Homogeneity and High Concordance of ALK Translocation in Primary Lung Adenocarcinoma and Paired Lymph Node Metastasis

Translocation of anaplastic lymphoma kinase (ALK) gene is an important determinator for the response to ALK tyrosine kinase inhibitor (TKI) in non-small-cell lung cancer (NSCLC) patients. The existence of genetic heterogeneity will affect the results of molecular testing, especially in biopsy samples from primary or metastatic sites of patients with advanced stage NSCLC. We intended to explore the heterogeneity of ALK gene translocation in excision specimens and to examine the existence of discordance of ALK status between primary tumours and corresponding lymph node metastases. A total of 106 ALK positive lung adenocarcinoma cases were collected for assessment of intratumour heterogeneity of ALK gene translocation, which were stained by the fully automated Ventana ALK D5F3 immunohistochemistry (IHC) analysis. In addition, the ALK gene translocations were evaluated in a series of 53 primary tumours and their paired lymph node metastases using ALK D5F3 IHC staining. The concordance rate between primary tumours and paired metastatic lymph nodes was 100%. ALK status was homogeneous in lung adenocarcinoma samples and was generally stable during metastasis. Therefore, ALK gene translocation can be measured reliably in material from either primary or metastatic tumours in lung adenocarcinoma patients.

ALK gene translocation status in primary tumour cells. All 106 ALK-positive cases were successfully examined by ALK D5F3 IHC. All cases showed diffused cytoplasmic staining pattern in the section of FFPE tissue tumour samples, without diverse signal intensities (Fig. 1B). All ALK-positive cases in this study showed a homogeneous strong ALK-expression of all neoplastic cells.
ALK gene translocation status in the lymph node metastases. Fifty three patients (53/196, 27%), including 37 ALK-positive cases and 16 ALK-negative cases, with paired primary tumours and lymph node metastases samples were further investigated for concordance of ALK status. Among these 53 paired sample patients, pathological N1 disease was confirmed in 10 (18.9%) patients, N2 disease in 42 (79.2%) patients and   Table 3. Lymph node metastases from the 37 ALK positive cases were examined and ALK expression with diffused cytoplasmic staining pattern exhibited in all lymph node metastases (Fig. 1D). ALK expression was absent in the metastases of the 16 patients with no ALK expression in their primary tumours. No discordant case of ALK expression was observed between the primary tumours and their corresponding lymph node metastases ( Table 4).

Analysis of EGFR, KRAS, BRAF and HER2 genetic alterations in 106 ALK-positive cases.
Genotyping for EGFR, KRAS, BRAF and HER2 mutations was conducted with direct sequencing in the ALK-positive 106 cases. EGFR mutations were detected in 5 (4.7%) cases. However, KRAS, BRAF and HER2 mutations were not detected in any specimens. EGFR mutation types and locations are shown in Table 5. Three cases with deletions in exon 19, whereas two had mutations in exon 21 (L858R).  Table 2. Histologic characteristics of ALK-positive lung adenocarcinoma samples. Abbreviation: IMA, invasive mucinous adenocarcinoma.

Discussion
Tumour heterogeneity refers to the existence of subpopulations of cells with distinct genotypes and phenotypes that may harbour divergent biological behaviors, within a primary tumour and its metastases; alternatively the divergence is between tumours of the same histopathological subtype, even between spatially separated regions within single biopsies (intra-and inter-tumour, respectively) 8 . Investigation of tumour heterogeneity will provide valuable information for diagnostic and/or therapeutic procedures. The selection of patients for crizotinib therapy relies on the ALK status of the tumour 18 . Thus, accurate determination of ALK status is crucial to ensure the potential clinical benefit of ALK inhibitors and to avoid the toxic effects in inappropriately selected patients 3 . The most commonly treatment-related adverse reactions in crizotinib therapy were vision disorder, gastrointestinal disorders and oedema 19 . Majority of the ALK gene translocation status are evaluated in primary tumours of NSCLC and in practice, most clinical decision-making for patients with advanced NSCLC depends on single-tumour biopsy samples from primary or metastatic sites, even from cytological specimens 20,21 . Therefore, the biomarker status is assumed to remain unchanged during metastatic progression.  Table 3. Clinicopathologic characteristics of 53 paired cases.

Patients ALK translocation status in primary tumours
Patients ALK translocation status in metastases

No translocation Translocation
No translocation 16 0 Translocation 0 37  In the present study, we evaluated the intratumour heterogeneity of ALK gene translocation in 106 ALK-positive excision specimens by the Ventana ALK D5F3 IHC and also investigated whether the ALK status changed during disease progression in 53 pairs of primary tumours and corresponding lymph node metastases. We found the ALK expression was homogenous in lung adenocarcinoma samples and there was no discordant case of ALK status between primary tumours and corresponding lymph node metastases. Therefore, primary tumours can clearly represent the ALK status of the metastatic tumours. The ALK status remained unchanged during metastasis.
Metastatic advanced lung cancer continues to represent a major health problem worldwide. Targeted therapies are the standard therapeutic options for these patients. Approximately 70-85% of new lung cancers are unsuitable for surgical therapy; thus, only small biopsy specimens are available for diagnosis or further genetic detection 20 . By examining the ALK status in the primary tumours and corresponding metastatic tumours in our series, we propose that ALK gene translocation is a stable genetic alteration once it has occurred. This hypothesis has important clinical implications. Firstly, fine needle biopsy or aspiration of primary and/or metastatic tumours and pleural effusions are feasible for the detection of ALK gene translocations. Secondly, despite the marked antitumour activity of crizotinib, some ALK-positive patients developed resistant to crizotinib on average within the first year or two of TKI therapy 6 . In general, crizotinib resistance in ALK-positive NSCLC occurs through secondary resistance mutations located in the ALK TK domain, amplification of the ALK fusion gene, and activation of bypass tracks 6,22 . Therefore, we suggested that a change in ALK status between the primary tumour and metastasis was not the cause of crizotinib resistance.
Previous studies have demonstrated the intratumour heterogeneity of ALK gene translocation in lung adenocarcinoma samples. Cai et al. 23 acquired 45 spatially separated tumour cell subpopulations using laser-capture microdissection from 20 patients with ALK-FISH positivity. A total of 45 lesions were tested for ALK translocation by RT-PCR and 34 lesions were positive for ALK translocation, but 11 lesions were ALK wild-type regions. Abe et al. 24 observed a total of 64 tumour areas in nine ALK-FISH positive cases using darkfield fluorescence microscopy; positive ALK was found in 50 areas. However, in the current study with the Ventana ALK-D5F3 IHC, we did not detect intratumour heterogeneity of ALK status because ALK protein expression was diffusely positive in all neoplastic cells. Several studies considered that the intratumoural ALK gene translocation heterogeneity may be associated with technical problems. Camidge et al. 25 recorded the percentage of positive cells, pattern of positivity (split, single red, or both), and copy number of fused, isolated red and green signals by evaluating 90 ALK FISH positive NSCLC cases. They concluded that the semivertical orientation of tumour cells and/ or nuclear truncation in paraffin-embedded tissue sections may make precise identification of break-apart signals difficult. Proietti et al. 21 reported that FISH on conventional cytology, in which tumour cells are arranged in a single layer, have much higher positive rates for FISH than histology specimens. Moreover, ALK FISH using small biopsy sections cannot examine as many fields and cells as is the case with excision samples. RT-PCR, proposed as an alternative screening method for ALK gene translocation, is also not recommended for the detection of ALK translocation. RT-PCR can not be able to detect all fusion transcript variants. Additionally, RT-PCR is more stringent requirements for the sample preparation due to the rapid degradation of RNA.
The Ventana ALK D5F3 IHC detection system, measuring the expression of the ALK protein containing the C-terminal active kinase domain, is a sensitive method for detecting ALK gene translocation, especially in paraffin-embedded tissues, which was approved as a CE-IVD (in vitro diagnostic products) in Europe, China, and US since 2012, 2013 and 2015, respectively 7, 26-30 . We described a case with a negative FISH result that was later identified as ALK-positive by Ventana IHC and responded well to crizotinib. Targeted next generation sequencing revealed a new ALK partner gene (BIRC6) and the paracentric inversion for generating this fusion gene 31 . Wekken A. et al. 30 evaluated tumour response rate and survival after crizotinib treatment of 29 consecutive patients with ALK-positive advanced NSCLC diagnosed by FISH and/or Ventana ALK D5F3 IHC on small biopsies or fine needle aspirations (FNA). They concluded that Ventana ALK D5F3 IHC was superior to ALK-FISH on small biopsies and FNA to predict tumour response and survival to crizotinib for advanced NSCLC patients. Therefore, the Ventana IHC assay may be more sensitive than FISH for detecting ALK status in metastases and/or small biopsies which often contain only few tumour cells.
In this study, we found that a portion of lung adenocarcinoma patients with concomitant EGFR and ALK alterations. The dual-positive cases accounted for 4.7% (5/106) of ALK-positive lung adenocarcinoma samples using direct sequencing, which was similar to the ratio of 4.4% (4/91) reported by Won et al. 13 but lower than the ratio of 18.6% (12/70) reported by Yang et al. 17 ; both studies used the same detection method. Won et al. 13 reported that the frequency of co-alteration detection increase in the same cohort when sensitive detection methods for EGFR mutation are applied, such as real-time PCR, targeted NGS, and mutant-enriched NGS. We reviewed the literatures and summarized the relevant clinicopathological and molecular characteristics of ALK/EGFR dual-positive cases; majority of cases were adenocarcinomas in the advanced stage and mainly involved Asian patients, most EGFR mutations were deletions in exon 19 and point mutations in exon 21, and patients showed differential sensitivities to EGFR-TKI and/or ALK-TKI 13,14,16,17 .
Intratumour heterogeneity (ITH) has recently been elucidated in several cancer types with the use of next-generation sequencing (NGS) approaches 23,[32][33][34][35] . Clonal analyses by de Bruin et al. 34 , and Zhang et al. 33 reported substantial intratumour heterogeneity (ITH) within lung adenocarcinoma samples, by using multiregion whole-exome sequencing (WES) and/or whole genome sequencing (WGS). Recent evidence supports a model of trunk-branched clonal evolution leading to variable ITH and complex clonal architecture of tumours 33,35,36 . The clonal structure of a tumour is visually represented as a phylogenetic tree, the ubiquitous alterations present in all tumour regions map to the trunk, whereas heterogeneous events present in only some regions of the tumour map to the branches. The findings of homogeneous ALK gene translocation in lung adenocarcinoma samples, and concordant ALK status between primary and lymph node metastasis in this study, suggested that ALK gene translocation is a ubiquitous event, must occur very early in lung adenocarcinoma pathogenesis 33 . The finding of ALK/EGFR coaltered lung adenocarcinoma cases may be explained by clonal evolutionary dynamics and the resulting complex clonal architecture of lung adenocarcinoma samples. In addition, in this study, we also demonstrated that ALK-positive cases more commonly showed lymph node metastasis and presented at higher clinical stage, and ALK gene translocations were significantly more common in lung adenocarcinoma samples with micropapillary and solid predominant patterns which associated with poorer prognosis 37 . We confirmed that tumour cells with ALK gene translocations contained enhanced metastatic potential and acquired the metastatic phenotype.

Conclusion
In summary, using the Ventana (D5F3) IHC, we found the homogeneity of ALK expression in lung adenocarcinoma samples and concordance in ALK status between primary tumours and corresponding lymph node metastases. Molecular testing for ALK translocation may be performed in either primary tumour or lymph node metastasis samples from the same patient. However, the main limitation in our study is that the metastases did not include different metastatic sites, especially distant metastases. Loco-regional or concurrent regional lymph node metastases are not considered to be biologically equivalent to distant metastases 9 . We also found a small portion of lung adenocarcinoma samples have concomitant EGFR and ALK alterations by using direct sequencing. The clinical relevance of these concurrent alterations remains to be elucidated in future studies.

Methods
Patients and tumour samples. All formalin-fixed and paraffin-embedded (FFPE) tissue sections were obtained from patients with histologically confirmed primary lung adenocarcinoma. The patients previously underwent curative surgery at the Cancer Hospital, Chinese Academy of Medical Sciences, Beijing, China, between February 2013 and February 2015. Written informed contents were obtained from all subjects before collecting the samples. All the methods were carried out in accordance with the institutional guidelines and approved by the Ethical Review Committee of the Cancer Hospital, Chinese Academy of Medical Sciences, Beijing, China. Tumours were immunoassayed with ALK D5F3 antibody as described below. A total of 106 ALKimmunopositive cases were collected. Ninety ALK-immunonegative lung adenocarcinoma cases were randomly selected and included in the study as controls.
The ALK status between primary tumours and their metastases was compared in 37 ALK-immunopositive and 16 ALK-immunonegative cases, as well as their corresponding metastatic lymph nodes. Thirty-seven ALK-positive cases with their corresponding metastatic lymph nodes were selected among 106 ALK-positive cases, meanwhile, 16 ALK-negative cases with their corresponding metastatic lymph nodes were selected among 90 ALK-negative cases. Our study included only metastatic tumour tissues with diameters greater than 0.5 cm from lymph nodes. The lymph node metastases of 53 cases were conducted into three tissue microarrays (TMA). Lymph node metastases were sampled by collecting 2.0 mm-diameter cores from two different representative sites based on hematoxylin and eosin (HE) stained sections.
Age, gender, cancer stage, smoking status and treatment of all patients were recorded. All cases had one slide stained for hematoxylin and eosin to confirm the presence of adequate tumour tissue. For each case, multiple slides corresponding to whole tissue sections were reviewed by two pathologists according to the 2011 International Association for the Study of Lung Cancer (IASLC)/American Thoracic Society (ATS)/European Respiratory Society (ERS) International Multidisciplinary Classification of Lung Adenocarcinoma and the American Joint Committee on Cancer (7th edition) of tumour, node and metastasis staging criteria 38,39 . In mixed-subtype adenocarcinoma samples, we assessed the percentage of each histological pattern (Lepidic, Acinar, Papillary, Micropapillary, Solid) in 5% increments and recorded the predominant histological pattern 38 . Cases with differences between the two reviewers were reevaluated and a consensus interpretation was rendered. None of patients had received prior ALK or EGFR TKI therapy.
Ventana IHC staining and scoring. ALK-IHC was performed on 4 μm-thick formalin fixed, paraffin-embedded tissue sections or TMA slides using the Ventana ALK D5F3 CDx assay on a Ventana Benchmark XT automated slide-processing system (Ventana Medical Systems Inc., Tucson, AZ). Briefly, slides of lung adenocarcinoma tumours were subjected to deparaffinization using EZ Prep (Ventana Medical Systems Inc.). Tissue sections were incubated with anti-ALK antibody (clone D5F3, Ventana Medical Systems Inc.) for 20 min. Optiview DAB IHC detection kit (Ventana Medical Systems Inc.) and Optiview amplification kit (Ventana Medical Systems Inc.) were used according to the manufacturer's recommendations for the visualization of the bound primary antibody 27 . Tissue slides were counterstained with hematoxylin II and Bluing Reagent (Ventana Medical Systems Inc.). An ALK-positive cell lines embedded in agar/FFPE or sections of normal appendix containing ALK-positive ganglion cells, were used as ALK-IHC external controls in each run (Ventana Medical Systems Inc.). Two consecutive 4μm-thick FFPE tissue slides was cut, one slide was used in ALK-D5F3 IHC analysis, and the other for routine negative control staining for a matched rabbit monoclonal negative antibody (Ventana Medical Systems Inc.). For evaluating the staining results, a binary scoring system (positive or negative for ALK status) was used (package insert for VENTANA anti-ALK D5F3 Rabbit Monoclonal Primary Antibody, Cat.N0.790-4794/06679072001). Tumour cells with strong granular cytoplasmic staining (any percentage of positive tumour cells) were designated as ALK-positive, whereas tumour cells without strong granular cytoplasmic staining were designated as ALK-negative. Negative quality control sections were first evaluated for lack of staining 27 . Considering that Ventana ALK D5F3 IHC produced more intense cytoplasmic signals because of excessive chromogen deposition, which resulted in false-positive staining 40 , the ALK expression in our study was assessed independently by one trained scientist (J.Y) and one pathologist (N.L).