Selumetinib for the treatment of cancer

Abstract

Introduction: The MAPK pathway is essential for regulation of cellular proliferation, differentiation and survival. Multiple human cancers have demonstrated activation of Raf-mitogen-activated kinase kinase (MEK)–extracellular signal-related kinase signaling, a hallmark of these tumors. Efforts to inhibit various protein kinases in this pathway have led to the development of MEK inhibitors. Selumetinib is one such drug, functioning as an oral, selective non-ATP-competitive MEK1/2 inhibitor.

Areas covered: In this article, the authors discuss the underlying biology of MEK inhibition and its rationale in cancer treatment. Furthermore, the authors summarize the clinical development of selumetinib in various tumor types, from initial Phase I studies to randomized Phase II studies, both as monotherapy or in combination with other chemotherapeutics.

Expert opinion: Given the frequency of activated MAPK signaling in multiple tumor types, the potent MEK inhibitor selumetinib had strong preclinical and early clinical rationale, particularly in those tumors harboring KRAS or BRAF mutations. While efficacy signals have been seen in various tumor types treated with selumetinib, better biomarkers are needed to select patients most likely to respond favorably to this agent. Furthermore, combinatorial therapy with selumetinib and other targeted agents can likely be optimized to maximize the antitumor effect of inhibiting RAS/MAPK signaling.

Keywords:

• ARRY-142886
• AZD6244
• extracellular signal-related kinase
• MAPK
• mitogen-activated protein kinase
• selumetinib

1. Introduction

The MAPK signaling pathway constitutes a family of protein kinases that regulate cellular proliferation, survival and angiogenesis, among other important functions. These kinases transduce signals from extracellular stimuli such as growth factors, causing intracellular responses through a cascade of phosphorylation events [1,2]. As activation of the MAPK pathway is critical for neoplastic growth, efforts to inhibit various protein kinases in this pathway have been attempted to prevent signal transduction and decrease tumor growth. One such example of these MAPK pathway inhibitors is selumetinib (Box 1), a mitogen-activated protein kinase kinase (MAPKK, or MEK) inhibitor that targets MEK1 and MEK2 and is the focus of this article.

Box 1. Drug summary.

Drug name	Selumetinib (AZD6244; ARRY-142886)
Phase	Phase I – III
Indication	Cancer
Pharmacology description	Selective allosteric MEK1/MEK2 inhibitor
Route of administration	Oral
Chemical structure
Pivotal trial(s)	[52,61,66]

Pharmaprojects - copyright to Citeline Drug Intelligence (an Informa business). Readers are referred to Informa-Pipeline (http://informa-pipeline.citeline.com) and Citeline (http://informa.citeline.com).

2. Preclinical MEK inhibitor development and the biological characterization of selumetinib

Due to the known activation of Raf-MEK–extracellular signal-related kinase (ERK) kinase signaling in multiple human tumors and the lack of clinically viable Ras, Raf or ERK inhibitors in the early 2000s, a series of MEK1/2 inhibitors were developed for clinical testing. The first-generation MEK1/2 inhibitor CI-1040 did not show efficacy in lung, breast, colon or pancreatic malignancies, and therefore its development was halted after Phase II testing [3]. Two second-generation MEK1/2 inhibitors, PD 0325901 and selumetinib (previously known as ARRY-142886 andAZD6244) did appear more promising based on increased potency against MEK and improved efficacy in preclinical models [4,5], though development of the former was eventually discontinued by Pfizer.

Initial preclinical characterization of selumetinib demonstrated noncompetitive activity with respect to adenosine triphosphate (ATP), consistent with its binding to the allosteric inhibitor binding site in MEK1/2, and a high specificity of the drug for MEK1/2 [6,7]. Mechanistically, the presence of selumetinib appears to lock MEK1/2 into an inactive conformation, enabling ATP binding and substrate but disrupting the molecular interactions required for catalysis and access to the ERK activation loop [7]. Furthermore, this compound has an IC₅₀ against purified MEK1 of 14 nmol/l, and no inhibition was shown when tested against > 40 other kinases [7]. As further proof of target inhibition, basal and epidermal growth factor-induced ERK1/2 phosphorylation were shown to be inhibited by selumetinib in multiple cell lines, as well as in isolated peripheral blood mononuclear cells, but no inhibition of MEK1/2 phosphorylation was demonstrated, suggesting direct inhibition of enzymatic activity and not inhibition of RAF activation [7]. Interestingly, selumetinib was found to be more potent in cell lines containing activating BRAF and RAS mutations but had minimal effect on other cell lines [8]. Selumetinib did inhibit ERK1/2 phosphorylation and growth of HT-29 (BRAF mutant colorectal) xenograft tumors in nude mice, consistent with previous reports [9,10]. Tumor regressions were also seen in a KRAS wild-type BxPC3 pancreatic tumor xenograft model, as well as in other human xenograft models (colorectal, pancreatic, non-small cell lung, hepatocellular and melanoma [7,8,11-14]. On the basis of this intriguing preclinical data, clinical development of selumetinib was undertaken [15].

3. Phase I clinical trials of selumetinib

To first assess safety and tolerability of selumetinib in patients with advanced cancers, a Phase I pharmacokinetic and pharmacodynamics study was performed [16]. A total of 57 patients were treated with selumetinib at various dose levels (50, 100, 200, 300 mg twice daily (b.i.d.) in a free-base suspension), with melanoma (35.1%), breast (17.5%) and colorectal (8.8%) cancers being the most prevalent tumors in this study. Overall, selumetinib was well-tolerated, with the most significant and prevalent toxicities being maculopapular rash (74% of all patients, with seven patients experiencing dose-limiting toxicity), diarrhea (56%), edema (33%) and fatigue (39%). Based on the incidence and severity of toxicities at various dose levels, the maximum tolerated dose (MTD) and recommended Phase II dose of selumetinib were considered to be 100 mg orally b.i.d. The half-life of selumetinib was found to be 8.3 h after a single dose. With regard to pharmacodynamics parameters, inhibition of ERK phosphorylation in lymphocytes from 12-O-tetradecanoylphorbol-13-acetate-treated whole blood was nearly ubiquitous both within 1 h of the first dose as well as in trough samples on days 15 or 22 of treatment. Inhibition of ERK phosphorylation by immunohistochemistry of post-treatment tumor samples when compared to pretreatment samples was consistently seen, with geometric mean inhibition of 79% (90% CI, 50 – 91%). Ki-67 labeling of paired tumor samples was also reduced but less consistently, with reduction in nine post-treatment samples, with at least 50% reduction in five of those tumors.

As part of this Phase I study, patient tumors were also evaluated for potentially predictive genomic alterations. Of the 26 patients with sufficient tumor for mutational analysis, RAS or RAF mutations were seen in 10 (KRAS = 5, NRAS = 4, BRAF = 1). Four of the mutated tumors showed strong inhibition of ERK phosphorylation and Ki-67 labeling index, but the small sample size precluded statistically significant changes in biomarker breakdown or duration of time on study based on mutational status. Overall, 19 patients (33%) had stable disease after 2 months of therapy, and stable disease persisted for at least 5 months in 9 patients, 2 of whom were even more prolonged (medullary thyroid cancer, 19 cycles; concomitant melanoma and renal cell carcinoma, 22 cycles). In addition, one melanoma patient with an NRAS mutation had 70% tumor shrinkage after three cycles of therapy but developed new brain metastases prior to confirmatory scans. In all, this Phase I trial identified a tolerable and target-inhibiting dose of 100 mg orally b.i.d. and provided rationale for further clinical testing of selumetinib in the Phase II setting.

In an attempt to develop a more convenient dosing of selumetinib than the free-base solution described above, an oral capsule formulation incorporating a hydrogen sulfate (Hyd-Sulfate) salt was developed and tested in the Phase I setting [17]. A total of 59 patients participated in this two-part study. In part A, 30 patients received escalating doses of the Hyd-Sulfate formulation b.i.d. In part B, 29 patients were randomized to one dose of either the free-base formulation or the Hyd-Sulfate capsule, followed by a washout period and crossover to the other formulation; all patients in part B ultimately received the Hyd-Sulfate capsule b.i.d. at the MTD. The MTD for the Hyd-Sulfate formulation was found to be 75 mg b.i.d., with dose-limiting toxicities, including rash and pleural effusion. The most prevalent toxicities in this study overall were fatigue (67.9%), acneiform dermatitis (62.5%), nausea (53.6%) and diarrhea (48.2%). Exposure of the 75 mg Hyd-Sulfate capsule relative to the 100 mg free-base suspension was 197% (90% CI, 161 – 242%). Furthermore, target inhibition with the Hyd-Sulfate capsule was confirmed through inhibition of 12-O-tetradecanoylphorbol-13-acetate-induced ERK phosphorylation. Interestingly, a melanoma patient harboring a BRAF V600E mutation achieved a complete response with this therapy, which persisted beyond 15 months of therapy. Based on its favorable safety and tolerability profile, this oral capsule formulation was also considered acceptable for further clinical testing and in fact was incorporated into the Phase II setting.

4. Selumetinib for the treatment of gastrointestinal malignancies

In one study of 69 patients with cholangiocarcinoma, 31 (45%) had KRAS mutations and 15 (22%) had BRAF V600E mutations [18], whereas KRAS mutations have been observed in 10 – 57% of gallbladder carcinomas [19], indicating that a significant proportion of biliary tract cancers activate RAS/MAPK signaling. Given this rationale, a multi-institutional Phase II study of selumetinib was performed in patients with metastatic biliary cancers (Table 1) [20]. In this study, 28 evaluable patients were treated with oral selumetinib 100 mg b.i.d. (free-base solution), with the most common toxicities on treatment consisting of rash (90%), xerostomia (54%) and nausea (51%). A total of three patients had a confirmed partial response, with one of these patients having an unconfirmed complete response. Furthermore, 17 patients (68%) had stable disease, including 3 (12%) with duration of > 1 year. Median progression-free survival (PFS) was 3.7 months (95% CI, 3.5 – 4.9), and median overall survival (OS) was 9.8 months (95% CI, 5.97 to not available). Interestingly, of these tumors, only two were found to harbor KRAS exon 2 mutations (none were responders), and none had BRAF mutations. It is unclear whether this genomic profile is truly reflective of biliary cancers as a whole given the relatively small sample size. However, the clinical activity seen with selumetinib in this group of KRAS and BRAF wild-type biliary cancers raises the question of an alternative mechanism of action such as epigenetic alterations or immune modulation. Furthermore, there was a lack of clear correlation between phosphorylated ERK (pERK) and phosphorylated AKT (pAKT) immunostaining and survival in this study, suggesting that these markers may not have predictive relevance.

Table 1. Selected Phase II clinical trials of selumetinib in advanced gastrointestinal cancer patients.

Author	Tumor type	Number of patients (N)	Study design	Selumetinib dose and formulation	Comparator	ORR (S vs C)	Median PFS (S vs C)	Median OS (S vs C)
Bekaii-Saab et al. (2011) [20]	Biliary	N = 28	Single arm	100 mg po b.i.d. (free-base)	None	3/28 (12%)	3.7 months (95% CI 3.5 – 4.9)	9.8 months (95% CI 5.97 to not available)
O’Neil et al. (2011) [32]	HCC	N = 19	Single-arm	100 mg po b.i.d. (free-base)	None	0/19 (0%)	1.4 months (95% CI 1.2 – 2.5)	4.2 months (95% CI 1.9 – 6.0)
Bennouna et al. (2011) [34]	Colorectal (second or third line)	N = 69 (34 selumetinib, 35 capecitabine)	Randomized, parallel-group	100 mg po b.i.d. (free-base)	Capecitabine 1250 mg/m^2 po b.i.d. × 2 wks every 3 wks	0/34 (0%) versus 1/35 (2.9%)	81 versus 88 days (HR 1.08; two-sided 80% CI 0.76 – 1.52)	N/A
Bodoky et al. (2012) [35]	Pancreas (second line)	N = 70 (38 selumetinib, 32 capecitabine)	Randomized, parallel-group	100 mg po b.i.d. (free-base)	Capecitabine 1250 mg/m^2 po b.i.d. × 2 wks every 3 wks	2/38 (5.3%) versus 3/32 (9.8%)	2.1 versus 2.2 months (HR 1.24, two-sided 80% CI 0.88 – 1.75)	5.4 versus 5 months (HR 1.03, two-sided 80% CI 0.68 – 1.57)

b.i.d.: Twice daily; C: Comparator; CI: Confidence interval; HCC: Hepatocellular carcinomas; HR: Hazard ratio; N/A: Not applicable; ORR: Overall response rate; OS: Overall survival; po: Orally; PFS: Progression-free survival; S: Selumetinib; wks: Weeks.

Apart from the potential clinical efficacy of selumetinib in biliary cancer, another intriguing effect was also seen with this drug. From the study by Bekaii-Saab et al. described above, quantitative analysis of computed tomography images for loss and gain of muscle in cholangiocarcinoma patients was undertaken [21]. A total of 84.2% of patients gained muscle after selumetinib initiation, with a mean overall gain of total lumbar muscle cross-sectional area was 13.6 cm²/100 days (∼ 2.3 kg on a whole-body basis). In contrast, patients undergoing standard treatment with chemotherapy or radiation had an overall muscle loss of −7.3 cm²/100 days (∼ 1.2 kg), with only 16.7% of these patients gaining muscle with treatment. This effect of selumetinib is hypothesized to be a result of inhibition of cytokine secretion such as IL-6, IL-1β and TNF-α, all implicated in the promotion of cancer cachexia [22,23]. While preliminary, these results suggest a potentially interesting on-target effect of selumetinib unrelated to antitumor activity.

Subsequent preclinical studies in primary xenograft models derived from patients with gallbladder and distal bile duct cancer have also investigated the possibility of sequence dependence of selumetinib when given in combination with gemcitabine for the treatment of biliary cancers [24]. While a prior study had not shown increased efficacy of gemcitabine and selumetinib in a pancreatic cancer model [25], other evidence showed the depletion of S-phase cells in pancreatic xenograft models when treated with selumetinib [26]. Since gemcitabine is active in the S-phase, it was hypothesized that concurrent use of gemcitabine and selumetinib might be antagonistic and not synergistic. When evaluated in biliary cancer xenograft models, DNA synthesis was indeed suppressed during selumetinib treatment, and reentry into S-phase was delayed by 48 h after treatment [24]. Strong schedule dependence was seen in all the models tested, suggesting a clinical schedule of gemcitabine treatment after a 48-h interruption of selumetinib therapy [24]. While this schedule has not been clinically tested to our knowledge, it underscores the importance of considering biological interactions of targeted therapies with cytotoxic chemotherapeutic agents when developing new agents and combinatorial regimens.

In contrast to biliary cancers, hepatocellular carcinomas (HCC) rarely harbor RAS or RAF mutations [18,27,28], but they often demonstrate activation of the Raf/MEK/ERK pathway [29-31], perhaps through autocrine signaling through receptor tyrosine kinases such as the EGFR, the insulin-like GFR and c-MET [32,33]. To test whether MEK1/2 inhibition would be clinically useful in HCC, a single-arm, multi-institutional Phase II study was undertaken in patients with advanced or metastatic HCC (Table 1) [32]. A total of 19 patients were treated on this study, with each patient receiving selumetinib 100 mg orally b.i.d. in the free-base solution. No objective radiographic responses to selumetinib were observed, with best response of stable disease seen in 6 (35%) of 17 patients (95% CI, 14 – 62%). Median PFS was 1.4 months (95% CI, 1.2 – 2.5), and median OS was 4.2 months (95% CI, 1.9 – 6.0). Common toxicities included nausea and maculopapular rash. In this study, MEK phosphorylation was indeed inhibited by selumetinib, proving effective target inhibition though ultimately no clinical benefit. Correlative studies performed did raise the possibility of phosphorylated ERK expression being a favorable prognostic factor, but not an adequate selection factor for therapy with MEK inhibitors.

Selumetinib has also been tested in patients with colorectal cancer in the second- or third-line setting. Bennouna et al. performed a Phase II parallel-group study in which patients with disease refractory to oxaliplatin and/or irinotecan were randomized to either capecitabine monotherapy or selumetinib therapy (Table 1) [34]. A total of 69 patients were treated with either selumetinib 100 mg orally b.i.d. (free-base suspension) or capecitabine 1250 mg/m² orally b.i.d. for 2 weeks out of every 3-week period. Median PFS did not differ between the two groups (81 vs 88 days for selumetinib vs capecitabine), and the number of disease progression events was also similar between groups (28 patients per arm). For the selumetinib group, a best response of stable disease was seen in 10 patients, and toxicities were similar to other studies of selumetinib. Of note, patients enrolled in this study were not selected on the basis of RAS mutation, despite the prevalence of this genomic alteration in colorectal cancer and mutational analysis was not performed as part of this study. It was concluded from this study that single-agent selumetinib had comparable efficacy to capecitabine monotherapy in the second- or third-line setting in patients with unselected colorectal cancer; however, this efficacy was not remarkable.

In a similar manner, selumetinib has also been evaluated clinically compared to capecitabine monotherapy in patients with pancreatic cancer refractory to first-line gemcitabine. A Phase II open-label study by Bodoky et al. investigated single-agent selumetinib 100 mg orally b.i.d. (free-base solution) versus capecitabine 1250 mg/m² orally b.i.d. for 2 weeks of every 3-week cycle (Table 1) [35]. A total of 70 patients were randomized in this study, with OS as the primary end point. Median OS was 5.4 months in the selumetinib arm versus 5 months in the capecitabine arm (HR 1.03, two-sided 80% CI 0.68 – 1.57, p = 0.92). Disease progression events occurred in a similar rate among the selumetinib and capecitabine arms, respectively (84 vs 88%, HR 0.89, two-sided 80% CI, 0.6 – 1.31, p = 0.69). Median PFS was also similar between the selumetinib and capecitabine arms (2.1 vs 2.2 months, HR 1.24, two-sided 80% CI, 0.88 – 1.75, p = 0.41), though it must be noted that radiographic assessments were not undertaken at specified times per protocol. Best overall response by response evaluation criteria in solid tumors (RECIST) was investigator-assessed and included two partial responses for the selumetinib arm and three partial responses in the capecitabine arm; stable disease was seen as best response in 12 and 9 patients in the selumetinib and capecitabine groups, respectively. On the basis of these data, it was concluded that there was no statistically significant improvement in OS between selumetinib and capecitabine in the second-line setting for patients with advanced pancreatic cancer, though overall selumetinib was a well-tolerated therapy.

5. Selumetinib for the treatment of thyroid cancer

Approximately 70% of papillary thyroid cancers have been found to harbor mutations in RAS, BRAF, RET or NTRK1 [36-39], with resultant stimulation of MAPK signaling. It has been shown preclinically that selumetinib is particularly potent in BRAF V600E mutant papillary thyroid cancer cell lines and some with RAS mutations [8,40,41]. Therefore, a Phase II trial was performed by Hayes et al. to investigate the safety and efficacy of selumetinib in patients with advanced, iodine-refractory, progressive papillary thyroid cancer (Table 2) [42]. In this study, 39 patients were treated with selumetinib 100 mg orally b.i.d. (free-base formulation). Best responses of 32 evaluable patients were 1 patient with partial response (3%) and 21 patients with stable disease (54%); therefore, the primary endpoint of overall response rate of at least 20% was not met. Median PFS in this single-arm study was 32 weeks. Toxicities were as previously reported for selumetinib therapy. Of note, patients with BRAF V600E mutant tumors (12 of 26 evaluated, or 46%) had a longer median PFS than those with BRAF wild-type tumors, but this was not statistically significant (33 vs 11 weeks, HR 0.6, p = 0.3).

Table 2. Selected Phase II clinical trials of selumetinib in thyroid cancer, NSCLC, AML and ovarian cancer.

Author	Tumor type	Number of evaluable patients (N)	Study design	Selumetinib dose and formulation	Comparator	ORR (S vs C)	Median PFS (S vs C)	Median OS (S vs C)
Hayes et al. (2012) [42]	Iodine-refractory papillary thyroid cancer	N = 39	Single-arm	100 mg po b.i.d. (free-base)	None	1/32 (3.1%)	32 weeks (95% CI 8.4 – 56)	N/A
Hainsworth et al. (2010) [49]	NSCLC (second or third line)	N = 84 (40 selumetinib, 44 pemetrexed)	Randomized, parallel-group	100 mg po b.i.d. (free-base)	Pemetrexed 500 mg/m^2 IV every 3 wks	2/40 (5%) versus 2/44 (4.5%)	67 versus 90 days (HR 1.08, two-sided 80% CI 0.75 – 1.54)	N/A
Janne et al. (2013) [52]	NSCLC (KRAS mutant, second line)	N = 87 (43 selumetinib + docetaxel, 40 placebo + docetaxel)	Randomized, parallel-group (docetaxel 75 mg/m^2 on day 1 of every 21-day cycle + placebo or selumetinib)	75 mg po b.i.d. (hydrogen sulfate)	Placebo	16/43 (37%) versus 0/40 0%)	5.3 versus 2.1 months HR 0.58, 80% CI 0.42 – 0.79)	9.4 versus 5.2 months (HR 0.80, 80% CI, 0.56 – 1.14)
Jain et al. (2014) [71]	AML (stratified by FLT3 ITD status)	N = 47 (36 FLT3 wild-type, 10 FLT ITD)	Single-arm	100 mg po b.i.d. (free-base)	None	FLT3 wild-type: 6/36 (17%); LT3 ITD: 0/10 (0%)	N/A	N/A
Farley et al. (2013) [72]	Low-grade serous ovarian/peritoneal	N = 52	Single-arm	50 mg po b.i.d. (hydrogen sulfate)	None	8/52 (15%)	11 months (3.6 – 15.9)	NR

AML: Acute myeloid leukemia; b.i.d.: Twice daily; C: Comparator; CI: Confidence interval; HR: Hazard ratio; IV: Intravenous; N/A: Not applicable; NR: Not reached; NSCLC: Non-small cell lung cancer; ORR: Overall response rate; OS: Overall survival; po: Orally; PFS: Progression-free survival; S: Selumetinib; wks: Weeks.

Despite this negative study, interest in selumetinib for the treatment of thyroid cancer remained. Preclinical data clearly demonstrated that an increase in MAPK signaling inhibits the expression of thyroid hormone biosynthesis genes, some of which facilitate iodine uptake and organification [43-45]. As radioiodine (iodine-131) therapy is a mainstay for patients with metastatic thyroid cancer of follicular origin [45], interference with this mechanism leads to radioiodine resistance and ultimately a poor prognosis for the patient. Preclinical studies of mouse thyroid cancers with conditional BRAF activation treated with MEK or BRAF inhibitors showed that the tumors in question were able to regain the ability to trap radioiodine [46]. To test this preclinical rationale in a clinical setting, 24 patients with metastatic thyroid cancer refractory to radioiodine were treated with selumetinib 75 mg orally b.i.d. (Hyd-Sulfate formulation) for 4 weeks; an iodine-124 positron emission tomography/computed tomography (PET-CT) study was performed pre- and post-treatment. If the post-treatment iodine-124 PET showed an increase in iodine uptake to a prespecified dosimetry threshold, the patients continued to receive selumetinib therapy; once determination of the maximum tolerable activity was made, therapeutic doses of iodine-131 were then administered. Of the 24 patients, 9 had tumors with BRAF mutations and 5 had tumors with NRAS mutations [45]. Selumetinib increased the uptake of iodine-124 in 12 of the 20 evaluable patients (4 of 9 with BRAF mutations and 5 of 5 with NRAS mutations). A total of 8 of these patients reached the dosimetry threshold for radioiodine therapy, and 5 of these patients ultimately had confirmed partial responses and 3 had stable disease. This exciting study confirmed the reversal of radioiodine resistance of refractory thyroid cancer with selumetinib treatment through induction of iodine uptake and retention in these tumors, a novel way to approach and treat this resistant tumor subtype.

6. Selumetinib for the treatment of NSCLC

NSCLC is another tumor type with a significant incidence of KRAS mutations, ∼ 25% overall [47,48], and therefore an attractive target for MEK inhibition [8]. Hainsworth et al. performed a randomized Phase II study of selumetinib versus pemetrexed in NSCLC patients whose disease had progressed on one or two prior chemotherapeutic regimens (Table 2) [49]. In this study, 84 patients were randomized to selumetinib 100 mg oral b.i.d. (free-base suspension) or pemetrexed 500 mg/m² intravenous (IV) every 3 weeks. A similar number of patients on the selumetinib and pemetrexed arms experienced disease progression on this trial within the specified time period (70 vs 59%, HR 1.35, two-sided 80% CI, 0.93 – 1.94, two-sided 95% CI, 0.77 – 2.36, p = 0.30). Median PFS was also not statistically significantly different between the selumetinib and pemetrexed arms (67 vs 90 days, HR 1.08, two-sided 80% CI, 0.75 – 1.54, two-sided 95% CI, 0.62 – 1.86, p = 0.79), though the protocol did not require regular tumor assessments and time to progression may have been overestimated for some patients. Two patients in each treatment group did achieve a best response of complete or partial response, with two patients in the selumetinib arm having partial responses and one patient each with partial and complete responses on the pemetrexed arm. Toxicities were overall as expected for each group. Of note in this study, however, patients were not selected according to KRAS mutational status or other genomic alteration, perhaps diluting the potential effect of selumetinib on tumors with less MAPK pathway activation.

In contrast, another Phase II study of selumetinib was performed in patients with KRAS mutant NSCLC, this time in combination with docetaxel (Table 2). This particular combination was selected on the basis of preclinical data showing greater tumor growth inhibition or regression and apoptosis, as well as synergism, between docetaxel and selumetinib in in vivo models [8,25,50]. A Phase I study of the combination in advanced solid tumors did demonstrate a manageable safety profile as well [51]. Therefore, a randomized Phase II study of docetaxel 75 mg/m² on day 1 of every 21-day cycle plus either placebo or selumetinib 75 mg oral b.i.d. as second-line therapy was undertaken in patients with KRAS mutant advanced NSCLC [52]. A total of 87 patients were randomized, with the primary end point of median OS not improved with the addition of selumetinib to docetaxel (9.4 vs 5.2 months, HR 0.80, 80% CI, 0.56 – 1.14, one-sided p = 0.21). Median PFS was significantly improved with selumetinib (5.3 vs 2.1 months, HR 0.58, 80% CI, 0.42 – 0.79, one-sided p = 0.014). However, grade 3 or higher adverse events were more frequent in the selumetinib group (82 vs 67%), the most common in the selumetinib arm being febrile neutropenia, grade 3–4 neutropenia, dermatitis acneiform and asthenia. Of note, there was an improvement in patient-reported disease-related symptoms in the selumetinib arm. It is unclear whether response to selumetinib could be improved with better understanding of the underlying biology of MAPK pathway activation and more accurate predictive biomarkers. A Phase III second-line study of docetaxel plus either selumetinib or placebo for patients with KRAS-mutant NSCLC is currently underway.

7. Selumetinib for the treatment of melanoma

With advantages in genomic technology over the past 10 years, DNA mutational analysis of tumors such as melanoma has altered the therapeutic approach to these malignancies. In the case of melanoma, approximately half of these tumors harbor BRAF mutations, resulting in constitutive activation of the MAPK cascade [53,54]. Activation of the MAPK cascade also occurs in the setting of NRAS mutations, which are found in ∼ 15 – 25% of melanomas [55,56]. Solit et al. demonstrated preclinically that BRAF mutation is associated with enhanced and selective sensitivity to MEK inhibition when compared to wild-type cells or those with a RAS mutation [10]. MEK inhibition with either CI-1040 or PD0325901 potently abrogated tumor growth in BRAF mutant xenografts, providing strong rationale for testing MEK inhibitors in melanoma, given the high prevalence of BRAF mutations in this tumor type.

Selumetinib monotherapy in melanoma was initially tested in patients with advanced melanoma unselected for BRAF mutational status. A randomized Phase II study of temozolomide 200 mg/m²/day orally for 5 days, then 23 days off treatment versus selumetinib 100 mg orally b.i.d. was undertaken by Kirkwood et al. (Table 3) [57]. PFS did not differ between selumetinib and temozolomide (78 vs 80 days, HR 1.07, 80% CI, 0.86 – 1.32), but five of six partial responders to selumetinib were found to be BRAF mutated [57].

Table 3. Selected Phase II clinical trials of selumetinib in advanced melanoma patients.

Author	Tumor type	Number of patients (N)	Study design	Selumetinib dose and formulation	Comparator	ORR (S vs C)	Median PFS (S vs C)	Median OS (S vs C)
Kirkwood et al. (2014) [57]	Melanoma	N = 200 (104 selumetinib, 96 temozolomide)	Randomized, parallel-group	100 mg po b.i.d. (free-base)	Temozolomide 200 mg/m^2/day po × 5 days, then 23 days off tx	6/104 (5.8%) versus 9/96 (9.4%)	78 versus 80 days (HR 1.07, 80% CI 0.86 – 1.32)	N/A
Gupta et al. (2014) [58]	Melanoma (BRAF wild-type)	N = 83 (41 selumetinib + docetaxel, 42 placebo + docetaxel)	Randomized, parallel-group (docetaxel 75 mg/m^2 IV every 3 wks + placebo or selumetinib)	75 mg po b.i.d. (hydrogen sulfate)	Placebo	32 versus 14%, p = 0.059	4.23 versus 3.93 months (HR 0.75, 90% CI 0.50 – 1.14)	9.5 versus 11.37 months (HR 1.15, 90% CI 0.71 – 1.84)
Catalanotti et al. (2013) [59]	Melanoma (BRAF mutant, stratified by pAKT)	N = 15 (10 high pAKT, 5 low pAKT)	Single-arm	75 mg po b.i.d. (hydrogen sulfate)	None	High pAKT: 0/10 (0%); low pAKT: 1/5 (20%)	2.2 months (high pAKT) versus 7.1 months (low pAKT)	8 months (high pAKT) versus 18 months (low pAKT)
Robert et al. (2013) [61]	Melanoma (BRAF mutant)	N = 91 (45 selumetinib + dacarbazine, 46 placebo + dacarbazine)	Randomized, parallel-group (dacarbazine 1000 mg/m^2 IV every 3 wks + placebo or selumetinib	75 mg po b.i.d. (hydrogen sulfate)	Placebo	18/45 (40%) versus 12/46 (26%), one-sided p = 0.081	5.6 versus 3.0 months (HR 0.63, 80% CI 0.47 – 0.84)	13.9 versus 10.5 months (HR 0.93, 80% CI 0.67 – 1.28)

bid: Twice daily; C: Comparator; CI: Confidence interval; HR: Hazard ratio; IV: Intravenous; mg: Milligrams; N/A: Not applicable; ORR: Overall response rate; OS: Overall survival; pAKT: Phosphorylated AKT; po: Orally; PFS: Progression-free survival; S: Selumetinib; tx: Treatment.

For patients with BRAF wild-type melanoma, it was hypothesized that MEK inhibition by selumetinib might still be effective as selumetinib has been shown to suppress pERK levels independent of BRAF and NRAS mutation status [14]. Given synergy seen in xenograft models with the combination of selumetinib and docetaxel [14], a randomized Phase II trial was performed in which patients with BRAF wild-type advanced melanoma were treated with docetaxel 75 mg/m² IV every 3 weeks plus either selumetinib 75 mg or placebo orally b.i.d. (Table 3) [58]. A total of 83 patients were randomized in a 1:1 fashion, with the primary end point of PFS not significantly improved with selumetinib (4.23 vs 3.93 months, HR 0.75, 90% CI, 0.50 – 1.14, p = 0.130). Similarly, OS was not improved with selumetinib (9.5 vs 11.37 months, HR 1.15, 90% CI, 0.71 – 1.84, p 0.318), though there was a trend toward improvement in overall response rate with selumetinib (32 vs 14%, p = 0.059).

A Phase II study by Catalanotti et al. was conducted in BRAF mutant melanoma patients stratified by pAKT expression (high vs low), with all patients receiving selumetinib 75 mg orally b.i.d. (Table 3) [59]. The rationale for this stratification was the hypothesis that BRAF mutant melanomas with low PI3K/AKT activation would be most sensitive to MEK inhibition [60]. The high pAKT cohort was closed after no responses were seen in the first 10 patients, and the low pAKT was eventually closed to poor accrual, given the low incidence of this tumor characteristic. However, tumor regression was seen in three of the low pAKT, BRAF mutant melanoma patients, supporting the trial’s hypothesis and suggesting that perhaps there may be a role for concomitant MEK and PI3K/AKT inhibition in BRAF mutant melanoma with high pAKT levels.

A randomized Phase II study of dacarbazine 1000 mg/m² every 3 weeks plus either selumetinib 75 mg orally b.i.d. or placebo in patients with BRAF mutant melanoma was also done to investigate whether the addition of selumetinib could improve outcomes in patients treated with dacarbazine, an FDA-approved treatment, in the first-line setting (Table 3) [61]. Xenograft models had previously shown significant tumor regression with selumetinib and temozolomide, which has the same active metabolite as dacarbazine [25]. Interestingly, PFS was significantly improved in the dacarbazine plus selumetinib arm when compared to dacarbazine plus placebo (5.6 vs 3 months, HR 0.63, 80% CI, 0.47 – 0.84, one-sided p = 0.021). OS was not improved, however, with the addition of selumetinib (13.9 vs 10.5 months, HR 0.93, 80% CI, 0.67 – 1.28, one-sided p = 0.39).

Given that over 80% of uveal melanomas have been found to harbor GNAQ or GNA11 mutations, genes that encode for widely expressed G-protein α-subunits that when mutated were hypothesized to activate the MAPK pathway, this melanoma subset was also hypothesized to potentially benefit from MEK inhibition by selumetinib [7,17,62-65]. A randomized Phase II clinical trial of 101 patients treated with selumetinib 75 mg orally b.i.d. or chemotherapy (temozolomide 150 mg/m² orally daily for 5 of every 28 days, or dacarbazine 1000 mg/m² IV every 21 days) was undertaken to test this hypothesis (Table 3) [66]. PFS was significantly improved with selumetinib over chemotherapy (15.9 vs 7 weeks, HR 0.46, 95% CI, 0.30 – 0.71, p < 0.001), as was response rate (14 vs 0%). Interestingly, however, PFS was greater in the GNAQ and GNA11 exon 5 wild-type population in this study, perhaps explained by the presence of other mechanisms of MAPK activation in these tumors, such as exon 4 mutations found in retrospect. Overall, however, OS was not found to be improved with selumetinib over chemotherapy (11.8 vs 9.1 months, HR 0.66, 95% CI, 0.41 – 1.06, p = 0.09), perhaps in part due to crossover that allowed the patients in the chemotherapy arm to subsequently receive selumetinib upon disease progression.

Despite interest in MEK1/2 inhibition of melanoma with selumetinib in the trials discussed above, other MEK inhibitors have also been tested for superior efficacy. Trametinib, a more potent MEK inhibitor and the only MEK inhibitor currently FDA-approved for the treatment of melanoma, has demonstrated a 22% response rate in patients with BRAF V600E/K mutant melanoma, and has shown superiority in PFS and OS when compared with dacarbazine or paclitaxel [67]. Other newer generation MEK inhibitors such as cobimetinib are in development and hold promise for increased potency, particularly when combined with BRAF inhibitors [68]. It is possible that in addition to more potent MEK inhibition, further clinical success with other MEK inhibitors may rely on the development of better biomarkers that predict for MEK dependence, and rational combinatorial therapy that prevents compensatory upregulation of the MAPK pathway that develops through resistance mechanisms.

8. Selumetinib for the treatment of other tumors: acute myeloid leukemia and ovarian carcinoma

Similar to many solid tumors, the RAS/RAF/MEK/ERK signaling pathway is prominently dysregulated in acute myeloid leukemia (AML), with FLT3, c-KIT, KRAS and NRAS mutations seen in 2 – 37% of patients [69], and ERK activation in 70 – 80% of AML cell lines and primary AML samples [69]. Furthermore, MEK inhibition in vitro has shown growth arrest of primary AML cells [70]. On the basis of these data, a Phase II study by Jain et al. investigated the efficacy of single-agent selumetinib in 47 patients with relapsed/refractory AML or 60 years old or more with untreated AML, and patients were stratified by FLT3 ITD status (Table 2) [71]. In the FLT3 wild-type cohort, 6 of 36 (17%) patients had a response (1 partial response, 3 minor responses, 2 unconfirmed minor responses), but surprisingly, no patient with FLT3 ITD mutation responded. NRAS and KRAS mutations were detected in 7 and 2% of patients, respectively. Baseline pERK activation was seen in 85% of patients analyzed but did not correlate with response, similar to the findings by Bekaii-Saab et al. in biliary cancer [20]. Perhaps most intriguing, a single-nucleotide polymorphism (SNP) rs3733542 in exon 18 of the KIT gene was seen in significantly higher number of patients with response or stable disease when compared with nonresponders (60 vs 23%, p = 0.027). These results underscore the need for better biomarker discovery predictive of response to MEK inhibition.

In a final example, single-agent selumetinib 50 mg orally b.i.d. was also tested in a single-arm Phase II study of woman with recurrent low-grade serous ovarian or peritoneal carcinoma early on in selumetinib development (Table 2) [72]. This study was based on data that demonstrated a higher frequency of KRAS and BRAF mutations in low-grade than high-grade serous carcinomas [73,74]. Of a total of 52 enrolled patients, 8 (15%) had an objective response to treatment, with 1 complete response and 7 partial responses. Furthermore, 34 (65%) had stable disease while on study. Median PFS for this study was 11.0 months, with 63% of patients having a PFS of longer than 6 months. Surprisingly, however, response to selumetinib did not seem to be related to KRAS or BRAF mutational status, and in this tumor type as well as others, better biomarkers for response to MEK inhibition are needed.

9. Conclusions

Selumetinib is an orally available MEK1/2 inhibitor developed to target activated ERK1/2, thereby inhibiting cellular proliferation and survival in various malignancies. As many tumor types have demonstrated marked activation of the Raf-MEK–ERK signaling pathway, an inhibitor that can potently and effectively target this molecular aberrancy is of great demand. Strong preclinical rationale for MEK inhibition via selumetinib in many tumor types has led to multiple Phase I and II clinical trials, some with encouraging signals of efficacy. However, these efficacy signals have not always been predicted accurately according to KRAS or BRAF mutational status, for example. Better understanding of selumetinib and de novo or acquired resistance mechanisms are needed to develop predictive biomarkers of response, thereby allowing clinicians to optimize selection of this therapy for appropriate patients. Furthermore, increasing knowledge of the MAPK pathway and selumetinib’s effect on this signaling pathway will enable the development of better therapy combinations with selumetinib and other targeted or cytotoxic agents.

10. Expert opinion

Due to the prevalence of activated Raf-MEK–ERK signaling in human malignancies, inhibition of this pathway is a widely sought after target. Given the lack of potent and selective Ras, Raf and ERK inhibitors for most tumors, the development of MEK inhibitors has been an obvious and exciting addition to the realm of targeted therapies in oncology. The first- and second-generation MEK inhibitors CI-1040 and PD 0325901, respectively, were disappointing in early-phase clinical trials, leading to abandonment of their further development. In contrast, selumetinib, a non-ATP-competitive, selective MEK1/2 inhibitor, demonstrated increased potency and efficacy in preclinical models and therefore was subsequently tested in multiple Phase I and II clinical trials. Phase I testing of selumetinib explored both a free-base solution (100 mg orally b.i.d. dosing) and an oral capsule formulation incorporating a Hyd-Sulfate salt (75 mg orally b.i.d. dosing). A manageable toxicity profile, demonstrated inhibition of ERK phosphorylation, and signals of preliminary efficacy seen in patients with RAS or RAF mutations treated with selumetinib in these Phase I trials led to continued development of the drug in multiple tumor types.

Selumetinib has been tested in multiple gastrointestinal tumor types in the Phase II setting, including biliary cancer, HCC, colorectal carcinoma and pancreatic cancer. While some objective responses were seen in biliary cancer patients, RAS and RAF mutational status were less prevalent than previously thought in these cancers, and modulation of phosphorylated ERK with selumetinib treatment did not correlate with survival. In addition, clinical efficacy of selumetinib in other gastrointestinal tumor types has been disappointing. However, hypothesis-generating data showing potential positive effects on muscle mass in cholangiocarcinoma patients treated with selumetinib and possible sequence dependence of selumetinib in combination with gemcitabine have generated interest in further study of this agent in gastrointestinal malignancies, and in particular, pancreaticobiliary cancers.

While a traditional Phase II trial of selumetinib in advanced thyroid cancer did not meet its primary end point, subsequent work suggested a novel mechanism of selumetinib in the treatment of radioiodine-refractory thyroid cancer. Pilot studies suggest that selumetinib can potentially reverse radioiodine resistance through induction of iodine uptake and retention, a possible mechanism that deserves further exploration.

Selumetinib in the treatment of NSCLC has been intriguing, though the early monotherapy study may have been negatively impacted by lack of patient selection by appropriate RAS and RAF mutational status. In combination with docetaxel in KRAS mutant NSCLC, selumetinib showed significant improvements in ORR and PFS and a trend toward improved OS, albeit in this small trial an increased toxicity burden as well.

MEK inhibitors as a class have been very effective in the treatment of melanomas, and while newer generation inhibitors such as trametinib have gained FDA approval for this indication, treatment with selumetinib has been less promising, either as monotherapy or in combination with chemotherapy. Most trials of selumetinib in this tumor type have focused on BRAF mutant cancers, and it is possible that inadequate patient selection may play a part in marginal PFS benefits and lack of OS benefit. However, there may be a role for combinatorial therapy with PI3K/AKT and MEK inhibitors in patients with BRAF mutant melanoma with high pAKT levels, causing investigators to surmise that negative trials of selumetinib may be due to poor patient selection but could be improved with the development of better predictive biomarkers of response to selumetinib.

While an intriguing agent based on its mechanism of inhibition of the MEK signaling pathway, the efficacy of selumetinib in the treatment of multiple tumor types with strong preclinical rationale has been somewhat disappointing. We surmise that this is a potentially efficacious drug that will require better understanding of underlying biology to maximally exploit its potential benefits. This includes a need for better predictive biomarkers of response for each tumor type, which may soon become available with advancing genomic technology and next-generation sequencing efforts. Furthermore, like prior examples of targeted therapies used as monotherapy with resultant limited efficacy, de novo or acquired resistance mechanisms will likely necessitate more rational combinatorial therapies, holding promise for the class of MEK inhibitors in the future. Optimization of efficacy of agents such as selumetinib will depend on both an enhanced biological understanding of the complex signaling pathways inherent in cancer and improved knowledge of ways to better exploit these aberrant pathways.

Declaration of interest

T Bekaii-Saab is a paid consultant for AstraZeneca. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.