Increased ALK Gene Copy Number and Amplification are Frequent in Non-small Cell Lung Cancer

Abstract

Introduction

Translocation of the anaplastic lymphoma kinase (ALK) gene is involved in the tumorigenesis of a subset of non-small cell lung carcinomas (NSCLCs) and identifies patients sensitive to ALK inhibitors. ALK copy number changes and amplification, which plays an oncogenic role in tumors such as neuroblastoma, are poorly characterized in NSCLC. We aimed to study the prevalence of ALK copy number changes and their correlation to ALK protein expression, epidermal growth factor receptor (EGFR) status, and clinicopathological data in patients with NSCLC.

Methods

ALK status was evaluated by fluorescence in situ hybridization (FISH). Specimens with ALK translocation were studied for echinoderm microtubule-associated protein-like 4 (EML4), KIF5B, and TFG status. ALK expression was assessed by immunohistochemistry. EGFR gene and protein status were evaluated in adenocarcinomas. Survival analysis was performed.

Results

One hundred seven NSCLC cases were evaluated. There were two cases of EML4-ALK translocation and one with an atypical translocation of ALK. Both cases of EML4-ALK translocation had ALK protein expression, whereas in the rest, ALK was undetected. Eleven cases (10%) exhibited ALK amplification and 68 (63%) copy number gains. There was an association between ALK amplification and EGFR FISH positivity (p < 0.0001) but not with prognosis. In conclusion, EML4-ALK translocation is a rare event in NSCLC.

Conclusion

The study reveals a significant frequency of ALK amplification and its association with EGFR FISH positivity in lung adenocarcinomas. Based on these findings, a potential role of ALK amplification in the response to ALK inhibitors alone or combined with EGFR inhibitors in NSCLC merits further studies.

Lung cancer is the leading cause in the cancer-related deaths in developed countries.¹ Although the prognosis of this disease remains poor, some clinically relevant advances have occurred in recent years. The identification of epidermal growth factor receptor (EGFR) mutations and amplification in a subset of patients with non-small cell lung cancer (NSCLC), coupled with the development of EGFR tyrosine kinase inhibitors, has opened new ways to treat this disease.^2,3 Nevertheless, there is a need to find additional molecular targets to further improve patients’ outcome.

Recently, the novel fusion transcript with transforming activity, formed by the translocation of echinoderm microtubule-associated protein-like 4 (EML4) (2p21) and anaplastic lymphoma kinase (ALK) (2p23), has been described in a subset of NSCLCs.⁴ ALK was originally involved in carcinogenesis in anaplastic large cell lymphoma as part of a chromosomal rearrangement as a fusion partner of nucleophosmin.⁵ Other fusion partners to ALK have been described since then and have been associated with other tumor types such as myofibroblastic tumors, B-cell lymphomas, and squamous cell carcinomas of the esophagus.^6,7 The products of these translocations are fusion proteins with constitutively activated ALK tyrosine kinase, which plays a role in carcinogenesis by the aberrant phosphorylation of multiple intracellular substrates downstream of ALK-chimerical oncoproteins.⁸ In preclinical models of activated ALK, specific ALK targeting drugs inhibit cell proliferation and induce apoptosis in models “addicted” to this genetic aberration.⁹

Moreover, the analysis of clinical specimens in a number of studies suggests that this fusion gene may define a novel subclass of tumors within NSCLC characterized by distinct clinicopathological features in both Western and Asian populations.^{10 –13} Nevertheless, other reports have described cases that represent exceptions to this rule.¹⁴ Therefore, there are still many issues about the role of ALK in lung cancer that remain to be determined. Most importantly, the identification of EML4-ALK has provided clinicians with a novel potential therapeutic target as demonstrated by the favorable results of a phase I trial with a dual MET and ALK small molecule inhibitor that has showed promising clinical responses in patients harboring ALK translocations.¹⁵

Additional mechanisms of aberrant activation of ALK have been described in neuroblastomas. In these tumors, activating mutations and ALK amplification are the underlying mechanisms for ALK-dependent tumorigenesis^{16 –19} and are associated with a specific clinical phenotype in this disease. Nevertheless, copy number changes of ALK and their significance are poorly characterized in NSCLC and may represent an additional mechanism of activation of ALK and, therefore, play a role in the carcinogenesis of a subset of lung tumors.

In this work, we sought to study the status of the ALK gene by fluorescence in situ hybridization (FISH) on a population of white patients and correlate it with the status of EGFR (mutations and copy number variations) in adenocarcinomas. We also correlated the genetic findings with clinical outcome.

PATIENTS AND METHODS

Patient Population

The samples included in this study were obtained from patients attending the Lung Cancer Unit in Hospital del Mar for diagnosis and treatment. The only inclusion criterion was the availability of tissue for biomarker studies. Clinical details of these patients were included in a database. This project was approved by the local ethics committee (CEIC-IMAS2009/3619/I).

Fluorescence In Situ Hybridization

Four-micrometer paraffin-embedded histologic sections were used for FISH analysis. To assess the genetic status of EGFR and ALK, we used LSI EGFR/CEP7 and break-apart ALK (2p23) probes (Abbott Molecular Inc., Des Plaines, IL). To determine the fusion partner for ALK for cases with ALK translocation, we examined previously identified fusion partners EML4 (2p23), KIF5B (10p11.22), and TFG (3q12.2).^20,21 Break-apart probes were designed using bacterial artificial chromosome (BAC) clones selected from the CHORI BAC/PAC resource (http://bacpac.chori.org): centromeric EML4: pooled RP11-804P20 and RP11-413N9, telomeric EML4: pooled RP11-798D22 and RP11-34L01; centromeric KIF5B: pooled RP11-633K11 and RP11-281A19 in spectrum Red, telomeric KIF5B pooled RP11-460H18 and RP11-166N17; and centromeric TFG: pooled RP11-320C17 and RP11-423M9, telomeric TFG: pooled RP11-49H3 and RP11-168G7. FISH was performed as described previously.²² Results were analyzed in a fluorescent microscope (Olympus, BX51) using the Cytovision software (Applied Imaging, Santa Clara, CA). A minimum of 50 nuclei was scored.

Patients were classified as EGFR positive/negative according to criteria described in detail elsewhere.²³ As criteria for copy number aberrations of ALK has not been established, we arbitrarily used the following cutoffs adapted from the criteria established for EGFR and HER2 in lung cancer specimens.^23,24 Briefly, gain (including both low and high genomic gain) was defined as a mean copy number of 3 to 5 fusion signals in ≥10% of cells and amplification as the presence of ≥6 copies of ALK per cell in ≥10% of analyzed cells. In cases where clusters were observed, we reported the percentage of cells with clusters and considered amplified, cases with ≥10% of ALK clusters. On cases with ALK amplification, FISH with CEP2 (a centromeric alpha-satellite specific for chromosome 2) was performed (Abbott Molecular Inc.) to exclude polysomy. Normal fusion ALK, EML4, KIF5B, and TFG signals presented as an overlapping orange/red and green (yellowish) signals. These probes were considered typically rearranged when separated green and orange/red signals (at least by three times the signal diameter) were identified and atypically rearranged when a single orange or green signal was seen.

Immunohistochemistry for EGFR and ALK

Four micrometers serial sections of paraffin blocks were used for the evaluation of EGFR and ALK done with the DAKO Cytomation EGFR PharmDxTM and ALK (D5F3) antibody (Cell Signaling, Danvers, MA), respectively. This antibody for ALK is able to detect the wild type and truncated proteins. Diaminobenzidine was used as the cromogen to identify positive results. The scoring for EGFR IHC was done considering staining intensity, as recommended by the manufacturer, from negative: 0, no detectable staining; to 3+, strong and continuous membranous staining. For ALK expression scoring, the percent of cells was determined within each staining intensity category 0 to 4+, and a hybrid score (H score) was calculated by the formula: % cells of 0 intensity+ (% cells of 1 intensity × 1) + (% cells of 2 intensity × 2)+ + (% cells of 3 intensity × 3) + (% cells of 4 intensity × 4).²⁵ We used the NCI-H2228 ALK translocated cell line as a positive control for ALK expression purchased from the American Type Culture Collection (http://www.ATCC.org).²⁶

EGFR Mutation Detection

DNA was extracted from macrodissected tumoral paraffin-embedded tissue using the QIAamp Tissue Kit (QIAGEN GMBH, Hilden, Germany) according to the manufacturer’s protocol. Mutational analysis of exons 18 to 21 of the EGFR gene was performed by direct sequencing using tumoral DNA with BigDye v3.1 following the manufacturer’s instructions and analyzed on an ABI3730XLSequencer (Applied Biosystems, Foster City, CA). Primers were designed using the Primer Express software (Applied Biosystems) and are available on request.

Statistical Analysis

Statistical analysis was carried out with SPSS version 13.0 (SPSS, Inc., Chicago, IL). To analyze correlations between ALK status and clinical-pathologic variables and EGFR status, we used the χ² test or Fisher’s exact test. Survival analysis was performed by the Kaplan-Meier method. Curves were compared by the log-rank test. All the statistical tests were conducted at the two-sided 0.05 level of significance.

RESULTS

Patients’ Characteristics

One hundred seven patients were included in this study. Clinical-pathologic characteristics are detailed in Table 1. Stage I patients are overrepresented because tissue availability is higher in these surgical cases.

TABLE 1.

Patients’ Characteristics (N: 107)

Median age, yr (range)	66 (40–85)
Gender
Females	25 (23%)
Males	82 (77%)
Smoking status
Never	16 (15%)
Former	39 (36%)
Current	52 (49%)
Performance status (ECOG)
0–1	98 (92%)
2–3	9 (8%)
Stage
I	49 (46%)
II	6 (6%)
III	25 (23%)
IV	27 (25%)
Histology
Adenocarcinomaa	69 (65%)
Squamous cell carcinoma	30 (28%)
NSCLC NOS	2 (2%)
LCC and BAC	6 (5%)

^aIncludes two cases of combined carcinomas (adenocarcinoma with large cell neuroendocrine carcinoma).

NSCLC NOS, non-small cell lung cancer not otherwise specified; BAC, bronchioalveolar carcinoma (noninvasive tumor with lepidic spread); LCC, large cell carcinoma (exclusion diagnosis for a poorly differentiated tumor with absence of adenocarcinoma, squamous, or small cell component).

ALK Translocations

The results of the FISH analysis of ALK are listed in Table 2. We found three cases with ALK gene translocation (Figure 1). Table 3 lists the characteristics of the translocated cases. Two cases showed a typical gene breakage signal of ALK, and for these, the FISH signal for EML4 was also aberrant, consistent with the translocation involving both genes. The third translocated case showed an atypical rearrangement with a deletion of the 5′ region of ALK in 70% of cells. In this case, the EML4 signal was normal, suggesting the existence of another fusion partner to ALK. We investigated the presence of TFG and KIF5B and found no translocation of these genes. Of note, this case was a combined carcinoma with an area of adenocarcinoma and an area of large cell neuroendocrine carcinoma (LCNEC), and the ALK aberration was restricted to the LCNEC area. Both patients with EML4-ALK translocation were elderly at diagnosis (>70 years), one of them was a never smoker female with a mucinous adenocarcinoma and the other was a heavy ex-smoker male with a solid adenocarcinoma.

TABLE 2.

ALK FISH Results

Histology	ADC	SCC	NSCLC NOS	LCC and BAC	All
ALK FISH (N: 107)
Normal	10	6	1	1	18 (17%)
Gain	45	18	1	4	68 (63%)
Amplification or <10% clusters	11	6	0	1	18 (17%)
Translocation	3a	0	0	0	3 (3%)

^aOne case was a combined carcinoma: ADC + large cell neuroendocrine carcinoma.

ADC, adenocarcinoma; SCC, squamous cell carcinoma; NOS, not otherwise specified; LCC, large cell carcinoma; NSCLC, non-small cell lung carcinoma; BAC, bronchioaveolar carcinoma; FISH, fluorescence in situ hybridization; ALK, anaplastic lymphoma kinase.

FIGURE 1.

A, Left (arrows): atypical anaplastic lymphoma kinase (ALK) rearrangement with deletion of 5′ region of ALK (loss of green-labeled probe centromeric to ALK). Right: normal echinoderm microtubule-associated protein-like 4 (EML4) fusion pattern. B, Left (arrows): typical gene breakage signal of ALK (yellow wild-type allele, and single red and green probe signals for the rearranged allele). Right (arrows): EML4 typical probe breakage.

TABLE 3.

Characteristics of Patients with ALK Aberrations

Patient	ALK FISH	Percentage Cells	ALK IHC	Gender	Age (yr)	Smoking	Histology	Stage
106	EML4-ALK	64	POS	M	74	Former	ADC	pIA
997	EML4-ALK	66	POS	F	72	Never	ADC	IV(Pl)
215	?-ALK	70	NEG	M	74	Current	ADC	pIIIB
1259	Amplified	15	NEG	M	59	Current	SCC	IV(L)
495	Amplified	70	NEG	F	73	Never	ADC	pIIIA
692	Amplified	86	NEG	M	55	Current	ADC	pIA
1123	Amplified	20	NEG	M	50	Current	SCC	pIIIB
956	Amplified	10	NEG	M	53	Current	ADC	pIIB
806	Amplified	40	NEG	M	71	Former	SCC	pIA
278	Amplified	90	NEG	M	82	Former	SCC	IIIB
180	Amplified	80	NEG	M	79	Former	SCC	pIA
1432	Amplified	90	NEG	M	48	Current	ADC	pIA
1493	Amplified	22	NEG	M	65	Current	ADC	pIB
524	Amplified	52	NEG	F	53	Current	BAC	pIA
1370	Clusters	3	NEG	M	66	Former	ADC	pIA
1328	Clusters	5	NEG	F	70	Never	ADC	pIB
1119	Clusters	8	NEG	M	54	Current	ADC	pIIIA
1062	Clusters	8	NEG	M	77	Former	SCC	pIA
1024	Clusters	5	NEG	M	59	Current	ADC	pIA
839	Clusters	7	NEG	F	64	Never	ADC	pIA
246	Clusters	6	NEG	M	51	Current	ADC	pIV(B)

M, male; F, female; ADC, adenocarcinoma; SCC, squamous cell carcinoma; IHC, immunohistochemistry; FISH, fluorescence in situ hybridization; ALK, anaplastic lymphoma kinase.

ALK Copy Number Gains and Amplification

In our series, 11 cases (10%) showed amplification of ALK (Figure 2). Table 3 lists the characteristics of these patients along with seven cases (7%) with a small percentage of cells (<10%) showing clusters. Within the samples with amplification, five cases were adenocarcinomas, one was a bronchioalveolar carcinoma, and five had squamous cell histology. There was no statistical association between amplification and gender, age, smoking history, or histology. We observed a significant association between the presence of amplification and early stage (I–II) (p = 0.04). Notably, 9 of 11 cases with amplification were surgical cases and two cases with squamous histology showed locoregional and exclusively lung dissemination, respectively. The greater tissue availability in the surgical cases might perhaps allow detection of amplification in those specimens with a low percentage of positive cells.

FIGURE 2.

A, Anaplastic lymphoma kinase (ALK) amplification (orange/green fusion signals) in a cell with three copies of CEP2 (red signals). B, Polysomic cells from the same amplified case showing 3 to 4 copies of ALK and 3 to 4 copies of the chromosome 2 centromere.

We found ALK copy number gains in 63% of cases. In the majority of these cases, gains were seen in a high proportion of cells (50 –95%). No statistically significant association was found between ALK copy number gain and any of the clinical or pathologic parameters analyzed.

ALK Protein Expression by Immunohistochemistry

We analyzed the ALK protein expression by immunohistochemistry in 90 cases where tissue was available (including all translocated and amplified cases). We observed ALK expression in the two cases with a typical ALK translocation pattern that also showed a rearrangement of EML4 (Figure 3). The case with a loss of the centromeric probe was negative by IHC. In the remaining specimens, ALK protein expression was undetected with the assay used.

FIGURE 3.

Immunohistochemical staining analysis of anaplastic lymphoma kinase (ALK) protein expression. Pictures are at ×40 magnification. A, Case with ALK over-expression H score: 260. B, Case with ALK overexpression H score: 80. Both cases correspond to the tumors from patients with echinoderm microtubule-associated protein-like 4 (EML4)-ALK translocations by fluorescence in situ hybridization (FISH).

EGFR and ALK in Adenocarcinomas

EGFR protein expression (1+ to 3+) was seen in 45 of 64 patients (70%). Fifty-six patients were evaluated for EGFR by FISH. Of these, 28 (50%) were FISH positive (15% amplifications) and 28 (50%) negative. EGFR mutation status was studied in 68 patients, and mutation on exons 18, 19, and 21 were found in one (2%), six (9%), and four (6%) patients, respectively. These cases were characterized by the clinical features typical of these patients subgroup (female, nonsmokers). Regarding the association between EGFR status and ALK, we found a statistically significant association between ALK amplification and FISH positivity for EGFR (p value <0.001). No statistically significant association with EGFR mutations or protein overexpression was observed (Supplementary Table 1, http://links.lww.com/JTO/A38). Two of the cases with ALK clusters in a low percentage of cells harbored an EGFR mutation on exon 21. No EGFR mutations were found in the patients with ALK translocations.

Survival Analysis

We sought to explore the potential impact on survival of ALK aberrations. Median follow-up for the whole series was 10.6 months (0.1– 61.2). As the majority of patients with ALK amplification or <10% of clusters were stage I patients, we stratified all stage I patients according to ALK status in two manners. First, we analyzed amplified (six patients with ALK amplification) versus nonamplified (33 patients with copy number gains, four with disomic tumors, and five patients with <10% clusters), excluding the patient with ALK translocation. Second, as the definition of ALK amplification we used is not validated and we had observed clusters in a low proportion of cells, we grouped these patients with amplification cases (N: 11) versus nonamplified patients (N: 37). Because of the low number of patients in each group, we continued the analysis with the second approach (grouping amplification with clusters). Median follow-up for stage I patients was 11.5 months (1.2–37.2). Five patients within the stage I subgroup received adjuvant chemotherapy, three of them within the ALK amplified group and two of them in the nonamplified group. Mean overall survival was 32.4 months and 28.6 months for patients with ALK amplification(+clusters) and for those without ALK amplification, respectively (median OS was not reached for either group). This difference was not statistically significant (p = 0.4) (survival curves as Supplementary Figure 1, http://links.lww.com/JTO/A39). All three cases with translocation were alive without disease progression at the timing of writing this work. Interestingly, one of the patients with a pathologically confirmed stage IV adenocarcinoma (pleural metastases) harboring an EML4-ALK translocation is alive after 5 years of follow-up.

DISCUSSION

Lung cancer is a devastating disease, and the need of developing better treatments is an urgent need for our patients. Targeted therapy strategies have recently demonstrated the importance of selecting patients according to the biologic characteristics of the tumors.²⁷

The discovery of a new potentially relevant oncogenic event in lung cancer, the EML4-ALK translocation, and the development of ALK inhibitors with promising results in preclinical models and early clinical trials provides the rationale for the comprehensive characterization of this new genetic aberration in patients with lung cancer.

Our work confirms the low incidence (3%) of ALK translocation in an unselected population of patients diagnosed with NSCLC. Both typical translocations involving EML4 and ALK showed a good correlation to FISH and IHC. Nevertheless, the atypically rearranged case was not detectable by IHC, probably resulting from a lack of protein expression derived from this translocation. This stresses the need of further characterizing these atypical translocations and their predictive value for sensitivity to ALK inhibitors. Moreover, the ability of IHC to detect the protein expression in tumors harboring the typical EML4-ALK rearrangement supports the use of IHC with this sensitive antibody as an easy and inexpensive clinical tool for mass screening of patients with advanced NSCLC for this particular rearrangement. Regarding the clinical-pathologic characteristics of patients with ALK translocations, our findings are in contrast with previously reported ones, which associated ALK translocations with nonsmokers and young age patients.^11–13 The patients in our series were elderly, two of them with a smoking history and one had LCNEC histology, a distinct subtype of lung tumor which has not previously been reported harboring ALK aberrations. Globally, although the number of patients with translocations is low and we lack the statistical power to draw definitive conclusions, we believe that screening of this genetic aberration at this moment should include the general population of patients with NSCLC, maybe except for those with EGFR mutations that seem to be mutually exclusive with ALK translocations.

Interestingly, we found an unexpectedly high prevalence of ALK gene amplifications and copy number gains in NSCLC. This is the first study reporting such high frequency of this genetic aberration. Other reports that have evaluated large populations of patients with NSCLC by FISH have been published. One of them screened more than 600 cases²⁸ and reported a prevalence of 0.5% of ALK amplification. More recent works evaluating ALK gene status by FISH^10,11 do not report amplification in their series. These differences with our findings may be explained, at least in part, by the fact that the previous studies used tissue microarrays (TMAs) or small biopsies from patients with advanced disease for FISH evaluation. Supporting this hypothesis, we observed that the percentage of cells within the tumor containing the amplification or clusters is frequently low, which may explain the possibility of this genetic aberration being missed in a TMA sample or in a small biopsy. Accordingly, we detected ALK amplification or clusters in surgical specimens (except for two cases) where more tissue is available for analysis. Another reason for the difficulty of detection of these alterations on a TMA section is that the cells harboring amplification were scattered within the specimen. Nevertheless, ALK amplification could also represent an early genetic event in a subset of lung carcinomas. In neuroblastoma, focal high-level amplification has been described as an oncogenic event and selects cells sensitive to ALK inhibitors^16,19,29. The clinical significance of ALK amplifications and increased copy number is not known in lung cancer. The lack of ALK protein expression by our immunohistochemical assay, recently validated as a highly sensitive antibody for detection of ALK translocated cases³⁰ together with the low percentage of cells with amplification in most cases, suggests that amplification might not be a biologically relevant event or predict response to ALK targeting molecules. Furthermore, ALK is contained in regions already reported as copy number polymorphisms (http://projects.tcag.ca/variation/). This could explain the large percentage of cases with copy number gain in our series. In addition, the association of EGFR FISH positivity with ALK amplification may define a subset of cases with more frequent aneuploidy as a result of genomic instability. Nevertheless, lack of protein detection with this novel antibody in the cases with copy number gain and amplification does not definitely rule out the potential benefit of ALK inhibitors in this population, as demonstrated in colorectal patients without EGFR protein expression that do respond to therapeutic monoclonal antibodies targeting EGFR.³¹ Moreover, genetic alterations present initially in a small percentage of the tumor cells have shown to have a relevant biologic role in subsequent phases of disease (i.e., resistant to initial targeted therapy).³² Future studies will be needed to see whether high ALK gene copy number and/or amplification represent important association to sensitivity to ALK fusion-targeted therapies.

In summary, we confirm that ALK translocations are infrequent in patients with NSCLC. Immunohistochemical analysis optimally detects EML4-ALK translocations and may represent a clinically easy way to screen patients with NSCLC for this alteration. Copy number gains and amplifications are a frequent event in NSCLC and ALK amplification typically coexists with EGFR amplification. Although these findings lack prognostic significance in our series, our results pave the way to study the predictive role of ALK amplification in NSCLC to potentially expand the spectrum of patients that benefit from ALK inhibitors, alone or combined with EGFR inhibitors.