A Phase I Study of the Safety, Pharmacokinetics, and Pharmacodynamics of Combination Therapy with Refametinib plus Sorafenib in Patients with Advanced Cancer

The MAPK signaling pathway is constitutively activated in many cancers, causing uncontrolled cellular proliferation, survival, and metastasis (1–7). Mutations in receptor tyrosine kinases, or downstream intracellular MAPK signaling components (e.g., RAS), are common in various tumors and present a long-studied therapeutic target. Activation of the tyrosine/threonine kinase MEK causes activation of downstream extracellular signal-related kinase (ERK), which subsequently activates a range of downstream cytoplasmic and nuclear targets, promoting cell growth, differentiation, and survival (8–10). MEK is therefore an attractive candidate for inhibition in cancer, primarily due to its critical activating role in MAPK signaling, but also due to its structure, which allows for the targeting of selective pharmaceutical inhibitors (11, 12). However, due to the multifactorial nature of tumor growth and development, combined therapeutic approaches that engage multiple targets simultaneously are often necessary to reverse or inhibit tumor growth.

Refametinib (BAY 86-9766/RDEA119; Bayer Pharma AG) is a selective, orally available, potent, allosteric (non-adenosine triphosphate competitive) inhibitor of MEK1/2 (13) that has shown activity in multiple tumor types, including colorectal cancer (14) and preclinical models of hepatocellular carcinoma (HCC; ref. 15). Sorafenib (Nexavar®; Bayer Pharma AG) is an oral multikinase inhibitor with potent activity against Raf-1 and wild-type and mutant BRAF, and with antiangiogenic activity mediated by inhibition of vascular endothelial growth factor receptors and platelet-derived growth factors (16). Sorafenib is currently approved as monotherapy for advanced HCC (17), advanced renal cell carcinoma (18), and radioactive iodine-refractory differentiated thyroid cancer (19).

ERK1/2 signaling has been shown to be regulated by homeostatic feedback controls that include the direct phosphorylation of inhibitory sites on Raf-1 (20), which helped to explain, in part, the lack of efficacy of single-agent MEK inhibitors in the majority of cells. This led to the hypothesis that combinations of a MEK inhibitor and a Raf inhibitor would be synergistic (8), which was confirmed in a preclinical study of refametinib in combination with sorafenib (15).

This phase I study (NCT00785226) assessed the safety and tolerability of escalating daily oral doses of refametinib combined with sorafenib in patients with advanced solid malignancies. Secondary objectives were to determine the pharmacokinetics (PK) and pharmacodynamics of the combination and to describe any observed tumor response. Additional secondary objectives were to correlate toxicity and tumor response profiles to selected refametinib and sorafenib biomarkers, and to evaluate the safety and tolerability of the maximum tolerated dose (MTD) of this combination in an expansion cohort of 30 patients with advanced cancer amenable to biopsy, including 20 patients with HCC.

This study was performed in accordance with the Declaration of Helsinki, and documented approval from the appropriate ethics committees and institutional review boards was obtained for all participating centers prior to the study, where required (ClinicalTrials.gov identifier NCT00785226).

All patients gave written, informed consent. Patients were eligible if they: were aged 18 years or older; had an Eastern Cooperative Oncology Group performance status of 0 or 1; had cardiac function within normal range (as measured by echocardiogram or multigated acquisition scan); and had a histologically or cytologically confirmed advanced solid malignancy. Patients in the MTD expansion cohort had to have an unresectable tumor, either HCC (Child–Pugh A status) or melanoma, head and neck, colorectal, breast, or thyroid, amenable to biopsy (optional for HCC patients). Other inclusion criteria for all cohorts were: amylase and lipase ≤2 × upper limit of normal (ULN); hemoglobin ≥8.5 g/L; absolute neutrophil count ≥1500/mm³; platelet count ≥75,000/mm³; total bilirubin ≤1.5 × ULN; aspartate aminotransferase/alanine aminotransferase ≤2.5 × ULN (or ≤5 × ULN for patients with liver involvement); pro-thrombin time international normalized ratio/partial thromboplastin time ≤1.5 × ULN; and creatinine ≤1.5 × ULN. Exclusion criteria were: a prior or concurrent distinct cancer in a primary site other than the one being evaluated in the study; swallowing dysfunction or malabsorption syndrome that may impair treatment with the study drug; cytotoxicity from chemotherapy or adverse events (AE) from administration of other treatments 4 or more weeks prior; major surgery within 4 weeks; cardiac disease; or HIV infection. The use of inhibitors and inducers of CYP3A4 and CYP2C19 (enzymes known to metabolize refametinib) was to be discussed with the study sponsor.

This open-label study comprised a dose-escalation phase to determine the MTD of refametinib combined with sorafenib, and an MTD expansion phase to further characterize safety and tolerability. Patients received a single dose of refametinib 3 days before continuous treatment with refametinib and sorafenib, both twice daily, for 28 days, which constituted 1 course. Standard 3+3 design was used for the dose-escalation phase. Dose escalation progressed if no patients experienced dose-limiting toxicities (DLT) as defined by the National Cancer Institute Common Terminology Criteria for Adverse Events, version 3.0, after 28 days (Fig. 1). Protocol-defined DLTs included the following drug-related toxicities: grade 4 neutropenia lasting more than 7 days; febrile neutropenia (grade 3 or 4); grade 4 thrombocytopenia; grade ≥3 laboratory or non-hematologic toxicity; grade ≥3 lipase and/or amylase elevation; grade ≥3 skin toxicity (not hand-foot skin reaction); grade 4 anemia; grade ≥2 pulmonary hemorrhage/bleeding within 2 weeks of initial dose; grade ≥3 diarrhea (if persists with anti-diarrheal medication); grade ≥3 international normalized ratio or partial thromboplastin time with associated bleeding; grade ≥3 hemorrhage/bleeding; and missing 7 or more consecutive refametinib daily doses because of drug-related toxicity. If a DLT was observed in 1 of 3 patients in a cohort, 3 further patients were enrolled at the same dose. Dose escalation terminated if a DLT was observed in 2 or more patients in cohorts of 3 or 6 patients, and the dose below was declared the MTD of refametinib, combined with the full recommended dose of sorafenib (21) (Fig. 1). Dose escalation was not to exceed refametinib 50 mg twice daily, the MTD identified in a previous phase I study (NCT00610194). Following determination of the MTD, an expansion cohort was opened.

DLTs were managed with dose interruptions until recovery, or until discontinuation if interruptions lasted more than 4 weeks. Dermatologic toxicities were expected, and patients were instructed to moisturize twice daily with alcohol-free emollient and to minimize exposure to sunlight.

Screening included baseline demographics and characteristics, laboratory tests, blood samples for circulating tumor cell (CTC) enumeration, and tumor biopsy for consenting non-HCC patients in the MTD expansion cohort (if tissue blocks were unavailable). Blood samples for PK analysis of refametinib and metabolite M17 were collected on days 1 to 4 of course 1, and day 1 of course 2, pre-dose and at 0.5, 1, 3, 6, 8, and 24 hours post-dose for all cohorts, and on days 1 to 4 of course 1 at 48 and 72 hours for cohorts 4 and 5 and the MTD expansion cohort. Additional PK samples were collected on day 15 of courses 1 to 3. Blood samples for CTC analysis and tumor biopsies were collected in the fourth week of course 1 or the first week of course 2. Tumor biopsies were assessed for mutational status of KRAS, BRAF, and PIK3CA, and for immunohistochemical labeling of phosphorylated ERK (pERK), along with CTC samples, and Ki67 and phosphorylated Akt. Tumors were assessed by computed tomography and magnetic resonance imaging scans at screening, every 8 weeks, and at study termination (Response Evaluation Criteria in Solid Tumors version 1.0). Confirmatory scans were performed 4 to 6 weeks after the documentation of an objective response. Brow or scalp hair follicles were collected for pERK analysis on day 1 of course 1 pre-dose and at 2, 4, 24, and 26 hours post-dose, and during courses 2 and 3.

The sample size for the dose-escalation phase was not based on formal power calculations because the study objectives were focused on initial safety and tolerability assessments. Thirty patients were planned to be enrolled in the MTD expansion cohort. Descriptive statistics were used for safety analyses for all patients who received at least 1 dose of refametinib. Efficacy analyses included all patients who completed at least 2 courses of treatment. Phosphorylated ERK was analyzed using a 1-way analysis of variance with Dunnett's multiple comparison to compare pre- and post-drug treatment values. Matched biopsy samples were analyzed using paired t-tests.

Sixty-two patients were enrolled overall, with 32 in the dose-escalation phase: 3 in cohort 1, 8 in cohort 2A, 3 in cohort 2B, 4 in cohort 3, 6 in cohort 4, and 8 in cohort 5 (Table 1). Enrollment and dose escalation proceeded according to design with the following exceptions: in cohort 2A, 1 patient experienced 2 DLTs (rash and pruritus), leading to an expansion of the cohort to 6 patients, 2 of whom required replacement, and no additional DLTs were observed; in cohort 4, 1 patient experienced 1 DLT (rash), leading to the enrollment of 3 additional patients; and in cohort 5, 2 patients had to be replaced, and 1 patient experienced a DLT (hyperuricemia), leading to enrollment of 3 additional patients. Treatment was tolerated in cohort 5 (refametinib 50 mg twice daily plus sorafenib 400 mg twice daily), and as refametinib dosing above the monotherapy MTD of 50 mg twice daily was not planned, this was declared the MTD, prompting enrollment of 11 non-HCC and 19 HCC patients into the MTD expansion cohort.

The mean age was similar between non-HCC and HCC patients, and most patients were male and white (Table 1). Frequent non-HCC malignancies included colorectal cancer (53.5%) and melanoma (14.0%). More HCC patients had an Eastern Cooperative Oncology Group performance status of 1 compared with non-HCC patients (63.2% vs. 32.6%).

Non-HCC patients received a mean of 156.2 doses (range, 1–868) of refametinib and 134.8 doses (range, 0–416) of sorafenib. Mean treatment duration was 99.5 days (range, 12–477) with a mean of 3.6 courses (range, 1–17). HCC patients received a mean of 125.9 doses (range, 11–403) of refametinib and 92.6 doses (range, 10–215) of sorafenib. Mean treatment duration was 87.9 days (range, 15–265) with a mean of 3.2 courses (range, 1–9).

Forty-three non-HCC patients and 19 HCC patients were eligible for safety analysis. Primary reasons for study discontinuation included disease progression and AEs (Table 2). All patients experienced at least 1 AE, and most experienced at least 1 AE considered possibly related to refametinib or sorafenib (Table 2). Serious AEs were more frequent in non-HCC patients (41.9% vs. 21.1%); 10 non-HCC patients had serious AEs considered possibly related to refametinib. Six non-HCC patients and 1 HCC patient had grade 5 AEs, although none was considered treatment-related.

Common AEs (≥20% incidence) in non-HCC patients included diarrhea, fatigue, nausea, dermatitis acneiform, and vomiting (Supplementary Table S1); no dose-related trends in AE incidence were observed. AE findings were generally similar in HCC patients, with the addition of increased aspartate aminotransferase and peripheral edema. Frequent treatment-related grade 3 or 4 AEs included diarrhea, rash, and increased blood alkaline phosphatase in non-HCC patients, and diarrhea, dermatitis acneiform, and increased aspartate aminotransferase in HCC patients (Supplementary Table S2).

Of the 4 DLTs reported in 3 non-HCC patients, 3 (pruritus and rash in 1 patient, and rash in another) were considered possibly treatment-related; grade 4 hyperuricemia in 1 patient was considered unlikely to be treatment-related (Supplementary Table S3). During dose escalation, 5 non-HCC patients experienced AEs leading to study discontinuation; grade 3 left ventricular dysfunction (1 patient) and grade 4 subarachnoid hemorrhage (1 patient) were deemed possibly treatment-related, whereas all others (grade 2 nausea, grade 2 vomiting, grade 3 gastrointestinal hemorrhage, grade 4 fall, and grade 5 convulsion) were either not or not likely to be treatment-related. Most AEs leading to study discontinuation in HCC patients (hyperbilirubinemia, increased aspartate aminotransferase, diarrhea, dermatitis acneiform, ascites, thrombocytopenia, and left ventricular dysfunction) were grade 2 or 3 and were considered possibly treatment-related, except for delirium (grade 2) in 1 patient.

All patients participated in the single-dose PK assessment of refametinib. Refametinib was readily absorbed, with maximum plasma concentrations typically observed approximately 3 hours after administration (Fig. 2A; Supplementary Table S4). At the MTD, plasma concentration declined, with a geometric mean half-life of 16.7 hours, with comparable values for HCC and non-HCC patients (17.6 and 15.7 hours, respectively). In the dose range studied, PK parameters appeared to increase in a manner broadly consistent with dose proportionality, with moderate variability.

Multiple-dose plasma refametinib concentrations on day 1 of course 2 increased with dose (Fig. 2B; Supplementary Table S5). At the MTD, refametinib multiple-dose exposure was comparable in HCC and non-HCC patients; geometric mean maximum drug concentration (C_max) values were 708 ng/mL and 674 ng/mL, and area under the curve (AUC)_0–12 values were 6528 ng×h/mL and 5631 ng×h/mL, respectively. Multiple-dose AUC_0–12 values were higher than single-dose AUC_0–12 values by an average of approximately 40%.

Plasma concentrations of the inactive refametinib metabolite M17 generally increased with dose and were lower compared with refametinib, with higher variability. After single and multiple dosing, metabolite M17 AUC_0–12 accounted for approximately 12% to 31% of refametinib exposure without relationship to the refametinib dose administered or disease type. At the MTD, metabolite M17 geometric mean C_max values were 176 ng/mL and 188 ng/mL, and AUC_0–12 values were 1505 ng×h/mL and 1315 ng×h/mL, in HCC and non-HCC patients, respectively. Geometric mean half-life of metabolite M17 was approximately 20 hours, and up to 2-fold accumulation was observed after twice-daily dosing.

Plasma sorafenib concentrations at 200 mg twice-daily and 400 mg twice-daily dose levels showed moderate to high variability, and exposure decreased with increasing doses of refametinib up to 30 mg twice daily (Supplementary Table S6). At the MTD, sorafenib exposure was lower in HCC patients (C_max 2076 ng/mL) compared with non-HCC patients (C_max 3866 ng/mL).

Mutational analysis was performed on tumor biopsies collected from 10 patients in the MTD expansion cohort, for 6 of whom pERK analysis was performed in matched biopsies (Table 3). Two patients had KRAS-activating mutations (G12V and G13D), 1 had a BRAF-activating mutation (V600E), and 1 had a PIK3CA mutation. Two further patients (1 with colorectal cancer and 1 with head and neck cancer) were wild-type for all mutations tested. There was a significant reduction (average 62%) in pERK levels post-dose with refametinib and sorafenib compared with baseline, with a 113-point reduction in H-score (Table 3). Matched biopsies from the 3 patients with MAPK-activating mutations showed the largest reduction (average 80%); 1 patient with colorectal cancer (wild-type status) showed no reduction. The level of pERK reduction did not correlate with clinical outcome (data not shown). There was no significant reduction in the percentage of cells stained positively for Ki67 or in phosphorylated Akt H-score (data not shown).

Total and pERK values were measured in 307 hair follicle samples from 62 patients over 2 to 3 courses of treatment. There was a trend toward reduced pERK levels in hair follicles post-dose in cohorts 1 to 5, although this was not significantly different from baseline (Supplementary Fig. S1). In the MTD expansion cohort in course 1, there was a significant reduction of pERK of 54% and 65% at 2 and 4 hours post-dose, respectively (P < 0.001).

CTC enumeration with simultaneous pERK analysis was performed in 7 samples from 11 patients in the MTD expansion cohort. There was a significant reduction (35%) in CTC pERK after day 28 of course 1 compared with baseline (P < 0.05), although there was no significant change in CTC number (Supplementary Fig. S2).

Thirty-eight (88.4%) non-HCC patients and 16 (84.2%) HCC patients were evaluable for response assessment. For non-HCC patients, the disease control rate (complete response, partial response, and stable disease) for refametinib plus sorafenib was 65.8% (25 out of 38 patients); no patients achieved a complete response, 1 (2.6%) patient achieved a partial response, 24 (63.2%) patients achieved stable disease, of whom 16 (42.1%) had stable disease for 15 or more weeks, and 12 (31.6%) had progressive disease (Fig. 3A). One patient was not evaluable for response. The patient with a confirmed partial response had colorectal cancer, was in the MTD expansion cohort, was wild-type for all mutations tested, and had a durable response of 358 days. The disease control rate in HCC patients was 43.8%; no patients had a complete response or partial response, 7 (43.8%) had stable disease, of whom 6 (37.5%) had stable disease for 15 or more weeks, and 7 (43.8%) had progressive disease (Fig. 3B). Two were not evaluable for response.

This phase I trial evaluated the safety and efficacy of the combination therapy of refametinib with sorafenib in patients with advanced malignancies.

The baseline demographics and disease characteristics of the non-HCC population were broadly consistent with a previous phase I study of refametinib monotherapy (14), although the HCC population comprised slightly more males. Only HCC patients with Child–Pugh A status were enrolled to ensure patients had stable liver function, similar to previous sorafenib trials (22, 23).

Although sorafenib dosing had been established elsewhere as 400 mg twice daily, because of safety concerns the initial dose-escalation cohorts approached this level cautiously. Lack of toxicity in cohort 2A (refametinib 5 mg twice daily plus sorafenib 400 mg twice daily) led to escalation with sorafenib 400 mg twice daily as the backbone. The MTD was determined to be refametinib 50 mg twice daily in combination with sorafenib 400 mg twice daily, consistent with the phase I refametinib monotherapy study (14). Overall, daily oral dosing of refametinib plus sorafenib was well tolerated, up to and including at the MTD, with indications of clinical activity, consistent with the phase I study (14). Common treatment-emergent AEs included gastrointestinal toxicity, fatigue, dermatologic toxicities, and anorexia, as reported in the phase I study and in a study of patients with unresectable HCC (14, 24). Dermatologic toxicities were expected based on previous trials of other MEK inhibitors (25–28). Most AEs were managed by temporary dose interruptions or modifications, concomitant treatments, and supportive care. Serious AEs were reported by both non-HCC patients and HCC patients, although most were either not or not likely to be treatment-related. The higher incidence of serious AEs in non-HCC patients may, in part, be explained by the slightly longer duration of treatment compared with HCC patients.

Overall, refametinib displayed a dose-related increase in exposure that was similar between non-HCC and HCC patients at the MTD (refametinib 50 mg twice daily). Refametinib PK characteristics and single-dose exposure were generally consistent with those reported in the phase I trial of refametinib monotherapy (14), although multiple-dose exposure was slightly lower in the current study. Refametinib PK were not affected by co-administration of sorafenib, although at increased doses of refametinib, sorafenib exposure appeared to be lower than historical monotherapy data (14) and was lower in HCC patients compared with non-HCC patients. Conclusive interpretations of PK results are difficult due to the small sample sizes, particularly at the lower dose levels (2 to 5 patients per cohort), and observed variability in PK parameters. In a phase II study of refametinib and sorafenib (24), multiple-dose PK results were consistent with historical data, without any apparent decrease in exposure.

Preliminary antitumor activity was observed with the combination therapy; biopsies from patients with colorectal and head and neck cancers showed significant reductions in pERK levels compared with baseline levels. MAPK pathway-activating mutations were identified in 4 out of 6 patients and, although the sample size was small, those with activating mutations tended to show greater increases in the magnitude of pERK reduction. These preliminary data suggest that the combination of refametinib plus sorafenib was effective in inhibiting ERK phosphorylation in advanced tumors to a similar level as seen in the phase I monotherapy study (14). However, in contrast to the monotherapy study, 1 patient with head and neck cancer with no identified mutations showed a 50% reduction in pERK levels following treatment. Further, there was no significant correlation between pERK reduction and clinical benefit in this small data set. Indeed, the patient with colorectal cancer who did not have MAPK pathway-activating mutations had a durable partial response lasting for 358 days.

Hair follicle and CTC pERK levels were examined as a potentially less invasive surrogate pharmacodynamic marker of the refametinib plus sorafenib combination. A significant reduction in pERK levels in hair follicles was identified in the MTD expansion cohort in course 1, further suggesting that refametinib was effective in inhibiting ERK phosphorylation, although the sample size was small (at least 3 patients per time point). A statistically significant, yet modest, reduction in CTC pERK level during course 1 was also identified, although total CTC levels were not decreased. High levels of variation in CTC number and pERK level were expected based on previous reports of CTCs in mixed tumor types (29). No changes in phosphorylated Akt levels or in Ki67 labeling were observed post-dosing, suggesting that refametinib does not cause wide-reaching inhibitory effects in MAPK signaling pathways. Thus, although there is a trend for reduction in pERK with the refametinib plus sorafenib combination in hair follicles or CTCs, the applicability of this approach for use as a surrogate pharmacodynamic marker for this drug combination remains questionable.

The low overall response rate (1.9%; 1 out of 54) was not unlike the activity of sorafenib monotherapy (response rate of 2%, and 71% of patients achieved stable disease) in a phase III trial in HCC (SHARP; ref. 23). However, in the present study, patients tended to have received a greater number of prior therapies than in SHARP. The disease control rate was 65.8% (25 out of 38) in non-HCC patients, indicating the ability of refametinib combined with sorafenib to achieve increased disease control in a range of solid tumor types. The disease control rate of 43.8% (7 out of 19) observed in HCC patients was similar to that of 44.8% reported in a study of Asian patients with HCC (24). Of note in this study, 4 patients had RAS mutations, of whom 3 had confirmed partial responses ranging from 128 to 382 days. In contrast, in the present study, the lone partial response was in a patient with colorectal cancer whose tumor was wild-type for KRAS and BRAF, whereas 3 colorectal cancer patients with activating mutations failed to respond to the refametinib plus sorafenib combination. However, direct comparisons with the study in Asian patients with HCC are difficult due to differences in patient demographics and the small sample size. Furthermore, the mutational status of the HCC patients in the present study was not determined.

Overall, refametinib plus sorafenib combination therapy is well tolerated in 28-day courses up to the MTD in non-HCC and HCC patients, with a dose-proportional PK profile. The clinical benefit observed indicates that the combination therapy may be favorable for use in a variety of advanced solid-tumor types, including those where no MAPK pathway mutations are identified. Further investigation into the safety and efficacy of refametinib plus sorafenib combination therapy is therefore warranted, and phase II studies of refametinib as monotherapy and in combination with sorafenib in HCC patients prospectively identified to have RAS-activating mutations are ongoing (NCT01915589 and NCT01915602, respectively).

A. El-Khoueiry is a consultant/advisory board member for Bayer. F. Braiteh reports receiving speakers bureau honoraria from Amgen, Bayer, Bristol-Meyers Squibb, Celgene, Incyte, Insys, Ipsen, Pfizer, and Sanofi and is a consultant/advisory board member for Amgen, AstraZeneca, Bayer, Bristol-Meyers Squibb, Incyte, Insys, Pfizer, and Sanofi. A.F. Hezel reports receiving speakers bureau honoraria from and is a consultant/advisory board member for Bayer and Novartis. H. Krissel has ownership interest (including patents) in Bayer AG. N.J. Clendeninn is a consultant/advisory board member for Ardea. No potential conflicts of interest were disclosed by the other authors.

Conception and design: A.A. Adjei, F. Braiteh, D.P. Leffingwell, L.-T. Yeh, K.J. Manhard, P. Rajagopalan, J.N. Miner, H. Krissel, N.J. Clendeninn

Development of methodology: A.A. Adjei, J.J. Stephenson Jr, D.P. Leffingwell, L.-T. Yeh, P. Rajagopalan, N.J. Clendeninn

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A.A. Adjei, D.A. Richards, A. El-Khoueiry, F. Braiteh, C.H.R. Becerra, A.F. Hezel, M. Sherman, C. Iverson, L.-T. Yeh, S. Gunawan, D.M. Wilson

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): F. Braiteh, J.J. Stephenson Jr, D.P. Leffingwell, C. Iverson, Z. Shen, L.-T. Yeh, P. Rajagopalan, J.N. Miner, H. Krissel, N.J. Clendeninn

Writing, review, and/or revision of the manuscript: A.A. Adjei, D.A. Richards, A. El-Khoueiry, F. Braiteh, C.H.R. Becerra, J.J. Stephenson Jr, M. Sherman, L. Garbo, D.P. Leffingwell, C. Iverson, L.-T. Yeh, S. Gunawan, D.M. Wilson, K.J. Manhard, P. Rajagopalan, J.N. Miner, H. Krissel, N.J. Clendeninn

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): A.A. Adjei, D.P. Leffingwell, N.J. Clendeninn

Study supervision: A.A. Adjei, C.H.R. Becerra, J.J. Stephenson Jr, D.P. Leffingwell, N.J. Clendeninn

The authors thank Tanja Torbica, PhD, at Complete HealthVizion for assistance in the preparation and revision of the draft manuscript, based on detailed discussion and feedback from all the authors. Editorial assistance was funded by Bayer HealthCare Pharmaceuticals.

This work was funded by Bayer HealthCare Pharmaceuticals and Ardea Biosciences, and by a Conquer Cancer Foundation Drug Development Research Professorship (A.A. Adjei).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.