibitors are highly selective for Tpl2 in our cell-free kinase assay panel of 13 kinases. The unique structure of Tpl2 may have facilitated our generation of highly selective Tpl2 inhibitors. Tpl2 is the only kinase in the human kinome that has a proline instead of a conserved glycine at the first glycine in the

Tpl2 Inhibitors Block Inflammation in Human Cell Types
GXGXXG motif of the ATP binding loop (23). This may allow Tpl2 to exclude inhibitors that cannot accommodate this unique primary structure. Indeed, Tpl2 is not inhibited by the broad-spectrum kinase inhibitor staurosporine, which binds in the ATP binding loop (24). A more complete understanding of both the selectivity and potency of our inhibitors will be possible once the Tpl2 crystal structure is obtained.
The high degree of cell-free kinase selectivity that characterizes our Tpl2 inhibitors translates into a high degree of selectivity for Tpl2 inhibition in isolated human monocytes. Our inhibitors selectively block LPS-induced MEK/ERK activation in human monocytes with IC 50 values Յ1 M, and they do not simultaneously inhibit JNK, p38, and NF-B signaling (as measured by phospho-c-Jun, phospho-p38, phospho-MK2, and phospho-IB␣) (Fig. 4). These results are consistent with the restricted defect in ERK activation in tpl2 Ϫ/Ϫ mouse macrophages (2), and they support the notion that the restricted preference of Tpl2 for activating only MEK/ERK in response to LPS in cells of the monocyte-macrophage compartment has been conserved throughout mammalian evolution. Moreover, our data demonstrate that Tpl2 is the physiological activator of MEK and ERK in LPS-stimulated primary human monocytes. This is demonstrated not only by the comparison of 1,7-naphthyridine-3-carbonitrile enantiomers in Fig. 4, but also by a wider examination of the Tpl2 and monocyte phospho-MEK IC 50 values of our inhibitors. As stated under "Results," of 140 compounds in this series with cell-free Tpl2 IC 50 values Յ0.2 M, 111, or 79%, had monocyte phospho-MEK IC 50 values Յ5 M. Thus, if potency in the monocyte phospho-MEK assay is decided by chance then ϳ79% of all compounds should have monocyte phospho-MEK IC 50 values Յ5 M. However, of 17 compounds tested that had cell-free Tpl2 IC 50 values Ͼ0.2 M none of them were active for inhibition of monocyte phospho-MEK at 5 M. As these 17 compounds do not have physical characteristics (e.g. permeability and solubility) that are significantly different from other compounds in this series, the most likely explanation for their inability to inhibit phospho-MEK in human monocytes is due to their inability to inhibit Tpl2. This high positive correlation between the potencies of our inhibitors in the cell-free Tpl2 assay with their potencies in the monocyte phospho-MEK assay, together with their high degree of selectivity for Tpl2, indicate that Tpl2 catalytic activity is required for the LPS-induced activation of MEK and ERK in primary human monocytes.
Our Tpl2 inhibitors block TNF␣ production in primary human monocytes and human blood ( Table 2). This is therapeutically important, because extravasated monocytes and synovial macrophages are primarily responsible for TNF␣ production in the inflamed joints of rheumatoid arthritis patients, and it is likely that monocyte-macrophage-derived TNF␣ is instrumental in the pathophysiology of other inflammatory diseases (11,12). The inhibition of blood TNF␣ provides a potential biomarker for tracking the effect of these inhibitors in a clinical trial, and their IC 50 values may establish initial target exposures necessary for achieving in vivo efficacy. It is important also to note that all of our compounds that inhibit monocyte phospho-MEK in LPS-stimulated monocytes also inhibit monocyte LPS-induced TNF␣ with similar potency (e.g. 0.5 and 0.6 M, respectively, for Compound 1). This is consistent with the notion that in human monocytes Tpl2 regulates TNF␣ production via activation of the MEK/ERK axis.
RA-FLS secrete cytokines, eicosanoids, and MMPs that are critical components in the pathogenesis of rheumatoid arthritis (25). The expression of these factors can be induced by proinflammatory cytokines, such as IL-1␤, which are found at increased levels in rheumatoid joints (Refs. 11, 12, and 25, and references therein). COX-2, the central enzyme in the PG biosynthetic pathway, is expressed at low levels in RA-FLS under basal conditions, but it is dramatically induced in RA-FLS in response to IL-1␤ (26,27). Compound 1 inhibited IL-1␤-induced COX-2 expression and PGE 2 secretion in cultured RA-FLS ( Fig. 6 and Table 3). The IC 50 for IL-1␤-induced PGE 2 was 0.4 Ϯ 0.1 M; however, complete inhibition of IL-1␤-induced PGE 2 was observed at concentrations Ն3 M, which is in close agreement with the near complete ablation of IL-1␤-induced COX-2 expression at 2.5 M (Fig. 6). Taken together with previous observations that COX-2 and PGE 2 expression are almost completely ablated in tpl2 Ϫ/Ϫ mouse cells (3), these data suggest that IL-1␤ induces COX-2 (and PGE 2 ) in RA-FLS through a signaling mechanism that requires Tpl2.
Studies with tpl2 Ϫ/Ϫ mouse fibroblasts indicate that Tpl2 is selectively required for ERK activation in that cell type following stimulation with IL-1␤, and that Tpl2 is not required for JNK, p38, and NF-B activation under those conditions (1). A similar restricted profile of Tpl2 selectivity may exist in RA-FLS: Compound 1 inhibited IL-1␤-stimulated ERK activity in RA-FLS as expected (Fig. 6), but preliminary immunoblot analyses indicate that our Tpl2 inhibitors do not block JNK and IKK␤ activation in RA-FLS under those conditions. It is well established that NF-B is a critical mediator of the expression of IL-6 and IL-8 in RA-FLS following stimulation with IL-1␤ (28), and Compound 1 dose dependently inhibited IL-6 and IL-8 in IL-1␤-stimulated RA-FLS (Table 3). However, with each RA-FLS isolate studied the dose-response profiles tended to level off at Ն50% inhibition (see Table 3). Thus, it is possible that a selective Tpl2 inhibitor cannot completely ablate IL-1␤induced IL-6 or IL-8 production in RA-FLS because of a coordinate, Tpl2-independent role for NF-B in the induction process. In addition, our pharmacologic inhibition of IL-1␤-induced IL-6 with Compound 1 in RA-FLS suggests that in tpl2 Ϫ/Ϫ mouse fibroblasts IL-1␤-induced IL-6 may be curtailed relative to TPL2 ϩ/ϩ controls. This possibility was not addressed in the earlier report (1). If, indeed, there is such a defect, it would help to differentiate these cells from the mouse macrophages described here in which a deficiency in TPL2 has no discernible effect on IL-1␤-induced IL-6 production, and it would thus highlight another important difference between fibroblasts and macrophages.
The excess MMPs in affected joints drives the progressive bone and cartilage destruction associated with rheumatoid arthritis and other arthritides (29). In the RA-FLS cultures studied here, MMP-1, -2, and -3 were detectable, but MMP-7, -8, -9, -12, and -13 were not, and only MMP-1 and -3 were induced in response to IL-1␤. Compound 1 inhibited IL-1␤-induced MMP-1 and MMP-3 expression (Table 3), but it did not inhibit MMP-2. Increased serum MMP-3 has been shown to correlate with increased radiographic scores in rheumatoid arthritis, and MMP-1 has been found at the sites of cartilage erosions (30 -32). Our results indicate that Tpl2 inhibition might slow or halt the progressive structural damage in arthritic disease by blocking the expression of these destructive enzymes. It is interesting to point out that Compound 1 was 10-fold more potent against PGE 2 than MMP-1 and MMP-3 (IC 50 values of 0.4 versus 4.0 and 2.3 M, respectively). These data suggest that in RA-FLS different thresholds of Tpl2-dependent ERK activation are required for optimal induction of these downstream events. It has been shown that the expression of several inflammatory genes in LPS-stimulated mouse macrophages require different levels of Tpl2-MEK-ERK signaling (33).
In future studies our compounds will be critical tools for discovering and delineating the many aspects of Tpl2 signaling biology that are currently unknown. In the present work they have provided valuable insights into various functions of Tpl2 in human cell types, and our data lend positive support to the notion that Tpl2 is a valid target for therapeutic intervention in a range of human inflammatory diseases, including rheumatoid arthritis.