inhibition of mutated, activated braf in metastatic melanoma

n engl j med 363;9 nejm.org august 26, 2010 809

The new england journal of medicineestablished in 1812 august 26, 2010 vol. 363 no. 9

Inhibition of Mutated, Activated BRAF in Metastatic MelanomaKeith T. Flaherty, M.D., Igor Puzanov, M.D., Kevin B. Kim, M.D., Antoni Ribas, M.D.,

Grant A. McArthur, M.B., B.S., Ph.D., Jeffrey A. Sosman, M.D., Peter J. O’Dwyer, M.D., Richard J. Lee, M.D., Ph.D., Joseph F. Grippo, Ph.D., Keith Nolop, M.D., and Paul B. Chapman, M.D.

A bs tr ac t

From the Abramson Cancer Center of the University of Pennsylvania, Philadelphia (K.T.F., P.J.O.); Massachusetts General Hospital Cancer Center, Boston (K.T.F.); Vanderbilt University, Nashville (I.P., J.A.S.); the University of Texas M.D. An-derson Cancer Center, Houston (K.B.K.); UCLA, Los Angeles (A.R.); Peter MacCal-lum Cancer Centre, East Melbourne, VIC, Australia (G.A.M.); Roche Pharmaceuti-cals, Nutley, NJ (R.J.L., J.F.G.); Plexxikon, Berkeley, CA (K.N.); and Memorial Sloan-Kettering Cancer Center, New York (P.B.C.). Address reprint requests to Dr. Flaherty at Massachusetts General Hospital Can-cer Center, 55 Fruit St., Yawkey 9E, Boston, MA 02114, or at [email protected].

N Engl J Med 2010;363:809-19.Copyright © 2010 Massachusetts Medical Society.

Background

The identification of somatic mutations in the gene encoding the serine–threonine protein kinase B-RAF (BRAF) in the majority of melanomas offers an opportunity to test oncogene-targeted therapy for this disease.

Methods

We conducted a multicenter, phase 1, dose-escalation trial of PLX4032 (also known as RG7204), an orally available inhibitor of mutated BRAF, followed by an extension phase involving the maximum dose that could be administered without adverse ef-fects (the recommended phase 2 dose). Patients received PLX4032 twice daily until they had disease progression. Pharmacokinetic analysis and tumor-response assess-ments were conducted in all patients. In selected patients, tumor biopsy was per-formed before and during treatment to validate BRAF inhibition.

Results

A total of 55 patients (49 of whom had melanoma) were enrolled in the dose-esca-lation phase, and 32 additional patients with metastatic melanoma who had BRAF with the V600E mutation were enrolled in the extension phase. The recommended phase 2 dose was 960 mg twice daily, with increases in the dose limited by grade 2 or 3 rash, fatigue, and arthralgia. In the dose-escalation cohort, among the 16 pa-tients with melanoma whose tumors carried the V600E BRAF mutation and who were receiving 240 mg or more of PLX4032 twice daily, 10 had a partial response and 1 had a complete response. Among the 32 patients in the extension cohort, 24 had a partial response and 2 had a complete response. The estimated median pro-gression-free survival among all patients was more than 7 months.

Conclusions

Treatment of metastatic melanoma with PLX4032 in patients with tumors that carry the V600E BRAF mutation resulted in complete or partial tumor regression in the majority of patients. (Funded by Plexxikon and Roche Pharmaceuticals.)

The New England Journal of Medicine Downloaded from www.nejm.org at UC SHARED JOURNAL COLLECTION on September 8, 2010. For personal use only. No other uses without permission.

Copyright © 2010 Massachusetts Medical Society. All rights reserved.

T h e n e w e ngl a nd j o u r na l o f m e dic i n e

n engl j med 363;9 nejm.org august 26, 2010810

Metastatic melanoma is an aggres-sive disease for which there are few ef-fective therapies. The two therapies ap-

proved by the Food and Drug Administration, high-dose interleukin-2 and dacarbazine, are each associated with response rates of only 10 to 20% and a small percentage of complete responses; neither is thought to improve overall survival.1,2 In randomized trials, the median survival among patients treated with dacarbazine was less than 8 months.3,4

A search for mutations in a component of the mitogen-activated protein (MAP) kinase pathway in a large panel of common cancers revealed that 40 to 60% of melanomas, and 7 to 8% of all can-cers, carry an activating mutation in the gene en-coding the serine–threonine protein kinase B-RAF (BRAF).5-15 Ninety percent of reported BRAF mu-tations result in a substitution of glutamic acid for valine at amino acid 600 (the V600E mutation). This BRAF mutation constitutively activates BRAF and downstream signal transduction in the MAP kinase pathway. BRAF mutations are also found in 40 to 70% of papillary or anaplastic thyroid cancers6-8,16-18 and in a small percentage of sev-eral other types of tumor.

PLX4032 (Plexxikon; RG7204, Roche Pharma-ceuticals) is a potent inhibitor of BRAF with the V600E mutation. Preclinical studies showed that PLX4032 and its analogue PLX4720 inhibit the kinase activity of BRAF with the V600E mutation at low nanomolar concentrations, abrogate signal-ing through the MAP kinase pathway, and block proliferation of cells carrying BRAF with the V600E mutation in vitro at high nanomolar con-centrations.17,18 Orally administered PLX4720 in-hibits the growth — and, at higher doses, induces the regression — of human melanoma tumors transplanted into immunocompromised mice. None of these effects are observed in normal tis-sues or in tumor cells that lack a BRAF mutation.

We conducted a trial of the use of PLX4032 in patients with metastatic cancer. The primary goals were to define the safety and pharmacokinetic characteristics of treatment with continuous, twice-daily administration of PLX4032, to deter-mine the maximum dose that could be adminis-tered until adverse effects prevented further dose increases (i.e., the recommended phase 2 dose), and to determine the objective response rate, the duration of response, and the rate of progression among patients who had melanoma tumors with

the V600E BRAF mutation and who were given the recommended phase 2 dose of PLX4032.

Me thods

Study Design

The study was sponsored by Plexxikon and Roche Pharmaceuticals, which provided the study drug. The study was designed by two academic authors and one industry author. All authors made the de-cision to submit the manuscript for publication. All authors analyzed the data, prepared the man-uscript, and vouch for the completeness and ac-curacy of the data and analyses. The study was conducted in accordance with the protocol.

Dose-Escalation PhasePLX4032 was initially in a crystalline formulation. In the dose-escalation phase of the study, which involved several consecutively enrolled groups of three to six patients, the first group received 200 mg of PLX4032 by mouth daily; subsequent groups received the drug at higher doses, according to a dose-escalation scheme. This formulation was found to have poor bioavailability (see the Results section), and enrollment for the dose-escalation phase was halted while the drug was reformu-lated as a highly bioavailable microprecipitated bulk powder, initially as a 40-mg capsule and sub-sequently as 80-mg and 120-mg capsules, as well as 240-mg tablets. Enrollment was resumed, with newly enrolled patients receiving the microprecip-itated-bulk-powder formulation at a dose of 160 mg (two 80-mg capsules) twice daily, with subsequent escalation.

Patients received continuous therapy with PLX4032 unless unacceptable side effects or dis-ease progression occurred. Doses were not esca-lated unless the patients receiving the highest current dose had been observed for at least 4 weeks and dose-limiting side effects had been reported in fewer than a third. Dose escalation in a given patient was permitted if the safety and adverse-effect profile had been established for the next highest dose. The recommended phase 2 dose was defined as the highest dose at which no more than one of six patients had dose-limiting side effects.

Extension PhaseOnce the recommended phase 2 dose had been identified, an extension cohort was treated at this



Inhibition of Mutated, Activated BR AF


dose. In this cohort, all patients had melanoma and a prospectively identified V600E BRAF muta-tion. The primary objective in the extension co-hort was to determine the response rate. Second-ary objectives were to define the toxicity and pharmacokinetics of PLX4032 more precisely and to obtain data on pharmacodynamic effects.

Patients

Eligibility criteria included the provision of writ-ten informed consent, an age of 18 years or older, solid tumors confirmed histologically that were refractory to standard therapy or for which stan-dard or curative therapy did not exist, Eastern Cooperative Oncology Group performance status score of 0 (able to be fully active and carry out all predisease activities without restriction) or 1 (un-able to perform physically strenuous activity but ambulatory and able to carry out work of a light or sedentary nature, such as light housework or office work),19 life expectancy of 3 months or longer, absence of known progressing or unsta-ble brain metastases, and adequate hematologic, hepatic, and renal function.

The dose-escalation phase of the trial was open to patients with any type of tumor, although pa-tients who had melanomas with the V600E mu-tation in BRAF were overrepresented because of the selective activity of PLX4032 against such tu-mors in preclinical testing. For the extension co-hort, eligibility was restricted to patients with melanomas harboring a V600E BRAF mutation, as ascertained by means of a polymerase-chain-reaction assay (TaqMan, Applied Biosystems). This assay involves hybridizing a probe specific to the 1799T→A substitution that results in the V600E BRAF mutation with DNA isolated from forma-lin-fixed, paraffin-embedded tumor tissue and determining the presence or absence of amplifi-cation after repeated chain-reaction cycles.20

Study Assessments

Safety evaluations were conducted at baseline, day 8, day 15, day 29, and every 4 weeks thereafter. These evaluations included a physical examina-tion, electrocardiography, laboratory studies that included a complete blood count, clinical chemi-cal testing, and urinalysis. Adverse events were graded according to the Common Terminology Criteria for Adverse Events (version 3.0) (http://ctep.info.nih.gov/protocolDevelopment/electronic_applications/docs/ctcaev3.pdf). During the trial,

squamous-cell carcinoma, keratoacanthoma type, was observed in several patients (see the Results section). As a result, the protocol was amended to ensure that patients underwent dermatologic evaluations at baseline and every 2 months dur-ing the study; computed tomographic (CT) scans of the chest were analyzed for the appearance of new lesions suggestive of a primary cancer.

CT studies were performed at 8-week intervals during therapy in all patients and at the end of the first 4 weeks of therapy in some patients. The findings were judged according to the Response Evaluation Criteria in Solid Tumors (RECIST), ver-sion 1.0. A complete response was defined as the disappearance of all target lesions, and a partial response as a decrease by at least 30% in the sum of the largest diameter of each target lesion, rela-tive to the corresponding sum at baseline. Pro-gressive disease was defined as an increase by at least 20% in the sum of the largest diameter of each target lesion, relative to the smallest corre-sponding diameter recorded since the start of treatment, or the appearance of one or more new lesions. Stable disease was defined as the absence of shrinkage sufficient for a partial response and the absence of enlargement sufficient for pro-gressive disease, relative to the corresponding sum at baseline. Progression-free survival was defined as the time from the first day of treatment to the first documentation of disease progression or death, whichever occurred first.

Pharmacokinetic assessments were made on day 1 and day 15 of the first 4 weeks of therapy, and single plasma samples were obtained every 4 weeks during the study. Plasma samples were analyzed by means of high-performance liquid chromatography, with detection by means of mass spectroscopy.

Patients in the dose-escalation phase who were receiving a dose greater than 160 mg twice daily and patients in the extension cohort underwent 18F-fluorodeoxyglucose–positron-emission tomog-raphy (FDG-PET) at baseline and on day 15 of the first 4 weeks of therapy. Selected patients un-derwent tumor biopsy before the start of therapy as well as on day 15. Biopsy specimens were im-mediately fixed in formalin for analysis of phos-phorylated extracellular signal-regulated kinase (ERK), the protein cyclin D1, and the monoclonal antibody Ki-67 by means of immunohistochem-istry. The sampled tumors were cutaneous or su-perficial lymph-node lesions; except in the case of





one patient, sequential biopsy specimens were not taken from the same lesion.

Statistical Analysis

For the primary end point of a partial or com-plete response in the extension cohort, we cal-culated that a sample of 32 patients would pro-vide 95% confidence (α = 0.05), with 80% power (β = 0.20), that an observed response rate of 40% would be consistent with a true response rate of more than 10%, which was considered justifica-tion for further study. For this report, January 31, 2010, was the cutoff date for the safety and effi-cacy follow-up.

R esult s

Patients

Fifty-five patients were enrolled in the dose-esca-lation phase of the study (see Fig. 2 in the Supple-mentary Appendix, available with the full text of this article at NEJM.org); 32 additional patients with metastatic melanoma that carried the V600E BRAF mutation were treated at the recommended phase 2 dose in the extension phase (Table 1). The majority of patients (49 of 55 [89%]) in the dose-escalation cohort had metastatic melanoma; 3 of the remaining 6 patients had papillary thyroid cancer that carried the V600E BRAF mutation. Screening for the V600E BRAF mutations was not a requirement for study entry during the dose-escalation phase, but throughout the trial, an in-creasing number of patients were prospectively identified as having the mutation (for a total of 16 of the 21 patients with melanoma who were enrolled in the groups receiving 240 mg twice daily to 1120 mg twice daily).

The initial crystalline formulation of PLX4032 was found to have no dose-limiting side effects or antitumor activity at doses of 200 mg daily to 1600 mg twice daily in consecutively enrolled groups of patients. Since the serum levels detect-ed were lower than the levels predicted in preclini-cal models to be effective, a microprecipitated-bulk-powder formulation with substantially higher bioavailability was developed and used for the remainder of the study; the lowest dose was 160 mg twice daily, and escalated doses were 240, 320 or 360, 720, and 1120 mg twice daily (Fig. 2 in the Supplementary Appendix). Patients who had been receiving the crystalline formula-tion were switched to the microprecipitated-bulk-powder formulation. The protocol allowed the dose to be reduced if side effects developed.

Once the recommended phase 2 dose was es-tablished for the microprecipitated-bulk-powder formulation in the dose-escalation cohort, the ex-tension cohort was enrolled. All patients in the extension cohort had melanoma carrying the V600E BRAF mutation (Table 1).

Adverse Events

During the latter part of the dose-escalation phase of the trial, when the microprecipitated-bulk-powder formulation of PLX4032 was used, dose-limiting toxic effects were not observed un-til a dose of 720 mg twice daily was given. At the next-highest dose given to one group of patients,

Table 1. Baseline Characteristics of the Patients, According to Study Cohort.*

Characteristic

Dose-Escalation Cohort (N = 55)

Extension Cohort (N = 32)

Age — yr

Median 63 52

Range 23–89 23–83

Sex — no. (%)

Male 34 (62) 19 (59)

Female 21 (38) 13 (41)

Tumor type — no. (%)

Melanoma 49 (89) 32 (100)

Thyroid 3 (5) 0

Other 3 (5) 0

Extent of metastatic melanoma — no. (%)†

Stage M1a 7 (14) 6 (19)

Stage M1b 6 (12) 2 (6)

Stage M1c 36 (73) 24 (75)

LDH >ULN — no. (%) 13 (41)

ECOG performance status score — no. (%)

0 28 (51) 15 (47)

1 27 (49) 17 (53)

No. of previous chemotherapy regimens — no. (%)‡

0 5 (10) 7 (22)

1 16 (33) 9 (28)

2 5 (10) 4 (12)

≥3 23 (47) 12 (38)

* ECOG denotes Eastern Cooperative Oncology Group, LDH lactate dehydroge-nase, and ULN upper limit of the normal range.

† The extent of metastatic melanoma is described as the American Joint Com mit-tee on Cancer stage. These data were not reported for the six patients in the dose-escalation cohort who did not have melanoma.

‡ The number of previous therapies was not reported for the six patients in the dose-escalation cohort who did not have melanoma.





1120 mg twice daily, four of the six patients had dose-limiting side effects: three patients with grade 3 rash (two of whom also had grade 3 fa-tigue) and one patient with grade 3 arthralgia (Table 2). A dose of 960 mg twice daily was eval-uated and determined to be tolerated in the first six patients given the dose, so it was established as the recommended phase 2 dose for the exten-sion cohort. Those six patients were included as

the first six patients in the extension cohort. In the extension cohort, 13 patients (41%) required a dose reduction during therapy (to 720 mg twice daily in 10 patients, to 600 mg twice daily in 1 pa-tient, and to 480 mg twice daily in 2 patients). The most common PLX4032-related grade 2 or 3 side effects observed were arthralgia, rash, nausea, pho-tosensitivity, fatigue, cutaneous squamous-cell car-cinoma, pruritus, and palmar–plantar dysesthe-

Table 2. Drug-Related Adverse Events of Grade 2 or Higher Reported in More Than 5% of the 87 Study Patients, According to the Dose of PLX4032 Given Twice Daily.*

Event<240 mg (N = 30)

240 mg (N = 4)

320 or 360 mg (N = 8)

720 mg (N = 7)

960 mg (N = 32)

1120 mg (N = 6)

Total (N = 87)

number (percent)

Arthralgia

Grade 2 0 1 (25) 2 (25) 0 10 (31) 1 (17) 14 (16)

Grade 3 0 0 0 0 1 (3) 1 (17) 2 (2)

Rash

Grade 2 1 (3) 0 0 1 (14) 7 (22) 1 (17) 10 (12)

Grade 3 0 0 0 0 1 (3) 3 (50) 4 (3)

Squamous-cell carcinoma, keratoacanthoma type

Grade 2 0 0 0 0 0 0 0

Grade 3 1 (3) 2 (50) 3 (38) 0 10 (31) 2 (33) 18 (21)

Nausea

Grade 2 1 (3) 0 1 (11) 1 (14) 4 (12) 1 (17) 8 (9)

Grade 3 0 0 0 0 1 (3) 0 1 (1)

Fatigue

Grade 2 0 0 0 0 2 (6) 1 (17) 3 (3)

Grade 3 0 0 0 0 2 (6) 2 (33) 4 (5)

Photosensitivity reaction

Grade 2 0 0 0 1 (14) 4 (12) 1 (17) 6 (7)

Grade 3 0 0 0 0 1 (3) 0 1 (1)

Palmar–plantar dysesthesia

Grade 2 0 0 0 0 2 (6) 1 (17) 3 (3)

Grade 3 0 0 0 0 2 (6) 0 2 (2)

Pruritus

Grade 2 0 0 0 0 4 (12) 0 4 (5)

Grade 3 0 0 0 0 0 1 (17) 1 (1)

Lymphopenia

Grade 2 0 0 2 (25) 0 2 (6) 0 4 (5)

Grade 3 0 0 0 0 0 1 (17) 1 (1)

* Adverse events were graded according to the Common Terminology Criteria for Adverse Events (version 3.0) (http://ctep.info.nih.gov/protocolDevelopment/electronic_applications/docs/ctcaev3.pdf). Four patients had a grade 4 adverse event: two had elevated γ-glutamyltransferase levels, one had fatigue, and one had reversible pancytopenia of uncertain attribution.





sia (Table 2). In total, 89% of all side effects were grade 1 or 2. Rashes were evenly distributed among the face or neck, trunk, and extremities.

Eight patients in the dose-escalation cohort (15%) and 10 patients in the extension cohort (31%) had cutaneous squamous-cell carcinomas, with a total of 35 carcinomas. These were re-viewed centrally, and all but one either were kera-toacanthoma type or had features of a kerato-acanthoma. The median time to the appearance of a cutaneous squamous-cell carcinoma was 8 weeks; the majority of the carcinomas were

resected, and in no case did they lead to discon-tinuation of treatment. No squamous-cell carci-nomas at other organ sites were observed dur-ing the study.

Pharmacokinetics

Pharmacokinetic analyses revealed that exposure increased with the dose throughout the range of doses of the microprecipitated-bulk-powder for-mulation that were administered, with exposure being proportional to dose for doses of 240 mg twice daily through 960 mg twice daily. At the recommended phase 2 dose, 960 mg twice daily, the mean (±SD) area under the plasma concentra-tion–time curve over a 24-hour period (AUC0–24) was 1741±639 μM×hour (Fig. 1A), and the mean maximum concentration at steady state was 86±32 μM (Fig. 1B). The mean PLX4032 level on day 15, after repeated dosing, was six to nine times the mean level on day 1, and its mean half-life was approximately 50 hours (range, 30 to 80). With the twice-daily dosing regimen, all patients were exposed to relatively constant daily levels of the drug at steady state.

Pharmacodynamics

Tumor-biopsy specimens that had been obtained at baseline and at day 15 were available for seven of the patients in the extension cohort who were receiving PLX4032 at the recommended phase 2 dose. Tumor levels of phosphorylated ERK, cyclin D1, and Ki-67 were markedly reduced at day 15 as compared with baseline in all specimens test-ed (Fig. 2A). This finding suggests that PLX4032

Mea

n PL

X40

32 A

UC

0–24

(µM

×hr

)

4000

3000

2000

1000

00 200 400 600 800 1000 1200

Twice-Daily Dose (mg)

B

A

Mea

n PL

X40

32 C

once

ntra

tion

(µM

)

160

120

80

40

00 6 12 18 24

Hour

Day 15

Day 1

Figure 1. Mean PLX4032 Concentrations over a 24-Hour Period.

Data are shown for the microprecipitated-bulk-powder formulation. Panel A shows the mean area under the plasma concentration–time curve (AUC0–24), according to the twice-daily dose. Panel B shows the mean concentration after the administration of a single dose, on day 1, or multiple doses at the steady state, on day 15, at the recommended phase 2 dose of 960 mg twice daily. I bars indicate standard deviations.

Figure 2 (facing page). Representative Findings of the Effect of PLX4032 at the Recommended Phase 2 Dose in Study Patients with Melanoma That Carried the V600E Mutation.

The recommended phase 2 dose was 960 mg twice daily. Panel A (hematoxylin and eosin) shows immunohisto-chemical analyses of the expression of phosphorylated extracellular signal-regulated kinase (ERK), cyclin D1, and Ki-67 in tumor-biopsy specimens obtained at base-line and on day 15 of treatment. Panel B shows the up-take of 18F-fluorodeoxyglucose (FDG) at baseline and on day 15 of treatment, as assessed by means of posi-tron-emission tomography (PET). Panel C shows com-puted tomographic images of lesions (arrows) in lung, liver, and bone (with each pair of images shown for a different patient) at baseline and at 8 weeks.





B FDG-PET

C Computed Tomography

A Tumor Biopsy

Baseline

Phosphorylated ERK Cyclin D1 Ki-67

Day 15

Baseline

Lung Liver Bone

8 Wk

Baseline Day 15





inhibited the MAP kinase pathway, resulting in decreased cyclin D1 levels and decreased prolifera-tion. In virtually all patients, a marked decrease in tumor uptake of FDG was noted at day 15 (Fig. 2B).

Tumor Response

Dose-Escalation PhaseNo responses were observed at doses of 160 mg twice daily of the microprecipitated-bulk-powder formulation or at any dose of the crystalline for-mulation. Of the patients receiving doses of 240 mg or more twice daily, 16 had melanoma with tu-mors that harbored the V600E BRAF mutation. Among these 16 patients, a partial or complete re-sponse was seen in the 1 patient receiving 240 mg twice daily, 2 of the 4 patients receiving 320 or 360 mg twice daily, 4 of the 6 patients receiving 720 mg twice daily, and 4 of the 5 patients receiv-ing 1120 mg twice daily. The overall response rate was 69% (11 of 16 patients), with 10 partial responses and 1 complete response (Fig. 1A in the Supplementary Appendix). Responses were seen at all sites of metastatic disease, including the liver, small bowel, and bone (Fig. 2C). The dura-tion of the response ranged from 2 to more than 18 months, with 4 patients still having a partial or complete response at the data cut-off date (Fig. 1B in the Supplementary Appendix). In addition to the patients with melanoma, the 3 patients with papillary thyroid cancer had a partial or complete response, with the response lasting 8 months in 1 patient (who was progression-free for 12 months) and stable disease lasting 11 and 13 months in each of the other 2 patients.

A total of 5 patients with metastatic melanoma whose tumors did not have a BRAF mutation re-ceived doses of 240 mg or more twice daily. None had evidence of tumor regression during the study; 4 had progressive disease within the first 2 months of treatment.

Extension PhaseThe extension cohort consisted solely of patients who had melanoma with the V600E BRAF muta-tion. All were treated at the recommended phase 2 dose of 960 mg twice daily. A total of 26 of the 32 patients had a response (81%), with a com-plete response in 2 patients and a partial response in 24 patients (Fig. 3A). Among patients with symp-tomatic metastatic disease, improvement of symp-toms, such as a reduced need for narcotics for pain in 3 patients, was reported within 1 to 2 weeks. As in the dose-escalation cohort, we observed tu-

mor responses in visceral organs and bone me-tastases as well as more typically responsive sites such as the lungs and lymph nodes. Responses were also routinely observed in patients with el-evated lactate dehydrogenase levels (10 partial re-sponses among the 13 patients) and in patients who had received more than one previous type of therapy (11 partial responses among the 16 pa-tients). To date, 16 of the 32 patients are still in the study; the estimated median progression-free sur-vival among these patients is more than 7 months, on the basis of a Kaplan–Meier analysis. The es-timated median overall survival has not been reached.

Discussion

Our trial shows that therapy targeting tumors containing activating V600E BRAF mutations can induce complete or partial tumor regression in patients. PLX4032 induced complete or partial tumor regression in 81% of patients who had mel-anoma with the V600E BRAF mutation. Responses were observed at all sites of disease, including the bone, liver, and small bowel. During the dose-escalation phase of the trial, we also saw responses in patients who were receiving doses below the recommended phase 2 dose. These efficacy data are particularly encouraging in light of the high disease burden in most of our patients and the presence of symptomatic disease in many of them.

Most side effects related to PLX4032 appeared to be proportional to the dose and exposure to the drug. Cutaneous side effects, fatigue, and ar-thralgia predominated. In the extension cohort, at the dose of 960 mg twice daily, approximate-ly 40% of patients required a short- or long-term reduction in dose to 720 mg, 600 mg, or 480 mg twice daily, many for grade 2 side effects. Squa-mous-cell carcinoma, keratoacanthoma type, de-veloped in a total of 10 of 32 patients (31%); we also observed squamous-cell carcinoma, kerato-acanthoma type, in patients in the dose-escala-tion cohort. The characteristic rapid eruption of individual, dome-shaped, nonpigmented lesions and histologic findings were present in each case. Usually, squamous-cell carcinoma, kerato-acanthoma type, are well-differentiated tumors with very low invasive potential and no metastatic potential; our data do not allow us to determine their behavior in patients receiving PLX4032. Re-cent data show that BRAF inhibitors can activate the MAP kinase pathway in cells that lack a BRAF





% C

hang

e fr

om B

asel

ine

125

75

100

50

25

−25

−50

0

−75

−100

B Response over Time

A Best Overall Response

Melanoma stage M1a

Melanoma stage M1b

Melanoma stage M1c

0 2 4 6 8 10 12

Months

Melanoma stage M1a

Melanoma stage M1b

Melanoma stage M1c

Time of partial orcomplete response

Progressive disease

Patient remainingin study

Thresholdfor response

accordingto RECIST

Figure 3. Antitumor Response in Each of the 32 Patients in the Extension Cohort.

All 32 patients had melanoma tumors that carried the V600E mutation of the v-raf murine sarcoma viral oncogene ho-mologue B1 (BRAF). All were treated at the recommended phase 2 dose of 960 mg twice daily. Panel A shows the best overall response for each of the 32 patients, measured as the change from baseline in the sum of the largest diameter of each target lesion. Negative values indicate tumor shrinkage, and the dashed line indicates the threshold for a par-tial response according to Response Evaluation Criteria in Solid Tumors (RECIST) (i.e., shrinkage by 30%). Two pa-tients had a complete response. Panel B shows the duration and characteristics of the responses in each patient.





mutation.21-23 This activation may pertain to some of the side effects seen with PLX4032.

Though the early response to PLX4032 seems to occur reliably, responsive tumors can develop resistance to treatment. Among the patients in the dose-escalation cohort who had a response to treatment, the duration of the response ranged from 2 to more than 18 months. The mechanism of secondary tumor resistance is not yet known. We also observed that in some patients with V600E BRAF mutations, the tumors showed re-sistance without evidence of an early response. The mechanism of this primary refractory state is currently under investigation. To date, we have not seen “gatekeeper” BRAF mutations in resis-tant tumors, although this issue requires more investigation.

Sorafenib, which inhibits BRAF (both the wild type and the V600E mutant) and v-raf-1 murine leukemia viral oncogene homologue 1 (RAF1), has been studied in melanoma. In animal models of melanoma, sorafenib has not shown selective ac-tivity against tumors carrying BRAF mutations, and in clinical trials, sorafenib used either alone or in combination with chemotherapy has not had significant antimelanoma effects.24-27 It is possi-ble that the non-BRAF effects of sorafenib mediate side effects that limit the likelihood of achieving a drug concentration that is high enough to in-hibit the V600E BRAF mutation.

It is now clear that melanomas can be catego-rized by specific molecular changes that drive their proliferation.28 The overriding hypothesis

is that inhibition of the activated pathway in the individual tumor will lead to tumor regression. There is recent preliminary evidence that ima-tinib can induce regression in 33% of the small proportion of melanomas driven by mutations in KIT (the v-kit Hardy–Zuckerman 4 feline sarco-ma viral oncogene homologue).29 In our study, PLX4032 induced responses in the vast majority of melanomas caused by BRAF mutations, which constitute 40 to 60% of all melanomas. We do not yet know whether treatment with PLX4032 will improve overall survival; an ongoing phase 3 trial (ClinicalTrials.gov number, NCT01006980) is addressing that question.

Supported by Plexxikon and Roche Pharmaceuticals.Drs. Flaherty, Puzanov, Kim, Ribas, McArthur, Sosman,

O’Dwyer, and Chapman report the receipt by their institutions of grant support from Plexxikon to conduct this clinical trial; Drs. Flaherty, Puzanov, Kim, Ribas, Sosman, and Chapman re-port receiving consulting fees or reimbursement for travel ex-penses from Roche Pharmaceuticals; Dr. Sosman reports pend-ing receipt of grant support from Roche Pharmaceuticals; Dr. O’Dwyer reports receiving grant support from Plexxikon; Drs. Lee and Grippo report being employees of Roche Pharmaceuti-cals; and Dr. Nolop reports being an employee of Plexxikon, holding equity in the company, and receiving reimbursement for travel expenses from the company. No other potential conflict of interest relevant to this article was reported.

Disclosure forms provided by the authors are available with the full text of this article at NEJM.org.

We thank Drs. Katherine Nathanson and Xiaowei Xu (of the University of Pennsylvania) for leading the analysis of tumors for BRAF mutations and the immunohistochemical analysis of phosphorylated ERK and Ki-67 during the dose-escalation part of the study, Drs. Astrid Koehler and Michael Stumm (of Roche Pharmaceuticals) for leading these analyses during the exten-sion part of the study, and Dr. Ruben Ayala (of Roche Pharma-ceuticals) for the pharmacokinetic analysis.

References

Comis RL. DTIC (NSC-45388) in ma-1. lignant melanoma: a perspective. Cancer Treat Rep 1976;60:165-76.

Atkins MB, Lotze MT, Dutcher JP, et 2. al. High-dose recombinant interleukin 2 therapy for patients with metastatic mela-noma: analysis of 270 patients treated be-tween 1985 and 1993. J Clin Oncol 1999; 17:2105-16.

Middleton MR, Grob JJ, Aaronson N, 3. et al. Randomized phase III study of te-mozolomide versus dacarbazine in the treatment of patients with advanced met-astatic malignant melanoma. J Clin Oncol 2000;18:158-66. [Erratum, J Clin Oncol 2000;18:2351.]

Bedikian AY, Millward M, Pehamberg-4. er H, et al. Bcl-2 antisense (oblimersen sodium) plus dacarbazine in patients with advanced melanoma: the Oblimersen Mel-anoma Study Group. J Clin Oncol 2006; 24:4738-45.

Davies H, Bignell GR, Cox C, et al. 5. Mutations of the BRAF gene in human cancer. Nature 2002;417:949-54.

Nikiforova MN, Kimura ET, Gandhi 6. M, et al. BRAF mutations in thyroid tu-mors are restricted to papillary carcino-mas and anaplastic or poorly differenti-ated carcinomas arising from papillary carcinomas. J Clin Endocrinol Metab 2003;88:5399-404.

Fukushima T, Suzuki S, Mashiko M, 7. et al. BRAF mutations in papillary carci-nomas of the thyroid. Oncogene 2003;22: 6455-7.

Cohen Y, Xing M, Mambo E, et al. 8. BRAF mutation in papillary thyroid carci-noma. J Natl Cancer Inst 2003;95:625-7.

Yuen ST, Davies H, Chan TL, et al. 9. Similarity of the phenotypic patterns as-sociated with BRAF and KRAS mutations in colorectal neoplasia. Cancer Res 2002; 62:6451-5.

Oliveira C, Pinto M, Duval A, et al. 10. BRAF mutations characterize colon but not gastric cancer with mismatch repair defi-ciency. Oncogene 2003;22:9192-6.

Wang L, Cunningham JM, Winters JL, 11. et al. BRAF mutations in colon cancer are not likely attributable to defective DNA mismatch repair. Cancer Res 2003;63: 5209-12.

Tannapfel A, Sommerer F, Benicke M, 12. et al. Mutations of the BRAF gene in chol-angiocarcinoma but not in hepatocellular carcinoma. Gut 2003;52:706-12.

Cho NY, Choi M, Kim BH, Cho YM, 13. Moon KC, Kang GH. BRAF and KRAS mu-tations in prostatic adenocarcinoma. Int J Cancer 2006;119:1858-62.

Singer G, Oldt R III, Cohen Y, et al. 14. Mutations in BRAF and KRAS character-ize the development of low-grade ovarian serous carcinoma. J Natl Cancer Inst 2003;95:484-6.





Honecker F, Wermann H, Mayer F, et al. 15. Microsatellite instability, mismatch repair deficiency, and BRAF mutation in treat-ment-resistant germ cell tumors. J Clin On-col 2009;27:2129-36.

Kimura ET, Nikiforova MN, Zhu Z, 16. Knauf JA, Nikiforov YE, Fagin JA. High prevalence of BRAF mutations in thyroid cancer: genetic evidence for constitutive activation of the RET/PTC-RAS-BRAF sig-naling pathway in papillary thyroid carci-noma. Cancer Res 2003;63:1454-7.

Tsai J, Lee JT, Wang W, et al. Discov-17. ery of a selective inhibitor of oncogenic B-Raf kinase with potent antimelanoma activity. Proc Natl Acad Sci U S A 2008; 105:3041-6.

Sondergaard JN, Nazarian R, Wang 18. Q, et al. Differential sensitivity of mela-noma cell lines with BRAF V600E muta-tion to the specific Raf inhibitor PLX4032. J Transl Med 2010;8:39.

Oken MM, Creech RH, Tormey DC, et 19. al. Toxicity and response criteria of the Eastern Cooperative Oncology Group. Am J Clin Oncol 1982;5:649-55.

Koch WH. Technology platforms for 20.

pharmacogenomic diagnostic assays. Nat Rev Drug Discov 2004;3:749-61.

Heidorn SJ, Milagre C, Whittaker S, et 21. al. Kinase-dead BRAF and oncogenic RAS cooperate to drive tumor progression through CRAF. Cell 2010;140:209-21.

Poulikakos PI, Zhang C, Bollag G, 22. Shokat KM, Rosen N. RAF inhibitors transactivate RAF dimers and ERK sig-nalling in cells with wild-type BRAF. Na-ture 2010;464:427-30.

Hatzivassiliou G, Song K, Yen I, et al. 23. RAF inhibitors prime wild-type RAF to activate the MAPK pathway and enhance growth. Nature 2010;464:431-5.

Wilhelm SM, Carter C, Tang L, et al. 24. BAY 43-9006 exhibits broad spectrum oral antitumor activity and targets the RAF/MEK/ERK pathway and receptor ty-rosine kinases involved in tumor progres-sion and angiogenesis. Cancer Res 2004; 64:7099-109.

Eisen T, Ahmad T, Flaherty KT, et al. 25. Sorafenib in advanced melanoma: a phase II randomised discontinuation trial anal-ysis. Br J Cancer 2006;95:581-6.

McDermott DF, Sosman JA, Gonza-26.

lez R, et al. Double-blind randomized phase II study of the combination of sorafenib and dacarbazine in patients with advanced melanoma: a report from the 11715 Study Group. J Clin Oncol 2008;26: 2178-85.

Hauschild A, Agarwala SS, Trefzer U, 27. et al. Results of a phase III, randomized, placebo-controlled study of sorafenib in combination with carboplatin and pacli-taxel as second-line treatment in patients with unresectable stage III or stage IV melanoma. J Clin Oncol 2009;27:2823-30.

Curtin JA, Fridlyand J, Kageshita T, et 28. al. Distinct sets of genetic alterations in melanoma. N Engl J Med 2005;353:2135-47.

Carvajal RD, Chapman PB, Wolchok 29. JD, et al. A phase II study of imatinib mes-ylate (IM) for patients with advanced melanoma harboring somatic alterations of KIT. J Clin Oncol 2009;27:Suppl:15S. abstract.Copyright © 2010 Massachusetts Medical Society.

posting presentations at medical meetings on the internet

Posting an audio recording of an oral presentation at a medical meeting on the Internet, with selected slides from the presentation, will not be considered prior publication. This will allow students and physicians who are unable to attend the meeting to hear the presentation and view the slides. If there are any questions about this policy, authors should feel free to call the Journal’s Editorial Offices.



LETTERS

Clinical efficacy of a RAF inhibitor needs broad targetblockade in BRAF-mutant melanomaGideon Bollag1, Peter Hirth1, James Tsai1, Jiazhong Zhang1, Prabha N. Ibrahim1, Hanna Cho1, Wayne Spevak1,Chao Zhang1, Ying Zhang1, Gaston Habets1, Elizabeth A. Burton1, Bernice Wong1, Garson Tsang1, Brian L. West1,Ben Powell1, Rafe Shellooe1, Adhirai Marimuthu1, Hoa Nguyen1, Kam Y. J. Zhang1, Dean R. Artis1,Joseph Schlessinger2, Fei Su3, Brian Higgins3, Raman Iyer3, Kurt D’Andrea4, Astrid Koehler3, Michael Stumm3,Paul S. Lin1, Richard J. Lee3, Joseph Grippo3, Igor Puzanov5, Kevin B. Kim6, Antoni Ribas7, Grant A. McArthur8,Jeffrey A. Sosman5, Paul B. Chapman9, Keith T. Flaherty4, Xiaowei Xu4, Katherine L. Nathanson4 & Keith Nolop1

B-RAF is the most frequently mutated protein kinase in humancancers1. The finding that oncogenic mutations in BRAF are com-mon in melanoma2, followed by the demonstration that thesetumours are dependent on the RAF/MEK/ERK pathway3, offeredhope that inhibition of B-RAF kinase activity could benefit mela-noma patients. Herein, we describe the structure-guided discovery ofPLX4032 (RG7204), a potent inhibitor of oncogenic B-RAF kinaseactivity. Preclinical experiments demonstrated that PLX4032 selec-tively blocked the RAF/MEK/ERK pathway in BRAF mutant cells andcaused regression of BRAF mutant xenografts4. Toxicology studiesconfirmed a wide safety margin consistent with the high degree ofselectivity, enabling Phase 1 clinical trials using a crystalline for-mulation of PLX4032 (ref. 5). In a subset of melanoma patients,pathway inhibition was monitored in paired biopsy specimenscollected before treatment initiation and following two weeks oftreatment. This analysis revealed substantial inhibition of ERK phos-phorylation, yet clinical evaluation did not show tumour regressions.At higher drug exposures afforded by a new amorphous drug for-mulation4,5, greater than 80% inhibition of ERK phosphorylation inthe tumours of patients correlated with clinical response. Indeed, thePhase 1 clinical data revealed a remarkably high 81% response rate inmetastatic melanoma patients treated at an oral dose of 960 mg twicedaily5. These data demonstrate that BRAF-mutant melanomas arehighly dependent on B-RAF kinase activity.

PLX4032 belongs to a family of mutant B-RAF kinase inhibitorsdiscovered using a scaffold-based drug design approach6. Thecrystallography-guided approach allowed optimization of a com-pound with modest preference for the mutated form of B-RAF (B-RAF(V600E)) in comparison to wild-type B-RAF. SupplementaryTable 1 summarizes the differential ability for PLX4032 to inhibitthe activity of over 200 kinases. PLX4032 displays similar potency forB-RAF(V600E) (31 nM) and c-RAF-1 (48 nM) and selectivity againstmany other kinases, including wild-type B-RAF (100 nM). Whereasthe vast majority of kinases are minimally affected, several werefound that were also inhibited at ,100 nM concentrations in bio-chemical assays; to date, inhibition of these non-RAF kinases such asACK1 (also known as TNK2), KHS1 (also known as MAP4K5) andSRMS has not been tested in cellular assays. As previously demon-strated for the related B-RAF inhibitor PLX4720 (ref. 6), the bio-chemical selectivity of PLX4032 translates to cellular selectivity:

potent inhibition of ERK phosphorylation and proliferation occursexclusively in BRAF-mutant cell lines4.

PLX4032 was co-crystallized with a protein construct that con-tained the kinase domain of B-RAF(V600E). PLX4032 (Fig. 1a) bindsin the active site of one of the protomers in the non-crystallographic-symmetry related dimer (Fig. 1). As previously described for therelated RAF inhibitor PLX4720 (PDB ID: 3C4C)6, the PLX4032-bound protomer adopts the DFG-in conformation to enable theformation of a unique hydrogen bond between the backbone NHof Asp 594 and the sulfonamide nitrogen of PLX4032 (Fig. 1b). Inaddition, PLX4032-binding causes an outward shift in the regulatoryaC helix, which may explain why the effect of PLX4720 and PLX4032on RAF dimerization is in stark contrast to other RAF inhibitors suchas AZD-628 and GDC-0879 (Fig. 1c)7. The apo-protomer displaysthe DFG-in conformation with the activation loop locked away fromthe ATP-binding site by a salt-bridge between Glu 600 and Lys 507(Fig. 1d).

In BRAF(V600E)-mutant xenograft studies, PLX4032 demonstrateddose-dependent inhibition of tumour growth, with higher exposuresresulting in tumour regression (Fig. 2a and ref. 4). Efficacy could bedemonstrated in cell lines and xenografts bearing either homozygousor heterozygous BRAF mutations. By contrast, no effect was observedon melanoma xenograft growth if both BRAF alleles were wild-type4,6.Due to their consistent pharmacokinetics in rodents, PLX4032 andPLX4720 were prioritized over a panel of related compounds that allhad similar activities in vitro and in vivo. For further drug development,PLX4032 was chosen (over PLX4720) because its pharmacokineticproperties scaled more favourably in beagle dogs and cynomolgusmonkeys.

In order to estimate PLX4032 exposures (as defined by AUC0–24,the area under the plasma concentration time curve over the dosingperiod of 24 h) that correlated with tumour response, conventionallyformulated daily oral doses of PLX4032 were administered in theBRAF(V600E)-bearing colorectal cancer COLO205 xenograft model.In this model, tumour growth inhibition was modest at 6 mg kg21

(AUC0–24 , 50 mM h), tumour stabilization was seen at 20 mg kg21

(AUC0–24 , 200 mM h), and significant tumour regressions wereobserved at 20 mg kg21 twice daily (BID) (AUC0–24 , 300 mM h).BRAF(V600E)-bearing melanoma xenograft models, includingNCI-LOX and COLO829 are also sensitive to PLX4032 (ref. 4).

1Plexxikon Inc., 91 Bolivar Drive, Berkeley, California 94710, USA. 2Yale University, 333 Cedar Street, New Haven, Connecticut 06520, USA. 3Roche Pharmaceuticals, 340 KingslandStreet, Nutley, New Jersey 07110, USA. 4Departments of Medicine and Pathology and Laboratory Medicine, Abramson Cancer Center, University of Pennsylvania, 421 Curie Boulevard,Philadelphia, Pennsylvania 19104, USA. 5Vanderbilt University, PRB 777 Nashville, Tennessee 37232, USA. 6The University of Texas M. D. Anderson Cancer Center, 1515 HolcombeBoulevard, Houston, Texas 77030, USA. 7University of California, Los Angeles, 100 UCLA Medical Plaza, Los Angeles, California 90095, USA. 8Peter MacCallum Cancer Centre,Beckett Street, Melbourne 8006, Australia. 9Memorial Sloan Kettering Cancer Center, 1275 York Avenue, New York, New York 10065, USA.

doi:10.1038/nature09454

1Macmillan Publishers Limited. All rights reserved©2010

NOT FINAL PROOF

www.nature.com/doifinder/10.1038/nature09454

www.nature.com/nature


Rats and beagle dogs were dosed for 28 days with increasing dosesup to 1,000 mg kg21 day21, and no toxicity was detected at any doselevel. Likewise, no adverse effects were detected in a standard batteryof safety pharmacology studies. Subsequent toxicology studies oflonger duration, 26 weeks in rats and 13 weeks in dogs, further con-firmed the tolerability of the compound. This safety profile wasachieved in spite of very high compound exposures, reaching2,600 mM h in rats and 820mM h in dogs. The rat exposures exceededthose that were effective in patients (see below). Importantly, nohistological changes were observed in the skin in any animal at anydose or duration of treatment, contrasting to results observed withother RAF inhibitors7.

PLX4032’s performance in these preclinical toxicology studies pro-vided the necessary safety data to support initiation of Phase 1 clinicaltesting in cancer patients. Clinical and pharmacokinetic results fromthis Phase 1 study have been reported recently5. In the initial stage ofthe Phase 1 study, cohorts of patients with advanced solid tumourswere treated with escalating doses of PLX4032 (ranging from 200 to

1,600 mg), administered twice-daily (BID) as oral capsules. The initialcrystalline formulation yielded modest drug exposures, so PLX4032was reformulated as a micro-precipitated bulk powder (MBP),and doses ranging from 160 to 1,120 mg BID were sequentiallyevaluated. Preclinical experiments in mice (Fig. 2b and ref. 4) anddogs demonstrated that the MBP formulation substantially increaseddrug bioavailability, an approximately tenfold improvement. Thisimproved bioavailability was also observed in patients, with meanexposures at a 160 mg BID dose of the MBP formulation (day 15AUC0–24 5 185mM h) similar to a 1,600 mg BID dose of the originalformulation (day 15 AUC0–24 5 203mM h).

During the dose-escalation stage of the Phase 1 trial, 21 patients withmetastatic melanoma, 16 with and 5 without BRAF-mutations, weretreated at doses that achieved AUC0–24 . 300mM h (ref. 5). Tumourdimensions were measured by computed tomography (CT). Tenpatients with BRAF-mutant melanoma achieved tumour regressionsqualifying as partial responses (PR, by response evaluation criteria insolid tumours (RECIST 1.0), .30% reduction in tumour dimensions)and one patient had a complete response (CR); none of the patientswith melanomas lacking BRAF mutations achieved PR. These dataalong with preclinical evidence of selectivity for BRAF-mutant celllines strongly justified limiting all further enrolment to patients withBRAF-mutant tumours. Dose-limiting toxicities detected at the1,120 mg BID dose included fatigue, rash and joint pain5. Therefore,960-mg BID was identified as the maximum tolerated dose (MTD),and subsequently a cohort of 32 patients with BRAF-mutant mela-noma was enrolled at the MTD in an extension of the Phase 1 study.

At the 960-mg BID dose, the steady state PLX4032 concentrationwas 86mM and the AUC0–24 was 1,741mM h; the half-life was estimatedto be 50 h (ref. 5). Out of the 32 patients treated at this dose, 24experienced tumour regressions qualifying as PRs and two patientshad CRs. BRAF(V600E) mutation status was assessed by a real-timepolymerase chain reaction (PCR) assay as described in Methods5,8, andmany of the samples were sequenced for verification of the PCR result.

PLX4032

a

N N

O

F

FN S

O

O

Cl

D594

E501F595

W531

c

PLX4032-bound protomer Apo protomer

F595

αc-helix

shift

F595F595R509R509

E600K507

E600

K507

d

b

Figure 1 | Three-dimensional structure of PLX4032 binding toB-RAF(V600E). a, Chemical structure of PLX4032. b, Structure highlightsthe interactions of azaindole with the kinase hinge and the sulfonamide withthe DFG loop, with F595 rendered in balls and other key protein residuesshown as sticks. c, Structure of the asymmetric dimer of B-RAF(V600E) isshown with the PLX4032-protomer bound to PLX4032 coloured yellow(consistent with b). The surface outline of the other protomer (blue) isshown lightly shaded. Highlighted residues are R509 to reflect its role inanchoring the dimer and F595 to show that both protomers are in the DFG-in state. The ac-helix shown in magenta is overlaid on the PLX4032-boundprotomer to show its typical configuration in an unoccupied protomer; thebinding of PLX4032 causes a shift of the ac-helix as noted by the arrow.d, Magnified view of the salt bridge between Lys 507 and Glu 600 that helpsprevent compound binding to the apo protomer.

a

0 10 20 300

500

1,000

1,500

Days on treatment

Tu

mo

ur

vo

lum

e (m

m3) Vehicle

6 mg kg–1 QD,

20 mg kg–1 QD, 200 μM h

20 mg kg–1 BID, 300 μM h

Day 7 AUC

Vehicle

25 mg kg–1 BID

75 mg kg–1 BID

b

50 μM h

20 300

500

1,000

Days on treatment

Tu

mo

ur

vo

lum

e (m

m3)

25 35

Figure 2 | Effect of PLX4032 on COLO205 xenograft tumour growth.Tumour volume measurements of mice treated by oral gavage with theindicated doses of PLX4032 or vehicle (n 5 10 for all groups, error barsindicate standard error) are shown. a, Administration in conventionalformulation occurred daily. Exposures measured on day 7 are shown. At the6, 20 and 20 BID doses, 1/10, 1/10 and 8/10 animals achieved CR, respectively.b, Administration in the MBP formulation occurred twice daily. At the25 mg kg21 BID dose (blue), 7/10 animals achieved CR and 3/10 animalsachieved PR; at the 75 mg kg21 BID dose (red), all animals achieved CR.

LETTERS NATURE


NOT FINAL PROOF

The reliability of the PCR assay is currently being assessed in concur-rent Phase 2 and Phase 3 trials. The BRAF(V600E) allele was detectedby the PCR assay in 46 of the 48 BRAF-positive patients describedabove. Interestingly, subsequent sequencing revealed that tumoursfrom the two patients lacking the BRAF(V600E) mutation were foundto carry the BRAF(V600K) mutation and were among the betterresponders (71% and 100% reduction in tumour dimensions); anadditional BRAF(V600K) response has been recently published9.

During the dose escalation stage of the study, a cohort of patientshad paired tumour biopsies to evaluate pathway inhibition, the firsttaken before initiation of PLX4032 treatment and the second takenafter 14 days of PLX4032 treatment. This paired-biopsy cohort ofpatients captured a wide inter-patient range of steady state PLX4032exposures. In addition to expected inter-patient variability in drugclearance, these patients were treated at different doses and with thetwo different formulations (one crystalline and one amorphous). Tomonitor ERK pathway activity, phosphorylated-ERK (pERK) levelswere determined by immunohistochemistry (IHC), both in the nucleusand in the cytoplasm. To monitor proliferation, Ki67 levels also weremeasured.

As shown in Supplementary Table 2, levels of pERK and Ki67 weredecreased in most biopsies following 2 weeks of dosing with PLX4032,even in patients with modest drug exposure. Patients exposed toplasma levels of PLX4032 less than 300mM h experienced no mea-surable decreases in tumour burden. In contrast, patients exposedto higher plasma levels of PLX4032 experienced tumour regression,often achieving PRs as defined by RECIST criteria (SupplementaryTable 2). Representative pictures illustrating decreases in ERK phos-phorylation and Ki67 are shown in Fig. 3a and b. Interestingly,decreases in cytoplasmic pERK correlated well with tumour response(Fig. 3c), whereas changes in nuclear pERK correlated poorly (Fig. 3d).In general, nuclear pERK was more sensitive to compound levels thancytoplasmic pERK, consistent with the idea that nuclear ERKresponds very quickly to phosphorylation/dephosphorylation events,whereas cytoplasmic ERK phospho-events are buffered by the manycytoplasmic scaffolding proteins. As further evidenced in Supplemen-tary Table 2, the improved pathway inhibition and tumour responsescorrelate with higher plasma drug exposures. In patients with tumourregressions, pathway analysis typically showed greater than 80%inhibition of cytoplasmic ERK phosphorylation (Fig. 3c). This resultindicates that near-complete inhibition of ERK signalling may beneeded for significant tumour response.

A growing body of literature shows that oncogenic BRAF is animportant stimulator of metabolic activity10,11, and in preclinical studiesPLX4032 rapidly inhibits fluoro-deoxy-glucose (FDG) uptake speci-fically in BRAF(V600E) mutant melanoma cell lines12. Therefore,FDG uptake in patients on the PLX4032 trial was assessed usingpositron emission tomography (PET) imaging before treatment andfollowing 2 weeks of dosing. All of the assessable patients treated withMBP-formulated PLX4032 experienced major reductions in FDGuptake. Representative FDG-PET images are shown in Fig. 4.

Toxicities such as fatigue, rash and joint pain in the treated patientsare detailed separately5. Thirty-one percent of the patients treated at theMTD developed skin lesions described as cutaneous squamous cellcarcinomas, keratoacanthoma type5. This apparent drug-induced effectis of particular interest, because investigators studying three other RAFinhibitors, sorafenib13–15, XL281 (ref. 16) and GSK2118436 (ref. 17) alsohave noted these skin lesions in a subset of treated patients. Althoughthe occurrence of these treatment-emergent tumours warrants carefuldermatological monitoring of patients during PLX4032 treatment, itshould be noted that these lesions were resected, and no patients dis-continued PLX4032 due to this drug-induced effect5. These skin lesionsgenerally appear within a few months of treatment initiation in sun-exposed areas of skin, suggesting that pre-existing oncogenic mutationsmay potentiate the RAF inhibitor effects.

Recent publications suggest a potential mechanism that may inpart account for the keratinocyte proliferation noted in some patients

on study7,18,19. These reports follow up on prior descriptions of para-doxical activation of the RAF/MEK/ERK pathway by RAF kinaseinhibitors20–22. Current evidence suggests that wild-type RAF kinaseactivity can be activated by RAF dimerization23. RAF dimerizationcan be induced by RAF inhibitors: binding to one protomer—whileinhibiting the kinase activity of that protomer—concurrentlyinduces a conformational switch in the partner protomer via anundefined allosteric mechanism to activate RAF kinase activity7,19.This paradoxical activation occurs in cells in which RAS is activatedeither by mutation or by some other priming event. Modulation ofRAF dimerization may not be the only unexpected effect of RAFinhibitors, because multiple additional factors are involved in bothpositive (for example, KSR, SRC, CNK (also known as CNKSR)) andnegative (for example, ERK), 14-3-3 (also known as YWHAQ),DUSP, RKIP (also known as PEBP1), RASSF) regulation of theRAF/MEK/ERK signalling pathway24,25.

The ability of PLX4032 to cause tumour regression in a large pro-portion of patients with BRAF-mutant, advanced-stage, metastatic

pERK level change

(percentage decrease)

Tum

our

resp

onse

(perc

enta

ge d

ecre

ase)

pERK level change

(percentage decrease)

R2 = 0.79

–60

–40

–20

0

20

40

60

80

100

–20 0 20 40 60 80 100

R2 = 0.023

–60

–40

–20

0

20

40

60

80

100

20 40 60 80 100

pERK (cytoplasm) pERK (nucleus)

c d

pERK

Ki67

H&E

Baseline Day 15

pERK

Ki67

Baseline Day 15a

b

Figure 3 | Semi-quantitative immunohistochemistry (IHC) in pairedtumour biopsies. Matched baseline and day 15 tumour samples are at thesame magnification; the scale bar is 70 mm. a, Representative IHC for Ki67and pERK staining is shown for patient 12. B. Representative IHC for Ki67,pERK and H&E (haematoxylin and eosin) staining is shown for patient 42.The arrow indicates tumour breakdown with macrophages engulfing thereleased melanin in the day 15 sample. c, Summary graph showingcorrelation of reduction in cytoplasmic pERK with tumour responses (datafrom Supplementary Table 2). d, Summary graph indicating weakcorrelation of reduction in nuclear pERK with tumour responses.

NATURE LETTERS


NOT FINAL PROOF

melanoma provides strong support for the hypothesis that the onco-genic B-RAF protein is a dominant driver of tumour growth andmaintenance. These results are particularly interesting in that theBRAF mutation is likely an initiating event in melanoma tumorigenesis:the vast majority of benign nevi harbour the same BRAF(V600E) muta-tion26. Our current understanding of melanocyte biology suggeststhat the nevi are benign because the BRAF mutation alone inducessenescence27. Clinical evaluation of sporadic nevi in patients treated attherapeutic doses revealed no effect of PLX4032 on nevi progression orregression. Additional descriptions of PLX4032 kinase selectivity,PLX4032 structure, and clinical pharmacokinetics and efficacy areincluded in Supplementary Information.

The durability of response to PLX4032 is still under evaluation.Median progression-free survival (PFS) in the Phase 1 extensioncohort has not been reached but is currently estimated to be at least7 months5. Although this compares rather favourably with a PFS ofless than 2 months in historical analysis of large numbers of advancedmelanoma patients28, tumour re-growth occurs in many of thepatients and the mechanisms of resistance are currently under investi-gation. Therefore, improved durability of response will be an import-ant goal of further clinical trials. One potential strategy to meet thisgoal is to combine PLX4032 with other targeted agents, immuno-therapy or chemotherapy. With a promising safety and efficacyprofile, PLX4032 has the potential to anchor such combination treat-ments, and may thereby offer further improved treatment options forBRAF-mutant melanoma patients.

METHODS SUMMARY

PLX4032 was synthesized using the general procedures described previously6.

Expression and purification of B-RAF, structure determination, protein kinase

activity measurements and xenograft studies were carried out as described previ-

ously6. Clinical methods have also been described recently5. Melanoma patients

were selected for study using previously described TaqMan methodology8. Semi-

quantitative immunohistochemistry for pERK and Ki67 was performed on

5-mm-thick formalin-fixed paraffin-embedded tumour biopsies following hae-

matoxylin and eosin staining to determine pathologic diagnosis and tissue mor-

phology and integrity. The degree of phospho-ERK staining in the nucleus and

cytoplasm was interpreted semi-quantitatively by assessing the intensity and

extent of staining on the slides. For Ki67 staining, the percentage of positive cells

was determined.

Received 19 May; accepted 31 August 2010.Published online 7 September 2010.

1. Greenman, C. et al. Patterns of somatic mutation in human cancer genomes.Nature 446, 153–158 (2007).

2. Davies, H. et al. Mutations of the BRAF gene in human cancer. Nature 417,949–954 (2002).

3. Solit, D. B. et al. BRAF mutation predicts sensitivity to MEK inhibition. Nature 439,358–362 (2006).

4. Yang, H. et al. RG7204 (PLX4032), a selective BRAFV600E inhibitor, displays potentantitumor activity in preclinical melanoma models. Cancer Res. 70,5518–5527 (2010).

5. Flaherty, K. et al. Inhibition of mutated, activated BRAF in metastatic melanoma.N. Engl. J. Med. 363, 809–819 (2010).

6. Tsai, J. et al. Discovery of a selective inhibitor of oncogenic B-Raf kinase withpotent antimelanoma activity. Proc. Natl Acad. Sci. USA 105, 3041–3046 (2008).

7. Hatzivassiliou, G. et al. RAF inhibitors prime wild-type RAF to activate the MAPKpathway and enhance growth. Nature 464, 431–435 (2010).

8. Koch, W. H. Technology platforms for pharmacogenomic diagnostic assays.Nature Rev. Drug Discov. 3, 749–761 (2004).

9. Rubinstein, J. C. et al. Incidence of the V600K mutation among melanomapatients with BRAF mutations, and potential therapeutic response to the specificBRAF inhibitor PLX4032. J. Transl. Med. 8, 67 (2010).

10. Esteve-Puig, R., Canals, F., Colome, N., Merlino, G. & Recio, J. A. Uncoupling of theLKB1-AMPKa energy sensor pathway by growth factors and oncogenicBRAFV600E. PLoS ONE 4, e4771 (2009).

11. Zheng, B. et al. Oncogenic B-RAF negatively regulates the tumor suppressor LKB1to promote melanoma cell proliferation. Mol. Cell 33, 237–247 (2009).

12. Søndergaard, J. N. et al. Differential sensitivity of melanoma cell lines withBRAFV600E mutation to the specific B-Raf inhibitor PLX4032. J. Transl. Med. 8, 39(2010).

13. Arnault, J. P. et al. Keratoacanthomas and squamous cell carcinomas in patientsreceiving sorafenib. J. Clin. Oncol. 27, e59–e61 (2009).

14. Dubauskas, Z. et al. Cutaneous squamous cell carcinoma and inflammation ofactinic keratoses associated with sorafenib. Clin. Genitourin. Cancer 7, 20–23(2009).

15. Kong, H. H. et al. Keratoacanthomas associated with sorafenib therapy. J. Am.Acad. Dermatol. 56, 171–172 (2007).

16. Schwartz, G. K. et al. A phase I study of XL281, a selective oral RAF kinase inhibitor,in patients (Pts) with advanced solid tumors. J. Clin. Oncol. 27, 3513 (suppl.),(2009).

17. Kefford, R. et al. Phase I/II study of GSK2118436, a selective inhibitor of oncogenicmutant BRAF kinase, in patients with metastatic melanoma and other solidtumors. J. Clin. Oncol. 28, 8503 (suppl.), (2010).

18. Heidorn, S. J. et al. Kinase-dead BRAF and oncogenic RAS cooperate to drivetumor progression through CRAF. Cell 140, 209–221 (2010).

19. Poulikakos, P. I., Zhang, C., Bollag, G., Shokat, K. M. & Rosen, N. RAF inhibitorstransactivate RAF dimers and ERK signalling in cells with wild-type BRAF. Nature464, 427–430 (2010).

20. Courtois-Cox, S. et al. A negative feedback signaling network underlies oncogene-induced senescence. Cancer Cell 10, 459–472 (2006).

21. Dougherty, M. K. et al. Regulation of Raf-1 by direct feedback phosphorylation.Mol. Cell 17, 215–224 (2005).

22. Hall-Jackson, C. A. et al. Paradoxical activation of Raf by a novel Raf inhibitor.Chem. Biol. 6, 559–568 (1999).

23. Rajakulendran, T., Sahmi, M., Lefrancois, M., Sicheri, F. & Therrien, M. Adimerization-dependent mechanism drives RAF catalytic activation. Nature 461,542–545 (2009).

24. Pratilas, C. A. et al. V600EBRAF is associated with disabled feedback inhibition ofRAF-MEK signaling and elevated transcriptional output of the pathway. Proc. NatlAcad. Sci. USA 106, 4519–4524 (2009).

25. Kolch, W. Coordinating ERK/MAPK signalling through scaffolds and inhibitors.Nature Rev. Mol. Cell Biol. 6, 827–837 (2005).

26. Pollock, P. M. et al. High frequency of BRAF mutations in nevi. Nature Genet. 33,19–20 (2003).

27. Michaloglou, C. et al. BRAFE600-associated senescence-like cell cycle arrest ofhuman naevi. Nature 436, 720–724 (2005).

28. Korn, E. L. et al. Meta-analysis of phase II cooperative group trials in metastaticstage IV melanoma to determine progression-free and overall survivalbenchmarks for future phase II trials. J. Clin. Oncol. 26, 527–534 (2008).

Supplementary Information is linked to the online version of the paper atwww.nature.com/nature.

Acknowledgements We thank L. Andries and M. Knaapen from HistoGeneX forevaluating paired biopsies, and also our colleagues at the Molecular ImagingResearch division of Charles River Labs for conducting the xenograft studies. Wealso thank D. Heimbrook, S. Cheng, L. Burdette and B. Lestini for helpful commentson the manuscript. This research was funded in part by NIH grants to K.L.N.

Author Contributions G.B., P.H., C.Z., K.L.N. and K.N. designed studies, interpreteddata and wrote the manuscript. J.T., G.H., E.A.B., B.W., G.T., B.L.W., B.P., R.S., A.M.,

Patient 45 Patient 61

Patient 69Patient 59

Figure 4 | Representative PET scans for patients taken pre-dose andfollowing 2 weeks of dosing with PLX4032. Each of these image pairsdemonstrates significant reduction in FDG uptake following PLX4032treatment. Note that tumour regressions were later documented for each ofthese patients: best responses were 70% for patient 45, 70% for patient 59,68% for patient 61 and 37% for patient 69.

LETTERS NATURE


NOT FINAL PROOF


H.N., F.S., and B.H. conducted or managed biochemical or biological studies. J.Z.,P.N.I., H.C., W.S., D.R.A. and R.I. designed and conducted chemistry andformulation experiments. Y.Z. and K.Y.J.Z. conducted and interpreted structuralstudies. J.S. helped interpret data and write the manuscript. K.D., A.K., M.S. andX.X. designed, managed and interpreted biomarker studies. P.S.L., R.J.L., J.G., I.P.,K.B.K., A.R., G.A.M., J.A.S., P.B.C. and K.T.F. managed or conducted clinical andtranslational studies.

Author Information Atomic and structural data have been deposited in ProteinData Bank under accession number 3OG7. Reprints and permissions information isavailable at www.nature.com/reprints. Readers are welcome to comment on theonline version of this article at www.nature.com/nature. The authors declarecompeting financial interests: details accompany the full-text HTML version of thepaper at www.nature.com/nature. Correspondence and requests for materialsshould be addressed to G.B. ([email protected]).

NATURE LETTERS


NOT FINAL PROOF

www.nature.com/reprints



mailto:[email protected]

METHODS

Synthesis of PLX4032

PLX4032 was synthesized from commercially available 5-bromo-7-azaindole and 2,4-

difluoroaniline as described in WO2007002433. 5-Bromo-7-azaindole was reacted with

4(chlorophenyl)boronic acid under Suzuki coupling conditions to provide 5-(4-chloro-phenyl)-

1H-pyrrolo[2,3-b]pyridine (2). Synthesis of propane-1-sulfonic acid (2,4-difluoro-3-formyl-

phenyl)-amide (3) from 2,4-difluoroaniline was described elsewhere.6 Compounds 2 and 3 under

aldol reaction conditions provided a mixture of 4a and 4b which upon demethylation and

oxidation provided N-[3-[5-(4-chlorophenyl)-1H-pyrrolo[2,3-b]pyridine-3-carbonyl]-2,4-

difluoro-phenyl]propane-1-sulfonamide, 1 (PLX4032).

Scheme 1

Step 1

5-Br-7-Azaindole 2

F

FHN S

O

O

Step 2

N NH

RO

F NH

4a R = H

F

SO

O O

N NH

F NH

F

SO

O

3 4b R = CH3

N NH

2

+

Step 4

Cl Cl

Cl

N NH

O

F NH

F

SO

O

1 (PLX4032)

ClHO

Step 3

5

N NH

Cl

N NH

Br

Step 1 – Preparation of 5-(4-chloro-phenyl)-1H-pyrrolo[2,3-b]pyridine (2). To 5-bromo-7-

azaindole (70.6 g, 358 mmol) and 4(chlorophenyl)boronic acid (67.3 g, 430 mmol) in 700 mL of

1,2-dimethoxyethane was added a solution of potassium carbonate (59.4 g, 430 mol) in 350 mL

of water. The resulting mixture was purged with nitrogen for 30 min and then

bis(triphenylphosphine)dichloropalladium(II) (25.3 g, 35.8 mmol) was added. The reaction

SUPPLEMENTARY INFORMATIONdoi: 10.1038/nature09454

www.nature.com/nature 1

mixture was heated at reflux overnight and the volatiles were removed under reduced pressure.

The crude residue was taken up in 400 mL of brine and extracted with ethyl acetate (3 x 500

mL). The combined organic layers were dried over sodium sulfate, filtered, and the resulting

filtrate was evaporated in vacuo to give a crude solid that was purified on a silica gel column (1.5

kg) with 1:4 to 1:1 ethyl acetate:heptane as eluent to provide 66.6 g (81%) of 2 as a tan solid.

m.p.: 216-218 °C; 1H-NMR (400 MHz, DMSO-d6) δ 11.74 (s, 1H), 8.49 (s, 1H), 8.19 (s, 1H),

7.71 (d, J = 8.2 Hz, 2 H), 7.50 (s, 1H), 7.49 (d, J = 8.2 Hz, 2 H), 6.48 (m, 1H); 13C-NMR (100

MHz, DMSO-d6) δ 148.8, 142.1, 138.7, 132.4, 129.6, 129.3, 127.8, 127.5, 126.8, 120.3, 100.9;

MS(ESI)[M+H+] = 229.1.

Step 2 – Preparation of propane-1-sulfonic acid (3-{[5-(4-chloro-phenyl)-1H-pyrrolo[2,3-

b]pyridin-3-yl]-hydroxy-methyl}-2,4-difluoro-phenyl)-amide (4a) and Propane-1-sulfonic

acid (3-{[5-(4-chloro-phenyl)-1H-pyrrolo[2,3-b]pyridin-3-yl]-methoxy-methyl}-2,4-difluoro-

phenyl)-amide (4b). To a suspension of 5-(4-chloro-phenyl)-1H-pyrrolo[2,3-b]pyridine (2, 64.9

g, 284 mmol) and propane-1-sulfonic acid (2,4-difluoro-3-formyl-phenyl)-amide (3, 90.4 g, 343

mmol) in methanol (1.8 L) in a water bath was added potassium hydroxide (129 g, 2.3 mol). The

reaction was allowed to stir for 72 hours at room temperature and then adjusted to pH 7 with 4N

hydrochloric acid (580 mL). The resulting mixture was evaporated in vacuo to remove methanol

and extracted with ethyl acetate (3 x 800 mL). The combined organic layers were dried over

sodium sulfate and evaporated in vacuo to give a crude oil. The crude oil was triturated with 3:1

MTBE/heptane (500 mL) to give a 1:3 solid mixture of 4a and 4b that was used directly for the

next step. For compound 4b: 1H-NMR (400 MHz, DMSO-d6) δ 11.79 (s, 1H), 9.60 (s, 1H), 8.49

(s, 1H), 8.06 (s, 1H), 7.66 (d, J = 8.59 Hz, 2H), 7.51 (d, J = 8.21, 2H), 7.37 (m, 1H), 7.32 (s, 1H),

7.11 (m, 1H), 6.06 (s, 1H), 3.35 (s, 3H), 2.96 (m, 2H), 1.63 (m, 2H), 0.84 (m, 3H); 13C-NMR

(100 MHz, DMSO-d6) δ 158.0 (dd), 155.5 (dd), 148.8, 142.5, 138.5, 132.6, 129.7, 129.3, 128.3

(d), 127.6, 125.6, 125.3, 122.5 (d), 118.5, 118.0 (dd), 113.5, 112.6 (d), 71.0, 57.3, 54.2, 17.5,

13.2; 19F-NMR (376 MHz, DMSO-d6) δ -37.7, -43.0; MS(ESI)[M+H+] = 506.2.

Step 3 – Preparation of propane-1-sulfonic acid (3-{[5-(4-chloro-phenyl)-1H-pyrrolo[2,3-

b]pyridin-3-yl]-hydroxy-methyl}-2,4-difluoro-phenyl)-amide (5). To a solution of 4a and 4b

(~284 mmol) in acetic acid (750 mL) was added 48% (wt) aqueous solution of hydrobromic acid

doi: 10.1038/nature09454 SUPPLEMENTARY INFORMATION


(150 mL). The resulting mixture was allowed to stir overnight at room temperature and then

evaporated in vacuo. The crude residue was dissolved in ethyl acetate (1 L) and water (1 L), and

adjusted to pH 7 with solid potassium carbonate. The layers were separated and the aqueous

layer was extracted with ethyl acetate (2 x 1 L). The combined organic layers were dried over

sodium sulfate, filtered, and the resulting filtrate was evaporated in vacuo to give 5 as viscous oil

that was used directly for the next step. 1H-NMR (400 MHz, DMSO-d6) δ 11.66 (s, 1H), 9.55 (s,

1H), 8.47 (s,1H), 7.97 (s, 1H), 7.62 (d, J = 8.59 Hz, 2H), 7.50 (d, J = 8.59 Hz, 2 H), 7.32 (m,

1H), 7.28 (m, 1H), 7.06 (m, 1H), 6.38 (d, J = 5.08, 1H), 6.08 (d, J = 5.08, 1H), 2.95 (m, 2H), 1.63

(m, 2H), 0.83 (m, 3H); 13C-NMR (100 MHz, DMSO-d6) δ 158.0 (dd), 155.5 (dd), 149.0, 142.2,

138.6, 132.5, 129.7, 129.2, 127.8 (d), 127.3, 125.6, 124. 5, 122.3 (d), 121.5 (dd), 118.4, 116.5,

112.3 (d), 61.0, 54.2, 17.5, 13.2; 19F-NMR (376 MHz, DMSO-d6) δ -37.9, -43.5;

MS(ESI)[M+H+] = 492.0.

Step 4 – Preparation of Propane-1-sulfonic acid {3-[5-(4-chloro-phenyl)-1H-pyrrolo[2,3-

b]pyridine-3-carbonyl]-2,4-difluoro-phenyl}-amide, 1 (PLX4032) . To a solution of 5 in 1,4-

dioxane (700 mL) was added 2,3-dichloro-5,6-dicyanobenzoquinone (83.8 g, 369 mmol)

followed by water (35 mL). The resulting mixture was allowed to stir for 2 hours at room

temperature and then treated with saturated aqueous sodium bicarbonate (700 mL). The mixture

was evaporated in vacuo to remove 1,4-dioxane and extracted with ethyl acetate (3 x 1L). The

combined organic layers were dried over sodium sulfate, filtered, and the resulting filtrate was

evaporated in vacuo to give a crude solid that was purified on a silica-gel column (3 kg) with

95:5 dichloromethane:methanol as eluent to provide 78.7 g (45% for 3 steps) of 1 (PLX4032) as

a white solid. m.p.: 264 °C; 1H-NMR (400 MHz, DMSO-d6) δ 13.03 (s, 1H), 9.79 (s, 1H), 8.71

(s, 1H), 8.65 (br s, 1H), 8.25 (br s, 1H), 7.76 (d, J = 8.5 Hz, 2H), 7.60 (m, 1H), 7.52 (d, J = 8.4

Hz, 2H), 7.28 (m, 1H), 3.13 (m, 2H), 1.75 (m, 2 H), 0.95 (m, 3H); 13C-NMR (100 MHz, DMSO-

d6) δ 180.6, 156.0, 152.3, 149.0, 143.9, 138.9, 137.0, 132.5, 130.3, 129.0, 128.9, 128.7, 127.1,

121.9, 118.1, 117.5, 115.7, 112.3, 53.5, 16.8, 12.6; 19F-NMR (376 MHz, DMSO-d6) δ -38.7, -

44.1; IR (KBr) 3267, 3124, 1639, 1322, 1143 cm-1; MS(ESI)[M+H+] = 490.3; elemental analysis

(% calculated, % found for C23H18ClF2N3O3S): C (56.39, 56.18), H (3.70, 3.72), Cl (7.24, 7.52),

N (8.58, 8.60), O (9.80, 9.80), F (7.76, 8.02), S (6.54, 6.57).



Preclinical studies. Expression and purification of B-RAF, structure determination, and protein

kinase activity measurements were carried out as previously described.6 To obtain co-crystals of

B-RAFV600E with PLX4032, the protein solution was initially mixed with the compound

dissolved in DMSO at a final compound concentration of 1 mM. This complex was co-

crystallized by a sitting drop vapor diffusion experiment in which equal volumes of complex (at

10 mg/ml concentration) and reservoir solution (100mM BisTris at pH 6.0, 12.5% 2,5-

hexanediol, and 12% PEG3350) were mixed and allowed to equilibrate against the reservoir at

4°C. The crystal was soaked in cryosolvent, followed by flash-freezing in liquid nitrogen. The

data were collected at Beamline ALS831 (Lawrence Berkeley National Laboratory, Berkeley,

CA) with the wavelength of 1.11Å. The Ramachandran plot from the refined structure shows

that 94%, 5.6% and 0.4% residues are in the most favored, additional allowed and generously

allowed regions, respectively. A summary of the crystallography statistics is included in

Supplementary Table 3. COLO205 tumor xenograft studies (Molecular Imaging Research, Ann

Arbor, MI) were carried out as previously described either using a conventional formulation

(5%DMSO, 1%methylcellulose)6 or using the MBP formulation.4

Clinical methods have been recently described.5

Patient Population

All patients enrolled in the extension cohort of the Phase 1 study had provided melanoma tumor

tissue for centralized confirmation of BRAF mutation status. Specimens were analyzed with an

assay (investigational use only, Roche Molecular Systems, Inc., Pleasanton, CA) designed to

specifically detect the BRAFV600E (1799 T>A) mutation in DNA isolated from formalin-fixed,

paraffin-embedded tumor tissue using previously described TaqMan® methodology.8

Histology and Immunohistochemistry of Tumor Biopsies

Semi-quantitative immunohistochemistry was performed on 5 µm-thick formalin-fixed paraffin-

embedded tumor biopsies following H&E staining to determine pathologic diagnosis and tissue

morphology and integrity. Antibodies to phosphorylated ERK (Cell Signaling Technology) and

Ki67 (clone 30.9 rabbit mAb, Ventana or DakoCytomation, Carpinteria, CA) were used to

immunostain the slides. IHC staining was visualized using iView DAB Detection kit (Ventana),



the ultraView™ Universal DAB Detection Kit (Ventana) or the EnVision+ HRP DAB system

(DakoCytomation). Rabbit (DA1E) mAb IgG XP™ Isotype Control (Cell Signaling Technology,

diluted 1:30.000 in Dako REAL Antibody Diluent) and CONFIRM Negative Control Rabbit Ig

(Ventana) was substituted with the primary antibody as a negative control for pERK and Ki67

respectively. Alternatively, normal mouse serum (1:1000 dilution) was substituted for the

primary antibody in each case as a negative control.

The degree of phospho-ERK staining in the nucleus and cytoplasm was interpreted

semiquantitatively by assessing the intensity and extent of staining on the entire tissue sections

present on the slides. The percent area of positively staining tumor cells was multiplied by their

degree of staining (none [0], weakly [1], moderate [2], strong [3] staining cells). A staining

score (H-score) was then calculated (out of a maximum of 300) equaling the sum of 1 x

percentage of weak, 2 x percentage of moderate, and 3 x percentage of strong staining. The sum

of these groups equated with the overall percentage of positivity (H-score). The repeatability

and reproducibility of the phospho-ERK staining profile were assessed by comparing two runs

for five tumor samples and three positions were assessed for each sample. The concordance was

100%. Bland- Altman analysis of the H‐scores of the cytoplasmic and nuclear staining

compartments confirmed excellent reproducibility.29 For Ki67 staining, the percentage of

positive cells was determined. At the University of Pennsylvania, all slides from patients 12-48

were reviewed by a single dermatopathologist (X.X.). All slides from paired biopsies from

patients 56-93 were evaluated at HistoGeneX (Belgium) by L Andries (Ki67) and M Knaapen

(pERK). In order to make sure that there was consistency in readings across pathologists, five

paired (pre- and post-treatment) stained slides for Ki67 and phospho-ERK (nuclear, cytoplasm)

were provided in a blinded fashion from the University of Pennsylvania to HistoGeneX and vice

versa.



RESULTS

PLX4032 kinase selectivity

As mentioned in the text, when the kinase selectivity panel was expanded to over 200 members,

several additional kinases were found to be sensitive to PLX4032. Most of these kinases were

assayed at a lower ATP concentration (10 µM for the counter-screens versus 100 µM for the

RAF kinases); since PLX4032 is a competitive inhibitor assay at the lower ATP concentration

results in lower IC50 values. In a panel of over 150 chemical analogs of PLX4032, there was

good correlation between biochemical potency for B-RAFV600E and cellular activity against B-

RAF-mutant cells. This correlation did not depend on the relative potency against B-RAFV600E

and wild type B-RAF. Therefore, we believe that efficacy in melanoma primarily derives from

inhibition of mutant B-RAF; future studies may explore the role of off-targets in other

indications.

PLX4032 structure

When PLX4032 was co-crystallized with B-RAFV600E, two unique molecules of the kinase

domain in the asymmetric unit adopt a side- to-side dimer configuration as observed in previous

RAF crystal structures.6,23,30 Previously, PLX4720 was co-crystallized with wild type B-RAF,

and the protomer with only partial ligand occupancy (apo) adopts a DFG-out conformation

representing the inactive state of the kinase. However, the apo-protomer in the PLX4032 co-

structure with B-RAFV600E displays the DFG-in conformation with the activation loop locked

away from the ATP-binding site by a salt-bridge between Glu600 and Lys507 (Figure 1D).

Subsequent analysis of the structure of PLX4720 co-crystallized with B-RAFV600E revealed that

the apo-protomer displays the DFG-in conformation, suggesting that this property is determined

by the mutation. It is interesting to speculate that the conformation of the apo-protomer may

determine the paradoxical activation described in the main text. The conformational difference

captured by the crystal structures (Figure 1C) suggests that, although wild-type B-RAF is in a

dynamic equilibrium between the active (DFG-in) and inactive (DFG-out) configurations,

oncogenic BRAF mutations such as V600E induce constitutive kinase activity by shifting the

equilibrium toward the active (DFG-in) configuration. We believe that selective binding to the

DFG-in conformation may contribute to a wide safety margin because such inhibitors would



suppress the tumor growth but spare the important biological functions mediated by wild-type B-

RAF kinases.

Clinical pharmacokinetics and efficacy

At the 960 mg BID dose, the steady state PLX4032 concentration of 86 ± 32 µM (mean ± SD)

and the AUC 0-24 of 1741 ± 639 µM·h (mean ± SD are relatively high plasma drug

concentrations.5 For example, the efficacious AUC 0-24 values for the marketed kinase inhibitors

imatinib and erlotinib are 81.2 µM·h and 67.4 µM·h, and the steady state concentrations are 2.5

µM and 3.0 µM, respectively.31 Since the cellular potencies of PLX4032 are comparable to

those of imatinib and erlotinib, the requirement for a high PLX4032 exposure derives from the

very high plasma protein binding property: > 99.5% of the drug is bound to plasma protein, so

the free drug fraction is low. By contrast, the plasma protein binding values for imatinib and

erlotinib are ~95% and ~93%, respectively.31. The pharmacokinetic profile is also influence by

the very low aqueous solubility of PLX4032.

Some degree of tumor regression (including four patients whose regressions did not qualify as

PRs) could be measured in 30 of the 32 patients at the 960 mg BID dose, and 17 of the 32

patients had at least 50% reduction in tumor dimensions (30% being sufficient to qualify as PR

by RECIST criteria). Interestingly, the only two patients who did not have measurable tumor

regression at their first CT scan no longer required narcotics for pain shortly after beginning

PLX4032 treatment; one of these patients who had been hypoxemic at baseline no longer

required supplementary oxygen within the first week of dosing. These data suggest that all 32

patients in this cohort derived some benefit from B-RAF inhibition.

4. Ya ng, H., et al. RG7204 (PLX4032), a selective BRAFV600E inhibitor, displays potent

antitumor activity in preclinical melanoma models. Cancer Res 70, 5518-5527 (2010).

5. Flah erty, K., et al. Inhibition of Mutated, Activated BRAF in Metastatic Melanoma. N Engl

J Med 363, 809-819 (2010).

6. T sai, J., et al. Discovery of a selective inhibitor of oncogenic B-Raf kinase with potent

antimelanoma activity. Proc Natl Acad Sci U S A 105, 3041-3046 (2008).



8. Koch, W.H. Technology platforms for pharmacogenomic diagnostic assays. Nature reviews

3, 749-761 (2004).

23. Rajakulendran, T., Sahmi, M., Lefrancois, M., Sicheri, F. & Therrien, M. A dimerization-

dependent mechanism drives RAF catalytic activation. Nature 461, 542-545 (2009).

29. Bland, J.M. & Altman, D.G. Statistical methods for assessing agreement between two

methods of clinical measurement. Lancet 1, 307-310 (1986).

30. Wan , P.T., et al. Mechanism of activation of the RAF-ERK signaling pathway by

oncogenic mutations of B-RAF. Cell 116, 855-867 (2004).

31. van Erp, N.P., Gelderblom, H. & Guchelaar, H.J. Clinical pharmacokinetics of tyrosine

kinase inhibitors. Cancer Treat Rev 35, 692-706 (2009).



Supplementary Table 1. Biochemical IC50 determinations of the kinase inhibitory activity of

PLX4032 versus a panel of kinases

Assay IC50 nM*

B-RAF-V600E 31 C-RAF 48 B-RAF 100 SRMS 18 ACK1 19 MAP4K5 (KHS1) 51 FGR 63 LCK 183 BRK 213 NEK11 317 BLK 547 LYNB 599 YES1 604 WNK3 877 MNK2 171 7 FRK (PTK5) 1884 CSK 233 9 SRC 238 9

*A list of over 200 kinases minimally affected by PLX4032 is included below. Note that all RAF enzymes and SRMS were assayed at an ATP concentration of 100 µM, while all other kinases in the table above were assayed at an ATP concentration of 10 µM. Kinases with <20% Inhibition at 1 µM: ABL1, ABL2, ADRBK1, AMPK_A2, ARK5, Aurora_A-C, BMX, CDC42_BPA, CAMK2A, CDK5_p35, CSF1R, DYRK1B, EPHA5, EPHA8, EPHB4, FES, FLT3, FYN, GSK3beta, JAK1, KDR, KIT, MAP4K2, MAPK3, MARK2, MARK4, MATK, MET, MINK1, NEK1, NEK2, PAK3, PAK6, PDGFRbeta, PHKG1, PKBalpha, PKC_beta_I, PKC_beta_II, PKC_delta, PKC_gamma, PKC_zeta, SRC, STK4, STK24 Kinases with <10% Inhibition at 1 µM: ACVR1B_(ALK4), ADRBK2_(GRK3), ALK, AMPK_A1/B1/G1, ASK1, AXL, BRSK1_(SAD1), BrSK2, BTK, CAMK1, CAMK1D, CAMK2B, CAMK2D, CaMKIdelta, CaMKIIbeta, CaMKIIdelta, CaMKIIgamma,,CDC42_BPB, CDK1/CyclinB, CDK2/CyclinA, CDK2/cyclinE, CDK3/cyclinE, CDK5_p25, CDK6/cyclinD3, CDK7/CyclinH/MNAT1,CDK9/CyclinT1, CHEK1, CHEK2, CK1delta, CK1gamma1, CK1gamma2, CK1gamma3, CK2alpha2, CLK1, CLK2, CLK3, CSNK1A1, CSNK1D, CSNK1E, CSNK1G1, CSNK1G2, CSNK1G3, CSNK2A1, CSNK2A2, DAPK1, DAPK2, DAPK3_(ZIPK), DCAMKL2_(DCK2), DDR2, DMPK, DRAK1, DYRK1A, DYRK2, DYRK3, DYRK4, EEF2K, EGFR, EPHA1, EPHA2, EPHA3 EPHA4, EPHA7, EPHB1, EPHB2, EPHB3, ERBB2, ERBB4, FER, FGFR1, FGFR2, FGFR3, FGFR4, FLT1, FLT4, FRAP1, GCK, GRK4, GRK5, GRK6, GRK7, GSK3A, HCK, HIPK, HIPK2, HIPK3, HIPK4, IGF1R, IGF-1R, IKBKB, IKBKE, IKKalpha, IKKbeta, INSR, INSRR, IRAK1, IRAK4, ITK, JAK2, JAK2_JH1_JH2, JAK3, JNK1alpha1, JNK2alpha2, LCK, LIMK1, LKB1, LOK, LTK, MAP2K1, MAP2K2, MAP2K6, MAP3K8, MAP3K9, MAP4K4, MAPK1, MAPK10, MAPK11, MAPK12, MAPK13, MAPK14, MAPK2, MAPK8, MAPK9, MAPKAPK2, MAPKAPK3, MAPKAPK5, MARK1, MARK3, MELK, MERTK, MKK7beta, MLCK, MRCKalpha, MRCKbeta, MST1R, MST4, mTOR/FKBP12, MUSK, NEK3, NEK4, NEK6, NEK7, NEK9, NLK, NTRK1, NTRK2, NTRK3, PAK2, PAK4, PAK7_(KIAA1264), PAR-1Balpha, PASK, PDGFRalpha, PDK1, PHKG2, PIK3CA/PIK3R1, PIK3CG, PIM1, PIM2, PIM-3, PKBbeta, PKBgamma, PKCalpha, PKCepsilon, PKCeta, PKCiota, PKCmu, PKCtheta, PKG1alpha, PKG1beta, PKN1, PLK2, PLK3, PRK2, PRKACA, PRKCA, PRKCE, PRKCH, PRKCI, PRKCN, PRKCQ, PRKD1, PRKD2, PRKG1, PRKG2, PRKX, PTK2, PTK2B, RET, RIPK2, ROCK1, ROCK2, ROS1, RPS6KA1, RPS6KA2, RPS6KA3, RPS6KA4, RPS6KA5, RPS6KA6, RPS6KB1, SGK, SGK2, SGKL, SIK, SNF1LK2, SNK, SRPK1, SRPK2, STK3, STK22B, STK22D, STK23, STK25, STK33, SYK, TAK1, TAO3, TAOK2, TBK1, TEC, TEK, TLK2, TXK, TYK2, TYRO3, ULK2, ULK3, VRK2, WNK2, WNK3, ZAP70



Supplementary Table 2. Semi-quantitative immunohistochemistry data Patient

#* Daily Dose Biopsy AUC

(µM·h)

pERK H-score nucleus

pERK H-score cytoplasm

Ki67 Patient tumor best response**

Pre - 100 10 30% 12 8 00 Post 16 5 40 10 5% + 3%

Pre - 50 20 20% 19 8 00 Post 54 10 20 5% + 32%

Pre - 60 10 60% 21 1 600 Post 11 8 11 10 20% + 47%

Pre - 0 105 60% 26 1 600 Post 16 8 18 120 25% + 29%

Pre - 70 50 30% 42 1 440 Post 11 01 2 1 5% - 54%

Pre - 100 70 50% 44 6 40 Post 39 9 2 2 2% Not assessable

Pre - NA NA NA 45 6 40 Post 66 2 0.5 70 9% - 70%

Pre - 10 160 30% 48 6 40 Post 79 1 0 60 1% - 14%

Pre - 12 70 NA 56 1 440 Post 10 79 9 2 NA - 59%

Pre - 125 120 18% 60 1 920 Post 18 59 0 15 10% - 21%

Pre - 90 190 18% 62 1 920 Post 29 18 5 30 3% - 79%

Pre - 70 220 22% 65 1 920 Post NA 3 20 20% - 86%

Pre - 165 260 40% 76 1 920 Post 25 52 10 20 2% - 100%

Pre - 155 205 64% 79 1 920 Post NA 30 45 2% - 60%

Pre - 28 145 10% 81 1 920 Post 12 41 1 3 1% - 73%

- 2 40 290 53% 85 1 920 Pre Post 2717 0 117 20% - 23%

- 27 126 7% 93 1 920 Pre NA 0 5 3% - 50%

*Tumor biopsies from patients 12-48 were assessed at the University of Pennsylvania; assessment of the remaining biopsies was organized by Roche. ** Maximum percent reduction from baseline measurement of tumor dimensions; reductions of 30% or greater qualify as PRs by RECIST criteria. Note that tumors of all patients in this table had BRAFV600E mutations.

NA, not available



Supplementary Table 3. Data collection and refinement statistics§ PLX4032 in B-RAFV600E §Data collection Space group P212121 Cell dimensions a, b, c (Å) 50.8, 104.4, 110.1 α, β, γ (°) 90.0, 90.0, 90.0 Resolution (Å) 21.3-2.45 (2.58-2.45)* Rsym or Rmerge 7.2 % (52.4%) I/σI 8.8 (1.4) Completeness (%) 99.8 (99.8) Redundancy 4.5 (4.5) Refinement Resolution (Å) 21.3-2.45 No. reflections 21,223 Rwork/ Rfree 21 .3/25.8 No. atoms Protein 4002 Ligand/ion 33 Water 65 B-factors Protein 65.7 Ligand/ion 43.4 Water 44.1 R.m.s deviations Bond lengths (Å) 0.003 Bond angles (º) 0.7 Each asymmetric unit contains two BRAFV600E molecules. § These data were collected from a single crystal. * Highest resolution shell is shown in parenthesis.



inhibition of mutated, activated braf in metastatic melanoma

Documents