ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2016

Digital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Medicine 1229

An integrative strategy for targetedevaluation of biomarker expressionin non-small cell lung cancer

JOHANNA MATTSSON

ISSN 1651-6206ISBN 978-91-554-9584-8urn:nbn:se:uu:diva-285844

Dissertation presented at Uppsala University to be publicly examined in Fåhraeussalen,Rudbecklaboratoriet, Dag Hammarskjöldsväg 20, Uppsala, Saturday, 11 June 2016 at 09:15for the degree of Doctor of Philosophy. The examination will be conducted in English.Faculty examiner: Åslaug Helland (Department of Cancer Genetics, Oslo University).

AbstractMattsson, J. 2016. An integrative strategy for targeted evaluation of biomarker expressionin non-small cell lung cancer. Digital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Medicine 1229. 58 pp. Uppsala: Acta Universitatis Upsaliensis.ISBN 978-91-554-9584-8.

Despite improvements in therapy, the prognosis for non-small cell lung cancer (NSCLC)patients remains poor, and cure is only possible in localized tumors after surgical resection. Anew generation of targeted cancer drugs has led to the expectation that lung cancer therapy canbe significantly improved, but these drugs are today only an option in a small subset of NSCLCpatients, and their effect is temporary. Therefore, the aim of this thesis was to characterizeNSCLC in order to find new treatment targets and to evaluate biomarkers that further optimizetherapy selection.

In Paper I, the expression of the potential treatment targets claudin 6 and claudin 18.2were evaluated based on immunohistochemical- and gene expression analysis. High ectopicprotein and gene expression were demonstrated for both claudins in small subgroups of NSCLC.Clinical trials using humanized monoclonal antibodies against both proteins are ongoing in othercancer forms and may be extended to NSCLC.

In Paper II, the prognostic impact of the inflammatory mediator cyclooxygenase 2 (COX-2)was evaluated. No prognostic significance was found in a meta-analysis incorporating geneexpression data of 1337 NSCLC patients. Likewise, COX-2 protein expression in tumor cellswas not associated with survival in two independent NSCLC cohorts. However, in one of theanalyzed cohorts, higher COX-2 expression in the tumor stroma was associated with longersurvival and may therefore be a subject for further investigation.

In Paper III, tumor and stromal COX-2 protein expression was examined in patients treatedwith the COX-2 inhibitor celecoxib in order to evaluate if COX-2 expression is a predictivebiomarker for benefit of celecoxib therapy. Celecoxib did not prolong overall survival neitherin the whole cohort nor in patients stratified according to COX-2 expression in tumor or stromalcells. Noteworthy, a tendency towards longer survival was again demonstrated in patients withhigh COX-2 stromal expression.

In Paper IV, the diagnostic methods for identification of ALK rearrangements were assessedin a large representative Swedish NSCLC population. Fluorescence in situ hybridization (FISH),as the diagnostic standard, was compared to two immunohistochemical assays. ALK geneexpression levels were incorporated to supplement the molecular data. The frequency of ALKrearrangements was lower than previously reported. The different methods to detect the ALKfusion demonstrated overlapping results. However, the overlap was poor, so the methods cannotbe regarded as interchangeable and should thereby be interpreted with caution when used inclinical diagnostics.

In summary, this thesis applied an integrative translational approach to characterize potentialnew treatment targets and to evaluate the detection of existing predictive biomarkers in NSCLC.

Keywords: non-small cell lung cancer, prognostic biomarkers, predictive biomarkers,immunohistochemistry, gene expression, COX-2, claudin, ALK

Johanna Mattsson, Department of Immunology, Genetics and Pathology, Molecular andMorphological Pathology, Rudbecklaboratoriet, Uppsala University, SE-751 85 Uppsala,Sweden.

© Johanna Mattsson 2016

ISSN 1651-6206ISBN 978-91-554-9584-8urn:nbn:se:uu:diva-285844 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-285844)

Till mamma och pappa

Supervisors and Examining Board

Main supervisor Patrick Micke, M.D., Associate Professor Department of Immunology, Genetics and Pathology Uppsala University, Uppsala

Assistant supervisors Johan Botling, M.D., Associate Professor

Department of Immunology, Genetics and Pathology Uppsala University, Uppsala

Fredrik Pontén, M.D., Professor Department of Immunology, Genetics and Pathology Uppsala University, Uppsala

Faculty opponent Åslaug Helland, M.D., Associate Professor

Department of Cancer Genetics Oslo University, Oslo

Examining committee Marie-Louise Fjällskog, M.D., Associate Professor

Department of Medical Sciences, Endocrine Oncology Uppsala University, Uppsala

Marika Nestor, M.D., Associate Professor Department of Immunology, Genetics and Pathology Uppsala University, Uppsala

Katalin Dobra, M.D., Associate Professor Department of Laboratory Medicine Karolinska Institute, Stockholm

List of Papers

This thesis is based on the following papers, which are referred to in the text by their Roman numerals.

I Micke P*, Mattsson JS*, Edlund K, Lohr M, Jirström K, Ber-glund A, Botling J, Rahnenfuehrer J, Marincevic M, Pontén F, Ekman S, Hengstler J, Wöll S, Sahin U, Türeci O. (2014) Aber-rantly activated claudin 6 and 18.2 as potential therapy targets in non-small-cell lung cancer. Int J Cancer. 135:2206-14.

II Mattsson JS, Bergman B, Grinberg M, Edlund K, Marincevic

M, Jirström K, Pontén, F, Hengstler JG, Rahnenführer J, Karls-son MG, Karlsson C, Helenius G, Botling J, Micke P, Gulyas M. (2015) Prognostic impact of COX-2 in non-small cell lung cancer: a comprehensive compartment-specific evaluation of tumor and stromal cell expression. Cancer Lett. 356:837-45.

III Gulyas M*, Mattsson JS*, Lindgren A, Ek L, Lamberg K, Behndig A, Holmberg E, Micke P, Bergman B. COX-2 expres-sion and effects of celecoxib in addition to standard chemother-apy in advanced non-small cell lung cancer. Submitted.

IV Mattsson JS, Brunnström H, Jabs V, Edlund K, Jirström K, Mindus S, La Fleur L, Ponten F, Karlsson MG, Karlsson C, Koyi H, Brandén E, Botling J, Helenius G, Micke P, Svensson MA. Inconsistent results in the analysis of ALK rearrangements in non-small cell lung cancer. Submitted.

*Contributed equally to the work.

Reprints were made with permission from the respective publishers.

Related publications

Micke P, Mattsson JS, Djureinovic D, Nodin B, Jirström K, Tran L, Jönsson P, Planck M, Botling J, Brunnström H. (2016) The impact of the fourth edition of the WHO classification of lung tumours on histological classification of resected pulmonary NSCCs. J Thorac Oncol. [Epub ahead of print].

Lohr M, Hellwig B, Edlund K, Mattsson JS, Botling J, Schmidt M, Hengstler JG, Micke P, Rahnenführer J. (2015) Identification of sample an-notation errors in gene expression datasets. Arch Toxicol. 89(12):2265-72.

Tran L, Mattsson JS, Nodin B, Jönsson P, Planck M, Jirström K, Botling J, Micke P, Brunnström H. (2015) Various antibody clones of napsin A, thy-roid transcription factor 1, and p40 and comparisons with cytokeratin 5 and p63 in histopathologic diagnostics of non-small cell lung carcinoma. Appl Immunohistochem Mol Morphol. [Epub ahead of print].

Moens LN, Falk-Sörqvist E, Ljungström V, Mattsson JS, Sundström M, La Fleur L, Mathot L, Micke P, Nilsson M, Botling J. (2015) HaloPlex Targeted Resequencing for Mutation Detection in Clinical Formalin-Fixed, Paraffin-Embedded Tumor Samples. J Mol Diagn. 17(6):729-39.

Haraldsson E, Swahn F, Verbeke C, Mattsson JS, Enochsson L, Ung KA, Lundell L, Heuchel R, Löhr JM, Arnelo U. (2015) Endoscopic papillectomy and KRAS expression in the treatment of adenoma in the major duodenal papilla. Scand J Gastroenterol. 50(11):1419-27.

Lindskog C, Edlund K, Mattsson JS, Micke P. (2015) Immunohistochemis-try-based prognostic biomarkers in NSCLC: novel findings on the road to clinical use? Expert Rev Mol Diagn. 15(4):471-90.

Noguchi S, Saito A, Horie M, Mikami Y, Suzuki HI, Morishita Y, Ohshima M, Abiko Y, Mattsson JS, König H, Lohr M, Edlund K, Botling J, Micke P, Nagase T. (2014) An integrative analysis of the tumorigenic role of TAZ in human non-small cell lung cancer. Clin Cancer Res. 20(17):4660-72.

Mattsson JS, Imgenberg-Kreuz J, Edlund K, Botling J, Micke P. (2012) Consistent mutation status within histologically heterogeneous lung cancer lesions. Histopathology. 61(4):744-8.

Contents

Introduction ................................................................................................... 13 Lung cancer .............................................................................................. 14

Epidemiology ....................................................................................... 14 Etiology ............................................................................................... 14 Histology and differences at the molecular level ................................. 15 Diagnosis and staging .......................................................................... 18 Treatment ............................................................................................. 20 Prognostic and predictive biomarkers.................................................. 23

Immunohistochemistry (IHC) .................................................................. 24 Fluorescence in situ hybridization (FISH) ............................................... 26 Tissue microarray (TMA) ........................................................................ 29 Gene expression microarray ..................................................................... 30

Present Investigation ..................................................................................... 32 Aim ........................................................................................................... 32 Molecular targets studied ......................................................................... 32

Claudin 6 and claudin 18.2 .................................................................. 32 Cyclooxygenase 2 (COX-2) ................................................................ 33 Anaplastic lymphoma kinase (ALK) ................................................... 33

Material and Methods ............................................................................... 34 Patient cohorts and study design .......................................................... 34 Methodological considerations ............................................................ 38

Results and discussion .............................................................................. 40 Paper I: ................................................................................................. 40 Paper II: ............................................................................................... 41 Paper III: .............................................................................................. 42 Paper IV: .............................................................................................. 44

Concluding remarks and future perspectives ................................................ 46

Acknowledgements ....................................................................................... 48

References ..................................................................................................... 50

Abbreviations

ABC ALK AP BRAF cDNA cRNA cTNM CK5 COX CT DAB DAPI DDR2 DNA EGFR EML4 ERBB2 FDA FFPE FGFR1 FISH H2O2 HE HQ HR HRP IASLC Ig IHC ISH KEAP1 KRAS mRNA n/a NOS NSAID

Avidin-biotin complex Anaplastic lymphoma kinase Alkaline phosphatase v-raf murine sarcoma viral oncogene homolog B Complementary DNA Complementary RNA Clinical tumor, node, metastasis Cytokeratin 5 Cyclooxygenase Computed tomography 3,3'-diaminobenzidine 4',6-diamidino-2-phenylindole Discoidin domain-containing receptor 2 Deoxyribonucleic acid Epidermal growth factor receptor Echinoderm microtubule associated protein like 4 v-erb-b2 avian erythroblastic leukemia viral oncogene homolog 2 Food and Drug Administration Formalin-fixed paraffin embedded Fibroblast growth factor receptor 1 Fluorescence in situ hybridization Hydrogen peroxide Hematoxylin and eosin 3-hydroxy-2-quinoxaline Hazard ratio Horseradish peroxidase International Association for the Study of Lung Cancer Immunoglobulin Immunohistochemistry In situ hybridization Kelch-Like ECH-Associated Protein 1 Kirsten rat sarcoma viral oncogene homolog messenger RNA Not applicable Not otherwise specified Non-steroidal anti-inflammatory drugs

NSCLC PCR PD1 PD-L1 PET PIK3CA PTEN PTGS2 pTNM RET RNA ROS1 SCLC STK11 TKI TMA TNM TP53 TTF1 VEGF WHO

Non-small cell lung cancer Polymerase chain reaction Programmed cell death protein 1 Programmed cell death ligand 1 Positron emission tomography Phosphatidylinositol-4,5-bisphosphate 3-kinase, catalytic subunit alpha Phosphatase and tensin homolog Prostaglandin-endoperoxide synthase 2 Pathological tumor, node, metastasis Ret proto-oncogene Ribonucleic acid ROS proto-oncogene 1 , receptor tyrosine kinase Small-cell lung cancer Serine/threonine kinase 11 Tyrosine kinase inhibitor Tissue microarray Tumor, node, metastasis Tumor protein p53 Thyroid transcription factor 1 Vascular endothelial growth factor World Health Organization

13

Introduction

Lung cancer is the leading cause of cancer death in Sweden, with more deaths each year than breast and colon cancer combined [1-2]. The overall prognosis of lung cancer is poor [3] and treatment options are limited. Cure is only possible in early stages when surgical resection is an option [2]. Chemotherapy remains the standard treatment for the majority of patients, but response is limited and the impact on survival is modest [4]. During the last decade, treatment options have progressed beyond chemotherapy and the discovery of targetable oncogenic driver mutations has revolutionized the treatment for a small subgroup of advanced lung cancer patients [5-6]. More recently, immunotherapy has emerged as a new potential treatment strategy, demonstrating long term responses in patients with late stage disease [7]. Unfortunately, these new treatment alternatives only show benefit in a subset of patients. Thus, the identification of other therapy targets or biomarkers that may refine the selection of existing treatment modalities are still war-ranted.

This thesis is based on four papers. An integrative translational approach has been adopted to describe and identify new potential treatment targets as well as to evaluate predictive biomarkers in non-small cell lung cancer (NSCLC).

14

Lung cancer Epidemiology Lung cancer is the most common cancer type worldwide, both regarding the incidence as well as the number of deaths. In 2012, 1.8 million (12.9% of total) were diagnosed and 1.6 million (19.4% of total) died of this disease. Lung cancer was the most common cancer type among men, with 1.2 million new cases (16.7%) and almost 1.1 million deaths (23.6%), whereas in wom-en the incidence was lower, with 583 000 new cases (8.7%) and 491 000 deaths (13.8%) [8]. Lung cancer rates vary by ethnicity, across countries, and between males and females within each country, mainly due to smoking trends [9]. In 2014 in Sweden, lung cancer was the fifth most common can-cer type in men and the fourth most common cancer type in women, but as in the rest of the world, it was the most common cause of cancer-related death. In 2014, 4063 patients were diagnosed with lung cancer and 3647 patients died from the disease [2]. The prognosis for patients diagnosed with lung cancer remains poor, with a 5-year survival rate of 13.6% among men and 19.4% among women across all stages [3].

Etiology Cigarette smoking is the main cause of lung cancer [10], and in Sweden, approximately 90% of all lung cancer cases are due to smoking [2]. Smokers have a 20-fold higher risk of developing lung cancer compared to never smokers and the risk increases with the duration of smoking as well as with the number of cigarettes smoked per day. By ceasing smoking, the risk of developing lung cancer decreases, but former-smokers still have a higher risk of developing lung cancer than never-smokers [11]. Although smoking is the most common cause of lung cancer, never-smokers still suffer from this type of cancer. Globally, about 300,000 lung cancer deaths per year are non-smoking related. The lung cancer risk in never-smokers varies geo-graphically, with the highest risk observed in Asia [12]. Lung cancer in non-smokers is most often of the adenocarcinoma histology, more common in women [13] and in younger patients (<60 years) [14]. In addition to smok-ing, other known risk factors include exposure to ionizing radiation e.g. ra-don gas, as well as exposure to occupational agents such as asbestos, chro-mium, nickel and arsenic. Furthermore, indoor and outdoor air pollution, pre-existing chronic lung disease, lack of physical activity, dietary factors, and second-hand smoke exposure are also associated with an increased lung cancer risk [11]. However, far from everyone who is exposed to cigarette smoke, or to other environmental risk factors, develop lung cancer. The risk of developing lung cancer most likely depends on genetic predisposition together with exposure to respiratory carcinogenic factors [15].

15

Histology and differences at the molecular level Lung cancer is a heterogeneous disease, both at the molecular and the histo-logical level. Historically, lung cancer is divided into two main histological subtypes: small-cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC). These two subtypes differ in their biological behavior as well as in treatment and prognosis [16]. SCLC accounts for 10-15% of all lung cancer cases and is believed to origi-nate from neuroendocrine cells in the lung [17]. This type of lung cancer is predominantly caused by substantial cigarette consumption and, compared to NSCLC, is characterized by a faster growth pattern, earlier spread, and high-er sensitivity to chemotherapeutic agents. Although SCLC initially responds well to chemotherapy, resistance to treatment is developed and SCLC re-mains fatal for the majority (95%) of patients [18-19].

NSCLC, which is the focus of this thesis, is the most common type of lung cancer, representing approximately 85% of all lung cancer cases [20]. Based on tumor histology, NSCLC is further sub-classified into three main sub-groups: adenocarcinoma, squamous cell carcinoma and large cell carcinoma [21]. Previously, there was no need to differentiate between these histologi-cal subtypes when considering the choice of treatment. However, this changed after the recognition that patients with non-squamous histology responded better to certain cytotoxic drugs as well as harbored distinct muta-tions applicable for targeted therapy [22]. The histological classification is based on an expert recommendation of pathologists appointed by the World Health Organization (WHO). The stud-ies in this thesis applied the histological classification based on the WHO classification of lung tumors from 2004 [21].

Adenocarcinoma Adenocarcinoma is the largest histological subgroup, comprising approxi-mately 40% of all lung cancer cases in Sweden [20]. Adenocarcinomas are epithelial tumors characterized by a glandular structure, pneumocyte marker expression (e.g. thyroid transcription factor 1 (TTF1) and napsin A) or mu-cin production. They most often occur in the lung periphery [23], where they are derived from small bronchi, bronchioles and alveoli [24]. Recently the WHO has proposed a further subclassification of invasive adenocarcinoma with regard to the architectural structures of acinar, papillary, micropapillary, lepidic or solid patterns [23]. These subgroups are of rele-vance regarding prognosis, but their potential role in guiding treatment deci-sions and surgical management is still not clear [25]. Among never-smokers, adenocarcinoma is the most frequent histological type [12]. Moreover, a

16

higher frequency of adenocarcinomas has been observed in patients from Southern and Eastern Asia compared to other parts of the world [24].

The histological subtypes also differ at the molecular level. In adenocarci-noma patients, mutations are commonly found in kirsten rat sarcoma viral oncogene homolog (KRAS), epidermal growth factor receptor (EGFR), v-raf murine sarcoma viral oncogene homolog B (BRAF) and v-erb-b2 avian erythroblastic leukemia viral oncogene homolog 2 (ERBB2). Rearrange-ments of anaplastic lymphoma kinase (ALK), ROS proto-oncogene 1, recep-tor tyrosine kinase (ROS1) and ret proto-oncogene (RET) are also common [26]. Furthermore, mutations are frequently observed in the tumor suppres-sor genes tumor protein P53 (TP53), kelch-like ECH-associated protein 1 (KEAP1) and serine/threonine kinase 11 (STK11) [27]. The frequencies of common genomic aberrations in adenocarcinoma are outlined in Table 1.

Squamous cell carcinoma Squamous cell carcinoma comprises approximately 20% of all lung cancer cases in Sweden [20]. During the first half of the 20th century, squamous cell carcinoma was the most common subtype of NSCLC, but due to decreased smoking consumption as well as changes in cigarette filters and composition, a decline in patients diagnosed with this subtype of lung cancer has occurred [29,30]. Tumors of squamous cell carcinoma histology are characterized by intracellular bridges and/or keratin structures. The tumors originate from bronchial epithelial cells and are often centrally located in the lung, i.e. in the main or lobar bronchus [21]. The expression of the immunohistochemical markers p40, p63, cytokeratin 5 (CK5) and cy-tokeratin 5/6 (CK5/6) are commonly used to characterize this subtype of NSCLC [23]. Squamous cell carcinomas display a very high mutation rate, reflecting the carcinogenic effects of tobacco usage [28,30]. Common genetic alterations in squamous cell carcinoma include mutations in the tumor suppressor genes TP53 [31] and phosphatase and tensin homolog (PTEN). Other common aberrations include mutations in discoidin domain-containing receptor 2 (DDR2) and phosphatidylinositol-4,5- bisphosphate 3-kinase, catalytic sub-unit alpha (PIK3CA), as well as amplifications of fibroblast growth factor receptor 1 (FGFR1) [26]. Frequencies of common genomic aberrations in squamous cell carcinoma are outlined in Table 1.

17

Table 1. Common genomic aberrations in NSCLC.

Adenocarcinoma Frequency

KRAS 25-30%

EGFR 15-20%

BRAF 1-3%

TP53 46%

KEAP1 17%

STK11 17%

ERBB2 1-3%

ALK 3-7%

ROS1 1-3%

RET 1%

Squamous cell carcinoma

TP53 81%

PIK3CA 5-10%

PTEN 4-8%

DDR2 3-5%

FGFR1 20%

Large-cell carcinoma Large-cell carcinomas constitute approximately 10% of all lung cancer cases in Sweden [20]. These tumors lack histological features of squamous cell carcinoma or adenocarcinoma differentiation [21]. However, based on immunohistochemical and mucin staining, the majority of large cell carci-nomas can today be allocated to the adenocarcinoma or squamous cell carci-noma subgroup [32]. Only cases that are negative with additional staining, or when staining is unavailable, are confined to the large cell carcinoma sub-group. Therefore, the proportion of diagnosed large cell carcinomas is de-creasing [33]. Not surprisingly, mutations in large cell carcinoma are dis-persed between adenocarcinoma and squamous cell carcinoma clusters and likewise, based on molecular features, large cell carcinomas can be allocated to the adenocarcinoma or squamous cell cancer group [34]. In addition to these three major sub-categories, other very rare entities of NSCLC exist, e.g. adenosquamous carcinoma and sarcomatoid carcinoma [23].

18

In the studies included in this thesis, all entities that were not assigned to the adenocarcinoma or squamous cell carcinoma subgroup were considered as large cell carcinoma, not otherwise specified (NOS).

Diagnosis and staging Patients presenting symptoms indicative of lung cancer are initially referred to a chest x-ray. If cancer indication remains, a computed tomography (CT) scan and positron emission tomography (PET) to assess the tumor extension is performed [35]. To ultimately confirm the lung cancer diagnosis, a histo-logical or cytological analysis is necessary [35-36]. The tissue or cell speci-men is obtained by either a bronchoscopy or, if the tumor is located periph-erally, a transthoracic puncture biopsy. To exclude mediastinal lymph node metastases, a mediastinoscopy, endobronchial ultrasound or endoscopic ul-trasound is performed [35]. Staging of lung cancer is important in order to determine the individual can-cer burden in a patient and to guide treatment decisions. The staging system applied in most types of cancer is the TNM system, which describes the size of the primary tumor and whether it has invaded nearby tissue (T), the pres-ence of lymph node metastasis (N) and the presence of distant metastasis (M) [37] (Table 2). Currently, the 7th edition of the TNM system is in use [38]. The TNM staging in NSCLC can be based on the clinical or radiologi-cal extent of the cancer, referred to as clinical TNM (cTNM), or after sur-gery based on the pathological findings, referred to as pathological TNM (pTNM) [39].

19

Table 2. TNM classification of carcinomas of the lung. Reproduced from [38].

Primary tumor (T) Description

T1 Tumor ≤ 3 cm, surrounded by lung or visceral pleura, no broncho- scopic evidence of invasion more proximal than the lobar bronchus T1a Tumor ≤ 2 cmT1b Tumor > 2 cm but ≤ 3 cm

T2 Tumor > 3 cm but ≤ 7 cm or tumor with any of the following:

- Involvement of main bronchus, ≥ 2 cm distal to the carina - Invading visceral pleura - Atelectasis or obstructive pneumonitis that extends to the hilar

region, but does not involve the entire lungT2a Tumor > 3 cm but ≤ 5 cmT2b Tumor > 5 cm but ≤ 7 cm

T3 Tumor > 7 cm or tumor that invades any of the following:

- Chest wall, diaphragm, phrenic nerve, mediastinal pleura, parie-tal pericardium

- Tumor in the main bronchus < 2 cm distal to the carina, but without involvement of carina

- Associated atelectasis or obstructive pneumonitis of the entire lung

- Separate tumor nodule(s) in the same lobe

T4 Tumor of any size that invades any of the following:

- Mediastinum, heart, great vessels, trachea, recurrent laryngeal nerve, esophagous, vertebral body, carina

- Separate tumor nodule(s) in a different ipsilateral lobe

Lymph nodes (N)

N0 No regional lymph node metastasisN1 Metastasis in ipsilateral peribronchial and/or ipsilateral hilar lymph

nodes and intrapulmonary nodes, including involvement by direct extension

N2 Metastasis in ipsilateral mediastinal and/or subcarinal lymph node(s)N3 Metastasis in contralateral mediastinal, contralateral hilar, ipsilat- eral or contralateral scalene, or supraclavicular lymph node(s)

Distant metastasis (M)

M0 No distant metastasisM1 Distant metastasisM1a Separate tumor nodule(s) in a contralateral lobe; tumor with pleu- ral nodules or malignant pleural (or pericardial) effusionM1b Distant metastasis

By the combination of T, N and M, an overall stage (I-IV) is obtained (Table 3), which gives an estimate of the individual prognosis and basis for therapy [39]. Stage I denote small tumors (≤ 5 cm) with no evidence of invasion,

20

while stage II tumors are larger and may have metastasized to regional lymph nodes. Stage III includes large tumors that may infiltrate adjacent structures or have lymph node metastasis in the mediastinal or contralateral lymph nodes. Stage IV denotes the most advanced stage of lung cancer, where the tumor has developed distant metastasis [40].

Table 3. Stage grouping. Reproduced from [38].

Stage I IA IB T1a N0 M0 T2a N0 M0

T1b N0 M0

Stage II IIA IIB T1a N1 M0 T2b N1 M0

T1b N1 M0 T3 N0 M0T2a N1 M0

T2b N0 M0

Stage III IIIA IIIB T1 N2 M0 T4 N2 M0

T2 N2 M0 Any T N3 M0T3 N1 M0T3 N2 M0T4 N0 M0

T4 N1 M0

Stage IV IV Any T Any N M1a Any T Any N M1b

Treatment Approximately 25% of the NSCLC patients in Sweden are diagnosed with localized disease (stage I or stage II) [20]. These patients are eligible for surgery, with the aim of complete resection and potential to cure the disease [41]. However, even with complete resection, many stage I and stage II NSCLC patients relapse (30-55%) [42]. The 5-year survival rate in patients with stage IA and IB is around 50% and 45%, respectively, while stage IIA and stage IIB present 5-year survival rates of around 35% and 25%, respec-tively [38]. With adjuvant chemotherapy, stage II patients demonstrate a decreased recurrence rate translating into a moderately (~5%) improved five-year overall survival rate [43-44]. Patients with stage I or II disease who are not eligible for surgery, e.g. because of medical comorbidities, are recom-mended radiotherapy with curative intent [41].

21

About 10% of the Swedish NSCLC patients present with stage IIIA disease [20]. This stage describes a heterogeneous group, ranging from resectable tumors with limited nodal metastases to unresectable nodal disease. Patients with limited lymph nodal involvement (N1) are considered for primary sur-gery followed by adjuvant chemotherapy. In contrast, patients with extended lymph node involvement (N2) may receive neoadjuvant chemoradiotherapy prior to surgical resection [4]. The 5-year survival rate in patients with stage IIIA disease is about 19% [38]. The majority of patients (~65%) are diagnosed with unresectable disease (stage IIIB and stage IV), with extended nodal disease or distant metastasis [20]. These tumors are not curable, and only palliative treatment is available. Depending on the performance status and individual tumor extension, stage IIIB-patients receive platinum-based chemoradiotherapy [4], leading to a 5-year survival of approximately 7% [38]. Stage IV patients usually receive platinum-based chemotherapy to improve the symptoms as well as to in-crease the survival time [4,45]. The 5-year survival rate in stage IV patients is approximately 2% [38]. Recently, for minor subgroups of advanced stage NSCLC patients, new individualized treatment options in the form of so-called targeted therapies have become available.

Targeted therapies The increasing understanding of cancer biology has identified genetic aber-rations believed to be important in tumor growth or progression. This has led to the development of targeted therapies against these specific molecular structures [46]. This personalized approach is applicable in small subsets of patients and is less toxic and have fewer side effects than conventional chemotherapy [47]. EGFR-inhibitors were the first class of targeted compounds approved in advanced NSCLC [5]. Mutations in EGFR induce intracellular signaling cascades, causing increased cell proliferation and cell survival [48]. Patients harboring activating EGFR mutations are likely to respond to treatment with EGFR tyrosine kinase inhibitors (TKIs) [49]. The first EGFR TKIs approved for treatment of advanced NSCLC were gefitinib and erlotinib. With these small molecule inhibitors, a prolonged progression-free survival has been demonstrated compared to standard chemotherapy (e.g. progression-free survival of 10.8 months with gefitinib compared to 5.4 months with chemo-therapy, and 9.7 months with erlotinib compared to 5.2 months with chemo-therapy) [50-51].

A second genomic aberration recently discovered in NSCLC is rear-rangement of the ALK gene. A fusion involving ALK and most commonly echinoderm microtubule associated protein-like 4 (EML4) results in a consti-tutively active ALK, affecting cell proliferation and survival [52]. Targeting

22

of ALK with a small molecule inhibitor, crizotinib, has shown impressive response rates in patients with ALK rearrangement [6], with a median pro-gression-free survival of 7.7 months compared to 3.0 months with chemo-therapy alone [53].

Other molecular aberrations being studied as potential future treatment targets in NSCLC are for example ROS1 [54], RET [55] and BRAF [56]. The specific inhibition of molecules involved in tumor angiogenesis has also been extensively investigated in clinical trials. Bevacizumab is a monoclonal antibody inhibiting vascular endothelial growth factor (VEGF) showing ben-efit in NSCLC patients (median survival time of 12.3 months for patients receiving bevacizumab in combination with chemotherapy compared to 10.3 months for patients receiving chemotherapy alone) and is approved for first-line treatment of advanced non-squamous lung cancer patients [57-58]. Although targeted therapies have shown impressive response rates in subsets of patients with advanced NSCLC, many patients rapidly develop resistance and experience relapse. Tumors with activating EGFR mutation develop secondary mutations (most commonly T790M), leading to decreased affinity for first-generation TKIs [59]. Similarly, patients receiving ALK-inhibitors eventually develop resistance, e.g. through acquired mutations in the ALK kinase domain, leading to impaired inhibitor function [60]. New agents and strategies to overcome resistance have been developed. Second and third generation EGFR- and ALK-inhibitors are demonstrating effect in patients resistant to first-generation inhibitors, but, resistance is eventually developed to these inhibitors as well [61-64].

Immunotherapy Over the past few years, immunotherapy has demonstrated encouraging re-sults in patients with NSCLC. Instead of targeting the cancer cells, immuno-therapy utilizes and enhances the patient’s own immune system to produce a response against cancer cells [7]. Today, the most promising immunomodulating drugs in NSCLC are so called check-point inhibitors, modulating T-cell activation. The best estab-lished example is antibody-based targeting of the programmed cell death-1 (PD-1) axis [7]. A phase III trial using the PD-1 inhibitor nivolumab in pa-tients with squamous cell carcinoma reported an increased median overall survival of 9.2 months compared to 6.0 months with chemotherapy (NCT01642004), leading to the United States Food and Drug Administration (FDA) approval of nivolumab as second-line therapy in previously treated squamous NSCLC patients [65]. Other clinical trials with inhibitors of PD-1 and the corresponding ligand of PD-1 (PD-L1) are ongoing, e.g. NCT02142738, NCT02220894, NCT02031458, NCT02008277.

23

Anti-inflammatory agents While the specific immune response against cancer might inhibit tumor pro-gression, unspecific inflammation preceding or following malignant trans-formation is considered to have tumor-promoting effects [66]. It is well known that chronic inflammation increases the risk of various cancer types [67], indicating that suppressing inflammation may represent a strategy for cancer prevention and therapy [68]. Already in 1983 it was demonstrated that nonsteroidal anti-inflammatory drugs (NSAIDs) reduced the number of colorectal adenomas in patients with familial adenomatous polyposis [69]. NSAIDs have also been used as chemopreventive agents to decrease the risk of colorectal cancers [70-71]. The initial NSAIDs targeted both of the rate-limiting enzymes involved in the prostaglandin synthesis, cyclooxygenase 1 (COX-1) and COX-2, but due to gastrointestinal side-effects caused by inhibition of COX-1, specific COX-2 inhibitors have been developed [72]. COX-2 is the most frequently evaluated anti-cancer anti-inflammation target. High levels of this enzyme have been demonstrated in various cancer forms [73-76], and the expression is reported to be associated with poor patient outcome [76-78]. Although many studies have tried to elucidate the role of COX-2 in lung cancer, the prognostic impact as well as the clinical benefit of therapeutic inhibition of this enzyme is still not clarified [79-80].

Prognostic and predictive biomarkers In order to provide information regarding the outcome of a cancer disease, prognostic biomarkers are utilized [81]. The currently most important factors to predict survival and to subsequently guide treatment decisions in NSCLC patients are TNM stage together with performance status and age [82-84]. However, patients with comparable tumor stage, age and performance status may experience strikingly different outcomes with regard to long-term sur-vival as well as early local or metastatic recurrence and death. This strongly suggests that additional factors influence the aggressiveness of the tumor and thus determine the prognosis. Therefore, a systematic correlation of molecu-lar alterations in the tumor tissue to clinical outcome may help to identify biomarkers that aid in prognostication. This is of particular importance for early stage patients, who are potentially curable by complete surgical resec-tion, but have a high risk of tumor recurrence [82]. The addition of adjuvant treatment in patients with stage II disease decreases this risk only modestly [43]. Thus, a majority of stage II patients could be spared from unnecessary treatment, while some stage I patients might benefit from adjuvant therapy. Biomarkers that predict recurrence and that are applicable in routine clinical diagnostics are therefore greatly needed [82].

24

Predictive biomarkers, on the other hand, are used to predict and monitor the response, or absence of response, to a specific treatment [85]. In lung cancer, oncogenic driver mutations present such valuable predictive biomarkers [86], and are often also targets for therapy [87]. Currently, EGFR mutations and ALK rearrangements are predictive for tumor response in patients treated with the corresponding inhibitor [4,86].

Immunohistochemistry (IHC) Immunohistochemistry (IHC) is a commonly used technique in pathological diagnostics as well as in research. With this technique, antibodies are utilized for visualization of protein expression in tissue sections. When performing IHC, the tissue morphology is preserved and information about the distribu-tion and location of the protein expression within the tissue is obtained [88]. The concept of IHC is based on the interaction between a specific antibody and the precise position (epitope) on its respective protein (antigen) [89]. Antibodies are a group of proteins called immunoglobulins (Ig) that are pre-sent in the blood. They are produced by plasma cells in response to invading pathogens in the body [90]. Immunoglobulins can be divided into five major classes: immunoglobulin G (IgG), IgA, IgD, IgE and IgM. IgG is the most abundant immunoglobulin in the serum, and is therefore most commonly used in the production of antibodies for immunoassays. In immunohistochemical reactions, two types of antibodies are used; monoclo-nal and polyclonal. Monoclonal antibodies are created by the same plasma cell clone, which make all the antibodies identical and directed against a single epitope on an antigen [91]. Polyclonal antibodies are produced by several plasma cell clones and comprise a heterogeneous mixture of antibod-ies directed against multiple epitopes on the same antigen [91-92]. There are two ways of visualizing the immune reaction, directly or indirectly (Figure 1A and B). With the direct method, the antigen is detected by the binding of a primary antibody that is directly labeled with an enzyme, biotin or fluorophore [89,91]. However, the most common way to visualize the immune reaction is by the indirect method. With this method an unlabeled primary antibody reacts with the antigen in the tissue, followed by the bind-ing of a labeled secondary antibody to the primary antibody. The secondary antibody can be labeled with different enzymes, e.g. alkaline phosphatase (AP) or horseradish peroxidase (HRP) [89]. HRP is most commonly used, and when adding hydrogen peroxide (H2O2) and the chromogen 3,3’-diaminobenzidine (DAB), a brown staining is produced where the primary and secondary antibodies are attached. This is known as DAB staining [93]. For low density antigens, it is recommended to employ further amplification

25

of the IHC signal. Amplification of a signal occurs when more label is de-posited per primary antibody compared to when using a standard primary-secondary antibody arrangement. Amplification can be accomplished by for instance the avidin-biotin complex, a chain polymer-conjugated technology such as the “EnVision” system developed by Dako-Cytomation, the “OptiView” hapten-based system from Ventana Medical Systems or with tyramide signal amplification. With the avidin-biotin complex, the secondary antibody is labeled with biotin. Streptavidin (or avidin) labeled with HRP is added and binds to biotin on the secondary antibodies, and when adding a chromogen, e.g. DAB, a brown staining is produced (the ABC method) [91,94]. The chain polymer-conjugated technology uses an enzyme-labeled dextran backbone to which several enzyme-molecules and secondary anti-bodies are attached. As many as 100 enzyme molecules (HRP or AP) and up to 20 secondary antibodies can be linked to a single backbone, and when a chromogenic substrate and H2O2 are added, a several-fold higher antigen detection is obtained compared to when using standard indirect protocols [95] (Figure 1C). The hapten-based technology from Ventana also uses a similar principle - the secondary antibody is conjugated to several non-endogenous 3-hydroxy-2-quinoxaline (HQ) haptens [96]. A hapten is a small chemical that is not antigenic in itself, but has to be conjugated to an endog-enous or exogenous protein, usually an antibody, to become antigenic [97]. To the haptens, tertiary antibodies linked to several HRP enzymes binds, and when adding DAB and H2O2, a brown staining is produced [96] (Figure 1D). Another way of amplifying the IHC-signal is by tyramide signal amplifica-tion. In this method, HRP-labeled secondary antibodies or HRP-avidin com-plexes are used. When H2O2 and labeled tyramine is added to the secondary antibodies, the tyramine is oxidized by HRP, leading to the binding of tyramine molecules in the immediate surrounding of the immunoreaction. The tyramine can be labeled with a fluorochrome, which can be visualized in a fluorescence microscope, or with biotin, which is visualized via labeled streptavidin or avidin in a light microscope [91,98]. To amplify the signal even further, the OptiView technology offers an optional step based on the tyramide signal amplification system. After the binding of secondary anti-bodies with conjugated HQ-haptens, a HQ-tyramide complex can be added. This complex reacts with H2O2 and HRP on the tertiary antibodies, leading to the binding of tyramine molecules in the surrounding of the immunoreaction. This is followed by the addition of HRP-linked quaternary antibodies bind-ing to the haptens in the complex, and when adding DAB, a highly amplified signal is obtained [96,99].

26

Polymer backbone

HRP

HQ-Hapten

Tertiary antibody

Secondary antibody

Primary antibody

Tissue sample with antigen

D) Hapten-based technology e.g. ”Optiview”

H2O2

+ DAB Brown color

Tissue sample with antigen

C) Polymer-based technology e.g. ”EnVision”

H2O2

+ DAB Brown color

Brown colorH2O2

+ DAB

Tissue sample with antigen

A) Direct detection

Tissue sample with antigen

H2O2

+ DAB

B) Indirect detection

Brown color

Figure 1. The principle of immunohistochemical staining with detection by the di-rect method (A) and by the indirect method (B). Further amplification of the IHC signal can be obtained by using the polymer-based technology here exemplified by “EnVision“ (C) or with the hapten-based technology, such as “OptiView” (D).

Fluorescence in situ hybridization (FISH) Another way of studying biological events in their morphological localiza-tion is by in situ hybridization (ISH). ISH is used for visualization of nucleic acids sequences in their “original place” e.g. in their cellular context [100]. The technique was first reported in 1969 [101], then with the use of radio-labeled probes, but was later further developed and the probes were instead labeled with fluorescent dye (fluorochromes), hence the name fluorescence in situ hybridization (FISH) [102]. One of the first applications of FISH was chromosome painting, where entire chromosomes are visualized [103-104], but later single-locus probes were developed making it possible to detect chromosomal aberrations such as deletions, duplications or rearrangements of genetic material [104-105]. For the detection of a specific genomic region, a probe that is complemen-tary to a part of that region is used. To visualize the probes, they have to be labeled. The labeling can be performed in two ways, directly or indirectly. When labeled directly, the probes are conjugated to a fluorochrome, e.g.

27

Texas red or fluoroscein, while with indirect labeling the molecule attached to the probe is non-fluorescent, for example either biotin or a hapten, and then incubated with avidin or an antibody that is fluorescently labeled [102,106]. Many probes, directly or indirectly labeled, are commercially available. The probes are applied on the tissue section and after denaturation the probes hybridize to the complementary target DNA [102]. If the probe is fluorescently labeled, the sample is directly counterstained with 4',6-diamidino-2-phenylindole (DAPI), which is a dye binding to AT-rich re-gions in the DNA and commonly used for nuclear counterstaining [91]. The signals are then observed with a fluorescent microscope equipped with filters that visualize the different wavelengths of the fluorescent probes. If the la-beling of the probe is indirect, the addition of a fluorescently labeled anti-body is necessary before the sample can be counterstained with DAPI and the probes visualized [102]. The principle of direct and indirect FISH is out-lined in Figure 2.

28

Target DNA on a microscopic slide

Denaturation of DNA Denaturation and labeling of probe

Direct labeling Indirect labeling

Probe

Hybridization of probe

Probe binds to complementary target DNA sequence

Fluorochrome-conjugated molecule

Hapten-conjugatedmolecule

Fluorescently labeled antibody

Addition of fluorescently labeled antibody

Evaluation in fluorescent microscope

Figure 2. The principle of direct and indirect FISH. Probes are applied to the tissue section. The probes and the target DNA are then denatured, followed by hybridiza-tion of the probes to the complementary DNA sequences. If the probe is directly labeled, DAPI is added and the sample is evaluated in fluorescence microscope. If the probe is indirectly labeled, the addition of fluorescently labeled antibodies that bind to the hapten-conjugated probes, are necessary before the sample can be coun-terstained with DAPI and the probes visualized.

In order to detect gene rearrangements, two different types of probes can be utilized: fusion probes and break-apart probes. With fusion probes, each of the two genes involved in the translocation are detected with differently la-beled probes (e.g. red and green). When no rearrangement has occurred, two separate color signals are visualized. However, if rearranged, the differently labeled probes will be very close or on top of each other and only one color will be observed (yellow) [107]. With break-apart probes, differently labeled probes (e.g. red and green) are hybridized on each side of the break point on one of the genes involved in the rearrangement. If no rearrangement has occurred, the probes will be very narrow or on top each other (yellow sig-nal), but if the sample is rearranged, there will be a split between the signals and the green and red signals will be separated. With fusion probes only known genetic partners can be detected, while with break-apart probes, the

29

fusion partner remains unidentified [108]. The principle of fusion and break-apart probes is illustrated in Figure 3.

No translocation Translocation

No translocation Translocation

FISH break-apart probes

FISH fusion probes

3’

5’

2p23.23’

5’

2p23.2

3’

5’

2p23.23’

5’

2p23.2

Figure 3. The principle of FISH fusion probes and break-apart probes. With fusion probes, each of the two genes involved in the translocation are detected with differ-ently labeled probes (e.g. red and green). If the sample is not rearranged, two sepa-rate color signals are visualized. If the sample is rearranged, the differently labeled probes will be very close or on top of each other and only one color will be observed (yellow). With break-apart probes, differently labeled probes (e.g. red and green) are hybridized on each side of the break point on one of the genes involved in the rear-rangement. If the sample is not rearranged, the probes will be very narrow or on top each other (yellow signal), but if the sample is rearranged, a split between the sig-nals will be displayed (the red and green signal will be separated).

Tissue microarray (TMA) At pathological laboratories patient tissue is most often available in the form of formalin fixed paraffin embedded (FFPE) material. FFPE material is a valuable resource, since fixation in formalin and embedding in paraffin pre-serves the protein and the DNA as well as the tissue morphology. With par-affin embedding, the samples are also easy to store [109-110]. Sections from the FFPE blocks can then be used in different analyses. However, to save

30

reagents and tissue, a tissue microarray (TMA) can be utilized. A TMA is composed of cylindrical cores from many different primary FFPE tumor blocks collected in a single “recipient” paraffin block. Therefore, it is possi-ble to simultaneously examine a large number of tissues on a single slide [111-113]. In order to collect representative tissue cores, the primary tissue blocks are cut in 4 μm sections, placed on glass slides, and stained with hematoxylin and eosin (HE) to visualize the morphology of the tissue. Rep-resentative areas are then marked on the glass slide, and cores from that area are punched from the tumor block using a manual or automated arraying instrument and deposited in the recipient block [113]. The generation of a TMA is presented in Figure 4.

HE slide with marked

areas

TMA slide with IHC staining

Figure 4. Generation of a tissue microarray (TMA). Hematoxylin and eosin (HE) stained sections from primary tissue blocks are marked for representative tissue areas. Tissue cores from these primary blocks are then collected in a recipient block by the use of an arraying instrument. The generated TMA-block is sectioned, and the sections are placed on glass slides and used in immunohistochemical analyses.

By assembling tissue cores from primary donor blocks into a TMA, the ma-jority of the tissue in primary donor blocks is saved even though new anal-yses are performed on the patient sample [112]. Up to 300 sections can be cut from a single TMA block, and the sections can be utilized for several different in situ analyses, for example IHC, mRNA in situ hybridization and FISH [112-113]. TMAs containing tumor specimens are particularly useful for linking protein expression from various markers to clinical parameters in order to screen for their prognostic, predictive or diagnostic significance in patients with a certain cancer disease [114].

Gene expression microarray Gene expression microarrays make it possible to simultaneously analyze mRNA expression levels of thousands of genes from a biological sample at a certain time point [115-116]. A probe on an Affymetrix Human Genome U133 Plus 2.0 array consists of a sequence of 25 oligonucleotides, and each

31

probe is synthesized on the surface of a chip. Commonly, 11 different per-fect-match probes, which bind to different positions on a single mRNA se-quence, constitute a probe set [117]. The Human Genome U133 Plus 2.0 array contains more than 54,000 probe sets, representing over 47,000 tran-scripts [118]. In order to perform a microarray analysis, RNA is first extract-ed from a sample and converted to complementary DNA (cDNA) by the use of reverse transcriptase. Thc cDNA is then converted to complementary RNA (cRNA) by an in vitro transcription reaction where biotinylated nu-cleotides are incorporated. The cRNA is then added to a glass chip and hy-bridized to the complementary gene sequences present on the chip. Non-bound material is washed away. The fluorescence marker streptavidin-phycoerythrine, which binds to biotin, is then added to the samples, the chip is scanned, and the signals emitted are detected [119]. This procedure is vis-ualized in Figure 5.

Sample collection RNA extraction

Analysis of gene expression data

Conversion to cDNA by reverse transcriptase

Conversion to cRNA with incorporation of biotin-labeled nucleotides, fragmentation of transcripts, and hybridization

to gene chip

biotin

GeneChip

Affymetrix

Figure 5. Gene expression microarray analysis using Affymetrix GeneChip. RNA is extracted from frozen tissue samples and converted to cDNA, followed by conver-sion to cRNA with subsequent incorporation of biotinylated nucleotides. The cRNA is then fragmented and hybridized to the Affymetrix GeneChip. Signals from fluo-rescent markers which have bound to biotin are detected and analyzed.

32

Present Investigation

Aim The main objective of this thesis was to evaluate the expression of possible new treatment targets and of biomarkers that can guide therapy in NSCLC. This was done on a multi-molecular level integrating genomic, gene expres-sion and protein data with clinicopathological parameters.

The specific aims for each study were: Paper I: To evaluate expression levels and survival associations of the poten-tial treatment targets claudin 6 and claudin 18.2 in NSCLC. Paper II: To evaluate the prognostic influence of tumor and stromal protein expression of the inflammatory mediator cyclooxygenase 2 (COX-2) in two independent NSCLC cohorts. Paper III: To determine if elevated tumor and stromal protein expression of COX-2 is associated with improved survival in patients treated with the se-lective COX-2 inhibitor celecoxib.

Paper IV: To compare the results of two different ALK detection methods, FISH and IHC, and to analyze the overall frequency of ALK rearrangements in a large representative Swedish NSCLC population.

Molecular targets studied Claudin 6 and claudin 18.2 Claudins are essential components of tight junctions and are responsible for the regulation of paracellular activity between cells [120-121]. Individual cells express more than one of the 26 human claudin family members [122], suggesting that each tissue has its own combination of claudin proteins and regulation of tight junction function [123-124]. Many of the claudins are broadly expressed, while others such as claudin 6 and 18.2 are selectively expressed. The expression of claudins has been demonstrated to be altered in

33

several tumor types [124]. Claudin 6 expression is normally limited to the embryonic development, but it is expressed in various cancer types [124-125]. Claudin 18 has two isoforms, claudin 18.1, which is constitutively expressed in lung tissue [126], and claudin 18.2, which is expressed only in differentiated epithelial cells of the gastric mucosa. Claudin 18.2 has also been demonstrated to be ectopically expressed in various cancer types [127].

Cyclooxygenase 2 (COX-2) Cyclooxygenase 2 (COX-2) is the most frequently studied anti-cancer anti-inflammatory target today [68]. COX-2 is encoded by the gene prosta-glandin-endoperoxide synthase 2 (PTGS2), and is the rate-limiting enzyme responsible for the conversion of free arachidonic acid to prostaglandins, prostacyclins and thromboxanes [128-129]. COX-2 is induced by growth factors, cytokines, tumor promoters, oncogenes and other extracellular stim-uli [130], and subsequently, high levels of COX-2 have been demonstrated in various malignancies [131-134]. Due to its pro-tumorigenic properties, COX-2 has been evaluated in the context of cancer prevention and treatment [135-137], and based on the promising results of COX-2 inhibition in colo-rectal cancer [138], COX-2 inhibitors have also been evaluated for the treatment of NSCLC patients [139].

Anaplastic lymphoma kinase (ALK) Anaplastic lymphoma kinase (ALK) is a receptor tyrosine kinase that be-longs to the insulin growth factor receptor superfamily [140]. The function of ALK is still unknown, but ALK is believed to play a role in the develop-ment and maintenance of the nervous system since the protein normally is expressed only in a small number of cells within the developing and mature nervous system [140-141]. Rearrangements of ALK are found in 3-7% of NSCLC patients [26], and this rearrangement leads to a constitutive active ALK, affecting downstream signaling and promoting malignant growth and proliferation. Most commonly, ALK forms a fusion protein with EML4 [52], but several other fusion partners have been described. All fusion proteins contain the ALK kinase domain, subsequently leading to comparable kinase activation [52,142-144]. ALK rearrangements can be targeted by small syn-thetic ALK-inhibitors, leading to impressive tumor shrinkage and improved survival in patients harboring this rearrangement [6,53].

34

Material and Methods Patient cohorts and study design In this thesis, four NSCLC patient cohorts were studied: the Uppsala I cohort (Paper I, II and IV), the Uppsala II cohort (Paper IV), the Örebro cohort (Paper II and IV) and the Cyclus cohort (Paper III). The Uppsala I cohort consists of operated patients with primary NSCLC reported to the Uppsala-Örebro Regional Lung Cancer Registry between 1995 and 2005, with available fresh frozen tissue in the Uppsala frozen tis-sue Biobank. When the Regional Lung Cancer Registry was cross-linked to the frozen tissue Biobank database, 382 patients with corresponding frozen tissue blocks were identified. Study inclusion was based on a) confirmed NSCLC histology in the fresh frozen section, b) a tumor sample size of ≥ 5mm, and c) a tumor cell fraction of ≥ 50%. Patients who had received neoadjuvant treatment were excluded. In total, 194 patients fulfilled the cri-teria. From these samples RNA was previously extracted and analyzed by Affymetrix gene expression microarrays [145-146]. Additionally, FFPE tissue was available for 189 of these patients plus 165 additional tumor sam-ples (where the corresponding frozen tissue specimen did not meet the pre-viously mentioned inclusion criteria), ending up with in total 354 patients with available FFPE tissue. Duplicate tissue cores from each patient sample were previously included in a TMA [147-148]. Information about histology was attained from the pathology reports at Uppsala University Hospital. Clinical patient characteristics, e.g. performance status according to WHO, stage and smoking history, was obtained from the records of the Uppsala-Örebro Regional Lung Cancer Registry. The Uppsala II cohort is an extension of the Uppsala I cohort, consisting of patients with primary NSCLC that had undergone surgical resection at the Uppsala University Hospital between 2006 and 2010. When the Regional Lung Cancer Registry was crosslinked to the Laboratory Information System (Sympathy), a total number of 390 patients were identified. After confirming histopathological examination and exclusion of patients who had received neoadjuvant treatment, 354 samples remained. Tumor cores from these FFPE samples were included in a TMA, with duplicate tissue cores from each patient. Clinical patient data were obtained from the records of the Uppsala-Örebro Regional Lung Cancer Registry [33,149]. The Örebro cohort consists of 262 patients that had undergone surgical re-section at the Örebro University Hospital between 1990 and 1995. A TMA comprising triplicate cores from each tumor specimen was available. Clini-cal data, including histology, stage, age, smoking status and overall survival had previously been obtained from the clinical, surgical and pathological records at the Örebro University Hospital [150-151].

35

The study design of the Uppsala I cohort, the Uppsala II cohort and the Örebro cohort is described and visualized in Figure 6. Clinical characteristics of patients included in the three cohorts are presented in Table 4. In the sta-tistical analyses, overall survival was computed from the date of diagnosis to the date of death or last follow-up until spring 2011 (Uppsala I cohort), summer 2015 (Uppsala II cohort) or autumn 2010 (The Örebro cohort).

TMA NSCLC

Protein expressionn = 354

mRNA gene expression n = 194

Fresh frozen tissueFFPE tissue TMA

Uppsala I cohort 1995-2005

Protein expressionn = 354

FFPE tissue TMA

Uppsala II cohort 2006-2010

Örebro cohort1990-1995

TMA NSCLC

Protein expressionn = 262

FFPE tissue TMA

TMA NSCLC

A) B) C)

Figure 6. Study design of the cohorts. (A) The Uppsala I cohort consists of NSCLC patients that had undergone surgical resection between 1995 and 2005. FFPE tissue samples from 354 patients were utilized for protein expression analysis using IHC on TMAs, while 194 fresh frozen tissue samples were subjected to gene expression analysis. (B) The Uppsala II cohort consists of NSCLC patients that were operated between 2006 and 2010. FFPE tissue samples from 354 patients were subjected to protein expression analysis using IHC on TMAs. (C) The Örebro cohort consists of NSCLC patients that underwent surgical resection between 1990 and 1995. FFPE tissue samples from 262 patients were subjected to protein expression analysis using IHC on TMAs.

36

Table 4. Clinical characteristics of the lung cancer patients included in the gene expression patient cohort (Uppsala I cohort), the Uppsala I TMA cohort, the Uppsala II TMA cohort as well as the Örebro TMA cohort.

Uppsala I TMA

Uppsala I Affymetrix

Uppsala II TMA

Örebro TMA

IHC, n (%) GE, n (%) IHC, n (%) IHC, n (%)

All patients 354 (100) 194 (100) 354 (100) 262 (100)

Gender Male 188 (53.1) 105 (54.1) 173 (48.9) 165 (63.0) Female 166 (46.9) 89 (45.9) 181 (51.1) 97 (37.0)

Age at diagnosis < 70 years 244 (68.9) 150 (77.3) 232 (65.5) 189 (72.1) > 70 years 110 (31.1) 44 (22.7) 122 (34.5) 73 (27.9)

Smoking history Current + former smoker 318 (89.8) 179 (92.3) 313 (88.4) 206 (78.6) Never smoker 34 (9.6) 15 (7.7) 41 (11.6) 19 (7.3 Missing data 2 (0.6) 0 0 37 (14.1)

TNM-stage IA 87 (24.6) 38 (19.6) 146 (41.2) 86 (32.8) IB 150 (42.4) 90 (46.4) 76 (21.5) 33 (12.6) IIA 12 (3.4) 6 (3.1) 41 (11.6) 34 (13.0) IIB 44 (12.4) 29 (14.9) 34 (9.6) 50 (19.1) IIIA 35 (9.9) 21 (10.8) 50 (14.1) 43 (16.4) IIIB 16 (4.5) 6 (3.1) 0 (0.0) 2 (0.8) IV 10 (2.8) 4 (2.1) 7 (2.0) 2 (0.8) Missing data 0 0 0 12 (4.6)

Histology Adenocarcinoma 195 (55.1) 106 (54.6) 210 (59.3) 134 (51.1) Squamous cell carcinoma 117 (33.1) 63 (32.5) 104 (29.4) 124 (47.3) Large cell cancer/NOS1 42 (11.9) 25 (12.9) 40 (11.3) 4 (1.5)

Performance status 0 187 (52.8) 105 (54.1) 214 (60.5) n/a2 1 134 (37.9) 73 (37.6) 137 (38.7) n/a 2 27 (7.6) 12 (6.2) 3 (0.8) n/a 3 5 (1.4) 4 (2.1) 0 (0.0) n/a 4 1 (0.3) 0 0 (0.0) n/a

Follow-up mean, months 59.9 58.4 49.9 49.3

1NOS = not otherwise specified 2n/a = not applicable

The Cyclus cohort comprises tissue from NSCLC patients from a Swedish randomized phase III trial [152]. The trial was designed to investigate if the addition of the COX-2 inhibitor celecoxib to standard chemotherapy would prolong survival in patients with advanced NSCLC. Chemotherapy-naïve patients with advanced (stage IIIB and IV) NSCLC were included in the trial (n=316). From the original trial population of 316 patients, biopsies and surgical specimens from 122 (38.6%) patients prior to the administration of celecoxib were available for protein expression analysis. The Cyclus study

37

design is described and visualized in Figure 7. Clinical characteristics of the patients with available tissue for protein analysis are presented in Table 5. Overall survival was computed from the date of diagnosis to the date of death or last follow-up until summer 2015.

n=316Cyclus trial

Randomization

n=158Standard chemotherapy +

celecoxib

n=158Standard chemotherapy +

placebo

n=67Pretreatment tissue

samples available for COX-2 analysis

n=55Pretreatment tissue

samples available for COX-2 analysis

Figure 7. The Cyclus trial study design. In the original phase III trial, 316 patients were included. These patients were randomly allocated to the standard chemothera-py + placebo arm (n=158) or to the standard chemotherapy + celecoxib arm (n=158). Of these 316 patients, pretreatment tissue specimens from 122 patients were available for COX-2 protein expression analyses: 55 patients from the chemo-therapy + placebo arm, and 67 patients from the chemotherapy + celecoxib arm.

Table 5. Clinical characteristics of the 122 lung cancer patients with available tissue specimens included in the Cyclus trial.

Cyclus, IHC Placebo (n=55) Celecoxib (n=67)

Age, median (range) 64 (47-84) 66 (38-85)

Gender Male 31 (56%) 31 (46%) Female 24 (44%) 36 (54%)

Performance status 0 14 (25%) 16 (24%) 1 32 (58%) 37 (55%) 2 9 (16%) 14 (21%)

TNM-stage IIIB 15 (27%) 21 (31%) IV 40 (73%) 46 (69%)

Histology Adenocarcinoma 32 (58%) 42 (63%) Squamous cell carcinoma 18 (33%) 21 (31%) Large cell carcinoma/NOS* 5 (9%) 4 (6%)

*NOS = not otherwise specified

38

Methodological considerations

Immunohistochemistry Immunohistochemistry (IHC) was used to evaluate protein expression in all Papers (I-IV) included in this thesis. The staining was performed on FFPE slides using an automated staining instrument, making the staining more reproducible compared to when manual staining is performed. The IHC- staining was manually evaluated on scanned slides using a semi-quantitative annotation method where staining intensity and fraction of cells stained were taken into consideration. The intensity and fraction of stained cells were assigned ordinal scores, which were multiplied to obtain an overall staining score for each sample. The overall score was used in the subsequent analy-sis. As the evaluation of the staining is subjective, we attempted to reduce this uncertainty by the independent evaluation by two observers. In case of any discordancy, a shared review of the images was performed to obtain a common interpretation. Since staining results can be influenced by many factors, we also compared two staining protocols using the same antibody clone, but differing with regard to staining instrument, pre-treatment method, incubation time of pri-mary antibody and signal amplification method (Paper IV). One main advantage of IHC is that protein expression can be allocated to a cell type or a tissue compartment. This was used to determine compartment-specific expression in Paper II and Paper III, where we analyzed protein expression in tumor cells and stroma cells separately.

Tissue microarrays In Paper I, II and IV, the protein expression was evaluated on TMAs. One potential problem with TMAs is that the included tissue cores may not be representative for the whole tumor. However, many studies have shown that two or three cores of the original tissue compensate for this heterogeneity and give reliable results [114,153-154]. To address this aspect, in Paper IV we also analyzed the corresponding whole section with IHC and FISH to directly compare the results. In Paper III, IHC was on the other hand per-formed on whole sections since no TMA was available for the patients in-cluded in this study.

FISH In Paper IV, ALK rearrangements were detected using FISH break-apart probes on TMAs. FISH has the advantage to identify ALK rearrangements independent of fusion partner or break-point. In general, interpretation of

39

FISH results is considered to be clear-cut, but the ALK specific inversion of a small region on chromosome 2 leads to narrow split signals that can be hard to detect. In accordance with the guidelines published by the Interna-tional Association for the Study of Lung Cancer (IASLC) [155], we consid-ered a sample ALK FISH positive if ≥ 15% of a minimum of 50 cells evalu-ated are rearranged.

Statistics The Kaplan-Meier method was used to calculate the survival rates. Survival curves for analyzed parameters were compared using the log-rank test. Uni- and multivariate Cox-regression analysis, with inclusion of clinicopathological parameters (age, stage, performance status), was per-formed to estimate hazard ratios (HR). Protein expression was dichotomized into high or low expression in each paper. Gene expression values were di-chotomized into high or low expression by outlier sum statistics [156] in Paper I, and used as a continuous variable in Paper II and IV. Gene expres-sion and protein expression values were compared using Spearman correla-tion (Paper I) or the Wilcoxon test (Paper II). Gene and/or protein expression and clinical parameters were compared using Spearman correlation (Paper I, II, IV). In Paper II, meta-analyses (random effect and fixed effects models) were conducted for combining single estimates from previous independent studies into one combined overall estimate. All p-values were two-sided and a statistical significance level of p < 0.05 was used. The statistical analyses were performed using the statistical computing language R (https://www.r-project.org) (Paper I, II and IV) and the program Stata (version 13.1) (Paper III).

Gene expression analysis Gene expression data, generated using Affymetrix GeneChip HG-U133 Plus 2.0 arrays, was included as a supplement to IHC (Paper I, II and IV) and FISH (Paper IV) for a subset of patients in the Uppsala I cohort. However, it should here be noted that RNA was extracted from another part of the same tumor, which had been snap frozen and stored at -80ºC, and that one cannot exclude that intratumoral differences might influence the comparison. Fur-thermore, the gene expression analysis was performed based on RNA ex-tracted from crude tumor tissue and for each gene the signal thus represents the average gene expression of all cell types present. Consequently, gene expression levels are difficult to compare to compartment-specific protein expression levels determined by IHC.

40

Results and discussion

Paper I: Claudin 6 and claudin 18.2 are aberrantly expressed in subgroups of non-small cell lung cancer. Lineage-specific claudins, which are normally confined to a certain cell-type, have been proposed as therapy targets as they may be ectopically ex-pressed in neoplastic tissue and are not expressed in most normal tissues [124]. Clinical trials with therapeutic antibodies targeting claudin 6 in ovari-an cancer and claudin 18.2 in gastro-esophageal cancer are currently ongoing (NCT02054351 and NCT01630083). To evaluate if claudin 6 and claudin 18.2 are expressed in NSCLC, and as such, could represent potential treat-ment targets also in this cancer type, protein expression was analyzed by immunohistochemistry on tissue microarrays (Uppsala I cohort). Membranous expression of claudin 6 and claudin 18.2 was observed in 6.5% and 3.7%, respectively. Expression of claudin 6 and claudin 18.2 was ob-served in principally non-overlapping subgroups of non-squamous carcino-ma, while non-malignant bronchial epithelial cells and pneumocytes were negative. High expression of claudin 6 was associated with shorter overall survival in the adenocarcinoma subgroup in both the univariate (HR=1.8, CI 1.1-2.9, p=0.032) and multivariate Cox-regression model (HR=1.9, CI 1.2-3.2, p=0.017). High expression of claudin 18.2 was primarily observed in patients with early-stage disease and in tumors with a low Ki67-index, sug-gesting a less aggressive phenotype. However, a longer survival was not demonstrated in the Cox-regression model. Notably, none of the samples with claudin 6- or claudin 18.2-positivity demonstrated an EGFR mutation, which may indicate that claudin 6 and 18.2 expressing patients constitutes a subgroup that is not overlapping with one of the most important markers guiding treatment decisions in advanced NSCLC today. Our protein findings were supplemented with gene expression microarray-based mRNA data for claudin 6 and total claudin 18. Protein and gene ex-pression could be compared in a subset of 189 samples, and strong correla-tions were observed for both claudin 6 (rho 0.6-0.8) and claudin 18 (rho 0.4-0.6). However, the gene expression data did not distinguish between the two claudin 18 splice variants, and should therefore be interpreted with caution. High expression of claudin 6 was not associated with survival in the univariate analysis, but one of the analyzed probe sets (208474_at) demon-strated in the multivariate analysis an association with shorter survival (HR=1.6, CI 1.0-2.4, p=0.04). This association was more prominent in the

41

adenocarcinoma subgroup. For claudin 18, no association with survival was observed. Thus, the gene expression data supports the protein expression data. In conclusion, claudin 6 and 18.2 are both ectopically expressed in a subset of NSCLC patients. This finding might be of clinical relevance, as these claudins represent cancer specific targets and initial clinical trials with thera-peutic antibodies against both claudins may be extended to NSCLC. How-ever, due to the low number of positive cases, further analyses are needed in a larger sample set to evaluate the expression pattern of claudin 6 and 18.2 and if they are confided to a certain subgroup of NSCLC.

Paper II: COX-2 is expressed in both tumor and stromal cells in NSCLC. COX-2 pro-tein expression in tumor cells is not associated with survival, while the prog-nostic significance of stromal COX-2 expression remains unclear. COX-2 is an enzyme which is induced and expressed in response to growth factors, cytokines and other stimuli [130]. Strategies to target COX-2 with anti-inflammatory drugs have shown promising results in colorectal cancer prevention [157] and have been demonstrated to reduce the number of colo-rectal adenomas in patients with familial adenomatous polyposis [158]. However, the role of COX-2 in NSCLC is controversial, with contradicting results regarding its prognostic value. We therefore evaluated the COX-2 gene (PTGS2) and protein expression in NSCLC tissue. PTGS2 expression based on Affymetrix gene expression microarray data was evaluated in 194 patients (Uppsala I cohort). Additionally, a meta-analysis was performed, including nine publically available data sets, including in total 1337 patients. By immunohistochemical analysis, the in-situ COX-2 protein expression pattern was evaluated in a compartment-specific manner, distinguishing between tumor cell positivity and positivity in the tumor stroma. The IHC analysis was performed on tissue microarrays from two independent cohorts (Uppsala I and Örebro). In the Affymetrix gene expression microarray analysis, PTGS2 expression did not show any significant statistical association with overall survival or other investigated clinical parameters (all p > 0.1). Similarly, no association with survival was demonstrated in the meta-analysis comprising nine public-ly available data sets, either with all histologies included (p=0.57) or when the histological subgroups were analyzed separately. Hence, no prognostic relevance of PTGS2 expression could be demonstrated.

42

In the second phase, we evaluated immunohistochemical COX-2 protein expression in the tumor cell and the stromal cell compartment. Tumor COX-2 expression did not demonstrate any significant correlations with clinical parameters or overall survival in the Uppsala I (HR=0.91, CI 0.69-1.2, adj. p=1.0) or the Örebro cohort (HR=1.18, CI 0.81-1.72, adj. p=1.0). COX-2 expression in stromal cells did also not correlate with any clinical parameters or with overall survival in the Uppsala I cohort (HR=0.86, CI 0.64-1.17, adj. p=1.0). However, in the Örebro cohort, a significant association was found between high stromal COX-2 expression and longer survival. This was ob-served in both uni- (HR=0.56, CI 0.38-0.82, adj. p=0.018) and multivariate Cox-regression analysis (HR=0.56, CI 0.38-0.84, adj. p=0.02), and was more profound in the adenocarcinoma subgroup (HR=0.42, CI 0.23-0.75, adj. p=0.02). In summary, the present findings dismiss the prognostic relevance of PTGS2 expression and reject the hypothesis of COX-2 tumor cell expression being of prognostic value in NSCLC. On the other hand, COX-2 protein expres-sion in the stromal compartment was associated with longer survival in one of the analyzed cohorts, which makes COX-2 stromal expression an interest-ing subject for further investigation.

Paper III: The addition of celecoxib to standard chemotherapy did not prolong overall survival in advanced NSCLC patients, regardless of COX-2 expression. Clinical trials have failed to demonstrate a benefit of adding COX-2 inhibi-tors to standard chemotherapy in NSCLC [159-161]. In Paper III, we aimed to analyze COX-2 expression in tumor and stromal cells as a predictive bi-omarker for longer survival upon COX-2 inhibition.

Between 2006 and 2009, 316 NSCLC patients were enrolled in a random-ized, double-blind, placebo-controlled, multicenter phase III trial (the Cyclus trial) with the primary objective to compare overall survival between two treatment arms: arm one (n=158) received the COX-2 inhibitor celecoxib in addition to standard chemotherapy and arm two (n=158) received placebo and standard chemotherapy. No survival difference was found between the two treatment arms [152]. However, COX-2 expression was not considered as a marker for celecoxib treatment. Therefore, we evaluated COX-2 expres-sion as a biomarker for survival benefit in 122 of the patients included in the Cyclus trial.

43

In the first step, an updated survival analysis of the patients enrolled in the original Cyclus trial (n=316) as well as of the patients with available tissue for COX-2 analysis (n=122) was performed. No significant survival differ-ence was seen between the two treatment arms, neither in the complete population (p=0.93) nor in the subpopulation available for COX-2 analysis (p=0.52). In the second step, we analyzed patient survival according to high and low COX-2 protein expression in tumor cells without taking the celecoxib treat-ment into account, however, no significant survival difference was observed (HR=1.07, CI 0.74-1.54, p=0.74). When comparing high and low COX-2 expression in stromal cells, there was a slight tendency toward a longer sur-vival in patients with positive COX-2 staining, but the difference was statis-tically non-significant (HR 0.80, CI 0.56-1.15; p=0.23).

In the third step, COX-2 expression was analyzed in accordance to treatment arm. Of the 122 patients available for COX-2 protein expression analysis, 67 received celecoxib in addition to standard chemotherapy. The overall surviv-al difference between patients with high and low COX-2 expression in tumor cells in this group was not significant (HR=0.91, CI 0.55-1.50, p=0.72). As expected, also in the group of patients who received placebo (n=55) no sur-vival difference was seen between patients with high and low COX-2 ex-pression (HR=1.30, CI 0.74-2.26, p=0.36). A similar pattern was seen in analysis of the COX-2 stromal expression. The survival did not differ between high and low stromal expression in patients receiving celecoxib (HR=0.97 for high vs. low expression, CI 0.59-1.58, p=0.89). In patients receiving placebo, there was a trend towards longer sur-vival time in patients with high COX-2 expression in stroma, although the difference did not reach statistical significance (HR=0.63, CI 0.36-1.10; p=0.10). In conclusion, the addition of celecoxib to standard chemotherapy did not prolong overall survival in advanced NSCLC patients, irrespective of COX-2 protein expression. However, the retrospective evaluation of COX-2 ex-pression had to be made on available tissue, including 122 patients, which is only 38.6% of the original cohort. Thus, with this small cohort of patients, minor differences in terms of survival may be undetected.

44

Paper IV: The frequency of ALK rearrangements in a Swedish NSCLC population is lower than previously reported. The two immunohistochemical assays evalu-ated demonstrated low sensitivity compared to the gold standard method fluorescence in situ hybridization. Fluorescence in situ hybridization (FISH) is regarded as the gold standard in the detection of ALK rearrangements and has been the only method guiding treatment decisions in ALK positive NSCLC patients. Nonetheless, FISH is expensive and difficult to interpret. Therefore, ALK IHC is often used as screening tool in clinical practice. Recently, the FDA approved an IHC assay (Ventana ALK (D5F3) CDx Assay, Roche Diagnostics Limited, UK) as a companion test to identify ALK positive patients likely to respond to ALK inhibitor therapy [162]. In Paper IV, we therefore compared the detection of ALK-positive patients with the recently FDA-approved IHC assay together with an in house Dako IHC protocol, using the same antibody clone, as well as with FISH. This was performed in three large Swedish NSCLC cohorts (in total 851 patients) incorporating clinical parameters and gene expression data for 194 patients. ALK status was first assessed on the TMAs for the three investigated co-horts. Tumors presenting ALK-positivity in at least one assay, or a contra-dicting result in one or two of the assays, were re-evaluated on whole tissue sections using all three methods (FISH, IHC Ventana and IHC Dako). With FISH, ALK rearrangement was detected in 13/754 (1.7%) of the NSCLC samples. With the Ventana assay, 14/791 (1.8%) were considered ALK positive, and with the in house Dako protocol, 11/806 (1.4%) were classified as ALK positive. When comparing the two IHC assays, 15 cases were positive in at least one of the IHC assays but only ten (66.7%) cases were positive with both protocols. Comparison between FISH and the Ventana IHC protocol demonstrated nine positive cases with both methods, however, additionally four cases were rearranged according to FISH, but not positive with the Ventana protocol. Furthermore, five cases were positive with the Ventana protocol, but not rearranged according to the FISH assay. Comparison between FISH and the Dako IHC protocol demonstrated eight positive cases with both methods. Additionally, five cases were rearranged with the FISH assay, but these cases showed negative protein expression. Moreover, three additional cases displayed positive ALK protein expression when using the Dako protocol, but were FISH negative.

45

If FISH is considered the reference method, the sensitivity and specificity of the Ventana IHC assay was 69% and 99%, respectively, while for the Dako assay it was 62% and 100%. To supplement the FISH and IHC assays, gene expression data from 194 patients (Uppsala I cohort) was available. Six cases were defined as gene expression positive, but only three of these were positive according to one of the three other ALK assays (FISH, Ventana IHC and Dako IHC). Notably, the two cases with the highest gene expression values were also positive with FISH. When incorporating clinical parameters, the correlation analysis revealed known clinical association with lower age, female sex, and non-smoking for all assays, but the grade of association was significant only for smoking sta-tus. A tendency towards a longer overall survival for ALK positive patients was observed with the three assays when combining all histologies, but sig-nificance was not reached either in the uni– or in the multivariate analyses. In conclusion, the results from this comparative study display a low frequen-cy of ALK aberrations in the Swedish population of NSCLC patients. Even though a FDA approved IHC assay was used to detect ALK fusion proteins, a low sensitivity was obtained, therefore, the use of this IHC assay when screening for ALK-positive patients is questionable. The poor overlap be-tween the different methods stresses the need for thorough validation of all ALK detection methods prior to implementation into clinical practice. Addi-tionally, we demonstrated that there are patients with positive ALK protein expression and high mRNA levels that are FISH negative, perhaps specify-ing a distinct group of patients that also would benefit from ALK inhibitor therapy.

46

Concluding remarks and future perspectives

Genetic aberrations, leading to constitutive protein activation, have been identified in subgroups of lung cancer patients and some of these aberrations are today targets of specific inhibitors. Unfortunately, these inhibitors are effective only in small subsets of patients and the response is often only temporary. Therefore, further efforts are warranted to better understand the molecular background of NSCLC in order to find new treatment targets or to refine existing therapy. The four papers presented in this thesis address this important aspect of clin-ical cancer research in different aspects – including evaluation of the expres-sion of potential new treatment targets and prognostic biomarkers, and the evaluation of detection strategy of an already established treatment target. In Paper I, ectopic expression of claudin 6 and claudin 18.2 was detected in small subsets of NSCLC patients. Since trials with monoclonal antibodies targeting claudin 6 (IMAB027) in ovarian cancer and claudin 18.2 (IMAB362) in gastro-esophageal cancer are ongoing (NCT02054351 and NCT01630083), these claudins may be targets in NSCLC as well. However, the effect of these antibodies has yet to be demonstrated. Furthermore, the frequency of claudin 6 and 18.2 expression in NSCLC is low, questioning a broader application of these therapeutic antibodies in this tumor entity. A demonstration of the functional importance of claudin 6 or 18.2 expression would be of interest for future studies, although antibody based therapy is not necessarily depending on the function of the protein. In Paper II and Paper III, we described the expression of the pro-inflammatory mediator COX-2 and the effect of the specific inhibition of COX-2 on survival of NSCLC patients. Since tissue inflammation possesses pro-tumorigenic properties, anti-inflammatory strategies may present valua-ble alternatives to standard treatment. To date, the most frequently studied anti-inflammatory target is COX-2 [68]. Also in lung cancer, many studies have evaluated the impact of COX-2 expression on prognosis or investigated the effect of specific COX-2 inhibitors in the therapy of NSCLC patients [79,159-160]. With the results from Paper II and III, we propose that COX-2 expression is not associated with prognosis and is of minor importance re-garding the treatment of NSCLC. It should be noted that other parameters

47

may be decisive for the response of COX-2 inhibitor therapy. For instance, the use of aspirin was associated with longer survival in patients with colo-rectal cancer harboring PIK3CA mutations [163]. Therefore, it would be interesting to analyze PIK3CA mutations in the tumor samples of the Cyclus cohort. Nonetheless, the minimal remaining tissue of these small biopsies makes this analysis difficult.

In Paper IV, we compared the results of different ALK-detection methods and analyzed the overall frequency of ALK rearrangements in NSCLC. Pa-tients harboring ALK translocations display a remarkable response to ALK-inhibitor therapy, and it is therefore of importance to accurately identify patients benefitting from this therapy. Our results stress the difficulties asso-ciated with the reliable diagnostic identification of ALK rearrangement in NSCLC samples. This study lead to the change of our in house IHC protocol to the FDA approved Ventana protocol in the routine diagnostic. Further-more, patients with clinical characteristics (non-smoker, younger age) indic-ative for ALK rearrangement should be evaluated directly using both FISH and IHC. Irrespective of the choice of analysis method we demonstrated a lower frequency of ALK rearrangements than previously reported. This stresses possible ethnic and geographical differences in the molecular back-ground of NSCLC.

The challenge of detecting ALK rearrangements has led to the introduction of other techniques in the diagnostic setting. Polymer chain reaction (PCR)-based techniques can be used, but these assays rely on the knowledge of the fusion transcript variants, and will miss novel or rare translocations and would therefore not be optimal [164]. Other studies have suggested next generation sequencing methods [165-166] as well as exon array profiling in the detection of translocations [167]. One of the most promising techniques for future diagnostics may be the NanoString technology, which is based on direct multiplexed measurement of gene targets in a single reaction [168]. The discordant tissue samples of our present study present the optimal vali-dation material for this new technique and analyses with the NanoString technology is planned in the near future. As a valuable consequence of the efforts to characterize the described bi-omarkers in NSCLC, the previously established Uppsala cohort (Uppsala I) was further expanded with molecular data. Even more important, a complete new NSCLC cohort (Uppsala II) was established. Thus, both cohorts, includ-ing altogether 708 patients, are a unique source for further studies integrating genomic, gene expression and protein data.

48

Acknowledgements

This work was performed at the Department of Immunology, Genetics and Pathology (IGP) at Uppsala University, and the Molecular Pathology unit at the Department of Clinical Pathology, Uppsala University Hospital. Finan-cial support was generously provided by the Lion’s Research Foundation, the Swedish Cancer Society and the Regional Lung Cancer Registry of the Uppsala-Örebro Region.

I would like to express my greatest gratitude to everyone who have encour-aged and supported me during my PhD studies. I would especially like to thank the following persons: All lung cancer patients for unselfishly donating their tissue for medical research. My greatest gratitude to my supervisor Patrick Micke – thank you for eve-rything! Thank you for sharing your wisdom, for believing in me, and for your patience and understanding. Thank you for putting things in perspective and for pushing me to do scary things. Thank you for giving me support and guidance when I needed it the most. My co-supervisor, Johan Botling, thank you for your encouragement and your valuable comments throughout these years. My second co-supervisor, Fredrik Pontén, thank you for always being so enthusiastic. Thank you for your valuable input in my projects. Special thanks to the former member in our group, Karolina Edlund. Thank you for all your great ideas and valuable comments and feedback. Thank you for all brainstorming sessions and for putting up with all my questions. I would like to thank the other members in our group. Firstly, Dijana, thank you for always supporting me, helping me put things in perspective and for your friendship. Secondly, Linnéa, thank you for being supportive and helpful throughout the years.

49

To all co-authors - thank you for good collaborations and for helping me with my research. It has been a pleasure working with you. The Molecular Pathology group: Monica Lindell, Åsa Håkansson, Mag-nus Sundström, Lotte Moens and Ulf Thunberg, thank you for being so friendly and for making me feel like one in the group! Thank you for an-swering all my questions and for helping me in the lab. Present and former room-mates at Rudbeck – thank you for providing such a friendly working environment, what would work be without you? Thank you Human Protein Atlas for letting me be a part of your research group. Thank you for your help with TMAs and scanning of slides! Thank you IH-lab for all your help. I would especially like to thank Dijana Cerjan and Inga Hansson for always helping me and for putting up with all my immunohistochemistry-questions. I would like to thank the whole Pathology Department here in Uppsala and especially Simin Tahmasebpoor for always being helpful and nice to me. Thank you Christina Magnusson for all your encouragement! We PhD students can not wish for a better administrator! Friends at Rudbeck! Thank you for your support and for sharing this fun but stressful time with me! Thank you friends outside of Rudbeck! Thank you for your encouragement and understanding! Now I hopefully will see you more often! Thank you Hanna! Without your support I don’t think I would have made it! Thank you Simona for your great pep-talks! Thank you Monica for your support and for helping me with my tense shoulders! Thank you Jennifer for all your valuable comments and your great support! Thank you Mi for your enthusiasm, and for pushing me and helping me throughout these years! You all mean a lot to me! Tack min kära familj. Tack Johan, mamma och pappa för att ni har stöttat och trott på mig alla år! Jag älskar er ♥ Daniel. Thank you for being so supportive and understanding. Thank you for always being there for me and for taking care of me during this stressful time. Jag älskar dig ♥

50

References

1. Cancerfondsrapporten 2015 (https://res.cloudinary.com/cancerfonden/image/upload/v1429886338/documents/cancerfondsrapporten-2015.pdf accessed 29th of February 2016)

2. Cancerfondsrapporten 2016 (https://res.cloudinary.com/cancerfonden/image/upload/v1458226058/documents/cancerfondsrapporten-2016.pdf accessed 18th of April 2016)

3. Cancer i siffror 2013 – Populärvetenskapliga fakta om cancer (http://www.socialstyrelsen.se/Lists/Artikelkatalog/Attachments/19108/2013-6-5.pdf accessed 29th of February 2016)

4. Nationellt vårdprogram för lungcancer (http://www.cancercentrum.se/globalassets/cancerdiagnoser/lunga-och-lungsack/vardprogram/natvp_lungcancer_2015-03-10.pdf accessed 10th of March 2016)

5. Blackledge G, Averbuch S. Gefitinib ('Iressa', ZD1839) and new epidermal growth factor receptor inhibitors. Br J Cancer. 2004;90:566-72.

6. Kwak EL, Bang YJ, Camidge DR, et al. Anaplastic lymphoma kinase inhibition in non-small-cell lung cancer. N Engl J Med. 2010;363:1693-703.

7. Carrizosa DR, Gold KA. New strategies in immunotherapy for non-small cell lung cancer. Transl Lung Cancer Res. 2015;4:553-9.

8. Ferlay J, Soerjomataram I, Dikshit R, et al. Cancer incidence and mortality worldwide: sources, methods and major patterns in GLOBOCAN 2012. Int J Cancer. 2015;136:E359-86.

9. Islami F, Torre LA, Jemal A. Global trends of lung cancer mortality and smok-ing prevalence. Transl Lung Cancer Res. 2015;4:327-38.

10. Ridge CA, McErlean AM, Ginsberg MS. Epidemiology of lung cancer. Semin Intervent Radiol. 2013;30:93-8.

11. Alberg AJ, Brock MV, Ford JG, et al. Epidemiology of lung cancer: Diagnosis and management of lung cancer, 3rd ed: American College of Chest Physicians evidence-based clinical practice guidelines. Chest. 2013;143:e1S-29S.

12. Sun S, Schiller JH, Gazdar AF. Lung cancer in never smokers--a different dis-ease. Nat Rev Cancer. 2007;7:778-90.

13. Pallis AG, Syrigos KN. Lung cancer in never smokers: disease characteristics and risk factors. Crit Rev Oncol Hematol. 2013;88:494-503.

14. Johnson DH, Schiller JH, Bunn PA Jr. Recent clinical advances in lung cancer management. J Clin Oncol. 2014;32:973-82.

15. Brennan P, Hainaut P, Boffetta P. Genetics of lung-cancer susceptibility. Lancet Oncol. 2011;12:399-408.

16. Larsen JE, Minna JD. Molecular biology of lung cancer: clinical implications. Clin Chest Med. 2011;32:703-40.

17. Park KS, Liang MC, Raiser DM, et al. Characterization of the cell of origin for small cell lung cancer. Cell Cycle. 2011;10:2806-15.

51

18. Rudin CM, Ismaila N, Hann CL, et al. Treatment of Small-Cell Lung Cancer: American Society of Clinical Oncology Endorsement of the American College of Chest Physicians Guideline. J Clin Oncol. 2015;33:4106-11.

19. Jackman DM, Johnson BE. Small-cell lung cancer. Lancet. 2005;366:1385-96. 20. Lungcancer. Årsrapport från Nationella lungcancerregistret (NLCR) 2014

(https://www.cancercentrum.se/globalassets/cancerdiagnoser/lunga-och-lungsack/kvalitetsregister/rapport/nlcr_2014_150930.pdf accessed 7th of April 2016)

21. Travis W.D, Brambilla E., Müller-Hermelink H.K., Harris C.C. (Editors.) World Health Organization (WHO) Pathology and Genetics of Tumours of the Lung, Pleura, Thymus and Heart. IARC Press, Lyon; 2004 (https://www.iarc.fr/en/publications/pdfs- online/pat-gen/bb10/BB10.pdf ac-cessed 7th of April 2016)

22. West H, Harpole D, Travis W. Histologic considerations for individualized systemic therapy approaches for the management of non-small cell lung cancer. Chest. 2009;136:1112-8.

23. Travis W.D, Brambilla E, Burke A.P, Marx A, Nicholson A. (Editors) WHO Classification of tumours of the lung, pleura, thymus and heart. Chapter 1, p 9-142. Lyon (France): IARC, 2015

24. Kadara H, Kabbout M, Wistuba II. Pulmonary adenocarcinoma: a renewed entity in 2011. Respirology. 2012;17:50-65.

25. Zugazagoitia J, Enguita AB, Nuñez JA, et al. The new IASLC/ATS/ERS lung adenocarcinoma classification from a clinical perspective: current concepts and future prospects. J Thorac Dis. 2014;6:S526-36.

26. Gerber DE, Gandhi L, Costa DB. Management and future directions in non-small cell lung cancer with known activating mutations. Am Soc Clin Oncol Educ Book. 2014:e353-65.

27. Cancer Genome Atlas Research Network. Comprehensive molecular profiling of lung adenocarcinoma. Nature. 2014;511:543-50.

28. Gandara DR, Hammerman PS, Sos ML, et al. Squamous cell lung cancer: from tumor genomics to cancer therapeutics. Clin Cancer Res. 2015;21:2236-43.

29. Stellman SD, Muscat JE, Thompson S, et al. Risk of squamous cell carcinoma and adenocarcinoma of the lung in relation to lifetime filter cigarette smoking. Cancer. 1997;80:382-8.

30. Heist RS, Sequist LV, Engelman JA. Genetic changes in squamous cell lung cancer: a review. J Thorac Oncol. 2012;7:924-33.

31. Cancer Genome Atlas Research Network. Comprehensive genomic characteri-zation of squamous cell lung cancers. Nature. 2012;489:519-25.

32. Rekhtman N, Ang DC, Sima CS, et al. Immunohistochemical algorithm for differentiation of lung adenocarcinoma and squamous cell carcinoma based on large series of whole-tissue sections with validation in small specimens. Mod Pathol. 2011;24:1348-59.

33. Micke P, Mattsson JS, Djureinovic D, et al. The Impact of the Fourth Edition of the WHO Classification of Lung Tumours on Histological Classification of Re-sected Pulmonary NSCCs. J Thorac Oncol. 2016. [Epub ahead of print]

34. Clinical Lung Cancer Genome Project (CLCGP); Network Genomic Medicine (NGM). A genomics-based classification of human lung tumors. Sci Transl Med. 2013;5:209ra153.

35. Socialstyreslen: Nationella riktlinjer för lungcancervård 2011 – stöd för styrning och ledning. (http://www.socialstyrelsen.se/Lists/Artikelkatalog/Attachments/18241/2011-3-2.pdf accessed 6th of March 2016)

52

36. Dietel M, Bubendorf L, Dingemans AM, et al. Diagnostic procedures for non-small-cell lung cancer (NSCLC): recommendations of the European Expert Group. Thorax. 2016;71:177-84.

37. Liam CK, Andarini S, Lee P, et al. Lung cancer staging now and in the future. Respirology. 2015;20:526-34.

38. Goldstraw P, Crowley J, Chansky K, et al. The IASLC Lung Cancer Staging Project: proposals for the revision of the TNM stage groupings in the forthcom-ing (seventh) edition of the TNM Classification of malignant tumours. J Thorac Oncol. 2007;2:706-14.

39. Tsim S, O'Dowd CA, Milroy R, et al. Staging of non-small cell lung cancer (NSCLC): a review. Respir Med. 2010;104:1767-74.

40. Goldstraw P. IASLC Staging Manual and Handbook in Thoracic Oncology. EditorialRx Press, Orange Park, FL (USA), 2009

41. Howington JA, Blum MG, Chang AC, et al. Treatment of stage I and II non-small cell lung cancer: Diagnosis and management of lung cancer, 3rd ed: Amer-ican College of Chest Physicians evidence-based clinical practice guidelines. Chest. 2013;143:e278S-313S.

42. Uramoto H, Tanaka F. Recurrence after surgery in patients with NSCLC. Transl Lung Cancer Res. 2014;3:242-9.

43. Pignon JP, Tribodet H, Scagliotti GV, et al. Lung adjuvant cisplatin evaluation: a pooled analysis by the LACE Collaborative Group. J Clin Oncol. 2008;26:3552-9.

44. NSCLC Meta-analyses Collaborative Group, Arriagada R, Auperin A, Burdett S, et al. Adjuvant chemotherapy, with or without postoperative radiotherapy, in operable non-small-cell lung cancer: two meta-analyses of individual patient da-ta. Lancet. 2010;375:1267-77.

45. Socinski MA, Evans T, Gettinger S, et al. Treatment of stage IV non-small cell lung cancer: Diagnosis and management of lung cancer, 3rd ed: American Col-lege of Chest Physicians evidence-based clinical practice guidelines. Chest. 2013;143:e341S-68S.

46. Sawyers C. Targeted cancer therapy. Nature. 2004;432:294-7. 47. Tseng HH, He B. Molecular markers as therapeutic targets in lung cancer. Chin

J Cancer. 2013;32:59-62. 48. Pedersen MW, Meltorn M, Damstrup L, et al. The type III epidermal growth

factor receptor mutation. Biological significance and potential target for anti-cancer therapy. Ann Oncol. 2001;12:745-60.

49. Lynch TJ, Bell DW, Sordella R, et al. Activating mutations in the epidermal growth factor receptor underlying responsiveness of non-small-cell lung cancer to gefitinib. N Engl J Med. 2004;350:2129-39.

50. Maemondo M, Inoue A, Kobayashi K, et al. Gefitinib or chemotherapy for non-small-cell lung cancer with mutated EGFR. N Engl J Med. 2010;362:2380-8.

51. Rosell R, Carcereny E, Gervais R, et al. Erlotinib versus standard chemotherapy as first-line treatment for European patients with advanced EGFR mutation-positive non-small-cell lung cancer (EURTAC): a multicentre, open-label, ran-domised phase 3 trial. Lancet Oncol. 2012;13:239-46.

52. Soda M, Choi YL, Enomoto M, et al. Identification of the transforming EML4-ALK fusion gene in non-small-cell lung cancer. Nature. 2007;448:561-6.

53. Shaw AT, Kim DW, Nakagawa K, et al. Crizotinib versus chemotherapy in advanced ALK-positive lung cancer. N Engl J Med. 2013;368:2385-94.

54. Shaw AT, Ou SH, Bang YJ, et al. Crizotinib in ROS1-rearranged non-small-cell lung cancer. N Engl J Med. 2014;371:1963-71.

53

55. Drilon A, Wang L, Hasanovic A, et al. Response to Cabozantinib in patients with RET fusion-positive lung adenocarcinomas. Cancer Discov. 2013;3:630-5.

56. Planchard D, Kim TM, Mazieres J, et al. Dabrafenib in patients with BRAF(V600E)-positive advanced non-small-cell lung cancer: a single-arm, multicentre, open-label, phase 2 trial. Lancet Oncol. 2016. [Epub ahead of print].

57. Lauro S, Onesti CE, Righini R, et al. The use of bevacizumab in non-small cell lung cancer: an update. Anticancer Res. 2014;34:1537-45.

58. Sandler A, Gray R, Perry MC, et al. Paclitaxel-carboplatin alone or with bevacizumab for non-small-cell lung cancer. N Engl J Med. 2006;355:2542-50.

59. Oxnard GR, Arcila ME, Sima CS, et al. Acquired resistance to EGFR tyrosine kinase inhibitors in EGFR-mutant lung cancer: distinct natural history of pa-tients with tumors harboring the T790M mutation. Clin Cancer Res. 2011;17:1616-22.

60. Doebele RC, Pilling AB, Aisner DL, et al. Mechanisms of resistance to crizotinib in patients with ALK gene rearranged non-small cell lung cancer. Clin Cancer Res. 2012;18:1472-82.

61. Miller VA, Hirsh V, Cadranel J, et al. Afatinib versus placebo for patients with advanced, metastatic non-small-cell lung cancer after failure of erlotinib, gefitinib, or both, and one or two lines of chemotherapy (LUX-Lung 1): a phase 2b/3 randomised trial. Lancet Oncol. 2012;13:528-38.

62. Cross DA, Ashton SE, Ghiorghiu S, et al. AZD9291, an irreversible EGFR TKI, overcomes T790M-mediated resistance to EGFR inhibitors in lung cancer. Can-cer Discov. 2014;4:1046-61.

63. Kim DW, Mehra R, Tan DS, et al. Activity and safety of ceritinib in patients with ALK-rearranged non-small-cell lung cancer (ASCEND-1): updated results from the multicentre, open-label, phase 1 trial. Lancet Oncol. 2016. [Epub ahead of print].

64. Shaw AT, Gandhi L, Gadgeel S, et al. Alectinib in ALK-positive, crizotinib-resistant, non-small-cell lung cancer: a single-group, multicentre, phase 2 trial. Lancet Oncol. 2016;17:234-42.

65. Brahmer J, Reckamp KL, Baas P, et al. Nivolumab versus Docetaxel in Ad-vanced Squamous-Cell Non-Small-Cell Lung Cancer. N Engl J Med. 2015;373:123-35.

66. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell. 2011;144:646-74.

67. Coussens LM, Werb Z. Inflammation and cancer. Nature. 2002;420:860-7. 68. Rayburn ER, Ezell SJ, Zhang R. Anti-Inflammatory Agents for Cancer Therapy.

Mol Cell Pharmacol. 2009;1:29-43. 69. Waddell WR, Loughry RW. Sulindac for polyposis of the colon. J Surg Oncol.

1983;24:83-7. 70. Ishikawa H, Wakabayashi K, Suzuki S, et al. Preventive effects of low-dose

aspirin on colorectal adenoma growth in patients with familial adenomatous polyposis: double-blind, randomized clinical trial. Cancer Med. 2013;2:50-6.

71. Cuzick J, Otto F, Baron JA, et al. Aspirin and non-steroidal anti-inflammatory drugs for cancer prevention: an international consensus statement. Lancet Oncol. 2009;10:501-7.

72. Cha YI, DuBois RN. NSAIDs and cancer prevention: targets downstream of COX-2. Annu Rev Med. 2007;58:239-52.

73. Sahin M, Sahin E, Gümüslü S. Cyclooxygenase-2 in cancer and angiogenesis. Angiology. 2009;60:242-53.

54

74. Hida T, Yatabe Y, Achiwa H, et al. Increased expression of cyclooxygenase 2 occurs frequently in human lung cancers, specifically in adenocarcinomas. Can-cer Res. 1998;58:3761-4.

75. Hwang D, Scollard D, Byrne J, et al. Expression of cyclooxygenase-1 and cy-clooxygenase-2 in human breast cancer. J Natl Cancer Inst. 1998;90:455-60.

76. Ogino S, Kirkner GJ, Nosho K, et al. Cyclooxygenase-2 expression is an inde-pendent predictor of poor prognosis in colon cancer. Clin Cancer Res. 2008;14:8221-7.

77. Jana D, Sarkar DK, Ganguly S, et al. Role of Cyclooxygenase 2 (COX-2) in Prognosis of Breast Cancer. Indian J Surg Oncol. 2014;5:59-65.

78. Denkert C, Köbel M, Pest S, et al. Expression of cyclooxygenase 2 is an inde-pendent prognostic factor in human ovarian carcinoma. Am J Pathol. 2002;160:893-903.

79. Zhan P, Qian Q, Yu LK. Prognostic value of COX-2 expression in patients with non-small cell lung cancer: a systematic review and meta-analysis. J Thorac Dis. 2013;5:40-7.

80. Yokouchi H, Kanazawa K, Ishida T, et al. Cyclooxygenase-2 inhibitors for non-small-cell lung cancer: A phase II trial and literature review. Mol Clin Oncol. 2014;2:744-750.

81. Oldenhuis CN, Oosting SF, Gietema JA, et al. Prognostic versus predictive value of biomarkers in oncology. Eur J Cancer. 2008;44:946-53.

82. Yu Y, He J. Molecular classification of non-small-cell lung cancer: diagnosis, individualized treatment, and prognosis. Front Med. 2013;7:157-71.

83. Berghmans T, Paesmans M, Sculier JP. Prognostic factors in stage III non-small cell lung cancer: a review of conventional, metabolic and new biological varia-bles. Ther Adv Med Oncol. 2011;3:127-38.

84. Tas F, Ciftci R, Kilic L, et al. Age is a prognostic factor affecting survival in lung cancer patients. Oncol Lett. 2013;6:1507-1513.

85. Conley BA, Taube SE. Prognostic and predictive markers in cancer. Dis Mark-ers. 2004;20:35-43.

86. Passiglia F, Bronte G, Castiglia M, et al. Prognostic and predictive biomarkers for targeted therapy in NSCLC: for whom the bell tolls? Expert Opin Biol Ther. 2015;15:1553-66.

87. Kisluk J, Ciborowski M, Niemira M, et al. Proteomics biomarkers for non-small cell lung cancer. J Pharm Biomed Anal. 2014;101:40-9.

88. Warford A, Howat W, McCafferty J. Expression profiling by high-throughput immunohistochemistry. J Immunol Methods. 2004;290:81-92.

89. Oliver C, Jamur MC (Editors). Immunohistochemical methods and protocols. 3rd edition. Chapter 1, 6, 24. Springer Science, London, 2010

90. Hoffman W, Lakkis FG, Chalasani G. B Cells, Antibodies, and More. Clin J Am Soc Nephrol. 2016;11:137-54.

91. Buchwalow, IB. Werner B. Immunohistochemistry. Basics and methods. 1st edition. Chapter 1, 4, 6, 10. Hamburg, Springer – Verlag Berlin Heidelberg; 2010

92. Thomas Boenisch. Immunohistochemical Staining Methods, 5th Edition, Chap-ter 1: Antibodies, pp. 1-9. (http://www.dako.com/08002_03aug09_ihc_guidebook_5th_edition_chapter_1.pdf accessed 17th of April 2016)

93. Leong TY, Cooper K, Leong AS. Immunohistology--past, present, and future. Adv Anat Pathol. 2010;17:404-18.

55

94. Heitzmann H, Richards FM. Use of the avidin-biotin complex for specific stain-ing of biological membranes in electron microscopy. Proc Natl Acad Sci U S A. 1974;71:3537-41.

95. Sabattini E, Bisgaard K, Ascani S, et al. The EnVision++ system: a new immuno histochemical method for diagnostics and research. Critical compari-son with the APAAP, ChemMate, CSA, LABC, and SABC techniques. J Clin Pathol. 1998;51:506-11.

96. Nitta H, Tsuta K, Yoshida A, et al. New methods for ALK status diagnosis in non-small-cell lung cancer: an improved ALK immunohistochemical assay and a new, Brightfield, dual ALK IHC-in situ hybridization assay. J Thorac Oncol. 2013;8:1019-31.

97. Chipinda I, Hettick JM, Siegel PD. Haptenation: chemical reactivity and protein binding. J Allergy (Cairo). 2011;2011:839682.

98. Bobrow MN, Harris TD, Shaughnessy KJ, et al. Catalyzed reporter deposition, a novel method of signal amplification. Application to immunoassays. J Immunol Methods. 1989;125:279-85.

99. Optiview Detection Chemistry. (http://www.ventana.com/documents/OptiView_F&B_Brochure.pdf accessed 23rd of February 2016).

100. Waters JJ, Barlow AL, Gould CP. Demystified ... FISH. Mol Pathol. 1998;51:62-70.

101. Pardue ML, Gall JG. Chromosomal localization of mouse satellite DNA. Sci-ence. 1970;168:1356-8.

102. Dyer SA, Green EK. Fluorescent in situ hybridization. Methods Mol Biol. 2002;187:73-86.

103. Pinkel D, Landegent J, Collins C, et al. Fluorescence in situ hybridization with human chromosome-specific libraries: detection of trisomy 21 and transloca-tions of chromosome 4. Proc Natl Acad Sci USA. 1998;85:9138-42.

104. Bishop, R. Applications of fluorescence in situ hybridization (FISH) in detect-ing genetic aberrations of medical significance. Bioscience Horizons (2010);3):85-95.

105. Nath J, Johnson KL. A review of fluorescence in situ hybridization (FISH): current status and future prospects. Biotech Histochem. 2000;75:54-78.

106. Morrison LE, Ramakrishnan R, Ruffalo TM, et al. Labeling fluorescence in situ hybridization probes for genomic targets. Methods Mol Biol. 2002;204:21-40.

107. Kulkarni, S., Pfeifer, J. (Editors). Clinical genomics. A guide to clinical next generation sequencing. 1st edition. Chapter 10, p 152-165. Washington (USA), 2015.

108. Bartlett, JMS, Shaaban, A., Schmitt F. (Editors). Molecular Pathology. A Prac-tical Guide for the Surgical Pathologist and Cytopathologist. 1st edition. Chapter 4, p 29-54. Cambridge (United Kingdom), 2015.

109. Berg D, Malinowsky K, Reischauer B, et al. Use of formalin-fixed and paraffin-embedded tissues for diagnosis and therapy in routine clinical settings. Methods Mol Biol. 2011;785:109-22.

110. Ludyga N, Grünwald B, Azimzadeh O, et al. Nucleic acids from long-term preserved FFPE tissues are suitable for downstream analyses. Virchows Arch. 2012;460:131-40.

111. Kononen J, Bubendorf L, Kallioniemi A, et al. Tissue microarrays for high-throughput molecular profiling of tumor specimens. Nat Med. 1998;4:844-7.

112. Nocito A, Kononen J, Kallioniemi OP, et al. Tissue microarrays (TMAs) for high-throughput molecular pathology research. Int J Cancer. 2001;94:1-5.

56

113. Kallioniemi OP, Wagner U, Kononen J, et al. Tissue microarray technology for high-throughput molecular profiling of cancer. Hum Mol Genet. 2001;10:657-62.

114. Torhorst J, Bucher C, Kononen J, et al. Tissue microarrays for rapid linking of molecular changes to clinical endpoints. Am J Pathol. 2001;159:2249-56.

115. Slonim DK, Yanai I. Getting started in gene expression microarray analysis. PLoS Comput Biol. 2009;5:e1000543.

116. Auer H, Newsom DL, Kornacker K. Expression Profiling Using Affymetrix GeneChip Microarrays. Methods Mol Biol. 2009;509:35-46.

117. Göhlmann H, Talloen W. Gene Expression Studies Using Affymetrix Microar-rays. Chapter 2-5. Chapman & Hall/CRC, Taylor & Francis Group (USA), 2009.

118. Datasheet: Affymetrix GeneChip® Human Genome Arrays. (http://media.affymetrix.com/support/technical/datasheets/human_datasheet.pdf accessed 17th of April 2016).

119. Haferlach T, Bacher U, Kohlmann A, Haferlach C. Discussion of the applicabil-ity of microarrays: profiling of leukemias. Methods Mol Biol. 2009;509:15-33.

120. Günzel D, Fromm M. Claudins and other tight junction proteins. Compr Physiol. 2012;2:1819-52.

121. Krause G, Winkler L, Mueller SL, et al. Structure and function of claudins. Biochim Biophys Acta. 2008;1778:631-45.

122. Günzel D, Yu AS. Claudins and the modulation of tight junction permeability. Physiol Rev. 2013;93:525-69.

123. Gupta IR, Ryan AK. Claudins: unlocking the code to tight junction function during embryogenesis and in disease. Clin Genet. 2010;77:314-25.

124. Hewitt KJ, Agarwal R, Morin PJ. The claudin gene family: expression in nor-mal and neoplastic tissues. BMC Cancer. 2006;6:186.

125. Ushiku T, Shinozaki-Ushiku A, Maeda D, et al. Distinct expression pattern of claudin-6, a primitive phenotypic tight junction molecule, in germ cell tumours and visceral carcinomas. Histopathology. 2012;61:1043-56.

126. Türeci O, Koslowski M, Helftenbein G, et al. Claudin-18 gene structure, regula-tion, and expression is evolutionary conserved in mammals. Gene. 2011;481:83-92.

127. Sahin U, Koslowski M, Dhaene K, et al. Claudin-18 splice variant 2 is a pan-cancer target suitable for therapeutic antibody development. Clin Cancer Res. 2008;14:7624-34.

128. Dubois RN, Abramson SB, Crofford L, et al. Cyclooxygenase in biology and disease. FASEB J. 1998;12:1063-73.

129. Zha S, Yegnasubramanian V, Nelson WG, et al. Cyclooxygenases in cancer: progress and perspective. Cancer Lett. 2004;215:1-20.

130. Dempke W, Rie C, Grothey A, et al. Cyclooxygenase-2: a novel target for can-cer chemotherapy? J Cancer Res Clin Oncol. 2001;127:411-7.

131. Maekawa M, Sugano K, Sano H, et al. Increased expression of cyclooxygenase-2 to -1 in human colorectal cancers and adenomas, but not in hyperplastic polyps. Jpn J Clin Oncol. 1998;28:421-6.

132. Chan G, Boyle JO, Yang EK, et al. Cyclooxygenase-2 expression is up-regulated in squamous cell carcinoma of the head and neck. Cancer Res. 1999;59:991-4.

133. Shamma A, Yamamoto H, Doki Y, et al. Up-regulation of cyclooxygenase-2 in squamous carcinogenesis of the esophagus. Clin Cancer Res. 2000;6:1229-38.

134. Gupta S, Srivastava M, Ahmad N, et al. Over-expression of cyclooxygenase-2 in human prostate adenocarcinoma. Prostate. 2000;42:73-8.

57

135. Steinbach G, Lynch PM, Phillips RK, et al. The effect of celecoxib, a cyclooxy-genase-2 inhibitor, in familial adenomatous polyposis. N Engl J Med. 2000;342:1946-52.

136. Pagliarulo V, Ancona P, Martines I, et al. Celecoxib for the prevention of nonmuscle invasive bladder cancer: results from a matched control study. Ther Adv Urol. 2015;7:303-11.

137. Smith MR, Manola J, Kaufman DS, et al. Celecoxib versus placebo for men with prostate cancer and a rising serum prostate-specific antigen after radical prostatectomy and/or radiation therapy. J Clin Oncol. 2006;24:2723-8.

138. Chan AT, Ogino S, Fuchs CS. Aspirin use and survival after diagnosis of colo-rectal cancer. JAMA. 2009;302:649-58.

139. Liu R, Xu KP, Tan GS. Cyclooxygenase-2 inhibitors in lung cancer treatment: Bench to bed. Eur J Pharmacol. 2015;769:127-33.

140. Iwahara T, Fujimoto J, Wen D, et al. Molecular characterization of ALK, a receptor tyrosine kinase expressed specifically in the nervous system. Onco-gene. 1997;14:439-49.

141. Hallberg B, Palmer RH. Mechanistic insight into ALK receptor tyrosine kinase in human cancer biology. Nat Rev Cancer. 2013;13:685-700.

142. Togashi Y, Soda M, Sakata S, et al. KLC1-ALK: a novel fusion in lung cancer identified using a formalin-fixed paraffin-embedded tissue only. PLoS One. 2012;7:e31323.

143. Takeuchi K, Choi YL, Togashi Y, et al. KIF5B-ALK, a novel fusion oncokinase identified by an immunohistochemistry-based diagnostic system for ALK-positive lung cancer. Clin Cancer Res. 2009;15:3143-9.

144. Rikova K, Guo A, Zeng Q, et al. Global survey of phosphotyrosine signaling identifies oncogenic kinases in lung cancer. Cell. 2007;131:1190-203.

145. Micke P, Edlund K, Holmberg L, et al. Gene copy number aberrations are asso-ciated with survival in histologic subgroups of non-small cell lung cancer. J Thorac Oncol. 2011;6:1833-40.

146. Botling J, Edlund K, Lohr M, et al. Biomarker discovery in non-small cell lung cancer: integrating gene expression profiling, meta-analysis, and tissue microar-ray validation. Clin Cancer Res. 2013;19:194-204.

147. Lohr M, Edlund K, Botling J, et al. The prognostic relevance of tumour-infiltrating plasma cells and immunoglobulin kappa C indicates an important role of the humoral immune response in non-small cell lung cancer. Cancer Lett. 2013;333:222-8.

148. Edlund K, Lindskog C, Saito A, et al. CD99 is a novel prognostic stromal mark-er in non-small cell lung cancer. Int J Cancer. 2012;131:2264-73.

149. Tran L, Mattsson JS, Nodin B, et al. Various Antibody Clones of Napsin A, Thyroid Transcription Factor 1, and p40 and Comparisons With Cytokeratin 5 and p63 in Histopathologic Diagnostics of Non-Small Cell Lung Carcinoma. Appl Immunohistochem Mol Morphol. 2015 [Epub ahead of print].

150. Fernandes OJ, Almgren SO, Thaning L, et al. Prognostic factors for the survival of surgically treated patients for non-small cell lung cancer. Acta Oncol. 2003;42:338-41.

151. Karlsson C, Helenius G, Fernandes O, et al. Oestrogen receptor β in NSCLC - prevalence, proliferative influence, prognostic impact and smoking. APMIS. 2012;120:451-8.

152. Koch A, Bergman B, Holmberg E, et al. Effect of celecoxib on survival in pa-tients with advanced non-small cell lung cancer: a double blind randomised clinical phase III trial (CYCLUS study) by the Swedish Lung Cancer Study Group. Eur J Cancer. 2011;47:1546-55.

58

153. Karlsson C, Bodin L, Piehl-Aulin K, et al. Tissue microarray validation: a methodologic study with special reference to lung cancer. Cancer Epidemiol Biomarkers Prev. 2009;18:2014-21.

154. Camp RL, Charette LA, Rimm DL. Validation of tissue microarray technology in breast carcinoma. Lab Invest. 2000;80:1943-9.

155. IASLC atlas of ALK testing in lung cancer. (https://www.iaslc.org/publications/iaslc-atlas-alk-testing-lung-cancer accessed 23rd of January 2016).

156. Tibshirani R, Hastie T. Outlier sums for differential gene expression analysis. Biostatistics. 2007;8:2-8.

157. Giovannucci E, Rimm EB, Stampfer MJ, et al. Aspirin use and the risk for colo-rectal cancer and adenoma in male health professionals. Ann Intern Med. 1994;121:241-6.

158. Giardiello FM, Hamilton SR, Krush AJ, et al. Treatment of colonic and rectal adenomas with sulindac in familial adenomatous polyposis. N Engl J Med. 1993;328:1313-6.

159. Groen HJ, Sietsma H, Vincent A, et al. Randomized, placebo-controlled phase III study of docetaxel plus carboplatin with celecoxib and cyclooxygenase-2 ex-pression as a biomarker for patients with advanced non-small-cell lung cancer: the NVALT-4 study. J Clin Oncol. 2011;29:4320-6.

160. Edelman MJ, Watson D, Wang X, et al. Eicosanoid modulation in advanced lung cancer: cyclooxygenase-2 expression is a positive predictive factor for celecoxib + chemotherapy--Cancer and Leukemia Group B Trial 30203. J Clin Oncol. 2008;26:848-55.

161. Lilenbaum R, Socinski MA, Altorki NK, et al. Randomized phase II trial of docetaxel/irinotecan and gemcitabine/irinotecan with or without celecoxib in the second-line treatment of non-small-cell lung cancer. J Clin Oncol. 2006;24:4825-32.

162. Food and Drug Administratiuon Approval Application (PMA) Number for VENTANA ALK (D5F3) CDx Assay. Summary data sheet: (www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfTopic/pma/pma.cfm?num=P140025 accessed 18th of January 2016).

163. Liao X, Lochhead P, Nishihara R, et al. Aspirin use, tumor PIK3CA mutation, and colorectal-cancer survival. N Engl J Med. 2012;367:1596-606.

164. Takeuchi K, Choi YL, Soda M, et al. Multiplex reverse transcription-PCR screening for EML4-ALK fusion transcripts. Clin Cancer Res. 2008;14:6618-24.

165. Abate F, Zairis S, Ficarra E, et al. Pegasus: a comprehensive annotation and prediction tool for detection of driver gene fusions in cancer. BMC Syst Biol. 2014;8:97.

166. Peled N, Palmer G, Hirsch FR, et al. Next-generation sequencing identifies and immunohistochemistry confirms a novel crizotinib-sensitive ALK rearrange-ment in a patient with metastatic non-small-cell lung cancer. J Thorac Oncol. 2012;7:e14-6.

167. Lin E, Li L, Guan Y, et al. Exon array profiling detects EML4-ALK fusion in breast, colorectal, and non-small cell lung cancers. Mol Cancer Res. 2009;7:1466-76.

168. Lira ME, Kim TM, Huang D, et al. Multiplexed gene expression and fusion transcript analysis to detect ALK fusions in lung cancer. J Mol Diagn. 2013;15:51-61.

Acta Universitatis UpsaliensisDigital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Medicine 1229

Editor: The Dean of the Faculty of Medicine

A doctoral dissertation from the Faculty of Medicine, UppsalaUniversity, is usually a summary of a number of papers. A fewcopies of the complete dissertation are kept at major Swedishresearch libraries, while the summary alone is distributedinternationally through the series Digital ComprehensiveSummaries of Uppsala Dissertations from the Faculty ofMedicine. (Prior to January, 2005, the series was publishedunder the title “Comprehensive Summaries of UppsalaDissertations from the Faculty of Medicine”.)

Distribution: publications.uu.seurn:nbn:se:uu:diva-285844

ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2016