SEARCH

SEARCH BY CITATION

Keywords:

  • epidermal growth factor receptor (EGFR);
  • anaplastic lymphoma kinase (ALK);
  • lung cancer;
  • cytological analysis;
  • Russia

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SUPPORT
  8. CONFLICT OF INTEREST DISCLOSURES
  9. REFERENCES

BACKGROUND

Although the molecular analysis of epidermal growth factor receptor (EGFR) and anaplastic lymphoma kinase (ALK) in archived lung cancer tissues is relatively well established, the genetic testing of cytological material has not yet become a routine.

METHODS

The current study used cell samples that were obtained by bronchial brushing, transthoracic needle aspiration, or biopsy imprint preparation between 1993 and 2008. Islets of malignant cells were visually located on the archived cytological slides, lysed in situ by a drop of sodium dodecyl sulfate-containing buffer, and subjected to the standard DNA and RNA extraction. Examination of paraffin-embedded tissue blocks (resection specimens or biopsy material) from the same patients was performed in parallel.

RESULTS

A total of 75 cytological/histological lung adenocarcinoma sample pairs underwent polymerase chain reaction analysis for the EGFR mutation. Two cytological samples and 1 morphological sample failed to produce DNA. Concordance for the wild-type and mutation status was observed in 54 of 72 and 14 of 72 informative pairs, respectively; 3 pairs and 1 pair, respectively, had mutation only in the cytological or histological material. The discrepancies could be explained by the failure to ensure a high percentage of lung cancer cells in the analyzed samples or, alternatively, by the genuine intratumoral molecular heterogeneity of some neoplasms. RNA extraction followed by reverse transcriptase-polymerase chain reaction analysis for the EML4-ALK translocation was performed for 44 EGFR mutation-negative sample pairs; failures were observed for 2 cytological and 6 histological specimens. All informative pairs were concordant either for the norm (32 of 36 pairs) or for the presence of EML4-ALK gene fusion (4 of 36 pairs).

CONCLUSIONS

Archived cytological slides appear to be well suited both for EGFR and ALK analysis. Cancer (Cancer Cytopathol) 2013;121:370–376. © 2013 American Cancer Society.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SUPPORT
  8. CONFLICT OF INTEREST DISCLOSURES
  9. REFERENCES

The incorporation of tyrosine kinase inhibitors (TKIs) in the management of metastatic lung cancer may be regarded as one of the most exciting success stories in medical oncology. Early clinical trials on the epidermal growth factor receptor (EGFR) TKIs gefitinib and erlotinib were based on the finding that virtually all non-small cell lung tumors demonstrate evident overexpression of EGFR. Contrary to some initial expectations, the overall response rates in nonselected patients with lung cancer was found to be relatively low; however, a fraction of treated patients experienced an indeed fascinating improvement in their health status.[1] Subsequent analysis of tumor DNA in responders led to the identification of previously unknown TKI-sensitizing mutations within the EGFR gene.[2-4] The EGFR mutation test has an unprecedented predictive accuracy: nearly all mutation-positive patients clearly benefited from the drug either by achieving an objective response or prolonged disease stabilization, whereas the shrinkage of mutation-negative tumors after treatment with gefitinib or erlotinib is exceptionally rare.[5, 6] Another target for TKIs, the echinoderm microtubule-associated protein-like 4-anaplastic lymphoma kinase (EML4-ALK) translocation, was discovered in the year 2007, when some therapeutic ALK inhibitors were already available for clinical trials. A subsequent study with the ALK-specific TKI crizotinib demonstrated disease control in approximately 90% of patients carrying ALK gene rearrangements.[7, 8]

EGFR and ALK tests have become a mandatory part of the diagnosis of lung cancer. While the molecular analysis of archived tumor tissues is relatively well established, the adaptation of the genetic investigation of cytological material into routine clinical use remains a subject for validation.[9] Some institutions already use the molecular analysis of cytological material in a regular diagnostic setting.[10-12] This practice is based on a significant number of reports demonstrating the suitability of cytological preparations for detection of the EGFR mutation[10-31]; however, only a few of these studies used paired histological (control) samples and included tumor series with a considerable frequency of mutations.[18, 24, 30, 31] To our knowledge, the results of ALK analysis in archived cytological samples have not been presented to date.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SUPPORT
  8. CONFLICT OF INTEREST DISCLOSURES
  9. REFERENCES

Both EGFR mutations and ALK gene rearrangements occur mainly in patients with adenocarcinomas,[32] and therefore we considered only lung cancer of this histological type for the current study. We examined the clinical records of the N.N. Petrov Institute of Oncology in St. Petersburg, Russia for the years 1993 through 2008, and identified 102 cases of lung adenocarcinoma that were evaluated by both morphological and cytological examination. We submitted the request for paraffin-embedded tissue blocks and archived cell smears, and received 65 sample pairs; in the remaining cases, either histological or cytological material could not be provided for the study because it was returned to the patient, forwarded to another hospital, or absent for other reasons. In addition, we considered 10 patients for whom only cytological slides were available. Although their histological samples were not available at the time of the current investigation, they had already been tested for EGFR status some years earlier, either in research[33] or within the diagnostic setting. Given that exactly the same methodology for detection of the EGFR mutation was used for these tissue blocks, we decided that these 10 cases could be added to the EGFR study (Table 1). The subsequent EML4-ALK translocation analysis retained only 44 cytological/histological doublets; 18 lung cancer cases were intentionally excluded because of the presence of an EGFR mutation,[34] 3 cases were considered noninformative because of failure of the EGFR test, and 10 pairs were left out of the reverse transcriptase-polymerase chain reaction (RT-PCR) analysis because of the above-mentioned lack of histological material (Table 1).

Table 1. Success Rate for DNA and RNA Isolation From Cytological and Histological Lung Cancer Samplesa
DNA/RNACytological SamplesHistological SamplesNo of Informative Sample Pairs
  1. a

    All samples were taken from primary lung lesions. Cytological material was obtained by bronchial brushing (48 cases), transthoracic needle aspiration (18 cases), or biopsy imprint preparation (9 cases). Histological samples were represented either by surgical resection specimens (38 cases) or biopsy material (27 cases); 10 tissue samples were not available at the time of this study, but were tested for the epidermal growth factor receptor (EGFR) mutation prior this investigation.

DNA isolation (n=75)73 (97%)74 (99%)72 (96%)
RNA isolation (n=44)42 (95%)38 (86%)36 (82%)

Archived hematoxylin and eosin-stained ethanol-fixed cell smears were obtained by bronchial brushing (48 cases), transthoracic needle aspiration (18 cases), or biopsy imprint preparation (9 cases). They were inspected by a cytologist, and the area containing predominantly tumor cells was labeled by a marker. DNA/RNA isolation was performed using 1 slide per case. A drop of the lysis buffer (1-× TE [pH 8.0] and 2% sodium dodecyl sulfate) was loaded onto the cells, pipetted back and forth, and collected into an Eppendorf tube (Eppendorf AG, Hamburg, Germany); subsequently, another drop was loaded onto the same spot. The volume of lysate in the tube was adjusted to 200 μL, and proteinase K digestion (500 ng/μL) was performed for 2 hours at 60°C.

While extracting nucleic acids from histological material, we used 2 to 3 20-μm thick sections of the tumor-containing areas of the tissue block. These sections were deparaffinized in 2 changes of xylene, washed by ethanol, and air-dried before the addition of 200 μL of the standard lysis buffer. The proteinase K digestion lasted for 6 hours.

Cell or tissue lysates were subsequently subjected to 2 rounds of organic extraction. The first extraction included equal volume of the TRIzol reagent (Life Technologies, Carlsbad, Calif) and 0.5 volume of chloroform. A second extraction was performed with a 0.5-volume of chloroform. Nucleic acids were precipitated by the addition of 2 volumes of ethanol, 0.1 volume of 3 M sodium acetate, and glycogen carrier up to 0.1 μg/μL and dissolved in 10 μL (for cytological material) or 40 μL (for histological material) of water. Approximately 10 μL of the sample (the entire amount of the former and a part of the latter) was subjected to the RT reaction in a final volume of 20 μL, and was eventually used as a template both for DNA and RNA (cDNA) analysis.

All reactions were performed without multiplexing (ie, only 1 target sequence was subjected to PCR amplification). PCR cocktails for the gel-based detection of the product included 1 μL of template, 0.5 U of hot-start polymerase, 1-× of PCR buffer (pH 8.3), 2.0 to 2.5 mM of magnesium chloride, 200 μM of deoxynucleoside triphosphates (dNTPs), and 0.3 μM of primers in a total volume of 10 μL. Real-time PCR had a slightly different mix composition: it contained 1 U of hot-start polymerase, was run in the presence of 0.2-× SYBR Green I, and had a total volume of 20 μL. Each reaction was initiated by the activation of Taq polymerase at 95°C for 10 minutes. The primers and PCR conditions for detection of the most common EGFR mutations and EML4-ALK translocations are summarized in Table 2. The quality of genomic DNA was assessed directly during the EGFR test (Table 2). The quality of cDNA was evaluated by PCR with primers specific to the succinate dehydrogenase complex, subunit A (SDHA) gene[35]; failed samples were excluded from the subsequent EML4-ALK analysis (Table 2). The example of EGFR mutation detection was presented in our earlier report.[33] EML4-ALK testing is exemplified in Figure 1.

image

Figure 1. Detection of the echinoderm microtubule-associated protein-like 4-anaplastic lymphoma kinase (EML4-ALK) translocation is shown. (A) Polymerase chain reaction amplification with EML4-ALK V.1 fusion-specific primers is demonstrated. Samples with a translocation produced a 134-base pair (bp) fragment (lane 3 and positive control [+]). (B) Control amplification is shown (SDHA [succinate dehydrogenase complex, subunit A, flavoprotein variant]-specific fragment, 165 bp; lanes 1-5 and +). M indicates molecular weight marker; lanes 1 to 5, analyzed samples; NEC, no enzyme control; NTC, no template control.

Download figure to PowerPoint

Table 2. PCR Analysis Of EGFR Mutations And EML4-ALK Translocations
Gene/MutationPrimersPCR ConditionsDetection Method
  1. Abbreviations: bp, base pairs; EGFR, epidermal growth factor receptor; EML4-ALK, echinoderm microtubule-associated protein-like 4−anaplastic lymphoma kinase; PAGE, polyacrylamide gel electrophoresis; PCR, polymerase chain reaction; SDHA, succinate dehydrogenase complex, subunit A, flavoprotein variant.

EGFR exon 19 deletion (ex19del)Forward: CTGTCATAGGGACTCTGGATReverse: CAGCAAAGCAGAAACTCACAT45 cycles (denaturation: 15 s, 95°C; annealing: 30 s, 55°C; synthesis: 30 s, 72°C)PAGE: normal fragment: 127 bp; mutation: lower size (112 bp for the most common 15-bp deletion)
EGFR L858RWild-type: CACCCAGCAGTTTGGCCAMutant: CACCCAGCAGTTTGGCCCCommon: GCATGAACTACTTGGAGGAC50 cycles (denaturation: 15 s, 95°C; annealing: 30 s, 63°C; synthesis: 30 s, 72°C)Allele-specific real-time PCR: 120 bp
V1 (EML4 exon 13/ALK exon 20)EML4: TGGAGCAAAACTACTGTAGAGALK: GTCGAGGTGCGGAGCTTG45 cycles (denaturation: 15 s, 95°C; annealing and synthesis: 60 s, 60°C)PAGE: 134-bp fragment
V2 (EML4 exon 20/ALK exon 20)EML4: CTAACTCGGGAGACTATGAAATALK: GTCGAGGTGCGGAGCTTG45 cycles (denaturation: 15 s, 95°C; annealing and synthesis: 60 s, 60°C)PAGE: 118-bp fragment
V3a/b (EML4 exon 6/ALK exon 20)EML4: CATAAAGATGTCATCATCAACCAALK: GTCGAGGTGCGGAGCTTG45 cycles (denaturation: 15 s, 95°C; annealing and synthesis: 60 s, 60°C)PAGE: 113-bp (V3a) or 146-bp (V3b) fragment
SDHA (RNA quality control)Forward: ACTGGCCACTCGCTATTG Reverse: CTTTGCTCTTATGCGATGGA50 cycles (denaturation: 15 s, 95°C; annealing: 30 s, 60°C; synthesis: 30 s, 72°C)PAGE: 165 bp

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SUPPORT
  8. CONFLICT OF INTEREST DISCLOSURES
  9. REFERENCES

DNA and RNA Isolation

The results of the nucleic acids extraction are summarized in the Table 1. Both ethanol-fixed cytological preparations and formalin-fixed histological samples produced reasonably high success rates for both DNA and RNA. The lowest value (38 of 44; 86%) was observed for the RNA isolated from archived paraffin-embedded tissue blocks. Cytological preparations yielded in average of 0.42 μg of nucleic acids (range, 0.17 μg-0.88 μg); corresponding estimates for histological material approached 3.74 μg (range, 0.50 μg-12.25μg).

EGFR Analysis

EGFR analysis yielded concordant results in 68 of 72 informative DNA sample pairs (94%) (Table 3). Concordance for mutational status was documented in 14 pairs, including 9 exon 19 deletions and 5 L858R substitutions. Discordant results were observed in 4 cases: 3 pairs demonstrated the mutation in cytological but not in histological samples (1 exon 19 deletion [ex19del] allele and 2 L858R alleles), whereas 1 doublet demonstrated the L858R allele only in the tissue-derived DNA. Testing of discordant samples by highly polymorphic genetic markers confirmed that each member within the pair belonged to the same individual (data not shown) (ie, a trivial confusion of the specimens could not explain the discrepancies between cytological and histological tests).

Table 3. EGFR and EML4-ALK Testing in Cytological and Histological Lung Cancer Samples
TestEGFR Mutations (n=72)EML4-ALK Translocations (n=36)
  1. Abbreviations: EGFR, epidermal growth factor receptor; EML4-ALK, echinoderm microtubule-associated protein-like 4−anaplastic lymphoma kinase.

Concordant pairs  
Wild-type/wild-type54 (75%)32 (89%)
Mutation/mutation14 (19%)4 (11%)
Total68/72 (94%)36 (100%)
Discordant pairs  
Cytology: mutation/histology: wild-type3 (4%)0
Cytology: wild-type/histology: mutation1 (1%)0
Total4 (6%)0

Possible causes of the 4 discordant results observed in the EGFR mutation test were subsequently reviewed on a case-by-case basis. Controversy for the first discordant pair (cytology: ex19del allele; histology: wild-type) was resolved simply by the repetition of molecular analysis: we performed PCR using an increased amount of tissue-derived DNA template and were able to detect a tiny ex19del-specific product. Two discordant pairs (cytology: L858R allele; histology: wild-type) allowed for the re-evaluation of the DNA source. Repetition of tumor cell microdissection and EGFR analysis confirmed the presence of the L858R allele in the tissue sample from one controversial case. Morphological re-evaluation of the remaining tissue material from another discordant pair revealed a prohibitively low percentage of the tumor cells. Visual re-examination of the fourth discordant doublet (cytology: wild-type; histology: L858R allele) was not possible due to lack of additional tumor cell smears.

ALK Analysis

EML4-ALK translocations were revealed in 4 of 36 EGFR mutation-negative cases (11%), and included 2 V1, 1 V2, and 1 V3a/b fusion variants.[36] Cytological and histological samples produced concordant results in all 36 pairs (Table 3). Given that, to the best of our knowledge, this is the first report on ALK translocations in Russian patients with lung adenocarcinoma, we considered sex and age characteristics of the analyzed cases. The study included 20 males, 6 of whom were aged≤50 years; EML4-ALK translocations were detected in 2 elderly men (aged 67 years and 73 years, respectively). Four of 16 women examined were aged<50 years, and 2 of these early onset cases contained the EML4-ALK gene fusion; it is interesting to note that this translocation was detected in the youngest participant of the study, a woman aged 30 years.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SUPPORT
  8. CONFLICT OF INTEREST DISCLOSURES
  9. REFERENCES

The current study has several strengths compared with the majority of previously published investigations.[10-31] In particular, both cytological and histological samples from each included patient were used, thus allowing for the direct comparison of the suitability of these types of biological material for molecular testing. Furthermore, several prior reports examined nonselected patients with lung cancer and therefore were able to reveal only a few mutation-positive samples within their data sets; this lung cancer series was intentionally limited to lung adenocarcinomas, which were shown to have a reasonably high frequency of EGFR mutations.[33] Therefore, the numbers of cases with normal and altered gene status were relatively well balanced. Most importantly, the analysis included “naturally” archived cytological slides and tissue blocks, which were stored for a prolonged period of time; it is worth mentioning that the examined cell smears were not protected by coverslips.

Although the detection of EML4-ALK in newly prepared cytological material has already been reported,[37, 38] the current study has demonstrated the suitability of retrospective cell smears for ALK testing. The current investigation used the RT-PCR technique, which is inexpensive and capable of detecting the fusion gene in a small number of cells but, contrary to the US Food and Drug Administration-approved fluorescence in situ hybridization test, cannot reveal some rare variants of ALK rearrangements.[39] Recently Wang et al[40] suggested a novel RT-PCR approach, which is based on the identification of the differences in the RNA expression level between 5′ and 3′ portions of ALK gene, and has the potential to reveal all activating ALK translocations.

The data in the current study confirm that the performance of cytology-based molecular testing is at least as effective as histology-based analysis, both with regard to success rates for DNA/RNA isolation and the ability to detect gene alterations (Tables 1 and 3). Virtually all prior studies in the field have led to similar conclusions[10-31, 37, 38]; however, the suitability of cytological material for routine molecular testing is still under discussion.[9] In particular, publication bias may play a serious role in this type of investigation; those research groups who fail to perform a genetic analysis of cytological samples are very unlikely to report their negative experience in the form of a scientific paper. It also must be acknowledged that cytological specimens often produce a lower yield of nucleic acids than tissue sections and this limitation may become critical when multiple DNA/RNA tests are required for proper treatment planning. Furthermore, many archives retain only a single smear for each individual patient; in contrast to histological blocks, this slide cannot be easily shared between various diagnostic procedures, and therefore the molecular testing may destroy the last source of tumor cells.

The results of the current study have demonstrated good concordance between cytology-based and histology-based testing for both EGFR and ALK (94% and 100%, respectively). The apparently better performance of ALK analysis can be explained by the design of the appropriate assay, which is aimed at detecting a chimeric transcript and is capable of amplifying a fusion PCR fragment even in the presence of a significant excess of normal cells. In contrast, the design of the EGFR test that was used required the presence of at least 5% of mutated DNA in the analyzed sample (unpublished data). Indeed, gel-based detection of the ex19del allele, although highly specific and able to identify all clinically relevant variants of EGFR deletions in a single PCR reaction,[33] may fail to reveal the ex19del-specific band in the case of a prohibitively small percentage of mutated cells. The detection of the L858R allele may be even more complicated, given that the mutated and normal EGFR sequences differ by only 1 nucleotide. As result, although the overall concordance between histological and cytological EGFR tests was formally high (68 of 72 pairs), it has to be acknowledged that as many as 4 of 18 doublets from apparently mutated tumors (22%) produced discordant results. The discrepancies are most likely to be the result of the failure to ensure a high percentage of lung cancer cells in the analyzed samples or, alternatively, may reflect a genuine intratumoral molecular heterogeneity of some neoplasms. Occasional instances of discordant EGFR status were repeatedly documented in studies using multiple tumor specimens from the same patient.[41] The mere existence of these situations may suggest the need to examine more than 1 sample from a patient, regardless of whether the examples of conflicting results of EGFR testing are attributed to errors in tumor cell enrichment or to specific biological properties of lung adenocarcinomas. One also may comment that all guidelines for molecular testing are likely to be subjected to major revision in the near future due to rapid advances in next-generation sequencing technologies.[42]

It is interesting to note that the frequency of ALK translocations in EGFR mutation-negative adenocarcinoma samples from Russian patients (4 of 36; 11%) appears to be slightly above the estimates observed in other patient series.[8] Similarly, our earlier investigation revealed that Russian patients with lung adenocarcinoma demonstrated the highest occurrence of EGFR mutations among white individuals.[33] It is important to consider that the majority of smokers in Russia consumed high-tar cigarettes in the past and therefore developed squamous cell lung cancer. Compared with Europe, North America, and some Asian countries, in which smokers clearly prevail among patients with all histological subtypes of lung cancer, nonsmokers constitute at least one-half of adenocarcinoma cases in Russia.[33] EGFR mutations and ALK translocations are strongly associated with abstinence from tobacco products,[2-8] and therefore the enrichment of Russian lung adenocarcinoma series by nonsmokers appears to be a plausible explanation for the elevated frequency of these lesions. Further studies are needed to validate this assumption.

FUNDING SUPPORT

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SUPPORT
  8. CONFLICT OF INTEREST DISCLOSURES
  9. REFERENCES

Supported by the Russian Foundation for Basic Research (grants 11-04-01643, 12-04-01490, and 12-04-00928), the Russian Federal Agency for Science and Innovations (contract 16 512 11 2237), the Presidential program for support of young Scientists (grants 768.2012.4 and 790.2012.4), and the Government of Moscow (grant 15/12).

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. FUNDING SUPPORT
  8. CONFLICT OF INTEREST DISCLOSURES
  9. REFERENCES