A novel B-RAF inhibitor blocks interleukin-8 (IL-8) synthesis in human melanoma xenografts, revealing IL-8 as a potential pharmacodynamic biomarker

The Raf kinases are key mediators of growth factor signals that effect cell survival, development, and proliferation. Of the three known isoforms of Raf, B-RAF is the sole family member known to be an oncogene (1–3). Indeed, mutations resulting in B-RAF activation, specifically B-RAF^V600E, have been identified in several cancers including melanoma, colorectal carcinoma, and papillary thyroid carcinoma (4–6). Mutational studies and crystallography have shown that the V600E mutation destabilizes the inactive conformation of the enzyme and produces a constitutively active kinase with ∼500-fold increase in activity (7). Cells expressing this form of mutated B-RAF have been shown to have very high basal extracellular signal-regulated kinase 1/2 (ERK1/2) phosphorylation. As a result, these cells produce large quantities of growth factors and chemokines dependent on mitogen-activated protein kinase signaling. One of these target genes encodes interleukin-8 (IL-8), a proinflammatory chemokine and potent autocrine factor known to promote tumor cell growth and angiogenesis (8). At least two known pathways are responsible for production of IL-8. ERK1/2 is known to activate AP1 was well as induce its expression, thus up-regulating IL-8 (9–11). In addition, phosphoinositide-3-kinase activation signals to nuclear factor-κB, thus inducing IL-8 expression (12).

IL-8 has already received attention as a prognostic factor for melanoma patients (13). Thus, it has been shown that patients with elevated levels of basic fibroblast growth factor, angiogenin, vascular endothelial growth factor, and IL-8 have poor survival rates. Interestingly, IL-8 was not found to be an early marker for melanoma development as it was only elevated in stage II and III patients. However, tumor burden and IL-8 did serve as independent predictive factors of progression-free survival. Consequently, IL-8 has emerged as an important measure of not only a patient's tumor burden but also their potential for disease-free survival.

The mitogen-activated protein kinase pathway is a well-accepted target for cancer therapy. Importantly, a compound with Raf inhibitory activity, Bay 43-9006 (sorafenib), recently gained approval for the treatment of renal cell carcinoma (14, 15). In addition, MEK inhibitors have shown early promise (16–18). Currently, clinical evaluation of drugs requires reproducible and robust biomarkers. The programs mentioned above have used phospho-ERK levels in ex vivo phorbol 12-myristate 13-acetate–stimulated peripheral blood mononuclear cells from drug-treated patients as a pharmacodynamic biomarker (19). These approaches are valid for evaluating compound doses and resultant bioavailability but do not necessarily correlate to intratumor, target-dependent activity. Analysis of a surrogate tumor-dependent biomarker is an important measure of a compound's ability to penetrate tumor tissue and therefore helps to guide dosing and potentially predict efficacy. There are several cases where tumor-dependent biomarkers are routinely used to assess disease progression. For example, prostate-specific antigen is a well-known indicator of benign prostatic hyperplasia as well as prostate cancer and has been used clinically since at least 1989 (20). Serum levels of P-selectin, cytokeratins, and sErbB1 are also gaining attention as potential tumor biomarkers (21–23). Other markers, MART1 and gp100, to name but a few, are commonly analyzed but require invasive biopsies and immunohistochemistry (24), thus making them far less convenient.

Described within is a series of diaryl imidazoles with potent anti-B-RAF kinase activity. These compounds have been shown to inhibit ERK1/2 phosphorylation and IL-8 production in a variety of melanoma cell lines harboring the B-RAF^V600E/D mutations. Importantly, compounds in this series reduce plasma and tumor IL-8 protein and mRNA levels when dosed in animals with established human melanoma xenograft tumors.

The RAF kinases were assayed at a concentration of 5 nmol/L in kinase reaction buffer [50 mmol/L HEPES (pH 7.4), 100 mmol/L NaCl, 10 mmol/L MgCl₂, Complete protease inhibitor cocktail (Roche Applied Science; used at a dilution of 1:200 when a tablet is dissolved in 1 mL water), 1 mmol/L NaF]. Briefly, compounds were diluted in 30% DMSO, 50 mmol/L HEPES and 10 μL were dispensed to reaction plates. Next, 50 μL of a buffer containing mouse MEK1-GST (100 nmol/L final) and [³³P]ATP (GE Healthcare; 10 μCi/mL final) were added to appropriate wells. Finally, 40 μL RAF kinase was added. Following 1-h incubation at room temperature, assays were stopped by the addition of 10 mmol/L ATP in 1% bovine serum albumin, 50 mmol/L EDTA. This mixture (80 μL) was transferred to 1% bovine serum albumin–blocked Flash Plates (Perkin-Elmer) coated previously overnight with a 1:2,500 dilution of anti-GST antibody (GE Healthcare). These plates were incubated for 2 h with shaking at room temperature. Plates were washed three times with PBS, 0.1% Triton X-100 and counted on a Top Count (Perkin-Elmer). IC₅₀s were determined by plotting percent of control against log M compound concentration and performing a nonlinear regression fit in GraphPad Prism (GraphPad Software).

An ELISA was developed to determine if Johnson & Johnson B-RAF inhibitors were ATP competitive. Briefly, kinase assays were set up as above with the following exceptions. Substrate solutions containing 2× mouse MEK were prepared with 2× cold ATP (200, 40, 8, 1.6, and 0.32 μmol/L). These substrate solutions were combined with dilution series of compounds before the addition of B-RAF. Assays were incubated for 15 min at room temperature and terminated by the addition of 10 μL 500 mmol/L EDTA. Assay mixtures were transferred to high binding ELISA plates (Corning) coated previously and blocked with anti-GST as above and incubated with shaking at room temperature for 2 h. Plates were washed three times as above and probed for 1 h with anti-phospho-MEK1/2 (Cell Signaling) at a 1:2,500 dilution in PBS with 1% bovine serum albumin and washed again. Finally, plates were incubated for 1 h with anti-rabbit horseradish peroxidase (Pierce) at 1:50,000 as above and washed. Plates were developed with TMB substrate (Chemicon) and stopped by the addition of 25 μL of 3 N HCl. Dose-response curves and competitive inhibition curves were produced with GraphPad Prism. K_i were determined from nonlinear regression analysis of data, where K_mapp = K_m * (1 + I / K_i) and Y = V_max * X / (K_mapp + X) (where X is concentration of ATP).

B-RAF-6HIS, B-RAF^V600E-6HIS, C-RAF-6HIS, and mouse MEK-GST were produced with a baculovirus expression system (Invitrogen). The RAF proteins were purified on nickel resin (Qiagen) and mouse MEK was purified on glutathione-Sepharose (GE Healthcare) followed by gel filtration on an S200 (GE Healthcare) to exclude endogenous insect glutathione-S-transferase. Isolated proteins were frozen at -80°C in the presence of 20% glycerol.

Melanoma cell lines A101D, A2058, A375, SW-MEL-28, WM-266-4 (B-RAF^V600D), colon carcinoma RKO, and glioblastoma multiforme DBTRG-05MG were obtained from American Type Culture Collection and cultured according to the recommended conditions. Cells were plated at 10,000 per well in a 96-well tissue culture plate the day before analysis. Growth medium was replaced by medium containing a dilution series of compound (previously diluted to 100× final concentration in 100% DMSO) resulting in a final DMSO concentration of 1%. Cells were returned to the 37°C incubator for 1 h after which they were washed once with ice-cold PBS and then lysed and assayed according to the manufacturer's instructions for the phospho-ERK2 ELISA kit (R&D Systems). Briefly, cell lysates were incubated in anti-ERK2-coated plates for 1 h, with shaking, at room temperature. Anti-phospho-ERK2 antibody was then applied to the plates for 1 h followed by washing and an anti-goat horseradish peroxidase–conjugated antibody for 30 min. Plates were washed and then developed with TMB substrate (Chemicon). IC₅₀ values were determined by plotting B/Bo versus log M and applying a one-site competition nonlinear regression fit in GraphPad Prism.

Soft agar colony formation assays were done as follows. Bottom layers were prepared first: 1% DNA grade agar was brought to boil in a microwave and cooled to 40°C in a water bath. DMEM (2×) was warmed to 40°C in water bath and allowed to equilibrate for at least 30 min. Equal volumes of the two solutions were mixed to give 0.5% agar plus 1× DMEM. Two milliliters were added to each well of six-well tissue culture plates and allowed to set. For top agar: 0.6% agar was brought to boil in a microwave and cooled to 40°C in a water bath. Two 2× RPMI plus 10% fetal bovine serum, penicillin, streptomycin, and glutamine were warmed to the same temperature. SK-MEL-28 cells were counted and adjusted to 500,000/mL. Dilution series of compounds were prepared at 2× in 1 mL of 2× medium in 15 mL tubes and kept at 40°C. The following steps were done in rapid succession. To compound containing medium 2× agarose, 0.1 mL of cell suspension was added then quickly mixed and poured onto bottom agar layer.

Plates were incubated at 37°C in humidified incubator for 10 to 14 days adding 2 mL 1× medium after 1 week. Plates were stained with 1.0 mL of 0.005% crystal violet for >1 h, and colonies were counted in five high-power fields on a dissecting microscope.

Melanoma cell lines A101D, A2058, A375, SW-MEL-28, WM-266-4 (B-RAF^V600D), colon carcinoma RKO, and glioblastoma multiforme DBTRG-05MG were exposed to compound 1 as for the phospho-ERK assay (described above) with the exception that the exposure time was 4 h. Supernatants were tested for IL-8 levels in an ELISA (M1918; Research Diagnostics). RNA was isolated from cell lysates (RNeasy; Qiagen) and quantitated (RNA LabChip; Agilent). Quantitative real-time PCR was done with IL-8 and GAPDH TaqMan gene expression assays (IL-8, Hs00174103_m1; GAPDH, 4333764T; Applied Biosystems) on a 7500 Fast Real-Time System according to the manufacturer's specifications.

WM-266-4 melanoma cells were resuspended in medium to 2 × 10⁷/mL, mixed 1:1 with Matrigel (BD 354234), and left on ice. nu/nu (nude) mice were injected in the right rear flank with 100 μL cell suspension. Tumors were allowed to grow to 200 to 500 mm³ before treatment. Animals were handled according to Institutional Animal Care and Use Committees protocol and regulations. Mice were sampled via retro-orbital bleed 18 h before dosing to obtain initial IL-8 levels. Before dosing, compound 1 was formulated at 6 mg/mL in a vehicle consisting of 5% pharmasolve and 30% SBE-cyclodextrin (CyDex). Mice received 100 μL (30 mg/kg) of compound or vehicle by s.c. injection. Plasma levels were maintained above 1 μmol/L over a 12-h period. At 9 and 12 h after the first dose, mice were euthanized via CO₂ asphyxiation and blood and tumor samples were taken. IL-8 levels in plasma were determined with an ELISA. For IL-8 mRNA studies, animals were treated as above with the following modifications. Mice received one or two doses of compound or vehicle at t₀ and t_2.5. Animals were euthanized as above at 2 or 4 h after the first dose. Tumors were placed immediately into liquid nitrogen. mRNA and cDNA were prepared according to standard procedures as above. IL-8 and GAPDH mRNA levels were quantitated with TaqMan RT-PCR (Applied Biosystems) as stated previously.

An in-house medicinal chemistry effort was initiated to identify B-RAF inhibitors with desirable drug-like qualities, including selectivity over other kinases, potency, and manageable pharmacokinetic properties. One example from this effort, referred to here as “compound 1,” contains pyrimidine and dichloro thiophene as the two aryl substituents of the imidazole core, is highlighted herein. Compound 1, with an ethylenediamine side chain, had K_i values versus recombinant B-RAF, B-RAF^V600E, and C-RAF in the low nanomolar range (Fig. 1). Compounds from this series were also found to be ATP-competitive inhibitors, showing increased IC₅₀ in the presence of increasing ATP concentrations (Fig. 2A). Compound 1, when docked in the ATP-binding pocket of the B-RAF crystal structure, is predicted to make donor-acceptor interactions with pyrimidine and aniline nitrogens and the protein backbone at Cys⁵³¹ (Fig. 2B). When analyzed for selectivity, compound 1 was found to block five kinases at >90% inhibition at 1 μmol/L compound and 10 μmol/L ATP (Supplementary Table S1).¹

As mentioned, B-RAF^V600E has ∼500-fold greater kinase activity than wild-type B-RAF. Consequently, many cells lines expressing mutant B-RAF have elevated phospho-ERK1/2 levels that are readily detected by standard ELISA techniques. Thus, phospho-ERK levels have served as a valuable measure of the intracellular activity of B-RAF inhibitors. For our first studies, we chose to evaluate the effect of B-RAF inhibitors in several different tumor types in an effort to establish how broadly applicable these compounds would be at blocking ERK phosphorylation. Therefore, melanoma cell lines A101D, A2058, A375, SK-MEL-28, and WM-266-4, colon cancer line RKO, and glioblastoma multiforme line DBTRG-05MG, all harboring activating B-RAF mutations, were exposed to compound 1 for 1 h and then the level of phospho-ERK was determined by ELISA. Compound 1 had a very similar effect in all of the cells tested, with an average IC₅₀ of 440 ± 122 nmol/L (Table 1). Because compound 1 showed no cellular selectivity, we focused our future work on the melanoma lines SK-MEL-28 and WM-266-4. Of these two, we chose to use SK-MEL-28 for routine screening of compounds in vitro and WM-266-4 for xenograft studies (as described below).

Recent studies with interfering RNA techniques have validated the role of B-RAF^V600E in cell proliferation, survival, and invasion (25). Here, small-molecule inhibitors are shown to possess similar properties. Thus, to evaluate the effect on anchorage-independent growth and cell proliferation, we tested compound 1 in a traditional soft colony agar formation assay as well as thymidine uptake proliferation assays. First, SK-MEL-28 were plated on agar in the presence of a range of compound doses and allowed to grow for 14 days. Compound 1 blocked colony formation with an apparent IC₅₀ of ∼3 to 5 μmol/L (Fig. 3A). In addition, compound 1 had an IC₅₀ of 2 μmol/L for inhibition of SK-MEL-28 proliferation over a 72-h period (Fig. 3B) with no appreciable loss in cell viability.

IL-8 expression is typically low to undetectable in resting cells. On exposure to proinflammatory or mitogenic signals, IL-8 levels rise dramatically (26). One key observation is that cells harboring the B-RAF^V600E/D mutation mimic a stimulated phenotype as they have high endogenous levels of phospho-ERK and constitutively produce IL-8 (27). These apparently ideal characteristics led us to evaluate IL-8 as a potential pharmacodynamic biomarker. Thus, we analyzed the effect of compound 1 on IL-8 production in SK-MEL-28, the same cells investigated for compound 1 effect on ERK phosphorylation. Cells were treated with compound 1 for 4 h and supernatants were analyzed for IL-8 levels. Interestingly, the colon cancer lines RKO, HT29, and Colo205 did not produce IL-8 at detectable levels, whereas the melanoma lines secreted variable amounts of IL-8 from 200 to 1,700 pg/mL (Table 1). Furthermore, there was no apparent correlation between phospho-ERK levels and amount of IL-8 produced. When evaluated for its potency in these cells, compound 1 showed dose-dependent inhibition of IL-8 production with IC₅₀s ranging from 109 to 889 nmol/L (Table 1). We sought to elucidate the mechanism of this reduction in IL-8 levels by analyzing IL-8 mRNA in compound 1–treated cells using quantitative real-time PCR. To simplify this experiment, we selected one cell line, SK-MEL-28, and exposed the cells to 10 μmol/L compound 1 for up to 6 h. As seen in Fig. 4, compound 1 dramatically reduced IL-8 message levels with the maximum effect being a 23-fold reduction observed at the 4-h time point. These results indicate that compound 1 blocked IL-8 synthesis at the transcriptional level, further elucidating the role of oncogenic B-RAF in regulating autocrine factors.

Clinical studies have shown that certain melanoma patients have elevated levels of IL-8 in their blood, which has served as a surrogate marker for tumor burden (13). In addition, the in vitro studies described herein show that IL-8 levels can be dramatically reduced in SK-MEL-28 cells treated with small-molecule B-RAF inhibitors. Taken together, these data suggest that IL-8 may serve as a predictive pharmacodynamic biomarker. Thus, an initial xenograft study was done to determine if mice harboring human melanoma xenografts also had elevated IL-8 in their blood and whether this correlated with tumor volume. For these experiments, the melanoma line WM-266-4 harboring the V600D mutation (equivalent in B-RAF kinase activity to V600E; ref. 7) was chosen because xenograft tumors from this line grew more consistently in size and rate (data not shown) and produced more IL-8 than SK-MEL-28, with levels after 4 h in culture of 730 versus 200 pg/mL for SK-MEL-28 (see Table 1). Furthermore, compound 1 blocked IL-8 production in WM-266-4 and SK-MEL-28 with the same IC₅₀ (Table 1). Hence, in this pilot xenograft study, nude mice were inoculated with either 1 × 10⁶ or 5 × 10⁶ WM-266-4 cells to identify optimal tumor growth conditions and evaluate IL-8 levels. Tumors were measured every 2 to 3 days and mice were bled over the course of the study to track IL-8 levels along with tumor growth (Fig. 5A and B). IL-8 plasma levels reached a maximum of 13 pg/mL in the mice that received 5 × 10⁶ cells and 8 pg/mL in the 1 × 10⁶ group. IL-8 plasma levels were well correlated with tumor size over the course of the study (Fig. 5C; r² = 0.61; P < 0.0001). Importantly, mice that were not inoculated with tumor cells had no detectable IL-8 in their plasma. Also, because the ELISA used for detection was highly specific for IL-8, it was concluded that only tumor-derived IL-8 was measured in these studies and not the endogenous mouse homologue, KC. Due to these findings, that IL-8 was readily detectable in mice with human melanoma tumors, experiments were done to evaluate IL-8 as a pharmacodynamic readout during compound treatment. Thus, mice with WM-266-4 xenograft tumors at an average volume of 500 mm³ received s.c. administration of compound 1 at 30 mg/kg over a 12-h period such that plasma drug levels would be >1 μmol/L (12 h was chosen due to the long half-life of plasma IL-8 as deduced from ref. 28). Further, s.c. administration was the preferred route of dosing as compound 1 had poor oral bioavailability. Blood and tumors were taken at 9 and 12 h after the first dose. Analysis of plasma samples revealed that treatment with compound 1 reduced IL-8 by 40% at 9 h (P = 0.05) and 84% at 12 h (P = 0.02; Fig. 6A). In support of these findings, tumor tissue was tested for drug levels at the end of the study and found to be, on the average, 3 μmol/L or >10-fold the IC₅₀ required for inhibiting ERK phosphorylation and IL-8 production in WM-266-4 melanoma cells in vitro. Taken together, these data suggest that compound 1 is able to penetrate tumor stroma and block B-RAF activity leading to the abolishment of ERK-dependent gene transcription. To substantiate this concept as well as shorten these studies, experiments were done to investigate the effect of s.c. compound administration on IL-8 mRNA levels in tumors. A shorter time course was chosen for these studies as in vitro experiments showed that a maximal effect on mRNA was observed 4 h after compound administration with a potential loss of effect at 6 h, whereas IL-8 clearance from plasma was modeled to take at least 6 to 9 h to reach baseline. In addition, the ultimate goal of these experiments was to establish a routine pharmacodynamic screen and dosing over a 12-h period was considered to be less desirable. Thus, mice received either one dose of compound 1 at 30 mg/kg at t₀ or two doses of compound 1 at 30 mg/kg, one at t₀ and the second at t_2.5, tumors were then harvested at 2 or 4 h after the first dose. TaqMan quantitative real-time PCR analysis of tumor cDNA revealed a 2-fold reduction in IL-8 mRNA for both dose groups, which was significant (P = 0.048) when compared with vehicle controls (Fig. 6B). Taken together, these results suggest that serum IL-8 protein and tumor IL-8 mRNA could serve as surrogate biomarkers for compound efficacy in xenograft pharmacokinetic/pharmacodynamic models, which may translate to clinical utility.

Mitogen-activated protein kinase inhibitors have been under development for the treatment of a variety of diseases for nearly 15 years. Recent advances in understanding the pathophysiologic role of dysregulated RAF signaling has led to the validation of this pathway in mediating disease and development of therapies for a variety of cancers. Indeed, since the identification of B-RAF mutations in a majority of melanomas in 2002 (1), several series of novel inhibitors have been produced and are heading towards (29, 30), or already in, the clinic. Described here is an example from our own B-RAF drug discovery program that has further elucidated the effect of inhibiting mutant B-RAF with in vitro and in vivo melanoma models.

RAF inhibitors have been available as tool compounds since the development of GW5074, ZM336372, and Bay 43-9006 (sorafenib) nearly a decade ago. These compounds show a variety of activities in kinase and cell-based assays. For example, ZM336372 exhibits potent inhibition of B-RAF kinase activity, with an IC₅₀ of 100 nmol/L, but, interestingly, induces rapid and strong activation of mitogen-activated protein kinase signaling independent of pathway agonism (31). Bay 43-9006, also a potent B-RAF inhibitor (∼20 nmol/L), has mixed allosteric and ATP competitive binding modes. This compound is ∼500-fold less potent in blocking ERK phosphorylation in cells that express B-RAF^V600E. This decrease in activity could be influenced by the fact that Bay 43-9006 binds preferentially to the inactive conformation of RAF kinases, a thermodynamically unfavored state for B-RAF^V600E (7). Described here is a purely ATP competitive compound that shows submicromolar kinase and cellular activity. Although a crystal structure of this compound bound to B-RAF is not available, docking of compound 1 suggests that a pyrimidine nitrogen and the aniline nitrogen perform donor-acceptor interactions with the protein backbone at Cys⁵³¹ (Fig. 2B). Further, the thiophene ring is favored to occupy a hydrophobic pocket adjacent to the ATP-binding site, whereas the ethylene diamine side chain points toward the solvent exposed space. This binding mode is similar to that found with other diaryl imidazole based kinase inhibitors, such as SB203580 bound to p38α (32).

The compound described above was profiled in several standard oncology related assays. For example, compound 1 was tested for its ability to inhibit SK-MEL-28 proliferation and colony formation (Fig. 3A and B). The results suggest that inhibition of B-RAF activity in cancer cells could lead to prevention of proliferation, metastasis, and invasion. Indeed, studies with small interfering RNA have shown that suppression of B-RAF^V600E expression in melanoma cells can block tumor growth and invasion and, in some cases, cause tumor shrinkage (25, 33, 34). Furthermore, recent evidence suggests that pharmacologic inhibition of B-RAF^V600E in melanoma xenografts has a similar effect (35). To a certain degree, these results could be explained by the ability of RAF pathway inhibitors to reduce the expression of integrins on the cell surface, thereby reducing their metastatic potential (26). It is also known that mitogen-activated protein kinase activity can regulate integrin expression as well as actin polymerization (36). Therefore, disrupting the anchorage-independent growth and invasiveness of melanoma cells are important properties of B-RAF inhibitors. These results are all related to the finding that B-RAF inhibition led to potent blockade of ERK1/2 phosphorylation (Table 1).

The pharmacokinetic properties of the diaryl imidazoles described here have prevented their evaluation in conventional xenograft studies. Therefore, a program was initiated to elucidate a viable pharmacokinetic/pharmacodynamic biomarker for activity of our compounds in mice bearing melanoma xenograft tumors. Preliminary DNA chip studies with compounds similar to compound 1 (data not shown) and literature review (8, 37) led us to IL-8 as a potential surrogate marker for antitumor activity. Indeed, it was found that IL-8 was produced at high levels by several melanoma cell lines. IL-8 was also readily detectable in plasma samples from mice bearing melanoma xenografts and IL-8 levels increased with tumor burden (Fig. 5B). Furthermore, when SK-MEL-28 cells were treated with compound 1 in vitro, IL-8 mRNA synthesis was dramatically reduced (Fig. 4). Importantly, it was found that the IC₅₀ for inhibition of ERK1/2 phosphorylation was similar to that for inhibiting IL-8 production, further suggesting the regulatory role of B-RAF in autocrine growth factor production. Subsequently, mice were dosed with compound 1 over a 4- or 12-h period to ensure a minimum level of 1 μmol/L inhibitor in the plasma. As shown in Fig. 6, significant reduction in tumor and IL-8 in plasma at 12 h (Fig. 6A) and IL-8 mRNA synthesis at 2 and 4 h (Fig. 6B) was observed. Taken together, these data strongly suggest that this compound not only penetrated tumor tissue but also potently inhibited intracellular B-RAF^V600E activity. As mentioned above, it has been shown that the concentration of IL-8 in plasma of melanoma patients can be representative of tumor burden (13). Thus, tracking IL-8 plasma levels would be an important measure in melanoma clinical trials where B-RAF inhibitors are employed. Further investigations should be carried out to characterize the effect of B-RAF activating mutations on IL-8 overexpression by other known B-RAF^V600E-containing tumors, such as colorectal and papillary thyroid carcinoma. In our experiments, only melanoma cells produced appreciable amounts of IL-8 (Table 1), which appeared to be independent of phospho-ERK levels.

In summary, we have disclosed a novel series of potent diaryl imidazoles that inhibit B-RAF^V600E in vitro and in vivo. These compounds should be useful for probing the mechanisms of B-RAF-dependent transformation and tumor biology. Furthermore, in vivo studies have revealed IL-8 plasma levels to be an important measure of a compound's ability to penetrate the tumor stroma and inhibit the B-RAF pathway. Finally, IL-8 could be an important noninvasive and quantitative pharmacodynamic biomarker for use in clinical trials where patients harbor tumors bearing B-RAF-activating mutations.