Oncogene Addiction and Overdose: Intermittent Treatment in Models of Drug-Resistant BRAF-Mutated Melanoma

Marian M. Deuker
Martin McMahon, PhD

Helen Diller Family Comprehensive Cancer Center and Department of Cell and Molecular Pharmacology University of California, San Francisco

Over the past 12 years, BRAF (v-raf murine sarcoma oncogene viral homolog B) has risen in stature: no longer an insufficiently studied regulator of cell signaling, BRAF has emerged as an important oncogenic driver of numerous cancers and a prominent target for successful pathway-targeted melanoma therapy.1 2 Indeed, the development of BRAF inhibitors (e.g., vemurafenib and dabrafenib), which selectively target the most common mutationally activated form of the protein in melanoma, has both improved treatment options for patients with BRAF-mutated melanoma and revealed unique features of RAS-regulated intracellular signaling.2 3 Unfortunately, dramatic melanoma regressions promoted by BRAFV600E-targeted therapies are often transient, with almost inevitable onset of recurrent disease driven by strikingly diverse mechanisms of drug resistance.4 5 These observations have fueled the development of preclinical systems to explore the causes and consequences of drug resistance and test novel therapeutic strategies to prevent the onset of drug-resistant melanoma.6

The Preclinical HMEX 1906 PDMX Model

The use of patient-derived melanoma xenografts (PDMX), serially passaged through immunocompromised mice, provides a complementary strategy to the use of either cell culture or genetically engineered mouse (GEM) models to study drug-resistant melanoma.7 While PDMX models are undoubtedly subject to selective pressure during establishment in mice, they avoid the detrimental stresses of cell culture, including: growth in plastic dishes, an oxygen-rich environment, and the use of growth supplements such as serum. Moreover, when used to model drug resistance, PDMX tumors are exposed to the drug via oral administration — a delivery method that more accurately reflects drug exposures in patients.

To that end, small fragments of a vemurafenib-naïve, lymph node-derived BRAF-mutated melanoma metastasis were subcutaneously implanted into immunocompromised mice. Some of these fragments grew into tumors that, when resected, could be serially propagated through larger numbers of mice, thereby generating the HMEX1906 PDMX model.6 At that point, a cohort of tumor-bearing mice was orally administered vemurafenib and, as expected, HMEX1906 melanomas displayed striking regression. However, when such mice were dosed continuously for greater than 50 days, stable drug-resistant tumors emerged that grew to end-stage. As expected, vemurafenib treatment of the original sensitive HMEX1906 tumors elicited profound inhibition of phosphorylation of the ERK1/2 (pERK1/2) MAP kinases, downstream effectors of oncogenic BRAFV600E. By contrast, drug-resistant HMEX1906 tumors displayed a higher basal level of pERK1/2 that was incompletely inhibited following exposure to vemurafenib. Hence, the HMEX1906 PDMX model proved useful for preclinical analysis of mechanisms of drug sensitivity and resistance in melanoma.

Extensive analysis was then performed on vemurafenib-sensitive or -resistant HMEX1906 melanomas to identify the mechanism(s) of resistance. Exome sequencing failed to detect any new mutations that might explain the properties of the resistant tumors. However, further analysis revealed that drug-resistant melanomas expressed elevated levels of mutated BRAF mRNA and BRAFV600E protein. Indeed, in one resistant tumor (45V-RT5), elevated BRAFV600E expression resulted from further amplification of the mutated BRAF gene. However, while all of the other resistant tumors displayed elevated BRAFV600E expression, none of them showed increased BRAF gene copy number. Hence, two different mechanisms of vemurafenib resistance — both of which resulted in increased BRAFV600E expression — arose from a small fragment of a PDMX. The continued dependence of resistant tumors upon BRAFV600E signaling was confirmed using either genetic or pharmacological inhibitors of the pathway. Hence, as first demonstrated by Robert Schimke at Stanford University in 1977, these data emphasize the importance of increased expression of the drug target as a relevant mechanism of cancer drug resistance.8 9

Drug-Resistant, Drug-Dependent Tumors

At this point, the study might have been concluded were it not for two provocative findings. First, it was noted that the transplantation efficiency of vemurafenib-resistant tumors was significantly higher if the mice were receiving vemurafenib at the time of the transplant, suggesting that the tumors were not simply drug-resistant, but also drug-dependent for optimal tumorigenesis. More telling was the observation that, while the original sensitive tumors readily gave rise to cell lines in vitro in the absence of vemurafenib, it was difficult to generate a cell line from the resistant melanomas unless cells were cultured in the presence of a moderate vemurafenib concentration (50 nM). These data confirmed that the drug-resistant tumor cells were also drug-dependent for growth and that, for these cells, culture either in too much or too little vemurafenib suppressed proliferation of the vemurafenib-resistant cells.

Thus, the tumor fitness benefit conferred by elevated expression of BRAFV600E in the presence of vemurafenib becomes a fitness deficit in the absence of vemurafenib. These observations are in accord with previous studies that showed how cells respond to both the quality (i.e., which pathways are on/off) and quantity (i.e., the integrated magnitude) of signal pathway activation. Specifically, it had previously been demonstrated that low levels of RAF pathway activation could promote the growth of cultured cells, whereas higher-level activation of the same pathway could elicit cell cycle arrest that, in primary cells, was irreversible and displayed features of cellular senescence.10 11 Hence, even though the HMEX1906 melanoma cells are addicted to oncogenic BRAFV600E signaling, they remain sensitive to the antiproliferative effects of high-level BRAFV600E→MEK→ERK pathway activation (Figure 1).

To test the relevance of this phenomenon in vivo, drug-resistant melanomas were implanted into mice that were dosed daily with vemurafenib. Whereas these continuously dosed mice showed sustained tumor growth, cessation of drug administration led to a highly reproducible decrease in melanoma cell proliferation accompanied by tumor regression, which correlated with a spike in the level of pERK1/2. However, these tumors eventually recommenced their growth in the absence of drug, with a concomitant decrease in pERK1/2. These studies indicated that, although elevated BRAFV600E expression confers vemurafenib resistance on HMEX1906 cells, upon drug removal the elevated BRAFV600E expression is now unopposed, thereby promoting a level of BRAFV600E→MEK→ERK signaling that elicits antiproliferative activity leading to tumor regression (Figure 2). However, eventually, a selective or adaptive response occurs, leading to reduced BRAFV600E→MEK→ERK activity, which allows for tumor regrowth.

Intermittent Dosing

The observation that vemurafenib-resistant tumors displayed a fitness deficit in the absence of drug led investigators to hypothesize that an intermittent drug-dosing schedule might forestall the onset of drug resistance by alternating the selective pressures on the drug-sensitive versus drug-resistant cell populations in the HMEX1906 tumors. To test this, mice were implanted with the original drug-sensitive HMEX1906 melanoma and, when tumors were readily measurable, randomized to receive vemurafenib either continuously or intermittently (4 weeks on, 2 weeks off). As expected, mice receiving continuous vemurafenib developed lethal drug-resistant disease within 100 days. By contrast, none of the mice on the intermittent dosing regimen developed drug-resistant disease within 200 days, even though at this point these tumors had been exposed to the same cumulative overall dose of vemurafenib as the continuously dosed mice.

A similar observation was obtained with a second chemo-naïve BRAF-mutated PDMX, in which intermittent dosing again forestalled the onset of drug resistance. Meanwhile, hints have emerged from the clinical literature suggesting that durable patient remission can be achieved using an intermittent dosing regimen.12

One question to arise is why elevated BRAFV600E→MEK→ERK signaling elicits antiproliferative effects in BRAF-mutated tumor cells. Certainly, elevation of BRAFV600E signaling in chemo-naïve HMEX1906 cells leads to cell cycle arrest, as is the case in other cancer cell lines.613 However, the potent cytostatic effects of elevated BRAFV600E signaling in vitro would not explain the frank regression of vemurafenib-resistant HMEX1906 tumors following cessation of drug administration in vivo. It is possible that the tumor regression observed under such circumstances is a composite of tumor cell autonomous effects combined with tumor cell-mediated activation of the innate immune system. Although PDMXs are generally propagated in immunocompromised mice, these strains of mice retain components of the innate immune system and natural killer (NK) cell function, which might be mobilized by excess BRAFV600E signaling within the tumor cells to promote tumor regression, as suggested previously.14

Next Frontiers

The major challenge ahead is to translate studies conducted in the well-controlled laboratory environment into the more complex world of clinical trials. This is indeed daunting due to the remarkable heterogeneity observed within BRAF inhibitor-resistant melanomas.5 15 Although drug resistance due to overexpression of BRAFV600E or the presence of truncated BRAFV600E splice variants may conform to the model described above, it seems unlikely that all mechanisms of BRAF inhibitor resistance will confer a fitness deficit in the absence of drug.6

Furthermore, even within a single melanoma, different resistance mechanisms may coexist — not all of which will benefit from intermittent dosing.9 16 Even the design of intermittent dosing clinical trials is complicated by the inability to predict the mechanisms of resistance that will emerge in any one lesion in any one patient. Hence, in the absence of a well-defined patient stratification strategy, it may be challenging to design a clinical trial that is sufficiently powered to detect a benefit in a sub-group of individuals.

An additional challenge is that the therapeutic landscape for patients with BRAF-mutated melanoma is rapidly evolving with the advent of new agents (dabrafenib, LGX818, trametinib, cobimetinib, etc.), some of which will be used in combination. Finally, when cancer cells are exposed to conventional or pathway-targeted chemotherapy, there is a selection for preexisting cell variants that have a fitness benefit in the presence of drug. Given the remarkable heterogeneity of melanoma identified by genome sequencing efforts, it is possible that intermittent dosing will simply select for those cells with mechanisms of resistance that are oblivious to intermittent dosing.

Despite the substantial challenges, a clinical trial is currently under way to test the utility of intermittent dosing with a BRAF inhibitor (LGX818, NCT01894672) in patients with stage IV or unresectable stage III BRAF-mutated melanoma. Planning is also under way to compare the ability of intermittent versus continuous dabrafenib plus trametinib to promote progression-free survival in patients (S1320).

Intermittent Combination Therapy: Sequential vs. Concurrent Administration

Assuming that such studies yield promising outcomes, what might be the best way to translate intermittent dosing of BRAFV600E pathway-targeted inhibitors into clinical practice? In the past few years, the response of melanoma patients to immunomodulators such as anti-CTLA-4, anti-PD1, or anti-PDL1 has generated considerable excitement.17 Since concurrent administration of both pathway-targeted and immunomodulatory therapies may be contraindicated either because of drug interference or unacceptable toxicity, an intermittent dosing regimen may allow such combinations to be employed in a rational sequential manner. For example, a patient with a BRAF-mutated melanoma might receive BRAFV600E pathway-targeted therapy for an initial period, until the deepest possible tumor regression has been achieved. At that time, the patient could be switched to an immunomodulatory agent to provoke a durable immune response against the residual disease. The patient might then be switched back and forth between the pathway-targeted and the immunomodulatory therapy, with the intent of eliciting the deepest and most durable remission. Clearly, the development of such a regimen would benefit from preclinical data supporting the concept. However, preclinical studies to test the utility of alternating pathway-targeted and immunomodulatory therapy are only feasible using either GEM models or PDMXs transplanted into immunocompromised mice reconstituted with a “humanized” immune system.18 19


The recent clinical successes of pathway-targeted and immunomodulatory therapies have converted advanced melanoma from a disease without efficacious treatment options into one with treatments affording remarkable disease control. This revolution in the treatment of BRAF-mutated melanoma also offers hope to patients with other recalcitrant cancers that have limited treatment options. However, attendant to each success are new challenges. Initial experience with pathway-targeted melanoma therapy has revealed the depth of heterogeneity that allows cancer cells to evade even the newest and most innovative treatments. Fortunately, the rapid rate of cancer evolution is paralleled by a rapid evolution in our ability to employ multiplex genetic and biochemical analyses to probe the inner workings of the cancer cell. Such detailed maps of the cancer cell promise to transform cancer from a disease treated with toxic chemotherapies to one treated with rational drug combinations designed to selectively ablate malignant cells and prevent lethal drug resistance.

The authors acknowledge our UCSF colleague, Dr. Adil Daud, for his comments on this manuscript.


  1. Davies H, Bignell GR, Cox C, et al. Mutations of the BRAF gene in human cancer. Nature 2002; 417(6892): 949-954.
  2. Chapman PB, Hauschild A, Robert C, et al. Improved survival with vemurafenib in melanoma with BRAFV600E mutation. New Engl J Med 2011; 364(26):2507-2516.
  3. Poulikakos PI, Zhang C, Bollag G, Shokat KM, Rosen N. RAF inhibitors transactivate RAF dimers and ERK signalling in cells with wild-type BRAF. Nature 2010; 464(7287):427-430.
  4. Wagle N, Emery C, Berger MF. Dissecting therapeutic resistance to RAF inhibition in melanoma by tumor genomic profiling. J Clin Oncol 2009; 29(22):3085-3096.
  5. Lito P, Rosen N, Solit DB. Tumor adaptation and resistance to RAF inhibitors. Nat Med 2013; 19(11):1401-1409.
  6. Das Thakur M, Salangsang F, Landman AS, et al. Modelling vemurafenib resistance in melanoma reveals a strategy to forestall drug resistance. Nature 2013; 494(7436):251-255.
  7. Siolas D, Hannon G. Patient-derived tumor xenografts: transforming clinical samples into mouse models. Cancer Res 2005; 73(17): 5315-5319.
  8. Schimke RT, Alt FW, Kellems RE, Kaufman RJ, Bertino JR. Amplification of dihydrofolate reductase genes in methotrexate-resistant cultured mouse cells. Cold Spring Harb Symp Quant Biol 1978; 42 Pt 2:649-657.
  9. Shi H, Moriceau G, Kong X, et al. Melanoma whole-exome sequencing identifies (V600E) B-RAF amplification-mediated acquired B-RAF inhibitor resistance. Nat Commun 2012; 3:724.
  10. Woods D, Parry D, Cherwinski H, Bosch E, Lees E, McMahon M. Raf-induced proliferation or cell cycle arrest is determined by the level of Raf activity with arrest mediated by p21Cip1. Mol Cell Biol 1997; 17(9):5598-5611.
  11. Zhu J, Woods D, McMahon M, Bishop JM. Senescence of human fibroblasts induced by oncogenic Raf. Genes Dev 1998; 12(19): 2997-3007.
  12. Callahan MK, Rampal R, Harding JJ, et al. Progression of RAS-mutant leukemia during RAF inhibitor treatment. N Engl J Med 2012; 367(24):2316-2321.
  13. McMahon M, Woods D. Regulation of the p53 pathway by Ras, the plot thickens. Biochim Biophys Acta 2001; 1471(2):M63-71.
  14. Krizhanovsky V, Yon M, Dickins RA, et al. Senescence of activated stellate cells limits liver fibrosis. Cell 2008; 134(4):657-667.
  15. Shi H, Hugo W, Kong X, et al. Acquired resistance and clonal evolution in melanoma during BRAF inhibitor therapy. Cancer Discov 2014; 4(1):80-93. doi:10.1158/2159-8290.CD-13-0642. Epub 2013 Nov 21.
  16. Poulikakos PI, Persaud Y, Janakiraman M, et al. RAF inhibitor resistance is mediated by dimerization of aberrantly spliced BRAF(V600E). Nature 2011; 480(7377):387-390.
  17. Callahan MK, Wolchok JD. At the bedside: CTLA-4- and PD-1-blocking antibodies in cancer immunotherapy. J Leukoc Biol 2013; 94(1):41-53.
  18. Dankort D, Curley DP, Cartlidge RA, et al. BRAF(V600E) cooperates with PTEN loss to induce metastatic melanoma. Nat Genet 2009; 41(5):544-552.
  19. Shultz LD, Brehm MA, Garcia-Martinez JV, Greiner DL. Humanized mice for immune system investigation: progress, promise and challenges. Nat Rev Immunol 2012; 12(11):786-798.
  20. Hanahan D, Weinberg RA. The hallmarks of cancer. Cell 2000; 100(1):57-70.
  21. Weinstein IB, Joe AK. Mechanisms of disease: Oncogene addiction — a rationale for molecular targeting in cancer therapy. Nat Clin Pract Oncol 2006; 3(8):448-457.