Clinical Challenges in Targeting Anaplastic Lymphoma Kinase in Advanced Non-small Cell Lung Cancer

Abstract

The revolution in individualized therapy for patients with advanced NSCLC has seen the emergence of a number of molecularly targeted therapies for distinct patient molecular subgroups. Activating anaplastic lymphoma kinase (ALK)-gene rearrangement has been detected in 3%–7% of NSCLC cases, and the ALK inhibitor crizotinib is now an approved treatment for patients with tumors harboring this event. However, resistance to ALK-targeted therapies is a ubiquitous problem in the management of advanced ALK-positive NSCLC, and can be mediated by secondary kinase mutations or the activation of compensatory alternative oncogenic drivers. New, more potent ALK inhibitors such as ceritinib (LDK378), alectinib (CH5424802), and AP26113, are now emerging, together with an increased knowledge of the molecular basis of resistance. There is therefore a need to evaluate the optimal clinical application of these new agents, either as sequential therapies, and/or in combination with other targeted agents, to combat resistance and prolong survival in patients with ALK-positive NSCLC. The remarkable clinical activity of ALK inhibitors also emphasizes the importance of optimal diagnostic testing algorithms, to ensure that all eligible patients receive these breakthrough therapies.

Introduction

Lung cancer is the leading cause of cancer-related deaths worldwide, with a dismal 5-year overall survival of approximately 18% [1]. Non-small cell lung cancer (NSCLC) accounts for 85% of cases, and >70% are diagnosed with advanced disease [2]. The 21^st century has witnessed a revolution in treatment for advanced NSCLC from a one-size-fits-all to a personalized approach. Traditionally, histologic subtypes have dictated the choice of chemotherapy, but now key oncogenic driver mutations are known, and there are increasing data regarding genetic alterations allowing adenocarcinomas to be further classified into clinically relevant molecular subtypes which predict response to novel agents. The first clinically relevant molecular alterations to be characterized were epidermal growth factor receptor (EGFR) mutations that respond to tyrosine kinase inhibitors (TKIs). Recently, the discovery of translocations, involving the anaplastic lymphoma kinase (ALK) gene has driven the investigation of novel treatment options for the 3%–7% of patients with NSCLC whose tumors harbor this event [3]. Crizotinib, the first-in-class small molecule ALK inhibitor, gained US Food and Drug Administration (FDA) approval for the treatment of ALK-rearranged (ALK-positive) NSCLC in 2011. In this review, we discuss the discovery of and testing for ALK rearrangements in NSCLC, and review data on crizotinib, upcoming agents, and trials for this patient population.

Discovery of ALK fusion genes

Oncogenic addiction is the phenomenon whereby tumor cells depend on an oncogene for survival and proliferation, making them attractive therapeutic targets [4]. The rearrangement of the ALK gene represents such a dependency in NSCLC. First reported as a fusion gene in a small proportion of anaplastic large cell lymphomas (ALCLs) [5], ALK rearrangement was subsequently discovered in NSCLC [6, 7], mostly in adenocarcinomas [8-11]. The most common alteration involves inversion on chromosome 2, leading to fusion of the protein encoded by the echinoderm microtubule–associated protein-like 4 (EML4) gene with the intracellular portion of the receptor tyrosine kinase encoded by the ALK gene [6, 12]. This EML4–ALK fusion protein constitutively activates a number of signaling cascades (Figure 1) [13]. These pathways promote initiation, progression, and survival of NSCLC [13, 14]. Four other ALK fusion proteins are also associated with NSCLC [15-18]. Several studies suggest that ALK rearrangements are largely independent of EGFR and KRAS mutations [19-24].

Figure 1.

Signaling cascades activated by the EML4-ALK fusion protein.

Optimal screening strategy

A subset of patients with NSCLC may possess clinicopathologic features that predict ALK-positivity. Most studies quoting a higher incidence of ALK-positivity involved patients who were light/never smokers (chance of carrying mutation 20% vs. 2% in smokers), were younger in age (median age 54 vs. 64 years for the ALK-negative [ALK wild-type] population), had acinar/signet ring histology, and had transcription termination factor 1 (TTF-1)-positive histology [8, 10, 25-29]. Furthermore, if we understand EGFR and ALK alterations to be mutually exclusive, the presence of an activating EGFR mutation or response to EGFR TKIs may predict for ALK-negative status[30]. However, ALK rearrangements are not entirely restricted to non-smokers or certain age groups. In the absence of strong data suggesting predictive factors, current National Comprehensive Cancer Network (NCCN) guidelines suggest screening all patients with advanced non-squamous NSCLC and patients with squamous disease if they are never smokers or were diagnosed based on small biopsy specimens [31]. The remarkable clinical activity of ALK inhibitors emphasizes the importance of testing for these mutations and ensuring that eligible patients receive appropriate targeted therapy.

Optimal testing modality

The discovery of ALK rearrangement and its potential as a therapeutic target triggered the co-development of diagnostic assays. The current FDA-approved break-apart fluorescence in situ hybridization (FISH) assay (AbbVie, Inc.), was clinically validated in Phase I/II trials involving crizotinib [10, 32]. The cut-off point for a positive result is >15% of tumor cells positive in ≥50 cell nuclei [10, 21]. The test can be performed on formalin-fixed paraffin-embedded specimens, and detects novel ALK fusion genes by targeting the tyrosine kinase domain of ALK, independent of the fusion partner [33]. Disadvantages include the need for specialized expertise to both perform the test and interpret the results, a risk of false negatives due to subtle splitting of colored signals, and associated costs [30, 33].

Other screening methods being evaluated include reverse transcription polymerase chain reaction (RT-PCR) assays and immunohistochemistry (IHC). Sanders et al. used multiplexed RT-PCR to detect 5 known EML–ALK variants, identified in 9% of specimens [34]. RT-PCR is highly sensitive and specific but requires high-quality RNA (unobtainable from many archived samples), and only detects known fusion variants, with the consequent potential of false negative results (in the setting of novel fusion genes), and lacks clinical validation.

ALK-directed IHC is an attractive alternative to FISH and may soon become an established diagnostic algorithm. IHC is quick, affordable, can be performed on a variety of tumor specimens, and also facilitates histologic comparison. Currently, the low degree of ALK expression in NSCLC makes the use of this technique challenging. More sensitive techniques using ALK monoclonal antibodies are being investigated. Yi et al. correlated IHC with FISH using the ALK1 antibody, and found >90% sensitivity and specificity when 2/3+ scores were considered IHC positive [35]. However, the poor transcriptional activity of EML–ALK in NSCLC leads to low staining intensity, and may impact the reliability of this assay [17, 36]. Results obtained using 5A4 and D5F3 antibodies have been more encouraging, and studies have suggested a sensitivity and specificity of 95%-100% for IHC using the 5A4 antibody [37-39]. One study supported a scoring algorithm in which ALK IHC scores of 0, 1, and 3+ were highly compatible with FISH results, while a score of 2+ was variable [37]. In another study, correlation between an IHC score of 0 with negative FISH status and between an IHC score of 1+ with positive FISH status were observed [39]. Both of these studies propose a two-tier system for evaluating ALK with an initial IHC screening followed by FISH assay for IHC 1+ and/or 2+ specimens. The D5F3 monoclonal antibody provided a sensitivity and specificity of 100% and 99%, respectively, using ALK FISH as gold standard [36]. The lack of clinical validation of these techniques prevents their widespread approval; nonetheless, an automated IHC companion diagnostic ALK assay (Ventana Medical Systems, Inc.) has now been launched in Europe [40]. The break-apart FISH assay remains the US standard and recommended testing strategy.

Treatment of ALK-positive lung cancer

The majority of ongoing trials involving ALK-positive patients are in the metastatic/advanced setting and this is therefore the focus of this review. Future trials will be needed in order to evaluate ALK inhibitors and other novel agents for early stage lung cancer treatment.

Chemotherapy

Retrospective analyses indicate that ALK status does not predict chemotherapy response [8, 27, 29]. Patients with ALK-positive NSCLC do not benefit from EGFR TKI therapy [29, 41, 42], and this decreased responsiveness highlights the mutual exclusivity of ALK rearrangements and EGFR mutations. Gandara et al. evaluated the expression of thymidylate synthase (TS) in 63 patients with ALK-positive lung adenocarcinoma and 1698 patients with ALK-negative disease. TS gene expression was low in ALK-positive tumors compared with ALK-negative tumors, supporting a rationale for pemetrexed therapy for ALK-positive NSCLC [43]. Retrospective analyses have evaluated the differential activity of pemetrexed in patients with ALK-positive NSCLC [44, 45]. In one study, multivariate analysis, adjusting for age, sex,smoking status, histology, line and type of therapy, ALK-positivity was associated with prolonged PFS on pemetrexed (hazard ratio 0.36) [45]. In contrast, a large multicenter, retrospective analysis did not support these findings, and the median PFS of ALK-positive patients treated with single-agent pemetrexed or non-platinum/pemetrexed combination was similar to that of ALK-negative patients. In the same series, among patients undergoing first-line platinum/pemetrexed therapy, the median PFS in patients with ALK-positive tumors was 7.3 months compared with 5.4 months for wild-type tumors. However, patients who were never/light smokers had a similar PFS to the ALK-positive group [46].The retrospective nature of the analysis, and the finding of improved sensitivity to chemotherapy among non-smoking patients [47], make interpretation of the data difficult.

ALK-targeted therapy

Preclinical studies have shown that ALK fusion gene products are oncogenic drivers of transformation, and ALK has therefore been extensively explored as a therapeutic target. Clinical investigation of crizotinib began as a c-Met inhibitor in patients with various malignancies. The subsequent discovery of ALK gene rearrangement in NSCLC, and promising results in patients with NSCLC, led to the addition of an expansion cohort to include this population, in which a response rate of 61% was seen, with a median PFS of 9.7 months [10, 48]. The singe-arm Phase II study PROFILE 1005 showed a comparable overall response rate (ORR) of 60% and median PFS of 8.1 months [32].A Phase III trial, PROFILE 1007, comparing crizotinib with standard chemotherapy in the second-line setting resulted in an improved ORR (65% vs. 20%), a shorter time to response (6.3 vs. 12.6 weeks), and an improved median PFS (7.7 vs. 3.0 months) with crizotinib. Overall survival benefit was not demonstrated on interim analysis, and this was likely related to crossover (64% of patients on chemotherapy crossed over to crizotinib after progression) [49]. In all these studies, toxicities were acceptable, with some visual disturbances, gastrointestinal side effects, fatigue, and edema. The PROFILE 1014 study is designed to answer the question of the superiority of crizotinib over front-line platinum/pemetrexed combination chemotherapy (NCT01154140). Despite the absence of mature randomized data, the NCCN panel recommends crizotinib in a front-line setting in advanced ALK-positive NSCLC [31]. However, in clinical practice, systemic chemotherapy may be started before genotyping results are available. Berge et al. reported that PFS benefit from crizotinib appears higher than with pemetrexed in patients with advanced ALK-positive NSCLC [50]. Pemetrexed exposure did not affect outcome with crizotinib; however, PFS benefit from pemetrexed was less after crizotinib use (4.5 months) compared with before crizotinib use (6 months).

Progression on crizotinib and emergence of resistance

Unfortunately, about 40% of ALK-positive patients show intrinsic resistance to crizotinib [10, 32, 48], while others derive dramatic responses initially but develop resistance within 12 months [51, 52]. Proposed mechanisms of acquired resistance include target gene alteration/amplification (≈30% of cases) and up-regulation of alternate cell-signaling pathways. One study describes four different mutations in the ALK tyrosine kinase (ALK-TK) domain that confer various degrees of resistance to crizotinib [51]. Most common is the L1196M amino acid substitution, similar to the gatekeeper mutations observed in EGFR (T790M) and BCR–ABL genes (T315I), originally identified as an independent mutation in a tumor from a patient with ALK-positive NSCLC [53]. Some other mutations, such as G1202R and S1206Y, are located close to the crizotinib-binding site on the ALK-TK domain and decrease the affinity of crizotinib for ALK, while the 1151T insertion may affect the affinity of ALK for ATP, conferring strong crizotinib resistance [51]. Additional studies have also identified novel mutations in the ALK-TK domain that predict for crizotinib resistance [52, 54]. Other mechanisms implicated in resistance include target gene amplification, with increase in ALK gene rearrangement copy numbers without a documented mutation [52] and up-regulation of alternate pathways including EGFR activation, and c-KIT amplification [51]. A recent study also suggests EGF-mediated HER family activation as a mechanism of ALK-TKI resistance [55]. There may be diverse and multiple mechanisms involved in resistance even within an individual patient, and these factors have emerged as major roadblocks in the transformative clinical impact of the ALK inhibitors.

Therapeutic advances in the setting of resistance

The identification of resistance mechanisms provides groundwork for the development of new ALK inhibitors to combat crizotinib resistance, including the development of combination therapies to attack bypass track pathways.

Novel ALK inhibitors

Next-generation ALK inhibitors currently under clinical evaluation include ceritinib (LDK378; Novartis), AP26113 (ARIAD), alectinib (CH5424802/RO5424802; Chugai/Hoffmann-La Roche), and ASP3026 (Astellas). Other agents are in earlier stages of development (Table 1).

Table 1. Anaplastic lymphoma kinase inhibitors in clinical and preclinical development.

Agent [references]	Targets (IC50)	Clinical trial phase/status	Common AEs	Response rate (%)	Median PFS (months)	Other targets
Crizotinib (Pfizer) [10, 32, 80]	WT-ALK (24 nM)	Approved	Visual disorders (59%), nausea (57%), vomiting (44%)	60 (ALK+ NSCLC [n=259])	8.1	c-MET, ROS-1
Ceritinib (LDK378; Novartis)[57, 58, 80]	WT-ALK (0.15 nM) G1269S, F1245C	Phase II/III	Nausea (82%), diarrhea (75%), vomiting (65%)	58 (ALK+ NSCLC [n=114]) 62 (crizotinib-naïve ALK+ NSCLC [n=34]) 56 (crizotinib-refractory ALK+ NSCLC [n=80])	10.4a 6.9b	IGF-1R, ROS1
Alectinib (CH5424802/RO54 24802; Chugai/Roche) [61, 62]	WT-ALK (1.9 nM) L1196M, F1174L	Phase I/II	Dysgeusia (30%), AST increased (28%), blood bilirubin increased (24%)	93 (crizotinib-naïve ALK+ NSCLC [Japan; n=46]) 59 (crizotinib-refractory ALK+ NSCLC [n=45])	–	GAK, LTK
AP26113 (ARIAD) [66, 80]	WT-ALK (0.62 nM) L1196M, F1171T	Phase I/II	Nausea (43%), fatigue (41%), diarrhea (35%)	62 (ALK+ tumors [n=21]) 71 (crizotinib-refractory ALK+ NSCLC [n=14])	–	EGFR, ROS1
TSR-011 (Tesaro) [81, 82]	WT-ALK (0.7 nM) L1196M, R1275Q	Phase I/IIa	QTc prolongation (20%; DLT), fatigue/asthenia (10%)	67 (ALK+ NSCLC [n=3])	–	TRKA, TRKB, TRKC
ASP3026 (Astellas) [80]	WT-ALK L1196M	Phase I	–	–	–	ROS1
X-396 (Xcovery) [83]	WT-ALK (<0.4 nM) L1196M, C1156Y	Phase I	–	–	–	–
EP-28122 (Teva/Cephalon) [80]	WT-ALK (1.9 nM) F1174L, R1275Q	Preclinical	–	–	–	–
AZD3463 (AstraZeneca) [84]	WT-ALK (Ki 0.75 nM) L1196M, T115Ins	Preclinical	–	–	–	IGF-1R

^aCrizotinib-naive patients;

^bCrizotinib-refractory patients.

AE, adverse event; ALK, anaplastic lymphoma kinase; ALK+, ALK-positive; AST, aspartate transaminase; DLT, dose-limiting toxicity; EGFR, epidermal growth factor receptor; GAK, cyclin G-associated kinase; IGF-1R, insulin-like growth factor-1 receptor; LTK, leukocyte receptor tyrosine kinase; NSCLC, non-small cell lung cancer; PFS, progression-free survival; ROS1, c-ros oncogene 1; TRK, tropomycin receptor kinase; WT-ALK, wild-type ALK.

Ceritinib is an oral ALK inhibitor with 20-fold greater preclinical potency than crizotinib, and activity against crizotinib-resistant mutations [56]. Ceritinib shows marked antitumor activity against both crizotinib-sensitive and crizotinib-resistant ALK-rearranged xenograft tumors [57]. An ongoing Phase I trial includes 130 patients with advanced cancers harboring genetic alterations in ALK [58]. Preliminary results have shown that in 114 patients with ALK-positive NSCLC treated with ceritinib ≥400 mg/day, the ORR and median PFS were 58% and 7.0 months, respectively. Significant clinical benefit was noted even in the crizotinib-pretreated group (n=80), including an ORR of 56%. Ceritinib was tolerated up to the maximum tolerated dose of 750 mg/day with primarily gastrointestinal side effects such as nausea, diarrhea, and vomiting (Table 1). Based on the encouraging results observed with this agent, the FDA granted it Breakthrough Therapy designation for the treatment of patients with ALK-positive metastatic NSCLC who have progressed on or are intolerant to crizotinib [59]. This status is intended to help expedite the drug's development and review, with Phase II trials currently underway [60]. Two Phase III trials comparing ceritinib with single-agent chemotherapy after progression on a platinum-based doublet and crizotinib (NCT01828112), and with a pemetrexed-platinum doublet in a first-line setting (NCT01828099) are currently recruiting patients.

Alectinib (CH5424802/RO5424802) is a potent ALK inhibitor that also targets the ALK L1196M gatekeeper mutation in vitro [61]. In a Phase II study in 46 Japanese patients with ALK-positive, crizotinib-naïve NSCLC, the objective response rate was 93.5%, including 2 (4%) complete responses and 41 (89%) partial responses; 40 of 46 patients continued to be on trial at the time of data reporting [62]. Adverse events included dysgeusia and increased aspartate aminotransferase (Table 1): visual disorders were rare and gastrointestinal toxicities were mild. Preliminary data from a Phase I study (n=45) in crizotinib-refractory patients indicated an ORR of 59% [63].

AP26113 is a dual ALK/EGFR inhibitor that also overcomes crizotinib resistance mediated by L1196M and other mutations in preclinical models [64, 65]. In a Phase I/II study in patients with advanced malignancies, preliminary responses have been reported in 13/21 (62%) patients with ALK-positive NSCLC, including responses in both crizotinib-naive and crizotinib-pretreated patients [66]. Phase II expansion cohorts will enroll both crizotinib-naïve and crizotinib-resistant patients [67].

These data indicate that new ALK inhibitors improve responses in patients who have progressed on crizotinib. For secondary mutations, knowledge of the precise resistance-inducing mutation may be important in selecting future salvage therapies since some crizotinib-resistance mutations have been found to show cross-resistance to other ALK inhibitors [68].

Alternative targets and combination therapies

With regards to alternative signaling pathways, ALK regulates downstream signaling such as the RAF/MEK/ERK and PI3K/AKT/mTOR pathways [13]. Combining targeted therapy against these pathways may help overcome crizotinib resistance; for example, combining an ALK inhibitor with a MEK, mTOR or EGFR inhibitor upfront may be explored.

Heat-shock protein 90 (Hsp90) is a molecular chaperone that facilitates correct folding and maturation of oncogenic client proteins, including ALK [69].The Hsp90 inhibitor ganetespib (Synta), exhibits single-agent activity against ALK-positive tumors in preclinical and clinical studies, with activity in resistant cells [54, 70]. In a Phase II study, the Hsp90 inhibitor retaspimycin (Infinity) demonstrated clinical activity in three heavily pretreated patients with ALK-positive NSCLC, two of whom had partial responses and the third had prolonged stable disease (7.2 months) [71]. Thus, Hsp90 inhibitors may represent an alternative strategy to overcome crizotinib resistance. Differential sensitivity of ALK-fusion variants to ALK inhibitors correlates with fusion protein stability [72], and combining Hsp90 and ALK inhibitors has provided synergistic cytotoxicity.[72] Clinical studies are underway involving AT13387 (Astex) plus crizotinib (NCT01712217) and AUY922 (Novartis) plus ceritinib (NCT01772797).

Although ALK-positive tumors are unresponsive to EGFR inhibitors, activation of a secondary pathway such as EGFR is a recognized resistance mechanism. Therefore, EGFR TKIs may improve sensitivity to crizotinib in combination regimens by targeting signaling pathways that contribute to resistance [51].

Clinical challenges in ALK inhibitor therapy – central nervous system (CNS) metastasis

Although some current data does not support an inherent propensity of ALK-positive NSCLC for CNS spread, [73] it has been encountered in the setting of progression after crizotinib therapy, and likely related to poor cerebrospinal fluid (CSF) penetration of the drug despite good systemic control [74]. In one reported case of CNS metastasis, the CSF concentration of crizotinib was 0.62 ng/mL, compared with a serum concentration of 237 ng/mL [75]. A retrospective study evaluating ALK inhibition after therapy for oligoprogressive NSCLC showed CNS to be the first site of progression in 46% of patients with ALK-positive disease [76]. In such patients, continuation of crizotinib after local therapy provided ongoing benefit. Otterson et al. showed that patients may be able to continue with crizotinib for a period of time following clinically documented progression [77]. Another retrospective analysis found that 30% of patients with ALK-positive NSCLC with isolated CNS failure on crizotinib were able to resume therapy after completion of radiotherapy, and continued to receive crizotinib for ≥4 more months without disease progression [78]. Thus, continuing ALK inhibitor therapy in such patients may be a valid option. Notably, favorable effects on brain metastases have been reported for alectinib [62, 63], ceritinib,[58] and AP26113 [66]. Combined high-dose pemetrexed and crizotinib also showed activity in an isolated case with miliary CNS metastases, suggesting that the synergistic effect of this combination may be beneficial in treating patients with ALK-positive NSCLC and brain metastases [79].

Conclusions

The emergence of targeted treatment options for ALK-positive NSCLC has revolutionized the care of patients with this disease. However, resistance to approved treatment often develops, and more research is required to further understand the molecular events associated with ALK-positive NSCLC as well as mechanisms of resistance. Future work will not only focus on optimal diagnosis and treatment at earlier stages of disease, but also on rational combinations of effective agents and the ideal sequence of therapy, particularly as more next-generation agents obtain regulatory approval. In addition, optimal supportive care and toxicity management is essential for patients who may hopefully live longer on sequential treatment.