EML4-ALK: Honing In on a New Target in Non–Small-Cell Lung Cancer

Lung cancer is the leading cause of cancer-related death in the United States and worldwide.¹ Despite recent advances in the treatment of the disease, the overall 5-year survival rate in the United States remains only 15%. Therefore, novel therapeutic strategies are needed.

Historically, the histologic appearance of lung cancer has been used to guide treatment decisions, primarily distinguishing patients with small-cell lung cancer (SCLC) from those with non–small-cell lung cancer (NSCLC). Three major subtypes of NSCLC—adenocarcinoma, squamous cell carcinoma, and large-cell carcinoma—were grouped together as a single entity for therapy and enrollment of patients onto clinical trials. With the recent regulatory approval of pemetrexed specifically for adenocarcinoma and large-cell carcinoma,² and bevacizumab for nonsquamous NSCLC,³ the discrimination of specific NSCLC subtypes has grown in significance.

During the past 5 years, it has also become evident that adenocarcinomas have distinct genomic changes that allow them to be further classified into clinically relevant molecular subsets. Each subset can be defined by a specific genomic alteration that is responsible for both the initiation and maintenance of the lung cancer (ie, a driver mutation). Importantly, specific mutations induce differential sensitivity of tumors to targeted therapeutic agents. For example, tumors highly sensitive to epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors (TKIs; ie, gefitinib or erlotinib) often contain dominant somatic mutations in exons that encode a portion of the tyrosine kinase domain of EGFR.^4–6 Conversely, tumors with somatic mutationsin KRAS, which encodes GTPase downstream of EGFR, display primary resistance to these drugs.^7,8

When knowledge of specific genetic changes is linked to a targeted therapy, dramatic improvements in response and clinical benefit can be observed without the need for initial treatment with cytotoxic chemotherapy. For example, in the phase III, randomized, open-label, first-line study of gefitinib compared with carboplatin plus paclitaxel in never or light East Asian smokers with advanced NSCLC (ie, the Iressa Pan Asia Study [IPASS]), patients with EGFR-mutant tumors experienced longer progression-free survival with gefitinib, and patients without mutations had longer progression-free survival with chemotherapy(EGFR mutation positive: hazard ratio = 0.48;95%CI, 0.36 to 0.64; P < .0001 [favors gefitinib]; EGFR mutation negative: hazard ratio = 2.85; 95% CI, 2.05 to 3.980; P < .0001 [favors chemo-therapy]).⁹ Thus, gefitinib is a reasonable first-line treatment for patients with EGFR-mutant lung cancer.

In 2007, Soda et al¹⁰ identified another potential driver mutation in NSCLC—fusion of the N-terminal portion of the protein encoded by the echinoderm microtubule-associated protein-like 4 (EML4) gene with the intracellular signaling portion of the receptor tyrosine kinase encoded by the anaplastic lymphoma kinase (ALK) gene. These fusions are the result of inversions within the short arm of chromo-some 2 (involving 2p21 and 2p23, approximately 12 Mb apart). Whereas other genetic alterations involving ALK have been identified in anaplastic large-cell lymphomas,¹¹ inflammatory myofibroblastic tumors,¹² and neuroblastomas,^13–15 none have yet been found in lung cancers.^16,17 The EML4-ALK fusion gene seems to be novel to NSCLC.^18,19

To date, multiple EML4-ALK variants have been identified in lung cancer (Fig 1).^10,19–23a Although the fusions contain variable truncations of EML4 (occurring at exons 2, 6, 13, 14, 15, 18, and 20), the ALK fusion in all of them starts at a portion encoded by exon 20 of the kinase gene. To date, all EML4-ALK fusions tested biologically demonstrate gain of function properties.^10,20,22

Fig 1.

Schematic representation of EML4-ALK translocations in non–small-cell lung cancer (NSCLC). Multiple EML4-ALK variants (V1 to V7) have been identified in NSCLCs. All involve the intracellular tyrosine kinase domain of anaplastic lymphoma kinase (ALK) starting at a portion encoded by exon 20. EML4, however, is variably truncated. Variant 1 fuses exon 13 to ALK¹⁰; variant 2 fuses exon 20¹⁰; variants 3a and 3b fuse exon 6a and 6b, respectively²⁰; variant 4 fuses exon 14 (with an additional 11 nucleotides of unknown origin [dark blue] and starting at nucleotide 50 in exon 20)¹⁹; variants 5a and 5b fuse exon 2 (with 5b containing an additional 117 nucleotides from intron 19)¹⁹; variant 6 fuses exon 13 (with an additional 69 nucleotides from intron 19 [dark blue])²²; and variant 7 fuses exon 14 (starting at nucleotide 13 in exon 20).²² Additional variants have been reported as V4, fusing exon 15 (minus 19 nucleotides and starting at nucleotide 21 in exon 20),²¹ and V5, fusing exon 18.²³ The nomenclatures used in the literature are depicted on the right. On the left is another potential nomenclature, slightly modified from that of Hiroyoki Mano (see http://atlasgeneticsoncology.org/Tumors/inv2p21p23NSCCLunglD5667.html).

How frequent are ALK fusions in lung cancer? Are there clinical characteristics unique to patients whose tumors harbor ALK fusions? Previous studies, mostly involving East Asian patients, have reported that 3% to 7% of lung tumors harbor EML4-ALK fusions (Table 1). Multiple histologic subtypes of NSCLC have been tested, but adenocarcinomas seem to be the major NSCLC cell type harboring EML4-ALK fusions. Preliminary data also indicate a significant relationship between smoking and EML4-ALK positivity, with the fusions more commonly found in light smokers (< 10 pack years) or never smokers.^21,23,24 EML4-ALK fusions have also been associated with a lack of EGFRorKRASmutations,^16,23–26youngerage,^23–25andadenocarcinomas with acinar histology.^24,25

Table 1.

Studies Evaluating EML4-ALK Positivity in Lung Cancer Patients

Study	No. of Patients	Ethnicity of Patients	EML4- ALK- Positive Patients		Detection Method	Clinical Associations
			No.	%
Soda et al10; NSCLC	75	Japanese	5	7	RT-PCR	No EGFR/KRAS mutations
Inamura et al25		Japanese			RT-PCR	AD; acinar histology (variant 2); no EGFR/KRAS mutations
Total	221		5	2
AD	149		5	3
SQ	48		0	0
LC	3		0	0
SCLC	21		0	0
Takeuchi et al19		Japanese			Multiplex RT-PCR	AD; not in 292 other cancer types
Total	364		11	3
AD	253		11*	4
ADSQ	7		0	0
SQ	71		0	0
LC	7		0	0
LCNE	3		0	0
PLEO	2		0	0
SCLC	22		0	0
Inamura et al24	11*	Japanese	NA		NA	Acinar histology; less differentiated; younger age; smaller size; never/light smokers; no EGFR/KRAS mutations
Perner et al28; NSCLC	603	Swiss (n = 507) and American (United States; n = 96)	16	3	FISH	NA
Koivunen et al21; NSCLC	305		8	3	RT-PCR	AD; never/light smokers; no KRAS/BRAF mutations
United States	138	American (United States)	2	1		(1 overlapping EGFR mutation)
Korea	167	Korean	6	4
Shinmura et al16; NSCLC	77	Japanese	2	3	RT-PCR	Adenocarcinoma; former smokers; no EGFR/KRAS/PIK3CA mutations
Martelli et al26; NSCLC	120	European	9	8	RT-PCR	No EGFR mutations (1 overlapping KRAS mutation)
Wong et al23; NSCLC	266	Chinese	13	5	RT-PCR	AD; female sex; never/light smokers; younger age; no
AD	209		11	5		EGFR/KRAS mutations
SQ	34		0	0
LEL	11		0	0
Other	12		2	17
Shaw et al27; enriched	141		19	13	FISH, IHC	AD; male sex; never/light smokers; younger age; no
NSCLC						EGFR/KRAS mutations
AD	130		18	14
ADSQ	4		1	25
SQ	2		0	0
LC/NOS	5		0	0

Abbreviations: NSCLC, non–small-cell lung cancer, RT-PCR, reverse transcriptase polymerase chain reaction; AD, adenocarcinoma; SQ, squamous cell carcinoma; LC, large-cell carcinoma; SCLC, small-cell lung carcinoma; ADSQ, adenosquamous cell carcinoma; LCNE, large-cell neuroendocrine carcinoma; PLEO, pleomorphic carcinoma; NA, not applicable; FISH, fluorescent in situ hybridization; LEL, lymphoepithelial-like carcinoma; IHC, immunohistochemistry; NOS, not otherwise specified.

^*Positive patients were identified from a prior study conducted by same group. Note that not all tumors were tested for all known EML4-ALK variants.

Importantly, lung cancers that harbor genomic ALK alterations may be clinically responsive to pharmacologic ALK inhibition. Just as treatment of lung cancer cells bearing drug-sensitive EGFR mutations with EGFR TKIs can lead to growth inhibition and cell death, treatment of tumor cells containing EML4-ALK fusions with ALK kinase inhibitors can lead to the potent suppression of downstream survival signaling pathways and an apoptotic response.^10,21,29 Furthermore, in an EML4-ALK mouse lung tumor model, treatment with an ALK inhibitor resulted in reduced tumor burden compared with controls.³⁰ Several novel selective inhibitors of ALK kinase are currently in preclinical or early clinical testing,³¹ including one compound, PF-02341066, which is a dual MET/ALK inhibitor. Preliminary data from a phase I trial of PF-02341066 presented at the American Society of Clinical Oncology annual meeting this year reported one partial response, two unconfirmed partial responses, and four patients with stable disease among 10 patients with EML4-ALK–positive NSCLC. The most common toxicities were nausea, emesis, fatigue, and diarrhea.³² An expanded cohort of the same trial including 13 patients with NSCLC, eight of whom were evaluable for response, reported six partial responses and one stable disease as the best response.³²

In this issue of Journal of Clinical Oncology, Shaw et al²⁷ confirm and extend our knowledge about the clinical features and outcomes of EML4-ALK–positive lung cancer in a North American population. In their study, for tumors to be analyzed for EML4-ALK fusions, they required that patients have two or more characteristics commonly associated with patients who have EGFR-mutant NSCLCs (ie, female sex, Asian ethnicity, history of no or light smoking, and adenocarcinoma histology). By using these criteria, they were able to enrich for positive tumors (13% of total analyzed). In 141 patients, 94% of whom were non-Asian, they found EML4-ALK fusions to be associated with younger age, male sex, minimal smoking exposure, and adenocarcinoma histology. Thirty-one percent of specimens had acinar growth patterns, and 61% had a solid growth pattern. EML4-ALK fusions were not found in tumors with EGFR or KRAS mutations. Remarkably, in never/light smokers without EGFR mutations, EML4-ALK fusions were found in 33% of patients. They further demonstrate thatEML4-ALKpositivityisassociatedwithresistanceto EGFR TKIs. Additionally, compared with patients whose tumors harbor neither EML4-ALK nor EGFR mutations, they showed that patients whose tumors harbor EML4-ALK display similar sensitivity to platinum-based combination chemotherapy and no differences in overall survival.

Currently, a major issue is defining the best way to assess for the presence of ALK fusions in lung tumors. A variety of methods have been used, including immunohistochemistry, fluorescent in situ hybridization, and reverse transcriptase polymerase chain reaction. None of the assays have proven to be easily adopted by diagnostic molecular pathology laboratories. ALK antibodies seem to give variable results.^24,26 Fluorescent in situ hybridization analysis involves use of two differently labeled break-apart probes on opposite sides of the breakpoint of the ALK gene; normal ALK appears as a yellow signal, whereas rearranged ALK at this locus results in separate red and green signals.²⁸ However, these signals can be difficult to interpret. Finally, reverse transcriptase polymerase chain reaction requires good-quality mRNA. Moreover, because fusions can occur at multiple points (Fig 1), multiple polymerase chain reaction primers are required to detect all known fusion transcripts. The variability of tumors testing positive in the literature is likely a result of the different methods of detection used in the various studies. Future efforts are needed to determine the optimal method for routine diagnostic detection of ALK fusions in lung cancers.

How should we find appropriate patients for treatment? One approach is to use clinical criteria to enrich for potential responders. However, clinical criteria can be misleading; for example, the first EML4-ALK fusion was found in a former smoker, not a light/never smoker.¹⁰Analogously,KRASmutationsareassociatedwithsmoking, but 15% of never smokers have tumors that harbor mutant KRAS.³⁴ Thus, we favor development of robust platforms for molecular diagnostic testing of lung adenocarcinomas. Ideally, tumors would be simultaneously tested for the three most common lesions to date–KRAS, EGFR, and ALK fusions. However, an alternative stepwise approach (Fig 2) is to test lung adenocarcinomas first for KRAS mutations because 15% to 30% of tumors harbor such alterations. If a tumor is positive for a KRAS mutation, no further molecular testing is required. Treatment recommendations would include KRAS-directed therapy (when it is developed); an EGFR TKI would be lower priority compared with other available treatments. (The sensitivity of KRAS-mutant tumors to ALK inhibitors remains to be established.) If the tumor is negative, it would be tested for EGFR mutations, which are found in approximately 10% of tumors. If the tumor is positive for an EGFR mutation, an EGFR TKI would be recommended. If negative, the tumor would be assessed for ALK translocations. If positive, treatment would include an ALK inhibitor; again, an EGFR TKI would be lower priority. If negative, then these triple-negative tumors should be considered for analysis of rarer mutations (eg, in BRAF, MEK1, AKT1, PIK3CA, and so on) and/or intensive research to identify new driver mutations.

Fig 2.

Suggested algorithm for molecular testing for patients with non–small-cell lung cancer with tumors of adenocarcinoma histology before treatment with oral targeted therapy. (*) “Other” includes BRAF, MEK1, AKT1, PIK3CA, and so on.

In summary, the work by Shaw et al²⁷ as well as multiple other groups regarding EML4-ALK validates the fusions as defining a distinct molecular subset of NSCLC. The promise of ALK inhibitors against these fusions brings us one step closer to personalized lung cancer therapy, in which patients are treated not empirically with cytotoxic chemotherapy but with targeted agents according to the genetic makeup of their tumors.