Ceritinib for the Treatment of Late-Stage (Metastatic) Non–Small Cell Lung Cancer

Since the original report by Soda and colleagues that the anaplastic lymphoma kinase (ALK) gene fuses with the echinoderm microtubule-associated protein like 4 (EML4), resulting in potent transforming activity in non–small cell lung cancer (NSCLC) limited to the adenocarcinoma histologic subtype, it is well recognized that between 3% and 7% of NSCLC is molecularly defined by the presence of an inversion or translocation of chromosome 2p involving the ALK gene, resulting in a transforming fusion gene and representing a distinct molecular subset of disease (1, 2). Generally, all the ALK rearrangements identified so far are constituted by two portions. The first is the highly conserved break point within ALK, located in the intron immediately upstream of the exons encoding the kinase domain; the second are the 5′-end partners containing a coiled-coil or leucine zipper domain responsible for oligomerization of fusion protein and ligand-independent activation of the ALK tyrosine kinase (TK) activity. Constitutive activation of downstream signaling pathways, such as the Ras/MAPK, PI3K/AKT, and JAK/STAT, results in uncontrolled cancer cell proliferation and survival. Thirteen variants of the EML4–ALK fusion have been described, according to the break point on EML4 (from exon 2 to exon 20; ref. 3). Furthermore, in addition to EML4, the TRK-fused gene (TFG; refs. 4–6) and the kinesin family member 5B (KIF5B; refs. 4–6) have been described to be fused to ALK in rare cases (Table 1).

Mice transduced with NIH3T3 cells forced to express the EML4–ALK fusion gene can be successfully treated with ALK inhibitors (ALKi). Apart from the EML4–ALK fusion gene, 11 other variants have been identified but it is still unclear whether these result in differential susceptibility to ALKi (3). In addition, chromosomal rearrangements involving the gene encoding ROS1 proto-oncogene receptor tyrosine kinase (ROS1) define a distinct molecular subgroup of NSCLCs that have shown sensitivity to ALKi, in particular crizotinib (ref. 7; Table 1).

The ALK tyrosine kinase inhibitor (ALK-TKI) crizotinib, which is a small-molecule potent inhibitor targeting cMET, ALK, and ROS1 tyrosine kinases, was the first in class shown to be effective for ALK-positive (ALK⁺) patients with advanced NSCLC demonstrating response rates of about 60% in a single-arm study, which led to accelerated FDA approval in October 2011 (8–10). This study was followed by a randomized trial in which crizotinib demonstrated superiority over standard chemotherapy in patients previously treated with platinum-based chemotherapy. ALK inhibitors can inhibit ROS1 kinase activity in cell lines, and crizotinib is also associated with clinically significant antitumor activity in patients with ROS1 NSCLC. A study investigating the use of crizotinib in ROS1 rearranged cancer demonstrated an overall response rate (ORR) of 72% [95% confidence intervals (CI), 58–84], with 33 partial responses (PR) and three complete responses (CR). The median duration of response (DOR) was 17.6 months (95% CI, 14.4 to not reached), with 25 patients (50%) still in follow-up for progression at the time of this report. The pharmacokinetics, antitumor activity, and safety profile of crizotinib in this group of patients were similar to those observed in patients with ALK⁺ NSCLC (7).

Despite dramatic initial activity of crizotinib in ALK⁺ NSCLC, invariably crizotinib resistance develops typically within 1 to 2 years from the beginning of treatment. The central nervous system (CNS) is a particularly common site of progressive disease in crizotinib-treated patients, suggesting the need for ALKi that not only can overcome acquired crizotinib resistance, but also penetrate the blood–brain barrier (11, 12). This review focuses on the use of ceritinib for the treatment of patients with ALK ⁺, metastatic lung adenocarcinoma.

Among the several highly potent next-generation ALK-TKIs currently under investigation, LDK378 (ceritinib) has demonstrated promising antitumor activity and was granted FDA approval in April 2014.

Ceritinib contains modifications in the chemical structure that promote a more favorable interaction with the mutant lipophilic residues at the gatekeeper position of the kinase domain with a 20-fold higher in vitro potency against ALK than crizotinib. Ceritinib led to suppression of ALK phosphorylation as well as the downstream PI3K–AKT, MEK–ERK, and mTOR signaling pathways at lower doses than crizotinib. Although ceritinib was potent against the two lung cancer cell lines with ALK rearrangements, it was not potent against NSCLC or breast cancer cell lines driven by KRAS, EGFR, PI3K, or HER2 (13). In addition, in treatment-naïve H228 xenograft models, ceritinib demonstrated more durable antitumor activity than crizotinib (13). The kinase selectivity has been tested in a cellular proliferation assay against 16 different kinases, and aside from ALK, no inhibition below 100 nmol/L was observed (14). Unlike crizotinib, ceritinib does not inhibit the activity of MET, a tyrosine kinase that can be overexpressed, amplified, or mutated in NSCLC, leading to cell progression and survival. However, ceritinib does inhibit insulin-like growth factor 1 receptor (IGFIR), insulin receptor (InsR), and ROS1 (15). The chemical architecture of ceritinib, including the chlorine in the 5-position of the pyrimidine, may interact more favorably with a methionine gatekeeper in crizotinib-resistant ALK and could explain the activity of this drug against several ALK resistance mutations such as L1196M, G1269A, I1171T, and S1206Y. However, it does not overcome the ALK resistance mutations G1202R and F1174C.

ASCEND-1, the first phase I in-human, single-arm study of ceritinib at 750 mg daily, included 255 patients, 246 of whom had ALK⁺ NSCLC, 67% with at least two prior treatment regimens, and 66% with prior ALKi treatment (16). After a median 7 months of follow-up, patients treated with ceritinib achieved an ORR of 58.5% (95% CI, 52.1%–64.8%) and a median progression-free survival (PFS) of 8.2 months (95% CI, 6.7–10.1). The median DOR was 9.7 months (95% CI, 7.0–11.4), with a median time to first response of 6 weeks.

Among 163 previously crizotinib-treated patients receiving 750 mg of ceritinib daily, the ORR was 54.6% (95% CI, 46.6%–62.4%) and PFS was 6.9 months (95% CI, 5.4–8.4). In 83 patients without prior ALKi, the ORR was 66.3% (95% CI, 55.1%–76.3%). At the time of data cutoff, the majority of these patients were still receiving ceritinib and the median PFS had not been reached.

In the 124 patients who started the study with brain metastases, ceritinib achieved an ORR of 54.0% (95% CI, 44.9%–63.0%) and a median PFS of 6.9 months (95% CI, 5.4–8.4). Tumor shrinkage was seen in patients with brain metastases with (50%; 95% CI, 39.7%–60.3%) and without (69.2%; 95% CI, 48.2%–85.7%) prior ALK-TKI. Discontinuation of treatment due to adverse events (AE) occurred in 10% of patients, and 59% of patients required at least one dose reduction. The most common AEs, occurring in more than half of patients, were diarrhea, nausea, vomiting, abdominal pain, and fatigue (15, 16).

The subgroup analysis between Asian and Caucasian patients showed that baseline demographics were similar but ALKi pretreatment had been received by 47 (29%) and 108 (66%) patients, respectively (17). Of 173 patients analyzed for efficacy, the ORR was 69% (95% CI, 55.2–80.9) in Asian patients (38/55) and 57% (95% CI, 47.3–65.9) in Caucasian patients (67/118). The median DORs among responders were 10.1 months (95% CI, 7.3 to not reached) and 6.9 months (95% CI, 4.5–11.4) in the Asian and Caucasian patients, respectively. The observed differences between Asians and Caucasians for ORR and DOR were not explained by differences in ALKi pretreatment (17).

Two phase III ongoing randomized trials are currently investigating the role of ceritinib both in chemotherapy and in patients with ALK⁺ NSCLS previously treated with crizotinib (NCT01828112) versus single-agent chemotherapy (pemetrexed or docetaxel), and in previously untreated (NCT01828099) ALK⁺ NSCLC versus platinum doublet chemotherapy (platinum/pemetrexed).

Adverse reaction data are based on 255 patients treated with ceritinib, 750 mg daily for ALK⁺ malignancies (n = 246 NSCLC; n = 9 other ALK⁺ diseases; refs. 16, 18).

Serious AEs were rare (≤2%), and included interstitial lung disease (ILD)/pneumonitis, convulsion, pneumonia, dyspnea, dehydration, hyperglycemia, and nausea (18). The most common AEs (≥1%) that resulted in discontinuation of therapy included ILD/pneumonitis, pneumonia, and anorexia. Fatal AEs were also rare (5% of patients), resulting from pneumonia (n = 4), respiratory failure (n = 1), ILD/pneumonitis (n = 1), pneumothorax (n = 1), gastric hemorrhage (n = 1), general physical health deterioration (n = 1), pulmonary tuberculosis (n = 1), cardiac tamponade (n = 1), and sepsis (n = 1). Neuropathies, including peripheral sensory or motor neuropathy, gait disturbance, paresthesia, hypoesthesia, dysesthesia, neuralgia, hypotonia, or polyneuropathy, occurred in 17% of patients. Vision abnormalities (9%) occurred infrequently but were clinically relevant. ILD/pneumonitis resulted in one death (0.4%). Concentration-dependent QTc prolongation has also occurred (18).

Mechanisms of acquired resistance are heterogeneous and may evolve dynamically in response to different ALK-TKIs and may be divided into two groups: ALK dominant or ALK nondominant (Table 1). ALK-dominant mechanisms include second mutations and C1156Y, within the kinase domain of the EML4–ALK fusion gene in the same patient who acquired resistance to crizotinib (19). L1196M is a gatekeeper mutation that interferes with the binding of crizotinib. Other resistance mutations in the ALK gene have been discovered in the clinical setting or in mutagenesis screening, including L1152R, 1151Tins, G1202R, S1206Y, F1174C, D1203N, G1269A, and L1196M. Eleven ALK⁺ patients with NSCLC with acquired resistance to crizotinib were reported to exhibit new-onset ALK copy number gain, which may occur in combination with resistance mutations.

The known ALK-nondominant mechanisms leading to crizotinib resistance are mutations of other oncogenes such as the EGFR and KRAS genes (19), amplification of the KIT gene (20), increased autophosphorylation of EGFR (20), and transformation to sarcomatoid carcinoma (ref. 21; Table 1). Recently several studies have suggested that ALK rearrangements cooccur with mutations in EGFR or KRAS at clinically relevant frequencies. Gainor and colleagues reported the genotyping data from 1,683 patients with NSCLC finding 4 out of 75 ALK⁺ patients with KRAS mutations (22). Won and colleagues have profiled 1,458 cases of lung cancer and found that 4 out of 91 cases had concomitant EGFR and ALK alterations (23). The possibility of coexistence of either EGFR or KRAS mutations has profound effect on therapeutic choices and highlights the need to extend the ALK testing to EGFR and KRAS-mutation positive cases. MET receptor expression but not MET gene amplification is significantly increased in ALK⁺ NSCLC compared with ALK-negative counterpart (24). Because crizotinib is a dual inhibitor of MET and ALK, it is possible that the status of MET expression may affect the efficacy of crizotinib in ALK⁺ NSCLC under therapy. However, second-generation ALKi has selective activity against ALK-TK and does not demonstrate activity against MET-TK.

Second-generation ALKi, such as alectinib (25, 26) and ceritinib (14), have been shown to be effective not only in crizotinib-naΐve patients, but also in those resistant to crizotinib.

Unfortunately ceritinib resistance has already been reported in 11 ALK⁺ NSCLC cases by FISH, showing mutations in two residues, G1202R and F1174C, respectively, in 3 out of 11 and 2 out of 11 post-ceritinib biopsies (13). In vitro resistance to both crizotinib and ceritinib was reported in less common ALK-resistance mutations such as C1156Y, 1151T-ins, and L1152P (13).

A novel ALK V1180L gatekeeper mutation from a cell line model and a second novel I1171T mutation from a patient who developed resistance to alectinib were recently reported (27). Both mutations demonstrated structural alterations with subsequent decrease binding affinity with alectinib and crizotinib. However, both mutations were sensitive to ceritinib and other next-generation ALK-TKIs and treatment of the patient with ceritinib led to a marked response.

After the acquisition of resistance to ALKi, regardless of the use of crizotinib or second-generation ALKi, specific treatment strategies should be considered directed to inhibition of the specific resistance mechanisms. The association of an ALKi beyond the state of progressive disease (PD) and an inhibitor of the specific resistance pathways (i.e., EGFR-TKI or KIT-TKI) would be appropriate. Also, chemotherapy with pemetrexed should be considered for ALK⁺ patients with any possible mechanism of resistance to crizotinib.

Among the other second-generation ALKi, alectinib and AP26113 are in more advanced development.

Alectinib is a potent, selective, and orally available ALK-TKI with 10-fold greater potency than crizotinib with activity against other kinases, including MET, IGFIR, and ALK with or without the gatekeeper mutation (L1196M; ref. 14). A single-arm, open-label, phase I/II trial was conducted in ALK⁺ NSCLC in Japan and demonstrated excellent efficacy for alectinib (25). In contrast with the trials of crizotinib, positive results based on both FISH and IHC or RT-PCR analysis were required for enrollment in that study. In the phase I portion, including 24 patients, a dose of 300 mg twice daily was chosen as the recommended dose in the phase II trial, which included 46 patients.

In the ongoing phase II portion of the study, 43 of 46 patients achieved an OR (93.5%; 95% CI, 82.1–98.6), including 2 CRs (4.3%; 0.5–14.8) and 41 PRs (89.1%; 76.4–96.4). Grade 3 treatment-related AEs were recorded in 12 of 46 (26%) patients, including 2 patients with decreased neutrophil counts and increased blood creatine phosphokinase. Serious AEs occurred in 5 patients (11%) but no grade 4 AEs or deaths were reported (25).

Intriguingly, no progression of CNS lesions was observed in 15 patients proved to harbor brain metastases by the time of data cutoff. The PFS rate at 1 year was 83% (95% CI, 68–92), although the median PFS was not reached.

Results from 47 patients enrolled in another phase I/II study showed alectinib to be well tolerated, with the most common AEs being fatigue (30%), myalgia (17%), and peripheral edema (17%). Dose-limiting toxic effects were recorded in 2 patients in the cohort receiving alectinib, 900 mg twice a day. At data cutoff (median follow-up of 126 days), investigator-assessed ORs were noted in 24 of 44 (55%) patients, with a confirmed CR in 1 (2%), a confirmed PR in 14 (32%), and an unconfirmed PR in 9 (20%). Sixteen (36%) patients had stable disease (SD); the remaining 4 (9%) had PD. Of 21 patients with CNS metastases at baseline, 11 (52%) OR, 6 (29%) CR (3 unconfirmed), 5 (24%) PR (1 unconfirmed), 8 (38%) SD, and the remaining 2 (10%) PD were reported. Alectinib, 600 mg twice a day, was chosen as the recommended dose for phase II (28).

AP26113 is a novel TKI that potently inhibits mutant activated forms of the ALK and EGFR genes as well as TKI-resistant forms, including L1196M of the ALK gene and T790M of the EGFR gene (29). Preliminary data for an ongoing dose-finding phase I/II study of AP26113 for advanced malignancy refractory to standard treatment showed the efficacy and safety of the compound in patients with NSCLC previously treated with ALKi or EGFR-TKIs. Among 57 evaluable ALK⁺ NSCLC patients, 41 (72%) responded. Among 51 evaluable ALK⁺ NSCLC patients with prior crizotinib exposure, 35 (69%) responded. The DOR was 1.6 to 14.7 months ongoing at the time of data cutoff. Among 49 patients with follow-up scans, the median PFS was 10.9 months. Nine out of 13 ALK⁺ patients with untreated progressing CNS lesions at baseline and with follow-up scans showed evidence of radiographic improvement in CNS disease, including 1 patient with improvement in leptomeningeal disease (30). The antitumor activity of at least two other second-generation ALKi, ASP3026 and X-396, has been shown in in vitro studies, and these agents are currently under clinical investigation (NCT01401504 and NCT01625234). Table 2 summarizes the trials with novel second-generation ALKi in clinical development.

ALK gene fusion can be detected by several methods, including RT-PCR, the first published method used (2); FISH, the currently accepted method approved by the FDA; and IHC.

FISH is a relatively expensive assay that can detect all types of ALK rearrangements known to date; however, its interpretation could be challenging even in the hands of experienced specialists. IHC is relatively inexpensive, fast, and familiar to most pathologists, and can also be used universally, similar to FISH, as it can potentially detect overexpressed ALK chimeric protein produced by any rearrangement type. However, the variability of chimeric proteins with different levels of expression raises questions about the correct choice of antibody and signal enhancement system to avoid false-negative results. Artifacts that may lead to false-positive results are also relatively frequent and cannot be underestimated. RT-PCR is a highly sensitive and specific technique that allows the detection of even a few molecules of chimeric ALK transcripts. However, the use of RT-PCR as a screening method for detecting ALK rearrangements may not be completely reliable for several reasons, including poor quality of RNA obtained from formalin-fixed, paraffin-embedded tissues, which are mostly available in the clinical setting, and the necessity of PCR multiplexing because of the wide variation in fusion types.

It is clear that each method has advantages and pitfalls, and an agreement about the optimal protocol for ALK testing has not yet been reached. IHC, FISH, and multiplex RT-PCR methodologies showed good sensitivity, specificity, and concordance when artifacts were characterized and excluded in a study that prospectively tested 36 patients with NSCLC who had adenocarcinoma and 10 ALK⁺ samples. However, all ambiguous cases had to be confirmed as ALK rearranged by at least two of the three methods. Blackhall and colleagues have recently reported a prevalence of 6.2% of IHC positivity with a 2.2% of FISH positivity in 1,281 European patients with NSCLC, showing that a screening strategy based on IHC or H-score might be feasible (31).

The concept of coapproval of therapeutic product and companion diagnostics has been encouraged as part of the strategy to promote personalized medicine (32). However, molecular diagnostic testing is complex and several assays with different merits may be suitable in that regard especially when testing molecularly heterogeneous tumors such as NSCLC.

The optimal choice for first-line therapy for ALK⁺ patients and the optimal sequence for therapies after progression to first-line therapy aiming at maximizing survival benefit and strategies to prevent or delay resistance remain to be determined. Although it is not clear at present whether crizotinib or second-generation ALKi will be the superior treatment for ALK⁺ patients, a head-to-head study comparing crizotinib with the second-generation ALKi would clarify this point. In fact, a randomized phase III trial comparing alectinib with crizotinib is currently being performed to address this issue (NCT02075840). The activity demonstrated by ceritinib and alectinib in patients with CNS disease involvement suggests the use of these agents preferentially in this clinical scenario. On the basis of already described mechanisms of resistance to first- and second-generation ALKi-specific treatment strategies, combination therapy with EGFR-TKI, KIT-TKI, MET-TKI, or pemetrexed-based therapy could be explored.

The challenge of future studies is in the identification of the mechanisms underlying acquired resistance to ceritinib and new ALKi. In addition, combinations between ALKi and other therapeutic strategies, such as inhibition of escape survival pathways and immunotherapy agents, are potential alternatives to increase survival in ALK⁺ NSCLC patients.

V. Papadimitrakopoulou reports receiving commercial research grants from Amgen, AstraZeneca, Bristol-Myers Squibb, Celgene, Clovis Oncology, Genentech, Janssen, MedImmune, Merck, and Pfizer and is a consultant/advisory board member for AstraZeneca, Clovis Oncology, Eli Lilly, Genentech, GlaxoSmithKline, Janssen, MedImmune, Merck, and Novartis. No potential conflicts of interest were disclosed by the other author.

Conception and design: E. Massarelli, V. Papadimitrakopoulou

Development of methodology: V. Papadimitrakopoulou

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): E. Massarelli

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): E. Massarelli, V. Papadimitrakopoulou

Writing, review, and/or revision of the manuscript: E. Massarelli, V. Papadimitrakopoulou