A Secondary Mutation in BRAF Confers Resistance to RAF Inhibition in a BRAFV600E-Mutant Brain Tumor

Abstract

BRAFV600E hyperactivates ERK and signals as a RAF inhibitor-sensitive monomer. Although RAF inhibitors can produce impressive clinical responses in patients with mutant BRAF tumors, the mechanisms of resistance to these drugs are incompletely characterized. Here, we report a complete response followed by clinical progression in a patient with a BRAFV600E-mutant brain tumor treated with dabrafenib. Whole-exome sequencing revealed a secondary BRAFL514V mutation at progression that was not present in the pretreatment tumor. Expressing BRAFV600E/L514V induces ERK signaling, promotes RAF dimer formation, and is sufficient to confer resistance to dabrafenib. Newer RAF dimer inhibitors and an ERK inhibitor are effective against BRAFL514V-mediated resistance. Collectively, our results validate a novel biochemical mechanism of RAF inhibitor resistance mediated by a secondary mutation, emphasizing that, like driver mutations in cancer, the spectrum of mutations that drive resistance to targeted therapy are heterogeneous and perhaps emerge with a lineage-specific prevalence.Significance: In contrast to receptor tyrosine kinases, in which secondary mutations are often responsible for acquired resistance, second-site mutations in BRAF have not been validated in clinically acquired resistance to RAF inhibitors. We demonstrate a secondary mutation in BRAF (V600E/L514V) following progression on dabrafenib and confirm functionally that this mutation is responsible for resistance. Cancer Discov; 8(9); 1130-41. ©2018 AACR.See related commentary by Romano and Kwong, p. 1064This article is highlighted in the In This Issue feature, p. 1047.

Trial registration: ClinicalTrials.gov NCT01677741.

Conflict of interest statement

Disclosure of potential conflicts of interest

Neal Rosen is on the scientific advisory board of and has ownership interests in BeiGene and Kura. Neal Rosen is also on the scientific advisory board of Chugai and Astra-Zeneca. Ira J. Dunkel is a consultant to Apexigen, Bayer, Bristol-Myers Squibb, Celgene, Eisai, Ipsen and Pfizer.

©2018 American Association for Cancer Research.

Figures

Figure 1. BRAF L514V emerges upon RAF inhibitor response and resistance in a pediatric glioma patient

A. Timeline of the patient’s clinical course. Significant events are shown along a timeline in weeks. Surgical resection was performed pre-enrollment on, and post-removal from, dabrafenib treatment. Dabrafenib was administered daily from weeks 0 to 40. Representative MRI images are shown, demonstrating pre-treatment tumor (Baseline, week 0), complete response of tumor to dabrafenib (week 12), and local tumor recurrence (week 40). Enhancing tumors of interest are marked with arrows. B. The fraction of cancer cells (cancer cell fraction, CCF) harboring key mutations in the pre/post treatment tumors and the SK-BT-DR short-term culture from the post-progression specimen inferred from whole-exome sequencing and IMPACT. C. Three-dimensional structure of BRAF V600E in complex with dabrafenib where the L514 residue is <4 angstroms away from inhibitor binding, in contact with the selectivity pocket. Protein coordinates are from the BRAF V600E/dabrafenib complex (PDB 4XV2).

Figure 2. BRAF V600E/L514V is an activating mutation and is resistant to dabrafenib

A. V5-tagged BRAF mutants [BRAF V600E (VE) or BRAF V600E/L514V (VELV)] were expressed in SKBR3 cells for 24 hr, followed by lapatinib treatment (1 μM for 1 hr), then followed by treatment with DMSO or dabrafenib (200nM) for another hour. Untransfected (UT). B. DBTRG-05MG cells were transiently transfected with V5-tagged BRAF V600E or BRAF V600E/L514V for 24 hr and treated with DMSO or dabrafenib for 1 hr. C – G. A375 (BRAF V600E melanoma) cells stably transduced with retrovirus carrying doxycycline-inducible GFP-tagged empty vector, or vector encoding V5-tagged BRAF mutants, were treated with doxycycline (150, 200, 150 ng/ml, respectively, in C and E; 50 or 200 ng/ml in D) for 24 hr, followed by treatment with dabrafenib (100nM in C, D and G; dose range as shown in E) for 1 hr (C – F) or over time course as shown in (G). In A – E and G, expression and/or phosphorylation of the indicated proteins were assessed by immunoblot. In F, p-ERK signal intensity was quantitated using densitometric analysis and graphed as a function of dabrafenib dose. H. A375 cells treated with dabrafenib at doses shown were assessed for cell viability and graphed as percentage relative to control. Doxycycline was replenished at 50% every 24 hr. Data represent mean ± SEM. I. A375 cells as in C – H were grown in soft agar with the indicated doses of dabrafenib (30 and 100 nM). The y-axis represents the number of colonies, expressed as an average of four fields per well, three wells per condition, relative to DMSO controls, for each cell line (data represent mean ± SEM; ****, P<0.0001, unpaired Student’s t-test).

Figure 3. BRAF V600E/L514V causes resistance to dabrafenib by formation of dimers

A. V5-tagged and FLAG-tagged wild-type BRAF and the indicated BRAF mutants were co-expressed in SKBR3 cells (lapatinib treated 1 hr prior to collection). The binding of V5-tagged BRAF proteins to FLAG-tagged BRAF proteins were determined by immunoprecipitation followed by immunoblotting analysis. B. Untransfected (UT) and transiently transfected SKBR3 cells expressing V5 or FLAG-tagged BRAF mutants with low (L), medium (M) and high (H) expression levels were treated with lapatinib for 1 hr. Homodimerization of BRAF V600E and V600E/L514V was evaluated by immunoblotting following immunoprecipitation, respectively. Relative FLAG signal (IP/WCL) in medium and high levels of BRAF mutants was determined by densitometry analysis. Whole cell lysate (WCL). Statistical analysis was performed using unpaired Student’s t-test in replicate experiments. C and D. V5-tagged BRAF [wild-type (WT); R509H; V600E (VE); V600E/R509H (VEH); V600E/L514V (VELV) or V600E/L514V/R509H (VELVH)] were expressed in SKBR3 cells for 24 hr, followed by lapatinib treatment (1 μM for 1 hr, C and D), then followed by treatment with DMSO (−) or dabrafenib (+, 200 nM) for 1 hr (D only). Untransfected (UT). Relative p-MEK and p-ERK levels were quantitated in replicate experiments. and shown as bar graph with statistical analysis. Not significant (n.s.) E. A375 (BRAF V600E melanoma) cells stably transduced with retrovirus carrying doxycycline-inducible vector encoding V5-tagged BRAF mutants were treated with doxycycline (150 or 25 ng/ml) for 24 hr, followed by treatment with dabrafenib (dose range as shown) for 1 hr. Expression and/or phosphorylation of the indicated proteins were assessed by immunoblot. p-ERK signal intensity was quantitated using densitometric analysis, and graphed as a function of dabrafenib dose.

Figure 4. Dabrafenib plus trametinib more potently inhibits signaling in BRAF V600E/L514V expressing cells than dabrafenib alone

A. V5-tagged BRAF mutants [BRAF V600E (VE) or BRAF V600E/L514V (VELV)] were expressed in SKBR3 cells for 24 hr, followed by lapatinib treatment (1 μM for 1 hr), then followed by treatment with DMSO or dabrafenib (200 nM) and/or trametinib (20 nM) for 1 hr. Untransfected (UT). Expression and/or phosphorylation of the indicated proteins were assessed by immunoblot. B. A375 (BRAF V600E melanoma) cells stably transduced with retrovirus carrying doxycycline-inducible GFP-tagged empty vector or vector encoding BRAF mutants [BRAF V600E (VE) or BRAF V600E/L514V (VELV)] were treated with doxycycline (150, 200, 150 ng/ml) for 24 hr, followed by treatment with dabrafenib (doses as shown) and / or trametinib (20 nM) for 1 hr. Expression and/or phosphorylation of the indicated proteins were assessed by immunoblot. C. A375 cells stably expressing GFP, VE and VELV treated with trametinib at doses shown were assessed for cell viability. Doxycycline was replenished at 50% every 24 hr. Data represent mean ± SEM. D. A375 cells as in B – C were grown in soft agar in the presence of doxycycline (150, 200, 150 ng/ml) with the indicated doses of dabrafenib (doses as shown, nM) and / or trametinib (20 nM). The y-axis represents the number of colonies, expressed as an average of four fields per well, three wells per condition, relative to DMSO controls for each cell line (data represent mean ± SEM; ****, P<0.0001, unpaired Student’s t-test).

Figure 5. RAF dimer inhibitors BGB3245 and BGB3290 overcome L514V-mediated resistance to dabrafenib

A. A375 cells expressing inducible BRAF mutants [BRAF V600E (VE) or BRAF V600E/L514V (VELV)] were cultured in doxycycline containing medium for 24 hr and then treated with increasing concentrations of the indicated RAF inhibitors for 1 hr. Cell lysates were examined by immunoblot using the antibodies indicated. B. BRAF V600E (VE) or V600E/L514V (VELV) were transiently transfected in SKBR3 for 24 hr. Then the cells were treated with 1 μM lapatinib for 1 hr, followed by treatment with DMSO or the RAF inhibitors indicated for 1 hr (the dose for vemurafenib, PLX7904, LY3009120, TAK632, BGB3245 and BGB3290 is 1 μM; the dose for dabrafenib and PLX8394 is 300nM). Untransfected (UT) SKBR3 cells were treated with DMSO or 1 μM lapatinib (Lanes 1 and 2). C. SKBR3 cells expressing VE or VELV were pretreated with lapatinib for 1 hr, followed by treatment with dabrafenib or BGB3245 over a time course. ERK signaling was assessed by immunoblot. D. A375 cells expressing inducible GFP or BRAF mutants (VE or VELV) treated with BGB3245 at concentrations shown were assessed for cell viability and graphed as percentage relative to control. Doxycycline was replenished at 50% every 24 hr. Data represent mean ± SEM.