The Biology and Clinical Features of Non–small Cell Lung Cancers with EML4-ALK Translocation
- First Online:
- Cite this article as:
- Pillai, R.N. & Ramalingam, S.S. Curr Oncol Rep (2012) 14: 105. doi:10.1007/s11912-012-0213-4
- 461 Views
The anaplastic lymphoma kinase (ALK) acts as a dominant oncogenic driver following chromosomal rearrangements in certain cancers including non–small cell lung cancer (NSCLC). NSCLC with ALK translocation occurs in a specific subset of patients and results in unique clinical features. Crizotinib is a small molecule inhibitor of ALK kinase that has recently been approved by the FDA for the treatment of patients with ALK-positive NSCLC. Treatment with crizotinib results in clinical benefit rate of 85%–90% and a median progression-free survival of 9–10 months for this molecular subset of patients. Ongoing studies will define the impact of crizotinib on overall survival and provide insights into the resistance mechanisms and potential activation of alternate pathways. Heat shock protein 90 inhibitors also appear promising in the treatment of ALK-positive NSCLC patients, based on early results. This article reviews the characteristics, treatment, and ongoing research in patients with ALK-positive NSCLC.
KeywordsEML4-ALK translocationNon-small cell lung cancer (NSCLC)CrizotinibHeat shock protein 90 (Hsp90)IPI-504Ganetespib
Anaplastic lymphoma kinase (ALK) is a receptor tyrosine kinase in the insulin receptor superfamily that was first described in the 1980s [1, 2]. It is part of the t(2;5) translocation, the most frequently found mutation in anaplastic large cell lymphoma (ALCL), a T-cell non-Hodgkin lymphoma. In ALCL, ALK is fused to nucleophosmin (NPM1), which has a dimerization domain that results in ligand-independent dimerization and constitutive kinase activation. ALK has other fusion partners which also cause oncogenic transformation that have been identified in neuroblastomas and inflammatory myofibroblastic tumors (IMT).
Non–small cell lung cancer (NSCLC) remains a major cause of death, with over 220,000 new cases estimated to be diagnosed in the United States in 2011 . Most lung cancer presents in advanced stages, and the mainstay of therapy is systemic chemotherapy, typically with a platinum-based regimen . However, a plateau has been reached in the efficacy of chemotherapy using nonspecific cytotoxic agents. In recent years, characterization of NSCLC by molecular typing, particularly in adenocarcinomas of the lung, has led to the investigation of therapeutic agents that target dominant oncogenic mutations. The use of epidermal growth factor receptor (EGFR) inhibitors in NSCLC patients with an activating mutation in exons 19 or 21 is associated with robust response rates and progression-free survival. In patients with a somatic EGFR mutation, front-line treatment with an EGFR inhibitor, gefitinib, results in improved response rate and superior progression-free survival . This benefit is limited to patients with EGFR mutation, as patients without the mutation have better overall response rates and progression-free survival with combination chemotherapy. In fact, the recently reported interim analysis of the EURTAC trial shows upfront therapy with EGFR blockade in patients with activating EGFR mutations to be superior to platinum-based chemotherapy, with median progression-free survival of 5.2 months with chemotherapy compared to 9.4 months with erlotinib (P < 0.0001) . This successful paradigm emphasizes the need for identifying dominant oncogenic events that can be targeted with specific therapeutic agents.
In 2007, Soda and colleagues [7•] from Japan reported a new molecular abnormality involving the fusion of anaplastic lymphoma kinase (ALK) with the echinoderm microtubule-associated protein-like 4 (EML4) in lung adenocarcinoma. The fusion protein results from an inversion in the short arm of chromosome 2, which fuses the N-terminal domain of EML4 to the intracellular kinase domain of ALK. The fusion protein causes constitutive activation of the kinase. Mouse fibroblasts transfected with EML4-ALK formed tumors when injected into nude mice, demonstrating the transforming activity of the fusion protein. In screening a cohort of NSCLC patients, there were two types of fusion products found: variant 1 consisted of intron 13 of EML4 fused to intron 20 of ALK, and variant 2 had intron 20 of EML4 fused to intron 20 of ALK.
The researchers went on to develop a transgenic mouse model to determine the oncogenicity of the EML4-ALK fusion protein . Transgenic mice expressing EML4-ALK developed hundreds of adenocarcinoma nodules in the lungs, detected on CT scan and on post-mortem examination. These nodules had positive cytoplasmic expression of ALK seen on immunohistochemistry using ALK antibodies and low level of lung-specific expression of EML4-ALK protein by immunoblot analysis. Treatment with a specific ALK inhibitor reduced the tumor burden in the mice, further demonstrating the “oncogene addiction” induced by EML4-ALK fusion protein. This observation has prompted further research into ALK-associated lung cancer, including development of targeted agents that inhibit the ALK kinase.
In the original report by Soda et al., 5 out of 75 NSCLC patients (6.7%) had the EML4-ALK fusion protein [7•]. However, other studies examining mostly Asian patients with early-stage disease demonstrate an incidence ranging from 1% to 5%. In a group of patients with adenocarcinoma that was selected for genetic screening based on female sex, Asian ethnicity, and light smoking history, 13% had EML4-ALK mutation [9•].
Although originally defined in a patient who was a smoker, the EML4-ALK fusion protein occurs more commonly in never or light smokers [7•, 10]. A review by Sasaki et al.  of multiple reports of the frequency of EML4-ALK translocations analyzed by smoking status showed that 9.4% of never or former light smokers (defined as less than 10 pack years and quit at least 1 year prior) had the fusion protein compared to 2.9% of current smokers. This is not unlike the incidence of EGFR mutations, which tend to be more frequent in never or light smokers .
Patients with EML4-ALK translocation have other distinguishing characteristics from patients with EGFR mutations. In the largest cohort of EML4-ALK positive patients described to date, Shaw et al. [9•] found that these patients were younger than patients with EGFR mutations, with a median age of 52 years compared to 66 years in EGFR mutation group and 64 years in wild-type group. In fact, four patients with EML4-ALK were under 40 years of age. EML4-ALK translocation is more commonly noted in male patients (58%) versus the EGFR mutants (26%) or wild-type group (32%).
EML4-ALK translocations are predominantly associated with adenocarcinoma histology. Two Japanese groups studied resected NSCLC tumor samples and found EML4-ALK solely in adenocarcinoma patients [12, 13]. In a review of ethnic Chinese patients who had resection for NSCLC, 11 of 13 patients with EML4-ALK fusion were adenocarcinoma; 2 patients of this group had bronchioloalveolar carcinoma and 9 patients had mixed tumor subtype . Two additional patients had atypical carcinoma with mixed squamous and glandular components. In the Asian population, the majority of EML4-ALK patients had an acinar pattern of adenocarcinoma (65%), often with mucin production . This contrasts with patients described in the US, who have over 60% solid or sheet-like pattern of growth [9•, 16]. In addition, 82% of these patients had signet-ring cells present, similar to those commonly seen in gastric, colon, and breast cancer.
Most reports show that EML4-ALK translocations are mutually exclusive of EGFR and KRAS mutations [9•, 15, 16]. However, a group in China found that 1% of the patients (4 out of 398 patients) had both EML4-ALK fusion protein and activating EGFR mutation . Therefore, until additional information is available, it is reasonable to test for both EGFR and ALK in patients with NSCLC.
Currently the diagnostic criterion for clinical trials enrolling patients with NSCLC with EML4-ALK translocation is the presence of the rearrangement detected by fluorescence in situ hybridization (FISH). FISH is a cytogenetic technique which uses fluorescent probes to bind to specific DNA sequences on chromosomes. The FDA approved crizotinib with FISH testing as a companion diagnostic test for ALK detection. Therefore, FISH is likely to be used as the preferred testing method for the near future. Other promising tests for either screening or confirmation of ALK translocation are currently under development. Immunohistochemistry (IHC) is a technique that is commonly utilized in pathology labs to stain for the presence of proteins of interest for diagnosis. Paik and colleagues  used a cohort of Korean patients with NSCLC to screen for ALK rearrangement by IHC. They used a mouse antibody against ALK on formalin-fixed and paraffin-embedded (FFPE) samples. The authors developed a semi-quantitative scoring system: a score of 0 indicated no stained cells, 1 if faint staining was present in more than 5% tumor cells or the presence of any staining intensity in ≤ 5% cells, 2 was associated with moderate staining intensity in more than 5% tumor cells, and 3 for strong intensity seen in more than 5% tumor cells. In the 465 patient samples screened, ALK expression was seen in 8.6% (40/465). The results of IHC were then correlated with standard FISH using a break-apart probe specific to the ALK locus; FISH-positivity was defined by greater than 15% split signals or isolated red signal (IRS). By FISH, the ALK rearrangement was detected in 4.2% of patients (19/453). All cases that were scored IHC 0 or 1 were negative by FISH; all cases with IHC score of 3 were positive. An IHC score of 2 was FISH positive in only 3 of 10 patients; this score was defined as equivocal. The overall sensitivity of the IHC assay was 100% and specificity was 95.2%. The authors then tested IHC as a screening test in a validation set of 187 patients with adenocarcinoma of lung with biopsied or surgically resected specimens. Fourteen patients had ALK positivity by IHC; of these, 2 cases had a score of 1 (both FISH negative), 6 had a score of 2 (50% FISH positive), and 6 had a score of 3 (100% FISH positive). By FISH there were 9 patients with ALK translocation; there were five false positives with IHC (2 with score of 1 and 3 with a score of 2). The validation set shows the feasibility of IHC as a screening tool to predict the presence of ALK gene rearrangement.
Yi and colleagues  also examined IHC in a group of 101 non-smokers with adenocarcinoma diagnosed in the Mayo Clinic Lung Cancer Cohort. They developed a similar algorithm correlating IHC score of 2 and 3 being positive with 0 and 1 being negative. The specificity of IHC was 97.8% and sensitivity was 90% in this patient sample.
In addition, reverse transcription polymerase chain reaction (RT-PCR) has the potential of identifying unknown fusion variants missed by FISH. A recent study by Takeuchi et al.  assessed multiplex RT-PCR on 253 lung adenocarcinoma samples and 403 solid tumor samples. None of the samples from the other organs yielded ALK fusion product. Approximately 4% (11 out of 253) of the samples were positive for EML4-ALK DNA among the lung adenocarcinoma specimens. Of these, two were found to be previously unidentified fusion variants of EML4-ALK.
Clinical data with crizotinib
Camidge and colleagues  presented an update of the response data of the second part of the phase 1 study at the 2011 American Society of Clinical Oncology (ASCO) annual meeting. To date, 119 patients have enrolled in the study, with 116 patients that were evaluable for response. The overall response rate (ORR) was 61% (71 of 116 patients) at a median follow-up period of 11 months. The majority of the responses seen were partial responses (59%), but two patients did achieve complete response (CR rate was 2%). The median time to response was 8 weeks, and the median response duration was 48 weeks. The preliminary median PFS was estimated to be 10 months. These data show even more convincingly the efficacy of ALK blockade in patients with ALK-rearranged NSCLC.
In a recent report, the survival of ALK-positive patients treated with crizotinib was compared to ALK-positive historical controls that did not receive crizotinib and ALK-negative controls . The authors identified 37 ALK-positive patients from the phase 1 study of crizotinib, as well as 253 ALK and EGFR mutant negative controls to compare with 82 patients who received crizotinib. The 1-year OS in the ALK-positive patients treated with crizotinib was 77% and 2-year OS was 64%, independent of age, sex, ethnicity, or smoking history. In the ALK-positive controls, OS at 1 year was similar at 73% but 2-year OS decreased to 33% without crizotinib therapy. Most of the ALK-positive controls were non-Korean; comparison with the crizotinib-treated cohort outside of Korea showed similar age, sex distribution, smoking history, number and types of prior therapies, and presence of brain metastases. In addition, patients who received second-line or third-line crizotinib (32 patients) had longer 1-year and 2-year OS (71% and 61%) compared to ALK-positive controls (46% at 1 year, 9% at 2 years, P < 0.004). The authors concluded that due to the improved survival benefit seen, patients with ALK rearrangement should receive crizotinib as a new standard of care. Though these results are interesting, this was not a prospective study and the control group was not homogenous.
Interim data from a second clinical study with crizotinib were also presented at this year’s ASCO meeting by Crino and colleagues . The PROFILE 1005 study examined 136 patients enrolled in a phase 2 study of crizotinib in recurrent, advanced, or metastatic NSCLC that progressed after previous chemotherapy. Crizotinib was administered at 250 mg twice daily in 3-week cycles; disease response was determined by RECIST criteria every 6 weeks. The median duration of treatment was 9 weeks, and 88% of patients remained on therapy at the time of interim analysis. A total of 83% of patients had tumor shrinkage. The most frequent adverse events reported were gastrointestinal toxicities (nausea in 46%, vomiting in 39%, and diarrhea in 29%), and 45% had vision disturbances. A total of 15% of patients had grade 3/4 adverse events (neutropenia, increased ALT, dyspnea). There were 9 deaths including 2 that were deemed treatment-related. The data from the PROFILE study show again that crizotinib is well tolerated and effective in patients with EML4-ALK rearranged NSCLC.
ALK inhibitors in development
Nerviano Medical Sciences
Patients with EML4-ALK do not appear to benefit from EGFR inhibition. In the largest group reviewed with EML4-ALK fusion kinase, patients who were wild type or had EML4-ALK had similar time to progression with EGFR tyrosine kinase inhibitor (5 months in ALK-positive group vs 6 months for wild type, P = 0.337) compared to a median time to progression (TTP) of 10 months in EGFR-mutant group [9•]. Patients treated with standard platinum-based chemotherapy had non-statistically significant lower response rate, with a median TTP of 8 to 10 months in all groups.
Unfortunately, similar to imatinib in chronic myelogenous leukemia, resistance to crizotinib has already been reported. A patient on the original phase 1 study of crizotinib who developed resistance after 5 months of therapy was evaluated with deep sequencing of cDNA, comparing EML4-ALK translocation before and after resistance by Choi et al. . There were two independent events noted: C1156Y and L1196M. Both of these mutant fusion proteins were transfected in BA/F3 cell lines. The mutant tyrosine kinases had decreased sensitivity to crizotinib in vitro. Structural analysis of these mutants showed that C1156Y was close to the N-terminal domain and close to the edge of the ATP-binding pocket. Interestingly, L1196M corresponds to the T315 mutation in BCR-ABL and the T790M gatekeeper mutation in EGFR and is in the active site of the ALK kinase.
The problem of resistance in EML4-ALK–rearranged NSCLC described in this patient was further explored in a study by Katayama et al. . This group developed a cell line model of crizotinib resistance by exposing the H3122, which expresses EML4-ALK variant 1, to increasing concentrations of crizotinib. The resistant cells expressed higher total protein of EML4-ALK and had increased number of EML4-ALK genes compared to the parental cell line. The resistant cell lines remained phosphorylated in the presence of crizotinib and behaved like cells without ALK rearrangement. This suggests that gene amplification may be an important mechanism of resistance. Then resistant cells exposed to even higher concentrations of crizotinib were shown to develop L1196M mutation. This model suggests that there is a stepwise process to the development of crizotinib resistance. Novel inhibitors of ALK kinase, such as NVP-TAE684, a 5-chloro-2,4-diaminophenylpyrimidine, and AP26113, which can decrease ALK phosphorylation, induce apoptosis, and reduce cell survival in both parental H3122 and resistant cell lines and in cells grown as xenografts in mice. In addition, Hsp90 inhibition with 17-AAG also overcomes crizotinib resistance. 17-AAG decreased ALK protein expression, phosphorylated ALK, and cell growth in both parental H3122 cells and crizotinib-resistant cells. These strategies represent potentially novel treatment options for patients with crizotinib resistance and require investigation in this patient population.
Sakamoto and colleagues  found a selective inhibitor to the ALK gatekeeper mutant, CH5424802. Cell lines with ALK fusion proteins, including NSCLC, anaplastic lymphoma, and neuroblastoma cell lines, showed growth inhibition with exposure to CH5424802. This compound retained its growth inhibitory effect in L1196M mutant cell lines, unlike crizotinib. In addition, there was a dose-dependent tumor regression in a mouse xenograft model of both Ba/F3 cells expressing native EML4-ALK and L1196M mutant. Evaluation of downstream signaling pathways by Affymetrix GeneChip analysis showed that inhibition of ALK by CH5424802 resulted in downregulation of the STAT3 pathway. In addition, x-ray crystal structure analysis revealed the specificity of CH5424802 was due to the presence of one hinge hydrogen bond, similar to interactions of other kinases with specific inhibitors, such as imatinib for BCR-ABL and erlotinib for EGFR. Crizotinib has two or three hinge hydrogen bonds unlike CH5424802 and loses a key bonding interaction when methionine is substituted for leucine in L1196M mutant. Due to its success in the lab, this compound is being investigated further under development code AF802 in Japan (Table 2).
Sequist and colleagues  published data from a phase 2 study of IPI-504 or retaspimycin, an analog of the Hsp90 inhibitor 17-AAG, in patients with advanced NSCLC who had previously received EGFR TKI therapy. A total of 76 patients were enrolled in the study; the primary end point was ORR assessed by CT scan every 2 cycles. Fifteen patients had ALK analysis performed on the tumor tissue. The overall response was 7% with a median PFS of 2.86 months for the overall population. However, in the three patients who had ALK rearrangements, 2 of 3 achieved partial response; all three patients remained on study for approximately 7 months. These surprising results have led to further interest in Hsp90 inhibition in ALK-positive NSCLC.
IPI-504 has been investigated further in the lab to characterize its effects in ALK-positive lung cancer due to findings of this phase 2 study . Hsp90 acts as a chaperone protein to stabilize its substrates or “client proteins,” which are degraded in the absence of Hsp90. EML4-ALK expressed in H3122 cell line is the most sensitive client protein to Hsp90 degradation to date; it is degraded in 3 h compared to EGFR, which takes 24 h. H3122 cells also show dose-dependent inhibition of cell growth when incubated with IPI-504. The expression of EML4-ALK sensitizes HEK293FT cells to the inhibitory effects of IPI-504; this was also shown in a mouse xenograft model. Unlike other mouse xenograft models, Hsp90 inhibition in EML4-ALK expressing tumors actually causes tumor regression instead of only growth stabilization. The growth inhibitory effects of IPI-504 are seen even in crizotinib-resistant H3122 cells.
Preliminary data from a phase 2 study of another Hsp90 inhibitor, STA-9090, or ganetespib, were presented at the 2011 ASCO meeting . Ganetespib was administered weekly for 3 weeks of a 4-week cycle to 76 patients with previously treated advanced NSCLC with PFS as the primary end point. The overall PFS at 16 weeks was 24; the EGFR and KRAS mutated patients had disease stabilization. Objective responses (complete and partial response) were seen in four patients; these were all ALK-positive patients. Ganetespib was not associated with hepatic and ocular toxicities seen with other Hsp90 inhibitors. Based on these data, Hsp90 inhibitors should be evaluated further either in combination with crizotinib or alone in patients with ALK-positive NSCLC.
The discovery of the EML4-ALK translocation in a unique subset of patients with adenocarcinoma of the lung has quickly led to development of targeted therapies against the constitutively active tyrosine kinase. Trials are ongoing to determine the efficacy of upfront treatment with crizotinib in ALK-positive NSCLC. In addition, the Hsp90 pathway appears to be a promising target for ALK rearranged NSCLC, particularly in patients who develop resistance to crizotinib. The success of the identification of this oncogene and resultant therapies should emphasize the need to better delineate the heterogeneity of NSCLC. In a recent report, the NCI Lung Cancer Consortium documented that 60% of the 830 patients evaluated with adenocarcinoma had driver mutations . This will no doubt lead to additional targeted therapies to individualized therapy of patients with advanced NSCLC.
Supported by the Georgia Cancer Coalition Distinguished Cancer Scholar Award to Suresh S. Ramalingam.
R. N. Pillai: none; S. S. Ramalingam: consultant to Pfizer, Infiniti.