ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2018

Digital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Medicine 1463

Transcriptomic and ProteomicAnalysis of Tumor Markers inTissue and Blood from Patientswith Lung Cancer

DIJANA DJUREINOVIC

ISSN 1651-6206ISBN 978-91-513-0328-4urn:nbn:se:uu:diva-348349

Dissertation presented at Uppsala University to be publicly examined in Rudbecksalen,Rudbecklaboratoriet, Dag Hammarskjölds väg 20, Uppsala, Friday, 8 June 2018 at 09:00 forthe degree of Doctor of Philosophy (Faculty of Medicine). The examination will be conductedin English. Faculty examiner: Professor Keith Kerr (University of Aberdeen).

AbstractDjureinovic, D. 2018. Transcriptomic and Proteomic Analysis of Tumor Markers in Tissueand Blood from Patients with Lung Cancer. Digital Comprehensive Summaries of UppsalaDissertations from the Faculty of Medicine 1463. 52 pp. Uppsala: Acta UniversitatisUpsaliensis. ISBN 978-91-513-0328-4.

Despite recent treatment advancements, the survival outcome remains poor for the majorityof patients with non-small cell lung cancer (NSCLC). The aim of this thesis was to evaluateprotein expression to predict prognosis and identify biomarkers that can be used as targets forimmunotherapy or for early detection of NSCLC.

In Paper I an optimized immunohistochemistry (IHC)-based prognostic model wasdeveloped for NSCLC. The prognostic performance of the model was compared to theclinicopathological parameters that are used in the clinical setting to predict outcome. Theprotein model failed to outperform clinicopathological parameters in predicting survivaloutcome questioning the potential of IHC-based assessment of prognostic markers in NSCLC.

In Paper II the human testis-specific proteome was profiled using RNA-sequencing (RNA-seq) data from testis and 26 other organs. More than 1000 genes demonstrated a testis-enrichedexpression pattern which makes testis the tissue with the most tissue-specific genes. Themajority of the testis-enriched genes were previously poorly described and were further profiledby IHC. This analysis provides a starting point to increase the molecular understanding oftesticular biology.

In Paper III the profiling of cancer-testis antigens (CTAs) was performed in NSCLC byusing RNA-seq data from 32 normal organs and NSCLC. Ninety genes showed CTA expressionprofiles. The transcriptomic data were validated by IHC for several CTAs. The comprehensiveanalysis of CTAs can guide biomarker studies or help to identify targets for immunotherapeuticstrategies.

In Paper IV the reactivity of CTAs was evaluated by measuring the abundance ofautoantibodies in plasma from patients with NSCLC and benign lung diseases. Twenty-nineCTAs demonstrated exclusive reactivity in NSCLC and six of them were reactive in anindependent NSCLC cohort. These findings suggest that some CTAs are immunogenic andcould be utilized in immunotherapy.

In Paper V an immunoassay was used on lung adenocarcinoma plasma samples and samplesfrom benign lung diseases. The plasma levels of 92 cancer related proteins were used to build amodel that discriminated lung adenocarcinoma from benign controls with a sensitivity of 93%and a specificity of 64%. The results indicate that this assay is promising for the early detectionof NSCLC.

In summary, this thesis presents an integrative analysis of lung cancer tissue and bloodsamples to characterize NSCLC on the transcriptomic and proteomic level and to identify cancerspecific proteins.

Keywords: non-small cell lung cancer, prognostic biomarkers, cancer-testis antigens,prediction model, tumor markers, autoantibodies, testis, screening

Dijana Djureinovic, Department of Immunology, Genetics and Pathology,Rudbecklaboratoriet, Uppsala University, SE-751 85 Uppsala, Sweden.

© Dijana Djureinovic 2018

ISSN 1651-6206ISBN 978-91-513-0328-4urn:nbn:se:uu:diva-348349 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-348349)

All cancers are alike but they are alike in a unique way.

~ Siddhartha Mukherjee

List of Papers

This thesis is based on the following papers, which are referred to in the text by their Roman numerals. Reprints were made with permission from the re-spective publishers. * Authors contributed equally.

I Grinberg M*., Djureinovic D*., Brunnström HRR., Mattsson

JSM., Edlund K., Hengstler JG., La Fleur L., Ekman S., Koyi H., Branden E., Ståhle E., Jirström K., Tracy DK., Pontén F., Botling J, Rahnenfuhrer J., Micke P (2017) Reaching the limits of prog-nostication in non-small cell lung cancer: an optimized bi-omarker panel fails to outperform clinical parameters. Mod Pathol. 30(7):964–977

II Djureinovic D., Fagerberg L., Hallström B., Danielsson A., Lindskog C., Uhlén M., Pontén F (2014) The human testis-spe-cific proteome defined by transcriptomics and antibody-based profiling. Mol Hum Reprod. 20(6):476-88

III Djureinovic D*., Hallström BM*., Horie M., Mattsson JSM., La

Fleur L., Fagerberg L., Brunnström H., Lindskog C., Madjar K., Rahnenfuhrer J., Ekman S., Ståhle E., Koyi H., Brandén E., Ed-lund K., Hengstler JG., Lambe M., Saito A., Botling J., Pontén F., Uhlén M., Micke P (2016) Profiling cancer testis antigens in non-small-cell lung cancer. JCI Insight.1(10):e86837

IV Djureinovic D., Dodig-Crnkovic T., Hellström C., Holgersson

G., Bergqvist M., Mattsson JSM., Pontén F., Ståhle E., Schwenk JM., Micke P (2018) Detection of autoantibodies against cancer-testis antigens in non-small cell lung cancer. Manuscript

V Djureinovic D., Pontén V., Landelius P., Al Sayegh S., Kappert

K., Kamali-Moghaddam M., Micke P., Ståhle E (2018) Multi-plex plasma protein profiling identifies novel markers to discrim-inate patients with adenocarcinoma of the lung. Manuscript

Related Publications

Fagerberg L., Hallström BM., Oksvold P., Kampf C., Djureinovic D., Odeberg J., Habuka M., Tahmasebpoor S., Danielsson A., Edlund K., Asplund A., Sjöstedt E., Lundberg E., Szigyarto CA., Skogs M., Takanen JO., Berling H., Tegel H., Mulder J., Nilsson P., Schwenk JM., Lindskog C., Danielsson F., Mardinoglu A., Sivertsson A., von Feilitzen K., Forsberg M., Zwahlen M., Olsson I., Navani S., Huss M., Nielsen J., Ponten F., Uhlén M (2014) Analysis of the human tissue-specific expression by genome-wide integration of tran-scriptomics and antibody-based proteomics. Mol Cell Proteomics.13(2):397-406 Gremel G., Bergman J., Djureinovic D., Edqvist PH., Maindad V., Bharambe BM., Khan WA., Navani S., Elebro J., Jirström K., Hellberg D., Uhlén M., Micke P., Pontén F (2014) A systemic analysis of commonly used antibodies in cancer diagnostics. Histopathology. 64(2):293-305 Uhlén M., Fagerberg L., Hallström BM., Lindskog C., Oksvold P., Mardi-noglu A., Sivertsson Å., Kampf C., Sjöstedt E., Asplund A., Olsson I., Edlund K., Lundberg E., Navani S., Szigyarto CA., Odeberg J., Djureinovic D., Takanen JO., Hober S., Alm T., Edqvist PH., Berling H., Tegel H., Mulder J., Rockberg J., Nilsson P., Schwenk JM., Hamsten M., von Feilitzen K., Forsberg M., Persson L., Johansson F., Zwahlen M., von Heijne G., Nielsen J., Pontén F (2015) Proteomics. Tissue-based map of the human proteome. Science. 347(6220):1260419 Micke P., Mattsson JS., Djureinovic D., Nodin B., Jirström K., Tran L., Jöns-son P., Planck M., Botling J., Brunnström H (2016) The Impact of the Fourth Edition of the WHO Classification of Lung Tumours on Histological Classi-fication of Resected Pulmonary NSCCs. J Thorac Oncol. 11(6):862-72 Bergman J., Botling J., Fagerberg L., Hallström BM., Djureinovic D., Uhlén M., Pontén F (2017) The human adrenal gland proteome defined by tran-scriptomics and antibody-based profiling. Endocrinology. 158(2):239-251

Gremel G., Djureinovic D., Niinivirta M., Laird A., Ljungqvist O., Johannes-son H., Bergman J., Edqvist PH., Navani S., Khan N., Patil T., Sivertsson Å., Uhlén M., Harrison DJ., Ullenhag GJ., Stewart GD., Pontén F (2017) A sys-tematic search strategy identifies cubilin as independent prognostic marker for renal cell carcinoma. BMC Cancer. 4;17(1):9 Brunnström H., Johansson A., Westbom-Fremer S., Backman M., Djurei-novic D., Patthey A., Isaksson-Mettävainio M., Gulyas M., Micke P (2017) PD-L1 immunohistochemistry in clinical diagnostics of lung cancer: inter-pathologist variability is higher than assay variability. Mod Pathol.30(10):1411-1421 Uhlen M., Zhang C., Lee S., Sjöstedt E., Fagerberg L., Bidkhori G., Benfeitas R., Arif M., Liu Z., Edfors F., Sanli K., von Feilitzen K., Oksvold P., Lundberg E., Hober S., Nilsson P., Mattsson J., Schwenk JM., Brunnström H., Glimelius B., Sjöblom T., Edqvist PH., Djureinovic D., Micke P., Lindskog C., Mardi-noglu A., Ponten F (2017) A pathology atlas of the human cancer transcrip-tome. Science. 18;357(6352) Mezheyeuski A., Bergsland CH., Backman M., Djureinovic D., Sjöblom T., Bruun J., Micke P (2018) Multispectral imaging for quantitative and compart-ment-specific immune infiltrates reveals distinct immune profiles that classify lung cancer patients. J Pathol. 244(4):421-431

Contents

Introduction ...............................................................................................13

Lung Cancer ..............................................................................................14Risk Factors ..........................................................................................14Diagnosis and Staging ...........................................................................14Histology ..............................................................................................17Treatment..............................................................................................18

Conventional Therapies ....................................................................18Targeted Therapy..............................................................................19Immunotherapy ................................................................................19

Prognostic Factors .................................................................................20Early Detection .....................................................................................21

Testis.........................................................................................................23Cancer-Testis Antigens .........................................................................24

Present Investigations ................................................................................26Aim ......................................................................................................26Methodological Background .................................................................27

Immunohistochemistry .....................................................................27Tissue Microarrays ...........................................................................28RNA-Sequencing ..............................................................................29Suspension Bead Array .....................................................................31Proximity Extension Assay ...............................................................32Statistics ...........................................................................................33

Patient Material .....................................................................................34Paper I - Reaching the Limits with Prognostication in NSCLC ..............36Paper II - The Human Testis Specific Proteome.....................................37Paper III - Profiling CTAs in NSCLC ....................................................38Paper IV - Detection of Autoantibodies against CTAs in NSCLC ..........38Paper V - Plasma Protein Profiling to Discriminate Patients with Lung Adenocarcinoma ...................................................................................39Concluding Remarks .............................................................................40

Acknowledgements ...................................................................................43

References .................................................................................................45

Abbreviations

ALK Anaplastic lymphoma kinase BAGE B melanoma antigen CT Computed tomography CTAs Cancer-testis antigens CTdatabase Cancer-testis database CTLA-4 Cytotoxic T-lymphocyte associated antigen 4 cTNM Clinical tumor-node-metastasis DAB 3,3'-diaminobenzidine DNA Deoxyribonucleic acid ECOG Eastern Cooperative Oncology Group EDTA Ethylenediaminetetraacetic acid EGFR Epidermal growth factor receptor EML4 Echinoderm microtubule associated protein-like 4 ERCC1 Excision repair cross-complementing rodent repair deficiency,

complementation group 1 FDA Food and Drug Administration FFPE Formalin-fixed-paraffin-embedded FPKM Fragments per kilobase of exon model per million mapped reads GAGE G antigen HE Hematoxylin-eosin HPA Human Protein Atlas HRP Horseradish peroxidase IgG Immunoglobulin G IHC Immunohistochemistry LDCT Low-dose computed tomography MAGE-A1 Melanoma antigen family A, 1 MFI Median fluorescence intensity mRNA Messenger ribonucleic acid NLST National Lung Screening Trial NSCLC Non-small cell lung cancer NSCLC-NOS Non-small cell lung cancer-not otherwise specified NY-ESO-1 New York esophageal squamous cell carcinoma-1 PCR Polymerase chain reaction PD-1 Programmed cell death protein-1 PD-L1 Programmed cell death ligand-1 PEA Proximity extension assay

PS Performance status pTNM pathological tumor-node-metastasis RET Ret proto-oncogene RIN Ribonucleic acid integrity number RNA-seq Ribonucleic acid-sequencing ROS1 ROS proto-oncogene 1, receptor tyrosine kinase RRM1 Ribonucleoside-diphosphate reductase SEREX Serological analysis of recombinant cDNA expression libraries SSX-2 Synovial sarcoma, X breakpoint 2 TMA Tissue microarray TNM Tumor-node-metastasis U.S United States

13

Introduction

Lung cancer is a global health problem that annually accounts for one fifth of all cancer related deaths (1). Based on histology, lung cancers are divided into two main entities, small-cell lung cancer and non-small cell lung cancer (NSCLC). The work presented in this thesis focuses on NSCLC, the histolog-ical subtype that comprises more than 80% of all lung cancer diagnoses (2). During the last two decades major treatment advances have been made, with the identification of several targetable genomic aberrations and the more re-cent introduction of immunotherapy for treatment of a subset of NSCLC pa-tients (3). Despite these improvements, the outcome for NSCLC patients re-mains poor across all stages and histological subgroups. Over half of patients die within one year from the time of diagnosis and only 15% of patients are still alive five-years after diagnosis. The dismal outcome can partly be ex-plained by delayed diagnosis and insufficient treatment options (4).

The work in this thesis aims to evaluate protein expression to predict progno-sis. Tissue and plasma-based assays were applied to identify molecules that can be used as biomarkers or targets for immunotherapy. Finally, a multi-plexed-protein based analysis was used for the early detection of cancer.

14

Lung Cancer

Risk Factors At the beginning of the 20th century lung cancer was a rare disease, however today it is the most commonly diagnosed cancer around the world (1, 5). The main reason for the dramatic increase in lung cancer incidence is attributed to the widespread usage of cigarettes that began towards the end of the 19th cen-tury (5). Cigarette smoke contains at least 70 different carcinogens that have the ability to induce changes in the genomic DNA (6). If not repaired, somatic mutations can occur in oncogenes and tumor suppressor genes, which can re-sult in loss of normal cellular growth and the development of cancer. Nine out of ten lung cancers are estimated to be caused by smoking (7). Compared to non-smokers, the relative risk of lung cancer in long-term smokers is 10-30 times higher and the risk increases with the duration of smoking and the num-ber of smoked cigarettes per day (8). Smoking cessation reduces lung cancer risk and is beneficial at all stages (9). Non-smokers constitute around ten per-cent of all lung cancer diagnoses (5). Exposure to second-hand smoke is also a risk factor with the strongest correlation evident for spousal and workplace-related exposure (10). After cigarette smoking, radon – a naturally occurring radioactive gas - is the main risk factor (11). When radon breaks down, ioniz-ing radiation is formed, which can induce carcinogenesis by a variety of ge-netic changes (12). Other causes of lung cancer include exposure to asbestos, arsenic, indoor and outdoor air pollution as well as other acquired lung dis-eases (9, 13). Hereditary factors play a minor role in the development of lung cancer. Multiple genetic polymorphisms have been identified underlying lung cancer risk, in particular three susceptibility loci, 15q25, 6p21 and 5p1, have been associated with lung cancer in multiple populations (14). The variation in cancer susceptibility is most likely influenced by both genetic and environ-mental components (15).

Diagnosis and Staging Symptoms of lung cancer may include cough, dyspnea, chest pain and hem-optysis. Unfortunately, lung cancer is often asymptomatic in its early stages and the majority of cases are diagnosed after the cancer has metastasized (16). Patients with suspected lung cancer are initially subjected to a chest x-ray. If

15

any abnormalities are found more detailed imaging techniques such as a com-puted tomography (CT) scan and positron emission tomography follow. To make a final diagnosis, a histological evaluation is required. A biopsy is ob-tained by a bronchoscopy for localized tumors and a transthoracic puncture biopsy for peripheral tumors. Further, molecular analysis of the tumor tissue is performed to guide treatment decisions by testing for epidermal growth fac-tor receptor (EGFR) mutations and rearrangements of anaplastic lymphoma kinase (ALK) or ROS proto-oncogene 1, receptor tyrosine kinase (ROS1) (17, 18). The diagnostic procedure is closely linked to the staging procedure. Stag-ing, which describes the size and the spread of the tumor, is an important guide for treatment decisions and is the most important prognostic factor in NSCLC (19). The staging of NSCLC uses the tumor-node-metastasis (TNM) classifi-cation system. First proposed in 1946, the TNM staging system has since un-dergone several revisions to the current 8th edition (20). The T in TNM de-scribes the size and the locoregional invasion of the primary tumor, the N in-dicates whether the tumor is present in nearby lymph nodes and the M refers to presence or absence of intrathoracic or distant metastases (21). Each of the three descriptors is individually assigned with a number that provides more detailed information where a higher number is indicative of more severe dis-ease (Table 1). Once the TNM categories have been scored, an overall stage for the tumor is assigned ranging from I-IV (Table 2). Staging is primarily performed before treatment mainly based on clinical imaging techniques and is therefore regarded as clinical TNM (cTNM). The tumor extent and the pres-ence of lymph node metastasis can only be determined histologically by a pathologist in patients who undergo surgery. This type of TNM is termed pathological TNM (pTNM) (22).

16

Table 1. TNM classification of carcinomas of the lung. Reproduced from (23) Primary tumor (T) Description

T1 Tumor ≤ 3 cm in diameter, no invasion more proximal than the lobar bronchus

T1a: Tumor ≤ 2 cm in diameter T1b: Tumor > 2 cm but ≤ 3 cm in diameter T2 Tumor > 3 cm but ≤ 7 cm in diameter or with the presence of any

following features: - Involvement of the main bronchus, ≥ 2 cm distal to the carina - Invasion of the visceral pleura - Atelectasis or obstructive pneumonitis, extended to the hilar

region but not involving the entire lung T2a: Tumor > 3 cm but ≤ 5 cm in diameter T2b: Tumor > 5 cm but ≤ 7 cm in diameter T3 Tumor > 7 cm or of any size with any of the following: - Direct invasion of the chest wall, diaphragm, phrenic nerve,

mediastinal pleura, parietal pericardium - Tumor in the main bronchus < 2 cm distal to the carina - Atelectasis or obstructive pneumonitis of the entire lung - Separate tumor nodule(s) in the same lobe T4 Tumor of any size with any of the following: - Invasion of the mediastinum, heart, great vessels, trachea,

recurrent laryngeal nerve, esophagus, vertebral body or carina - Separate tumor nodule(s) in a different ipsilateral lobe Lymph nodes (N) N0 No regional lymph node metastasis N1 Metastasis in ipsilateral hilar or peribronchial nodes N2 Metastasis in ipsilateral mediastinal or subcarinal lymph node(s) N3 Metastasis in contralateral mediastinal, contralateral hilar,

ipsilateral or contralateral scalene, or supraclavicular lymph node(s) Distant metastasis (M) M0 No distant metastasis M1 Distant metastasis M1a: Separate tumor nodule(s) in contralateral lobe or tumor with

pleural nodules or malignant pleural effusion M1b: Distant metastasis

17

Table 2. Stage grouping. Reproduced from (23) 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 M0 T2a N1 M0 T2b N0 M0 Stage III IIIA IIIB T1 N2 M0 Any T N3 M0 T2 N2 M0 T4 N2 M0 T3 N1 M0 T3 N2 M0 T4 N0 M0 T4 N1 M0 Stage IV IV Any T Any N M1a Any T Any N M1b

Histology

Based on the microscopic growth pattern, NSCLC is further divided into dif-ferent histological subtypes. The predominant subtypes, comprising more than 95% of cases, are adenocarcinoma, squamous cell carcinoma and large cell carcinoma (2). Less common subtypes include large cell neuroendocrine carcinomas, adenosquamous carcinoma and sarcomatoid carcinomas (24).

Adenocarcinoma is the largest histological subtype of NSCLC, comprising almost half of all lung cancers (8). In younger men (< 50 years old), women of all ages and in former and never smokers, adenocarcinoma is the most com-mon histology (25). Adenocarcinoma is considered to arise from small bron-chi, bronchioles or alveolar epithelial cells, and is typically located in the pe-riphery of the lung (26). Histologically, this subtype is characterized by the formation of glands, the production of mucin and the expression of protein markers such as thyroid transcription factor-1 and/or napsin-A (24). The growth pattern of adenocarcinoma demonstrates a high variability which has led to attempts of further morphological classification, including solid, acinic, lepidic, papillary and micropapillary forms and their combinations (27).

Squamous cell carcinoma is the second largest subtype of NSCLC account-ing for 25-30% of all lung cancers. Squamous cell carcinoma is strongly

18

associated with cigarette smoking and almost all patients with this histology are or have been smokers (28). Historically, this has been the most common subtype of NSCLC, but the incidence has decreased over the past few decades due to changing smoking habits and possibly due to changes in cigarette com-position and filtering (29). Squamous cell carcinoma is considered to originate in the central part of the lung from metaplastic bronchial epithelium (30). His-tologically, well-differentiated squamous cell carcinoma is characterized by keratinization, intracellular bridges and pearl formation (30). For poorly dif-ferentiated tumors, immunohistochemical staining of cytokeratin 5/6, p40 and/or p63 should be used to confirm diagnosis (27). Squamous cell carcino-mas exhibit complex genomic aberrations with a high frequency of mutations, as is expected from tobacco-associated carcinogenesis (28, 31, 32).

Large cell carcinoma presents an undifferentiated histological pattern with no microscopic evidence of squamous, adenocarcinoma or neuroendocrine differentiation (27). Since molecular markers that assist in further characteri-zation of this morphologically undifferentiated subtype have been established, most cases previously annotated as large cell carcinomas are now classified into the adenocarcinoma or squamous cell carcinoma subtypes (33). If a thor-ough morphological assessment cannot be performed, for example when only small biopsies and cytological samples are available, the histological classifi-cation NSCLC-NOS (not otherwise specified) is typically used (34).

Until recently, the histological classification of NSCLC did not have any ma-jor impact on therapy decisions. Today, however, it is evident that certain cy-totoxic drugs are more effective in cases with non-squamous histology. Fur-thermore, various histological subtypes harbor distinct mutations that can be exploited for targeted therapy (35, 36).

Treatment Depending on tumor stage, histological subtype and presence of molecular aberrations, NSCLC patients are eligible for different treatment options.

Conventional Therapies Surgery is the only type of treatment that gives the possibility to cure NSCLC. However, only 25% of NSCLC diagnoses are stage I-II and therefore qualify for surgery with curative intent (37). Patients with stage IIIA who are in good physical condition can also be operated. Despite curative intention, 30-55% will experience recurrence (38). To decrease the risk of recurrence, adjuvant chemotherapy is recommended for stage II-IIIA and is also considered for selected patients with stage IB (39, 40). For those with stage I-IIA where sur-gery is not a suitable option, e.g. due to comorbidities, curative radiation is an

19

effective alternative (41). Unresectable stage III patients are typically treated with a combination of radiotherapy and chemotherapy (42). Stage IV patients usually receive chemotherapy to reduce the symptoms and prolong survival. For a smaller subset of these patients, targeted therapy is an option.

Targeted Therapy The two main targetable genetic aberrations in NSCLC are mutations in EGFR and rearrangements of ALK, both of which predominantly occur in ad-enocarcinomas (43). EGFR is a cell-surface tyrosine kinase receptor that upon activating mutations initiates several pathways involved in cell proliferation, survival, invasion and angiogenesis. Activating mutations in EGFR are pre-sent in approximately 15% of Caucasian patients and 40-50% of Asian pa-tients (43). For patients with activating EGFR mutations there are four small molecule tyrosine kinase inhibitors currently approved by the Food and Drug Administration (FDA) (gefitinib, erlotinib, afatinib and osimertinib) (35, 44). ALK, also a tyrosine kinase receptor, acquires oncogenic properties upon fu-sion with other proteins. The most common fusion partner is echinoderm mi-crotubule associated protein-like 4 (EML4) gene. The fusion protein EML4 - ALK is detected in 2-5% of NSCLC patients predominantly with adenocarci-noma histology, younger age and no or rare smoking habits (45, 46). Four ALK inhibitors are approved by the FDA for ALK positive patients (crizo-tinib, ceritinib, alectinib and brigatinib) (3, 47). Similar to ALK, two other tyrosine kinase receptors, ROS proto-oncogene 1, receptor tyrosine kinase (ROS1) and ret proto-oncogene (RET), can become constitutively activated upon rearrangement with their respective gene partners. Both rearranged ROS1 and RET are present in 1-2 % of NSCLC (48, 49). Due to the high similarity between the kinase domains in tyrosine kinase receptors, the ALK inhibitor crizotinib can be used to treat ROS1 positive tumors. Several other inhibitors are in clinical studies for the treatment of ROS1 and RET positive tumors (50, 51).

Although impressive response rates are observed with tyrosine kinase inhibi-tors, the effect is only temporary. Resistance frequently occurs within one year of starting treatment due to additional mutations in the targeted genes, which affects the efficiency of these drugs (52, 53).

Immunotherapy It is well recognized that tumorigenesis is accompanied by immune cell infil-tration and an anti-tumor immune response is connected to a specific T-cell activation. This has been observed in several types of cancer, including NSCLC (54). Utilizing the immune system as an anti-cancer treatment has historically only had modest effect in solid cancers. However, recent

20

developments targeting T-cell regulatory elements, so called immune check-points, have profoundly altered this view (55).

Immune checkpoints are molecules that regulate T-cell activation and maintain self-tolerance during an immunological response to pathogens. A feature of tumor cells is that they are able to deregulate immune regulatory mechanisms and escape T-cell surveillance. Checkpoint inhibitors interfere with this pathway and are able to restore the natural anti-tumor immunity. One of the checkpoint molecules is the cytotoxic T-lymphocyte antigen 4 (CTLA-4), which is involved in the early stages of T-cell activation. Ipilimumab, which targets CTLA-4, was the first checkpoint inhibitor approved by FDA for treatment of metastatic melanoma in 2011 (56). Another checkpoint mol-ecule is the programmed cell death protein 1 (PD-1) that participates in the later stages of T-cell activation. Monoclonal antibodies, nivolumab and pem-brolizumab, both targeting PD-1, are approved for the treatment of advanced NSCLC. In addition, antibodies targeting the programmed cell death ligand 1 (PD-L1) are approved (atezolizumab) or are in development (durvalumab and avelumab) for NSCLC treatment (3). The treatment with these checkpoint in-hibitors as second line therapy results in a long-term response in approxi-mately 20% of patients (57, 58). On the other hand, most patients are inher-ently resistant or become resistant to checkpoint inhibition. Current chal-lenges include understanding the resistance mechanisms, developing predic-tive biomarkers to select for patients that most likely will benefit from checkpoint inhibition and combining immunotherapy with other treatment modalities for an optimal outcome (55).

Prognostic Factors Prognostication is an important aspect of the diagnostic work-up for NSCLC patients. Both patient-and tumor-related parameters provide prognostic infor-mation that can guide therapy decisions (59). The most important prognostic factor is tumor stage (60) and all three individual parameters of staging; T, N, M, are prognostic in themselves (61). Another important prognostic factor is the individual performance status (PS), which attempts to describe a patient´s overall well-being. The most frequently used PS score in NSCLC is the East-ern Cooperative Oncology Group (ECOG) score (also referred to as Zubrod score or World Health Organization score), which spans from grade 0 = full activity to grade 5 = dead (62). Age can also be used to aid prognostication in NSCLC (63-65). A number of other characteristics, such as gender and smok-ing status have been shown to influence prognosis, although their impact is modest and therefore these factors are usually not considered for treatment decisions. In addition, several histological parameters, such as tumor subtype, vascular invasion, necrosis and grade of differentiation provide prognostic in-formation, but are also generally not employed in routine clinical practice

21

(66). Therefore, only stage, PS and age are consistently used to stratify pa-tients in clinical trials. However, even with similar stage, age and PS, patients can develop very different disease courses. In particular, among those patients with localized tumors, accurate prognostication would help to determine which patients are likely to develop recurrence after radical tumor resection. Since adjuvant chemotherapy is recommended to reduce the risk of recur-rence, patients with an overall good prognostic profile could potentially be saved from the demanding treatment regime (39, 40). On the other hand, pa-tients with very small tumors but with high-risk profiles may benefit from an intensified therapy. Thus, the quest for prognostic markers has been an area of extensive research in NSCLC. Protein expression, evaluated by immuno-histochemistry (IHC), is among the most commonly studied type of biomarker on the molecular level (67). Several single proteins have shown promising results and have been validated in larger cohorts e.g. excision repair cross-complementing rodent repair deficiency, complementation group 1 (ERCC1) and ribonucleoside-diphosphate reductase (RRM1) (59, 68). Despite such nu-merous studies, none of the proteins have demonstrated sufficient power to be introduced in the clinical setting (67). Compared to single markers, the com-bination of several markers may improve the prognostic performance. This approach has mainly been applied in studies based on mRNA expression pro-filing, which have identified gene signatures associated with better or worse clinical outcomes (69, 70). Also, DNA methylation profiles and microRNA signatures have shown to have prognostic impact in NSCLC (71, 72). Cur-rently, however, none of the proposed molecular markers have been employed in routine clinical diagnostic practice.

Early Detection The majority of lung cancer cases are detected at late stages when only treat-ment with life-prolonging and palliative intention can be offered (16). Thus, to improve the dismal prognosis of lung cancer cases, earlier detection is cru-cial. Several large trials for screening for lung cancer have been conducted since the 1970s. However, these early trials did not demonstrate a reduction in lung cancer mortality in comparison to conventional radiograph screening (73-75). In the more recent screening programs, the development of low-dose CT (LDCT) has enabled better sensitivity for small nodules at acceptable lev-els of radiation exposure. In 2011 the National Lung Screening Trial (NLST) presented positive results regarding LDCT. The NLST included more than 50000 participants that were randomized to three annual LDCT scans or a plain chest radiograph. This study reported a significant reduction in lung can-cer mortality in the participants who received LDCT scans (76). Screening of high risk individuals was recommended in the U.S, primarily based on the positive outcome of NLST (77). Implementation of screening with LDCT is

22

still under evaluation in Europe (78). The high number of false positives ob-served, the high cost, as well as the potential harmful side effects have raised concern for widespread implementation of LDCT, and thus other screening methods are needed. Blood-based screening tests represent attractive alterna-tives as they are minimally invasive and relatively cost-efficient. However, no such test has yet reached implementation in the clinical setting.

23

Testis

The testes produce haploid germ cells necessary for sexual reproduction and androgens that are required for male sex characteristics. The majority of tes-ticular tissue is made up by seminiferous tubules where the production of hap-loid germ cells occurs in a stepwise and highly coordinated series of events through the process of spermatogenesis. The Leydig cells that are responsible for the production of androgens, mainly testosterone, are localized between the seminiferous tubules. At the basal region of the seminiferous tubules, sper-matogonial stem cells divide by mitosis into spermatogonia. After several rounds of mitosis, the spermatogonia develop into primary and secondary spermatocytes, which divide via meiosis to form spermatids. The final differ-entiation of spermatids into sperm takes place in the luminal region of the tubules. The various phases of spermatogenesis can be distinguished due to the morphologically different cell types present at each stage (Figure 1) (79). In addition to germ cells, the Sertoli cells are interspersed throughout the tu-bules. They have numerous important functions, including regulation and the development as well as the number and quality of spermatogenic cells (80).

Figure 1. A schematic representation of the various steps of spermatogenesis. Adapted from https://www.proteinatlas.org/humanproteome/testis

24

During spermatogenesis, the germ cells are subjected to a number of specific events such as the reduction of the number of chromosomes, condensation of the nucleus and removal of excess cytoplasm within the sperm. The sperm cells acquire unique morphological features and structures such as the flagel-lum and the acrosome which enable sperm motility and fertilization of the egg. The functions of testis involve complex processes and endow a unique-ness to this organ. This has been reflected in several studies that have analyzed the testis transcriptome. Compared to other organs, a particularly high number of genes are expressed in testis and testis is the tissue with the highest number of tissue-specific genes (81, 82).

The immune system in testis is tightly regulated to protect gametes from any kind of exogenous alteration and to avoid host immune responses as germ cells could be regarded as non-self (83). The downregulated immune system is of importance for the exploitation of cancer-testis antigens (CTAs) which are further presented below.

Cancer-Testis Antigens CTAs are a group of proteins that are expressed in various types of cancers, but not in normal tissues with the exception of testis and placenta (84). The first CTA was identified in 1991 when tumor cells and normal cells from a melanoma patient were used as a “tool” to analyze immune recognition by the patient´s own immune cells (85). A target antigen for a cytotoxic T-lympho-cyte clone was identified and termed melanoma-associated antigen 1 (MAGE1, later MAGE-A1) (85). Identification of MAGE-A1 represented a major advancement in tumor immunology as it was the first immunogenic tumor antigen described to have elicited autologous cytotoxic T-lymphocyte responses in a cancer patient (84). The autologous typing approach used to identify MAGE-A1 was later used to identify other tumor antigens MAGE-A3 and members of the BAGE (B melanoma antigen) and GAGE (G antigen) gene families (86-88). A few years after the identification of MAGE-A1, a serology-based approach was developed to analyze immune responses against tumor antigens referred to as the serological analysis of recombinant tumor cDNA expression libraries (SEREX) (89). SEREX identified additional tumor antigens, e.g. synovial sarcoma, X breakpoint 2 (SSX2) and New York esoph-ageal squamous cell carcinoma-1 (NY-ESO-1), which have elicited high-titer immunoglobulin G (IgG) humoral responses in cancer patients (90, 91). Fur-ther studies of the tumor antigens identified by these two methods demon-strated their expression in various cancers of different histological origin, in testis and placenta, but not in other normal tissues and therefore they were named cancer-testis antigens (91). Although several CTAs have shown to

25

elicit T-cell and/or B-cell responses in cancer patients (e.g. MAGE, BAGE, NY-ESO-1, SSX2, SCP1 and SLCO6A1) (85-87, 90-93), more recent expres-sion studies on cancer and a range of normal tissues have expanded the group of CTAs based on their gene expression patterns and not their antigenic ability (94, 95). Thus, today more than 200 genes are now categorized as CTAs in the Cancer Testis database (CTdatabase) (96). In testis, CTAs are expressed in both early and late stages of the spermatogenesis. Their biological role in both germline tissue and tumors is in general poorly understood. Nonetheless, their ability to induce immunogenic responses along with an almost cancer-specific expression pattern and the downregulated immune component in tes-tis has rendered CTAs attractive targets for immunotherapy, mainly in the form of therapeutic cancer vaccines (97). The most evaluated CTAs for vac-cination strategies are the members of the MAGE-A family and NY-ESO-1 (97). The largest vaccination study performed in NSCLC, MAGRIT, evalu-ated the efficacy and safety of a MAGE-A3 vaccine as adjuvant treatment in more than 2000 completely resected NSCLC patients (98). However, the trial was ended when the vaccine did not significantly improve the overall survival compared to placebo. Numerous clinical trials are currently ongoing in pa-tients with NSCLC and other cancer types with the aim of evaluating NY-ESO-1 and other members of the MAGE families as immunotherapeutic tar-gets.

26

Present Investigations

Aim The overall aim of this thesis was to develop and evaluate the prognostic per-formance of a protein panel and to identify tumor markers in NSCLC as po-tential targets for immunotherapy or as biomarkers for early detection. More specifically the aims were as follows:

I Develop and optimize an IHC-based prognostic model for NSCLC and to compare its performance to the clinicopathological parameters that are currently used in the clinical setting to predict patient out-come.

II Describe the human testis-specific proteome by using RNA-sequenc-ing (RNA-seq) data from normal human testicular tissue and other normal human tissue types.

III Profile CTAs in NSCLC by using RNA-seq data from NSCLC pa-tients and normal tissue sites.

IV Analyze reactivity of CTAs by measuring the abundance of autoanti-bodies in plasma from patients with NSCLC and patients with non-malignant lung diseases.

V Analyze the plasma levels of cancer related proteins in NSCLC and non-malignant lung diseases and build a discriminatory model for early detection of lung adenocarcinoma.

27

Methodological Background Immunohistochemistry IHC is a routinely used technique in clinical pathology and life science re-search that enables visualization and distribution of proteins in tissue. Histo-pathological tissue samples are usually fixed in formalin to preserve tissue morphology. After fixation, the tissue is dehydrated in graded alcohols and embedded in paraffin blocks resulting in formalin fixed paraffin embedded (FFPE) tissue. A section of the FFPE tissue is placed on a glass slide and can be used for staining. Fixation with formalin generates methylene bridges that crosslink proteins in tissue samples and modifies the structure of many protein epitopes. To break the cross-links and unmask hidden epitopes, antigen-re-trieval is performed prior to immunohistochemical staining (99). Thereafter, an antibody directed towards the protein of interest can be added. To visualize the antibody-antigen interaction, a detection method that produces a visible signal is necessary. In the direct detection method only one antibody is used, whereas in the more common indirect detection method, a primary and a sec-ondary antibody are used. The secondary antibody (or the primary antibody in the direct method) is coupled to a reporter molecule that can either be an enzyme or a fluorophore. In the enzymatic reaction, which is the most com-mon, the secondary antibody is linked to e.g. horseradish peroxidase (HRP) that in presence of the chromogen 3,3'-diaminobenzidine (DAB) and hydro-gen peroxide produces a brown precipitate at the site of the enzymatic activity, which is subsequently observable by light-microscopy (Figure 2) (100).

Figure 2. The basic principle of immunohistochemistry using the indirect detection method. A primary antibody targeting the antigen is recognized by a secondary anti-body that is linked to horseradish peroxidase (HRP). A brown precipitate is pro-duced in the presence of 3,3´-diaminobenzidine (DAB) and hydrogen peroxide.

28

In Paper I automated IHC was performed on FFPE tissue that was assembled in tissue microarrays (TMAs) for visualization and distribution of proteins. The staining of tumor cells was manually evaluated on scanned slides. An IHC score was obtained by multiplying the staining intensity with the fraction of stained tumor cells and this score was used for further analyses. The use of IHC in Paper II and Paper III was to complement the transcriptomics data and to provide compartment specific localization of the protein. The Human Protein Atlas (HPA) database was used to search for availability of reliable antibodies. Antibodies were chosen if the staining pattern was in accordance with the expected subcellular and histological localization in the scientific lit-erature.

Tissue Microarrays TMAs are arrays of small tissue samples, usually between 0.6-1 mm in diam-eter, which enable numerous tissue samples from different donors to be con-tained in a single tissue block, and after sectioning, on a single glass slide. For construction, the area of interest of the donor sample is selected by visual de-lineation on a hematoxylin and eosin (HE) stained slide. From this area a cy-lindrical tissue core is punched out from the donor block and placed in a re-cipient paraffin block at defined positions. This procedure is repeated from different donor blocks until all wanted tissues are placed in a final TMA that usually comprises up to 100 samples per block (Figure 3). Since the cylindri-cal core only represents a small part of the original tissue, usually two cores are taken from the same donor block to increase coverage of tissue heteroge-neity (101-103). Sections of 4-5 μm are cut from the TMA block to generate TMA slides for analysis (104). With the TMA technique, a large number of tissues can simultaneously be studied at different molecular levels. DNA can be analyzed with fluorescence in situ hybridization and RNA with mRNA in situ hybridization, but the most common application is the detection of protein antigens using IHC. In research, the TMA technique is widely used for bi-omarker studies (104).

A potential drawback with the TMA technique is that only small parts of the whole tumor are evaluated. The TMAs used in Paper I include two cores from the same patient in order to increase the coverage of tissue heterogeneity. Several studies have shown that two or three cores of the original tissue gives a reliable representation of the whole tumor (101, 102).

29

Figure 3. Basic steps in the production of a tissue microarray (TMA). Tissue cores from formalin-fixed paraffin embedded (FFPE) tissue are assembled in a recipient block. The recipient block is sectioned and the resulting TMA can be immunohisto-chemically stained for visualization of a protein in tissues from multiple donors.

RNA-Sequencing RNA-seq is a deep-sequencing technique that has become the method of choice for transcriptome analysis. RNA-seq data can be used to determine gene expression levels of each transcript in a sample and compared to previ-ous gene expression methods with microarrays, RNA-seq provides higher res-olution, is not limited to detection of known transcripts, and can be used as an exploratory tool for identification of novel transcripts (105). Transcripts can be analyzed at a single base-pair resolution, thus RNA-seq can be used to study the expression of disease-associated single-nucleotide polymorphisms, mutations, fusion genes and isoforms (106-108). Depending on sequencing purposes and the type of RNA analyzed, different techniques are utilized to select the RNA class of interest. For analysis of gene expression, alternative splicing and genetic variation detection, where the pro-tein-coding RNA is of interest, an enrichment for mRNA using poly(A)+RNA selection is performed prior to sequencing library preparation. As most RNA-seq platforms only are able to sequence relatively short sequences the RNA is fragmented into shorter pieces. The vast majority of RNA-seq experiments are carried out on instruments that sequence DNA molecules, therefore the RNA is typically converted to cDNA library in the library preparation. Adap-tors are added to the cDNA fragments and thereafter amplified by polymerase chain reaction (PCR). In the sequencing reaction, all cDNA fragments are simultaneously sequenced generating so called “reads”, comprising nucleo-tide sequences usually of 30 to 400 base pairs in length. The resulting reads are either aligned to a reference genome or if identification of novel transcripts is desired, the transcripts can be reconstructed from these reads, which is re-ferred to as a de novo assembly (105) (Figure 4).

30

In Paper II and Paper III, RNA-seq was used for gene expression analysis. For RNA-seq data of normal tissues, used in both Paper II and Paper III, RNA was extracted from fresh-frozen tissue obtained from histologically nor-mal surgical specimens. One cut of the fresh-frozen tissue was stained with HE to ensure accurate tissue morphology by a pathologist. All samples had high-quality RNA of 7.5 RIN (RNA integrity number). For RNA-seq anal-ysis of NSCLC samples used in Paper III, one cut of the fresh-frozen tissue was stained with HE and the tumor cell content was evaluated by a pathologist. Only samples with at least 10% of tumor were included. Five sec-tions from each patient were used for RNA extraction and the majority of cases (188/199) were of high-quality RNA with RIN values 7.5. The sam-ples that had less RIN values than 7.5 did not show any deviance in the sub-sequent analyses and were therefore kept. Samples were prepared for sequenc-ing using poly(A) selection and sequenced with a read length of 2 x 100 bases. The raw reads from 20050 transcripts obtained from the sequencing were pro-cessed and mapped to the human genome. Quantification for all human genes were obtained by calculating fragments per kilobase of exon model per mil-lion mapped reads (FPKM).

Figure 4. Simplified overview of the various steps in RNA-seq experiments per-formed in this thesis to determine gene expression. In brief, RNA is extracted and an enrichment for mRNA is performed using poly(A) selection. The RNA is converted to cDNA and a library is prepared that is used for sequencing. The resulting sequences are mapped to a reference genome and can be quantified to determine gene expression levels.

31

Suspension Bead Array The suspension bead array technique allows multiplex analysis of up to 500 analytes in hundreds of samples in parallel. This method uses magnetic beads (denoted microspheres) with a surface presenting functional groups that ena-ble different capture molecules e.g. antibodies or proteins to be coupled cova-lently. Depending on the molecule coupled onto the beads, various types of target analytes e.g. proteins or (auto)antibodies can be detected in body fluids. Each target molecule is coupled to unique bead sets that are color-coded with a mixture of different fluorescent intensities of three dyes. Each set of beads, usually hundred-thousand to a million being prepared per coupling batch, is modified by with a particular capture molecule which is then represented by the beads’ unique color-codes (109). Since each bead set can be distinguished, combining the different bead sets (ranging from 1-500) into a so called “sus-pension bead array” allows a multiplexed analysis in up to 384 samples per experiment. After incubation with sample, washing buffer and a reporter mol-ecule, the beads are analyzed in a flow cytometer-like instrument developed by Luminex (110). In the instrument, each individual bead is subjected to two lasers, one to determine the bead identity based on its specific color-code and the second to define the magnitude of the reporter molecule (Figure 5). The latter then collects the reported signals from 50 or more beads per bead set to then given rise to the relative quantification of the bound target analyte in the sample, commonly referred to median fluorescence intensity (MFI) (109).

Figure 5. Simplified overview of the suspension bead array technique using detection of autoantibodies as an example. Antigens (red) are coupled to unique bead sets (A). If present, autoantibodies in the sample will bind to the bead-antigen complex (B). A reporter molecule (green), coupled to anti-human IgG, is added and will bind to pre-sent autoantibodies. A flow-cytometer instrument identifies the antigen and measures the signal from the reporter molecule (C).

In Paper IV, plasma samples from patients with NSCLC and benign lung diagnoses were used for the detection of potential autoantibodies against CTAs. The detection of autoantibodies was performed using the suspension bead array technique. Protein fragments produced within the HPA project

1 2 500

A. B. C.

32

were used as antigens and were covalently coupled to magnetic beads (111). The protein fragments were selected from unique or low homology regions to other proteins and produced in Escherichia coli with an affinity tag consisting of six histidines and albumin binding domain from streptococcal protein G (111). The length of the protein fragments varied between 20 to 148 amino acids. Binding of the autoantibodies was defined by the intensity of the re-porter molecule measured on each bead by the Luminex instrument as MFI values. The MFI values were binned and the binned values were used in fur-ther analysis.

Proximity Extension Assay The proximity extension assay (PEA) uses pairs of antibodies that are directed towards different epitopes on the same target, and each antibody is conjugated to a unique DNA oligonucleotide. The DNA oligonucleotides on each pair of antibodies have complementary sequences. When the antibodies bind their target, the oligonucleotides are brought in proximity and are allowed to hy-bridize. After hybridization, the oligonucleotides are extended by DNA poly-merase and the newly formed DNA molecule acts as a surrogate marker for the specific target protein (Figure 6). This sequence can be amplified and quantified by quantitative real-time PCR, where the number of PCR templates formed is proportional to the initial concentration of the target in the sample (112, 113). With this assay it is possible to simultaneously measure 92 pro-teins and four controls in 90 samples such as serum, plasma and other types of biological fluids, using a microfluidic real-time PCR instrument (112).

Figure 6. Overview of the proximity extension assay. Pairs of antibodies are directed towards the same protein (A). When in proximity, the oligonucleotides linked to each antibody hybridize (B) and can be extended (C) resulting in newly formed DNA mol-ecules that can be measured and quantified (D).

In Paper V, PEA was used to assess plasma levels of cancer-related proteins that are included in an oncology panel which is commercially available from Olink Bioscience, Uppsala, Sweden. PEA is based by the use of two antibod-ies directed towards one target. Only when both antibodies from each pair have bound its target, the oligonucleotides that are linked to the antibodies are allowed to hybridize and get extended by a polymerase reaction. Since both

A. B. C. D.

33

antibodies are required to recognize the same target in order for a polymerase reaction to start, this assay offers a high level of specificity but also a high level of sensitivity from the use of a DNA polymerase which degrades non-proximal DNA strands resulting in a reduced background noise (114). The extended DNA molecule is amplified and quantified by real-time PCR and obtained Cq-values are normalized against extension controls, inter plate con-trol and a correction error. For data analyses, normalized protein expression values on a log2 scale were used where a high value corresponds to a high protein concentration.

Statistics Overall survival was calculated from the date of diagnosis to the date of death. Survival was analyzed by univariate and multivariate Cox models. For uni-variate analyses, IHC scores were used in Paper I and mRNA values in Paper III. Multivariate analyses were performed with the inclusion of clinicopatho-logical variables, stage, PS and age. The Cox models in Paper I were visual-ized by Kaplan-Meier plots, based on dichotomized IHC and clinicopatholog-ical variables that were dichotomized as follows: stage I vs stage II-IV, PS 0 vs PS I-IV and age at diagnosis ≤ 70 vs > 70 years. The performance of the prognostic models constructed in Paper I was as-sessed by the concordance-index (C-index). C-index is a rank-based method that compares the predicted survival times with the observed survival times of a model to estimate the survival probability. A model with a C-index of 1 implies perfect prediction whereas a model with a C-index of 0.5 or less has no predictive ability. The Fisher´s exact test was used to assess the correlation between methylation status and CTAs in Paper III and an association between reactivity and clin-ical characteristics in Paper IV. The comparative analysis of plasma protein levels between different patient groups was performed using the Wilcoxon test. Correction for multiple testing was performed on all tests using the Ben-jamini -Hochberg method or with Bonferroni correction. P values less than 0.05 were considered significant.

For the development of the classifier in Paper V, the TreeBagger function was used which is based on a random forest algorithm. A classifier was de-veloped using 80% of the data set and the algorithm based on thousands of decision trees, evaluated on the remaining 20%. The performance of the clas-sifier was calculated by obtaining values for sensitivity and specificity. To evaluate and visualize the performance, primarily sensitivity and specificity of the classifier, a receiver operating characteristic curve and confusion matrix were developed. All analyses were performed in Matlab or with the R statistical software.

34

Patient Material The Uppsala I cohort includes patients with primary NSCLC that were re-ported to the Uppsala-Örebro Regional Lung Cancer Registry between 1995-2005 and operated at the Uppsala University Hospital during these years. His-tological confirmation was done by a pathologist. Patients who received neo-adjuvant treatment were excluded. FFPE tissues were available for 354 pa-tients and were included in a TMA with duplicate tissue cores from each pa-tient sample (Paper I) (115, 116).

The Uppsala II cohort includes patients with primary NSCLC that were reported to the Uppsala-Örebro Regional Lung Cancer Registry between 2006-2010 and operated at the Uppsala University during these years. Histo-logical confirmation was done by a pathologist. Patients who received neoad-juvant treatment were excluded. FFPE tissues were available for 357 patients and were included in a TMA with duplicate tissue cores from each patient sample (Paper I) (117, 118). Fresh-frozen tissue was available for 199 of these 357 patients and extracted RNA was used for RNA-seq analysis (Paper III) (119). EDTA-plasma samples were available for 133 patients (Paper IV).

The Uppsala cohorts represented the main study material in this thesis. The material from the two cohorts is illustrated in Figure 7 and the patient´s char-acteristics are shown in Table 3.

Figure 7. Overview of the main study material – the Uppsala cohorts I and II. The number of cases (n) in this figure represents the cases that were analyzed in Paper I, Paper III or Paper IV.

!""#$%$ && cohort'(()*+ '(,(

-./*&&

0.90 1.00

1 0 100 0 0 0 100

100

00

100

Coordinate 1

Coo

rdin

ate

!"#$%&'(

!"#!$%&'" / NOS)

0.9

0

!""#$%a & cohort,001*+ '((1

-./*&

&23*$4$%5#6#4*7*826

&23*$4$%5#6#4*7*845

9:/-#;< 4*7*,00

=%$#>$*$4$%5#6# 4*7*,33

35

Table 3. Clinical characteristics of the patients included in the Uppsala cohorts I and II.

Uppsala I TMA I n (%)

Uppsala II TMA II n (%)

Uppsala II RNA-seq n (%)

Uppsala II Plasma n (%)

All patients 326 (100.0) 345 (100.0) 199 (100.0) 133 (100.0) Gender Female 150 (46.0) 176 (51.0) 103 (51.8) 70 (52.6) Male 176 (54.0) 169 (49.0) 96 (48.2) 63 (47.4) Age ≤ 70 years 226 (69.3) 229 (66.4) 120 (60.3) 82 (61.7) > 70 years 100 (30.7) 116 (33.6) 79 (39.7) 51 (38.3) Smoking history Ever 296 (90.8) 305 (88.4) 180 (90.5) 121 (91.0) Never 28 (8.6) 40 (11.6) 19 (9.5) 12 (9.0) Missing data 2 (0.6) 0 (0.0) 0 (0.0) 0 (0.0) Stage IA 80 (24.5) 141 (40.9) 70 (35.2) 48 (36.1) IB 137 (42.0) 74 (21.4) 45 (22.6) 27 (20.3) IIA 10 (3.1) 40 (11.6) 25 (12.6) 20 (15.0) IIB 43 (13.2) 34 (9.9) 23 (11.5) 15 (11.3) IIIA 33 (10.1) 46 (13.3) 33 (16.6) 22 (16.5) IIIB 15 (4.6) 0 (0.0) 0 (0.0) 0 (0.0) IV 8 (2.5) 10 (2.9) 3 (1.5) 1 (0.7) Histology Adenocarcinoma 177 (54.3) 205 (59.4) 108 (54.3) 74 (55.6) Squamous cell carcinoma 113 (34.7) 102 (29.6) 67 (33.7) 46 (34.6) Large cell ca. / NOS 36 (11.0) 38 (11.0) 24 (12.0) 13 (10.0) Performance status 0 172 (52.8) 207 (60.0) 120 (60.3) 84 (63.2) 1 121 (37.1) 133 (38.6) 77 (38.7) 48 (36.0) 2 27 (8.3) 5 (1.4) 2 (1.0) 1 (0.7) 3 5 (1.5) 0 (0.0) 0 (0.0) 0 (0.0) 4 1 (0.3) 0 (0.0) 0 (0.0) 0 (0.0) The original NSCLC Uppsala I and II cohorts included 354 and 357 patients, respectively. The number of patients in this table represents the cases that were analyzed in Paper I, Paper III or Paper IV. NOS = Not otherwise specified.

36

In addition to these two cohorts, EDTA-plasma samples were obtained from 34 patients with NSCLC that were treated with radiotherapy at the Uppsala University Hospital between 1983-1996 and were used in Paper IV. In Paper V, plasma samples from 50 of the patients included in Uppsala cohort II were analyzed and additionally 94 plasma samples obtained from patients with NSCLC, diagnosed during the same time period. Samples used as benign con-trols in Paper IV (n = 57) and Paper V (n = 68) were obtained from patients that were admitted to Uppsala University Hospital during 2006-2013 and di-agnosed with various benign lung diseases. Plasma samples from 83 patients diagnosed with colorectal cancer that had metastasized to lung and 48 patients diagnosed with typical carcinoid were also used in Paper V.

The histological diagnosis was reviewed by a trained pathologist. The pa-tients` clinical characteristics were obtained from the records of the Uppsala - Örebro Regional Lung Cancer Registry and by review of the patient’s charts.

Paper I - Reaching the Limits with Prognostication in NSCLC NSCLC patients bear a high risk of cancer recurrence even after radical sur-gical resection. Prognostic factors could provide useful information for this patient group and may help distinguish patients with a high risk of recurrence from those with low risk and thus refine therapy decision making. Currently there are no sufficiently good prognostic factors to do this and prognostic studies in NSCLC have been an area of intense research over the years. In Paper I we developed an IHC-based prognostic panel and evaluated its prognostic performance in comparison to the clinicopathological parameters (stage, PS and age) that currently are used to predict outcome. Five proteins (MKI67, EZH2, SLC2A1, CADM1 and TTF1) were included in the panel and chosen based on their prognostic association in the published literature, prog-nostic association in gene-expression data sets, the availability of reliable an-tibodies and the representation of diverse biological processes. The protein expression in tumor cells was analysed by IHC on a training cohort (n = 326) where all five proteins demonstrated independent prognostic value. Based on the staining of the tumor cells, an IHC score was obtained ranging from 0 – 24 for each tumor and each protein. All the different scores were evaluated for each protein and the score that best predicted survival was used together with other proteins` optimized scores and in combination with the clinical pa-rameters to develop an optimized prognostic model. When the optimized prognostic model was applied on an independent NSCLC cohort (n = 345), it did not predict survival better than the combination of clinicopathological pa-rameters. These results put into question the potential of IHC-based assess-ment of protein biomarkers for prognostication in NSCLC.

37

Paper II - The Human Testis Specific Proteome The main function of the testes is to produce haploid germ cells necessary for sexual reproduction. Although the testis has been extensively studied, a large fraction of the genes expressed in testis are not characterized. In Paper II we performed a comprehensive analysis of the human testis-specific proteome by using RNA-seq data of normal human testicular tissue from seven individuals and 26 other normal human tissue types obtained from 88 individuals. All 20050 putative protein coding genes were classified into categories based on their expression in testis. Of all putative coding genes, 77% were expressed in testis, which was more than in any of the other analyzed tissues. More than 1000 genes showed a testis-enriched expression pattern which also makes tes-tis the tissue with most tissue-specific genes in the human body. The majority of the testis-enriched genes identified here were previously poorly character-ized and were further profiled by IHC with respect to protein localization in the specific cell types (Figure 8). This study provides a starting point to in-crease our molecular understanding of biology and pathology of the testicular tissue.

Figure 8. Immunohistochemical staining of selected testis-specific genes. Protein lo-calization is visualized in spermatogonia (a), spermatocytes (b), spermatids (c) and sperm (d).

38

Paper III - Profiling CTAs in NSCLC CTAs are proteins specifically expressed in normal testis or placenta tissues but are also expressed in a wide range of cancers, including lung cancer. Sev-eral CTAs have been shown to encode immunogenic proteins that can induce antitumor immune responses in cancer patients. Because of their ability to in-duce an immune response and their almost cancer specific expression pattern, CTAs are considered to be attractive targets for immunotherapy. In Paper III we performed a comprehensive profiling of CTAs in NSCLC based on RNA-seq data of 142 samples from 32 different normal human organs and 199 NSCLC samples. The expression of the 232 currently annotated CTAs in the CTdatabase was evaluated. Ninety-six of these CTAs were confirmed to be CTAs in NSCLC while the remaining 136 previously reported CTAs were either not expressed in NSCLC or expressed at substantial levels in somatic tissues. By applying stringent criteria to our RNA-seq dataset, we additionally identified 55 CTAs that were previously not annotated in the CTdatabase, thus representing potential new CTAs. The transcriptomic data were validated on protein level by IHC for several CTAs where previously no data or only mRNA data was available. Furthermore, we analysed the gene expression of CTAs in relation to DNA methylation and confirmed the previously reported regulatory mechanism by methylation. However, we could not confirm the previously reported prognostic impact of CTAs in NSCLC, neither in our RNA-seq cohort nor in an independent meta-analysis consisting of more than 1000 NSCLC cases. This study presents a detailed catalogue of CTAs that can guide further studies on CTAs in NSCLC.

Paper IV - Detection of Autoantibodies against CTAs in NSCLC Several CTAs have been shown to encode immunogenic proteins that are able to induce anti-tumor responses in cancer patients. A few of these CTAs have been evaluated as targets in vaccination strategies. One approach to identify other immunogenic CTAs that could be used for immunotherapy, is through the detection of autoantibodies towards these proteins which can result from B - and T - cell responses in the cancer patient. In Paper IV we used suspen-sion-bead arrays to measure the abundance of circulating autoantibodies in plasma from 133 NSCLC patients and 57 patients with benign lung diseases. The analysis included 120 protein fragments representing 112 CTAs. Reac-tivity against 69 antigens, representing 81 CTAs, was demonstrated in at least one of the analysed samples. Of these, 29 antigens demonstrated exclusive reactivity in NSCLC. Reactivity against CT47A genes, PAGE3, VCX, MAGEB1, LIN28B and C12orf54 were only found in NSCLC patients at a frequency of 1-4% and the presence of autoantibodies towards these six

39

antigens was confirmed in an independent group of 34 NSCLC patients. This study identified autoantibodies against CTAs in plasma of NSCLC patients. The reactivity suggests an immunogenic potential that might be utilized in immunotherapeutic strategies.

Paper V - Plasma Protein Profiling to Discriminate Patients with Lung Adenocarcinoma Early detection of NSCLC is of high clinical relevance to provide curative treatment options. Many efforts have been made to establish non-invasive ap-proaches for screening purposes. In Paper V we used PEA to assess protein levels of 92 cancer-related proteins in 144 lung adenocarcinomas and 68 pa-tients with non-malignant lung disease and to build a discriminatory model for the early detection of lung adenocarcinoma. Several proteins showed sig-nificantly different levels between the two groups; CEACAM5, CXL17, VEGFR2 and ERBB3. Based on the difference in plasma protein levels be-tween the two groups a multi-parameter classifier was developed built on a decision tree algorithm. The classifier discriminated between lung adenocar-cinoma and benign controls with a high sensitivity (93%) and acceptable specificity (64%). The results indicate that this multiplex-immunoassay has the potential as a screening assay for early lung cancer detection.

40

Concluding Remarks The numerous attempts to improve prognostication in NSCLC have resulted in more than 500 reported proteins with prognostic impact (67). Nevertheless, not a single protein marker has been implemented in the clinical practice and overall survival is currently still best predicted by use of the clinicopatholog-ical parameters stage, PS and age (19). In Paper I we developed and evaluated an optimized protein panel based on the prognostic information from a multi-tude of gene expression data sets and several IHC-based studies. The panel, consisting of proteins representing different tumorigenic pathways, was opti-mized on a training cohort and was finally tested on a validation cohort where its performance was compared to the traditionally used clinicopathological parameters. Despite optimization of the protein panel, the clinicopathological parameters remained superior for predicting survival outcome in NSCLC cases. When the proteins were added to the clinicopathological parameters, they only provided limited value, which is of minor relevance in clinical prac-tice.

Although study 1 questions the use of IHC-based prognostic markers in NSCLC, we believe that prognostic studies have a value, e.g. to indicate if a protein has a clinical relevance. However, if a protein is suggested as a prog-nostic biomarker a head-to-head comparison to clinicopathological parame-ters should always be performed. Several guidelines for reporting of prognos-tic biomarker studies have been published (120, 121). However, most aspects of these guidelines address the completeness of formal reporting, rather than clinical significance. More specific recommendations in different cancer types may be helpful to define reference factors (e.g. stage, PS, age) that the poten-tial biomarkers are tested against. We believe that our study design can act as a prototypic model for how prognostic biomarkers should be validated and has the potential to be developed and implemented in guidelines for the eval-uation of prognostic biomarkers in NSCLC.

The testis is an organ with highly specialized functions. Previous gene expres-sion studies on testis have demonstrated an unusual and diverse set of genes compared to other organs (81, 82). In Paper II we used a transcriptomic and antibody-based approach to describe all protein-coding genes in the testis. Pu-tative protein coding genes were categorized based on their expression in tes-tis and other organs and more than 1000 genes demonstrated testis-specific expression. The majority of the testis-specific genes were poorly character-ized regarding their function and cell type specificity. The transcriptomic data was complemented with IHC and we could demonstrate protein expression in specific cell types and compartments for several of the previously uncharac-terized proteins. This analysis demonstrates the power of combining RNA-seq data with IHC to more deeply characterize the testis specific proteome. The spatial information of protein expression will help to understand the basic

41

biology and pathology of the testis. The complex transcriptome of the testis is most likely affected by meiosis-specific events. Since testis was the only tis-sue harboring cell types that undergo meiosis in this study, it would be of interest to also include oocytes in a similar analysis to discriminate spermato-genesis specific genes from genes generally involved in meiotic functions of gametocytes.

A large number of the testis-specific genes belong to the group termed cancer-testis antigens (CTAs) that are restricted to testis and placenta among normal tissues but frequently expressed in cancers of various histological types. Sev-eral CTAs have been shown to encode immunogenic proteins and are there-fore considered as attractive targets for immunotherapy (97). In Paper III we performed a comprehensive characterization of the CTAs in NSCLC using RNA-seq and IHC. Based on our extensive transcriptomic strategy, our study provided a refinement of the currently reported CTAs. Several of the CTAs included in the CTdatabase demonstrated expression in somatic tissues and we believe that their classification as CTAs should be reconsidered. By using our extensive transcriptomic data set of normal and NSCLC tissue, we iden-tified genes with a CTA expression pattern in NSCLC that were previously not reported as CTAs. For several of these genes, where previously only in-formation on mRNA data was available, the protein localization was visual-ized using IHC. Most importantly, we confirmed the expression of a set of CTAs in NSCLC that provides a starting point for more focused studies to evaluate CTAs as immunotherapeutic targets in NSCLC. The combination with novel T-cell activating agents give hope for vaccination strategies using single CTAs or combinations of CTAs as antigens. Since not all CTAs will demonstrate similar immunogenic competency, studies that help to identify the most promising antigens are clearly of clinical value. An in vivo marker of immunogenicity can be a measurement of T-cell reactivity or the occur-rence of autoantibodies against CTAs, as analyzed in paper IV. The presence of autoantibodies against CTAs was analyzed in Paper IV. By using a multiplexed approach, we analyzed the reactivity of 120 CTAs through the detection of autoantibodies in plasma from NSCLC patients and patients with benign lung diseases. Six of the CTAs (CT47A, PAGE3, VCX, MAGEB1, LIN28B and C12orf54) were exclusively reactive in NSCLC sam-ples and their reactivity was confirmed in a second NSCLC cohort. Autoanti-bodies towards these six CTAs were detected in the range of 1 – 4% of NSCLC patients. Despite the low frequency of reactivity these results confirm that certain CTAs provoke an immune response in NSCLC patients and sug-gest that they should be further evaluated as immunotherapeutic targets. The antigens selected for this study and the suspension-bead array technique are mainly suitable for screening purposes. These results therefore need to be con-firmed by another approach that has higher sensitivity and specificity. Besides

42

indicating possible immune reactions towards tumor antigens presented on cancer cells, autoantibodies hold the potential as diagnostic biomarkers as they are stable over time and potentially detected before clinical manifesta-tions. Therefore, the detected autoantibodies towards certain CTAs identified here could also be evaluated in a discriminative model for use in early cancer detection. In Paper V we evaluated a multiplex PEA immunoassay as a screening method for early detection of cancer. The levels of 92 proteins involved in various tumorigenic processes were measured in samples obtained from lung adenocarcinoma and benign lung diseases. Based on the difference in the pro-tein levels between these two patient groups, a classifier was developed that could discriminate between lung adenocarcinoma and benign controls with a high sensitivity (93%) and acceptable specificity (64%). The performance of the assay suggests a potential for its use as a screening assay for lung cancer. However, the discriminatory model needs to be validated in plasma samples of an independent cohort and should include even more controls. Also, the analysis should not be restricted to adenocarcinoma only. All types of lung cancer should be identifiable in the screening setting. There is room for improvement. The proteins included in the assay were not specifically developed for lung cancer, thus the assay could be modified by including proteins with higher discriminatory power. Other immunoassays of-ten use cytokeratins to identify epithelial tumors. Other markers could be CTAs or autoantibodies against those proteins that are potentially more spe-cific than general epithelial markers. Thus, we believe that the encouraging performance of the PEA technique provides a multitude of possibilities to de-tect proteins for clinical diagnostics.

43

Acknowledgements

This work was carried out at the Department of Immunology, Genetics and Pathology, Rudbeck laboratory, Uppsala University. Initially I would like to thank all lung cancer patients who donated tissue for medical research and everyone who contributed to sample collection and preparation, making this work possible.

My supervisor Patrick Micke, thank you for being there for me 110%. For all the time you have invested, for always taking your duties as a supervisor seriously, providing help and guidance whenever needed and for your fantas-tic personality. My co-supervisor Fredrik Pontén, thank you for giving me the opportunity to work with interesting projects that inspired me to do a PhD and for your support to actually make that happen. Thank you for providing an exciting atmosphere wherever you go, for looking at science and life with the brightest eyes and for being a source of inspiration. HPA – this grandiose group that has made Sweden a major contributor to the field of proteomics. I´m happy and thankful for having the opportunity to be part of this fantastic project. All the wonderful colleagues in HPA-Uppsala of which many have turned into friends, thank you for all the fun at and outside work during these years. To all the co-authors for your help and contributions. Especially, Elis-abeth Ståhle for your support and the collaboration that extended our anal-yses to also include plasma samples. Jochen, Tea and Cicci for introducing me to the world of plasma, beads and fancy plots. Cicci and Tea thank you for all your help in the lab and with data analysis, your kindness, for always making me feel welcome and for teaching me so much. Karolina, thank you for your on-spot comments, improvements of our manuscripts and for sharing your knowledge and experience. Marianna, thank you for all the fun time we spent in the office and at the gym. It was a match made in heaven. Julia, thank you for being a fantastic colleague and an even better friend. Johanna, you have been like a sister to me. Thank you for everything you are and for making me want to come to work all those rainy mornings. Max, thank you for every time you helped out in R, for always being kind and positive and having a heart of gold. Linnéa, thank you for our joint struggle with our favorite soft-ware, tips of the best sparkling wine and when we were abroad, for knowing where to go when no one else had a clue. Victor, thank you for bringing new skills to our group, all the help with our project and fun outside Rudbeck.

44

Hedvig, thank you for your sparking energy and for cheering up especially when things were going slow. Stina, Marcus and Helena, thank you for fun conversations and for being good colleagues. The Lund-pathology groups for inspiring meetings, scientific discussions and bath-tub occasions. Gäz, Ulf, Sanaz, Lotte, Åsa, Agnes, Linnea, Magnus, Monica and Monica for providing a nice working atmosphere and laughter in the corridor. Mi, my first supervisor that turned into one of my best friends. Thank you for the support, energy and love that you spread wherever you are. Jennifer, thank you for your support and kindness. Johan Botling, thank you for sharing your scien-tific knowledge whenever needed and for an epic karaoke experience in To-kyo. The members of the PhD council Svea, Verónica, Bene, Marcus, Anja, Loora, Sanna, Leonor, Emma and Viktor, thank you for your efforts organ-izing after-works and conferences. Christina, Helene and Tuulikki, thank you for helping out whenever needed. Ni är guld värda. Evelina, thank you for your energy boosts, plans that quickly turn into reality and for being my Chandler when I am Joey. Sara, thank you for being there no matter what and for putting a golden lining on everyday life. Lesley, thank you for all the in-telligent and funny things you say, for your friendship and for our adventures across the Atlantic. The Nordlund/Bergstrand family Cleo, Martin and Jessica, thank you for your friendship, dinners, trips near and far and love & support. Jessica and Fiona, thank you so much for all the extra help and sup-port during my thesis writing, all your proofreading, the highly needed PhD survival kit and for in general being absolutely amazing friends.

To my family – you are the best!

45

References

1. Ferlay J, Soerjomataram I, Dikshit R, Eser S, Mathers C, Rebelo M, et al. Cancer incidence and mortality worldwide: sources, methods and major patterns in GLOBOCAN 2012. Int J Cancer. 2015;136(5):E359-86.

2. Herbst RS HJ, Lippman SM. Lung cancer. N Engl J Med. 2008;359:1367–1380.

3. Herbst RS, Morgensztern D, Boshoff C. The biology and management of non-small cell lung cancer. Nature. 2018;553(7689):446-54.

4. Society AC. Cancer Facts and Figures 2012. 2012. 5. Proctor RN. The history of the discovery of the cigarette-lung cancer link:

evidentiary traditions, corporate denial, global toll. Tob Control. 2012;21(2):87-91.

6. Humans IWGotEoCRt. Tobacco smoke and involuntary smoking. IARC Monogr Eval Carcinog Risks Hum. 2004;83:1-1438.

7. Cancerfondsrapporten 2017 2017 [ 8. Dela Cruz CS, Tanoue LT, Matthay RA. Lung cancer: epidemiology, etiology,

and prevention. Clin Chest Med. 2011;32(4):605-44. 9. Alberg AJ, Brock MV, Ford JG, Samet JM, Spivack SD. 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(5 Suppl):e1S-e29S.

10. Planchard D, Besse B. Lung cancer in never-smokers. Eur Respir J. 2015;45(5):1214-7.

11. Pawel DJ, Puskin JS. The U.S. Environmental Protection Agency's assessment of risks from indoor radon. Health Phys. 2004;87(1):68-74.

12. Robertson A, Allen J, Laney R, Curnow A. The cellular and molecular carcinogenic effects of radon exposure: a review. Int J Mol Sci. 2013;14(7):14024-63.

13. Neuberger JS, Field RW. Occupation and lung cancer in nonsmokers. Rev Environ Health. 2003;18(4):251-67.

14. Okazaki I, Ishikawa S, Sohara Y. Genes associated with succeptibility to lung adenocarcinoma among never smokers suggest the mechanism of disease. Anticancer Res. 2014;34(10):5229-40.

15. Brennan P, Hainaut P, Boffetta P. Genetics of lung-cancer susceptibility. Lancet Oncol. 2011;12(4):399-408.

16. Spiro SG, Gould MK, Colice GL, American College of Chest P. Initial evaluation of the patient with lung cancer: symptoms, signs, laboratory tests, and paraneoplastic syndromes: ACCP evidenced-based clinical practice guidelines (2nd edition). Chest. 2007;132(3 Suppl):149S-60S.

17. Socialstyrelsen. Nationella riktlinjer för lungcancervård 2011-– stöd för styrning och ledning. 2011.

46

18. Dietel M, Bubendorf L, Dingemans AM, Dooms C, Elmberger G, Garcia RC, et al. Diagnostic procedures for non-small-cell lung cancer (NSCLC): recommendations of the European Expert Group. Thorax. 2016;71(2):177-84.

19. Sculier JP, Chansky K, Crowley JJ, Van Meerbeeck J, Goldstraw P, International Staging C, et al. The impact of additional prognostic factors on survival and their relationship with the anatomical extent of disease expressed by the 6th Edition of the TNM Classification of Malignant Tumors and the proposals for the 7th Edition. J Thorac Oncol. 2008;3(5):457-66.

20. Staging manual in thoracic oncology [Available from: https://www.iaslc.org/sites/default/files/wysiwyg-assets/8th_staging_manual_2016_hi-res.pdf.

21. Liam CK, Andarini S, Lee P, Ho JC, Chau NQ, Tscheikuna J. Lung cancer staging now and in the future. Respirology. 2015;20(4):526-34.

22. Tsim S, O'Dowd CA, Milroy R, Davidson S. Staging of non-small cell lung cancer (NSCLC): a review. Respir Med. 2010;104(12):1767-74.

23. Goldstraw P, Crowley J, Chansky K, Giroux DJ, Groome PA, Rami-Porta R, et al. The IASLC Lung Cancer Staging Project: proposals for the revision of the TNM stage groupings in the forthcoming (seventh) edition of the TNM Classification of malignant tumours. J Thorac Oncol. 2007;2(8):706-14.

24. Travis W.D BE, Burke A.P, Marx A, Nicholson A. (Editors). WHO Classification of tumours of the lung, pleura, thymus and heart. Lyon (France): IARC; 2015.

25. Brambilla E, Travis WD, Colby TV, Corrin B, Shimosato Y. The new World Health Organization classification of lung tumours. Eur Respir J. 2001;18(6):1059-68.

26. Wistuba, II, Gazdar AF. Lung cancer preneoplasia. Annu Rev Pathol. 2006;1:331-48.

27. Travis WD, Brambilla E, Nicholson AG, Yatabe Y, Austin JH, Beasley MB, et al. The 2015 World Health Organization Classification of Lung Tumors: Impact of Genetic, Clinical and Radiologic Advances Since the 2004 Classification. J Thorac Oncol. 2015;10(9):1243-60.

28. Gandara DR, Hammerman PS, Sos ML, Lara PN, Jr., Hirsch FR. Squamous cell lung cancer: from tumor genomics to cancer therapeutics. Clin Cancer Res. 2015;21(10):2236-43.

29. Thun MJ, Lally CA, Flannery JT, Calle EE, Flanders WD, Heath CW, Jr. Cigarette smoking and changes in the histopathology of lung cancer. J Natl Cancer Inst. 1997;89(21):1580-6.

30. Drilon A, Rekhtman N, Ladanyi M, Paik P. Squamous-cell carcinomas of the lung: emerging biology, controversies, and the promise of targeted therapy. Lancet Oncol. 2012;13(10):e418-26.

31. Cancer Genome Atlas Research N. Comprehensive genomic characterization of squamous cell lung cancers. Nature. 2012;489(7417):519-25.

32. Clinical Lung Cancer Genome P, Network Genomic M. A genomics-based classification of human lung tumors. Sci Transl Med. 2013;5(209):209ra153.

33. Rekhtman N, Ang DC, Sima CS, Travis WD, Moreira 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(10):1348-59.

34. Travis WD, Brambilla E, Nicholson AG, Yatabe Y, Austin JHM, Beasley MB, et al. The 2015 World Health Organization Classification of Lung Tumors: Impact of Genetic, Clinical and Radiologic Advances Since the 2004 Classification. J Thorac Oncol. 2015;10(9):1243-60.

47

35. Silvestri GA, Rivera MP. Targeted therapy for the treatment of advanced non-small cell lung cancer: a review of the epidermal growth factor receptor antagonists. Chest. 2005;128(6):3975-84.

36. Scagliotti GV, Parikh P, von Pawel J, Biesma B, Vansteenkiste J, Manegold C, et al. Phase III study comparing cisplatin plus gemcitabine with cisplatin plus pemetrexed in chemotherapy-naive patients with advanced-stage non-small-cell lung cancer. J Clin Oncol. 2008;26(21):3543-51.

37. Rami-Porta R, Crowley JJ, Goldstraw P. The revised TNM staging system for lung cancer. Ann Thorac Cardiovasc Surg. 2009;15(1):4-9.

38. Uramoto H, Tanaka F. Recurrence after surgery in patients with NSCLC. Transl Lung Cancer Res. 2014;3(4):242-9.

39. Howington JA, Blum MG, Chang AC, Balekian AA, Murthy SC. Treatment of stage I and II non-small cell lung cancer: Diagnosis and management of lung cancer, 3rd ed: American College of Chest Physicians evidence-based clinical practice guidelines. Chest. 2013;143(5 Suppl):e278S-313S.

40. Strauss GM, Herndon JE, 2nd, Maddaus MA, Johnstone DW, Johnson EA, Harpole DH, et al. Adjuvant paclitaxel plus carboplatin compared with observation in stage IB non-small-cell lung cancer: CALGB 9633 with the Cancer and Leukemia Group B, Radiation Therapy Oncology Group, and North Central Cancer Treatment Group Study Groups. J Clin Oncol. 2008;26(31):5043-51.

41. Dosoretz DE, Katin MJ, Blitzer PH, Rubenstein JH, Galmarini DH, Garton GR, et al. Medically Inoperable Lung Carcinoma: The Role of Radiation Therapy. Semin Radiat Oncol. 1996;6(2):98-104.

42. Giaj-Levra N, Ricchetti F, Alongi F. What is changing in radiotherapy for the treatment of locally advanced nonsmall cell lung cancer patients? A review. Cancer Invest. 2016;34(2):80-93.

43. Kumar M, Ernani V, Owonikoko TK. Biomarkers and targeted systemic therapies in advanced non-small cell lung cancer. Mol Aspects Med. 2015;45:55-66.

44. Mok TS, Wu YL, Ahn MJ, Garassino MC, Kim HR, Ramalingam SS, et al. Osimertinib or Platinum-Pemetrexed in EGFR T790M-Positive Lung Cancer. N Engl J Med. 2017;376(7):629-40.

45. Soda M, Choi YL, Enomoto M, Takada S, Yamashita Y, Ishikawa S, et al. Identification of the transforming EML4-ALK fusion gene in non-small-cell lung cancer. Nature. 2007;448(7153):561-6.

46. 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.

47. Mezquita L, Planchard D. The role of brigatinib in crizotinib-resistant non-small cell lung cancer. Cancer Manag Res. 2018;10:123-30.

48. Gainor JF, Shaw AT. Novel targets in non-small cell lung cancer: ROS1 and RET fusions. Oncologist. 2013;18(7):865-75.

49. Wang R, Hu H, Pan Y, Li Y, Ye T, Li C, et al. RET fusions define a unique molecular and clinicopathologic subtype of non-small-cell lung cancer. J Clin Oncol. 2012;30(35):4352-9.

50. Ferrara R, Auger N, Auclin E, Besse B. Clinical and Translational Implications of RET Rearrangements in Non-Small Cell Lung Cancer. J Thorac Oncol. 2018;13(1):27-45.

48

51. Shaw AT, Felip E, Bauer TM, Besse B, Navarro A, Postel-Vinay S, et al. Lorlatinib in non-small-cell lung cancer with ALK or ROS1 rearrangement: an international, multicentre, open-label, single-arm first-in-man phase 1 trial. Lancet Oncol. 2017;18(12):1590-9.

52. Oxnard GR, Arcila ME, Sima CS, Riely GJ, Chmielecki J, Kris MG, et al. Acquired resistance to EGFR tyrosine kinase inhibitors in EGFR-mutant lung cancer: distinct natural history of patients with tumors harboring the T790M mutation. Clin Cancer Res. 2011;17(6):1616-22.

53. Gainor JF, Dardaei L, Yoda S, Friboulet L, Leshchiner I, Katayama R, et al. Molecular Mechanisms of Resistance to First- and Second-Generation ALK Inhibitors in ALK-Rearranged Lung Cancer. Cancer Discov. 2016;6(10):1118-33.

54. Bremnes RM, Al-Shibli K, Donnem T, Sirera R, Al-Saad S, Andersen S, et al. The role of tumor-infiltrating immune cells and chronic inflammation at the tumor site on cancer development, progression, and prognosis: emphasis on non-small cell lung cancer. J Thorac Oncol. 2011;6(4):824-33.

55. Khanna P, Blais N, Gaudreau PO, Corrales-Rodriguez L. Immunotherapy Comes of Age in Lung Cancer. Clin Lung Cancer. 2016.

56. Anagnostou VK, Brahmer JR. Cancer immunotherapy: a future paradigm shift in the treatment of non-small cell lung cancer. Clin Cancer Res. 2015;21(5):976-84.

57. Brahmer J, Reckamp KL, Baas P, Crino L, Eberhardt WE, Poddubskaya E, et al. Nivolumab versus Docetaxel in Advanced Squamous-Cell Non-Small-Cell Lung Cancer. N Engl J Med. 2015;373(2):123-35.

58. Garon EB, Rizvi NA, Hui R, Leighl N, Balmanoukian AS, Eder JP, et al. Pembrolizumab for the treatment of non-small-cell lung cancer. N Engl J Med. 2015;372(21):2018-28.

59. Massuti B, Sanchez JM, Hernando-Trancho F, Karachaliou N, Rosell R. Are we ready to use biomarkers for staging, prognosis and treatment selection in early-stage non-small-cell lung cancer? Transl Lung Cancer Res. 2013;2(3):208-21.

60. Jeremic B. Advances in Radiation Oncology in Lung Cancer. Sartor ALBLWBH-PHMMK, editor. Berlin and Heidelberg, Germany: Springer 2005.

61. Birim O, Kappetein AP, van Klaveren RJ, Bogers AJ. Prognostic factors in non-small cell lung cancer surgery. Eur J Surg Oncol. 2006;32(1):12-23.

62. Oken MM, Creech RH, Tormey DC, Horton J, Davis TE, McFadden ET, et al. Toxicity and response criteria of the Eastern Cooperative Oncology Group. Am J Clin Oncol. 1982;5(6):649-55.

63. Tas F, Ciftci R, Kilic L, Karabulut S. Age is a prognostic factor affecting survival in lung cancer patients. Oncol Lett. 2013;6(5):1507-13.

64. Agarwal M, Brahmanday G, Chmielewski GW, Welsh RJ, Ravikrishnan KP. Age, tumor size, type of surgery, and gender predict survival in early stage (stage I and II) non-small cell lung cancer after surgical resection. Lung Cancer. 2010;68(3):398-402.

65. Bolton WD, Rice DC, Correa AM, Hofstetter W, Komaki R, Mehran R, et al. Influence of age on choice of therapy and surgical outcomes in patients with nonsmall cell lung cancer. Am Surg. 2009;75(7):598-603; discussion -4.

66. Woodard GA, Jones KD, Jablons DM. Lung Cancer Staging and Prognosis. Cancer Treat Res. 2016;170:47-75.

67. Tsao MS, Jablons DM. Molecular prognostication of non-small cell lung cancer. Semin Thorac Cardiovasc Surg. 2013;25(1):4-7.

49

68. Friboulet L, Olaussen KA, Pignon JP, Shepherd FA, Tsao MS, Graziano S, et al. ERCC1 isoform expression and DNA repair in non-small-cell lung cancer. N Engl J Med. 2013;368(12):1101-10.

69. Zhu CQ, Ding K, Strumpf D, Weir BA, Meyerson M, Pennell N, et al. Prognostic and predictive gene signature for adjuvant chemotherapy in resected non-small-cell lung cancer. J Clin Oncol. 2010;28(29):4417-24.

70. Kratz JR, He J, Van Den Eeden SK, Zhu ZH, Gao W, Pham PT, et al. A practical molecular assay to predict survival in resected non-squamous, non-small-cell lung cancer: development and international validation studies. Lancet. 2012;379(9818):823-32.

71. Hu Z, Chen X, Zhao Y, Tian T, Jin G, Shu Y, et al. Serum microRNA signatures identified in a genome-wide serum microRNA expression profiling predict survival of non-small-cell lung cancer. J Clin Oncol. 2010;28(10):1721-6.

72. Sandoval J, Mendez-Gonzalez J, Nadal E, Chen G, Carmona FJ, Sayols S, et al. A prognostic DNA methylation signature for stage I non-small-cell lung cancer. J Clin Oncol. 2013;31(32):4140-7.

73. Brett GZ. Earlier diagnosis and survival in lung cancer. Br Med J. 1969;4(5678):260-2.

74. Kubik A, Parkin DM, Khlat M, Erban J, Polak J, Adamec M. Lack of benefit from semi-annual screening for cancer of the lung: follow-up report of a randomized controlled trial on a population of high-risk males in Czechoslovakia. Int J Cancer. 1990;45(1):26-33.

75. Fontana RS, Sanderson DR, Woolner LB, Taylor WF, Miller WE, Muhm JR, et al. Screening for lung cancer. A critique of the Mayo Lung Project. Cancer. 1991;67(4 Suppl):1155-64.

76. National Lung Screening Trial Research T, Aberle DR, Adams AM, Berg CD, Black WC, Clapp JD, et al. Reduced lung-cancer mortality with low-dose computed tomographic screening. N Engl J Med. 2011;365(5):395-409.

77. Bach PB, Mirkin JN, Oliver TK, Azzoli CG, Berry DA, Brawley OW, et al. Benefits and harms of CT screening for lung cancer: a systematic review. JAMA. 2012;307(22):2418-29.

78. Oudkerk M, Devaraj A, Vliegenthart R, Henzler T, Prosch H, Heussel CP, et al. European position statement on lung cancer screening. Lancet Oncol. 2017;18(12):e754-e66.

79. Griswold MD. Spermatogenesis: The Commitment to Meiosis. Physiol Rev. 2016;96(1):1-17.

80. Griswold MD. The central role of Sertoli cells in spermatogenesis. Semin Cell Dev Biol. 1998;9(4):411-6.

81. Fagerberg L, Hallstrom BM, Oksvold P, Kampf C, Djureinovic D, Odeberg J, et al. Analysis of the human tissue-specific expression by genome-wide integration of transcriptomics and antibody-based proteomics. Mol Cell Proteomics. 2014;13(2):397-406.

82. Ramskold D, Wang ET, Burge CB, Sandberg R. An abundance of ubiquitously expressed genes revealed by tissue transcriptome sequence data. PLoS Comput Biol. 2009;5(12):e1000598.

83. Perez CV, Theas MS, Jacobo PV, Jarazo-Dietrich S, Guazzone VA, Lustig L. Dual role of immune cells in the testis: Protective or pathogenic for germ cells? Spermatogenesis. 2013;3(1):e23870.

84. Caballero OL, Chen YT. Cancer/testis (CT) antigens: potential targets for immunotherapy. Cancer Sci. 2009;100(11):2014-21.

50

85. van der Bruggen P, Traversari C, Chomez P, Lurquin C, De Plaen E, Van den Eynde B, et al. A gene encoding an antigen recognized by cytolytic T lymphocytes on a human melanoma. Science. 1991;254(5038):1643-7.

86. Boel P, Wildmann C, Sensi ML, Brasseur R, Renauld JC, Coulie P, et al. BAGE: a new gene encoding an antigen recognized on human melanomas by cytolytic T lymphocytes. Immunity. 1995;2(2):167-75.

87. Van den Eynde B, Peeters O, De Backer O, Gaugler B, Lucas S, Boon T. A new family of genes coding for an antigen recognized by autologous cytolytic T lymphocytes on a human melanoma. J Exp Med. 1995;182(3):689-98.

88. Gaugler B, Van den Eynde B, van der Bruggen P, Romero P, Gaforio JJ, De Plaen E, et al. Human gene MAGE-3 codes for an antigen recognized on a melanoma by autologous cytolytic T lymphocytes. J Exp Med. 1994;179(3):921-30.

89. Sahin U, Tureci O, Pfreundschuh M. Serological identification of human tumor antigens. Curr Opin Immunol. 1997;9(5):709-16.

90. Tureci O, Sahin U, Schobert I, Koslowski M, Scmitt H, Schild HJ, et al. The SSX-2 gene, which is involved in the t(X;18) translocation of synovial sarcomas, codes for the human tumor antigen HOM-MEL-40. Cancer Res. 1996;56(20):4766-72.

91. Chen YT, Scanlan MJ, Sahin U, Tureci O, Gure AO, Tsang S, et al. A testicular antigen aberrantly expressed in human cancers detected by autologous antibody screening. Proc Natl Acad Sci U S A. 1997;94(5):1914-8.

92. Tureci O, Sahin U, Zwick C, Koslowski M, Seitz G, Pfreundschuh M. Identification of a meiosis-specific protein as a member of the class of cancer/testis antigens. Proc Natl Acad Sci U S A. 1998;95(9):5211-6.

93. Lee SY, Williamson B, Caballero OL, Chen YT, Scanlan MJ, Ritter G, et al. Identification of the gonad-specific anion transporter SLCO6A1 as a cancer/testis (CT) antigen expressed in human lung cancer. Cancer Immun. 2004;4:13.

94. Scanlan MJ, Gordon CM, Williamson B, Lee SY, Chen YT, Stockert E, et al. Identification of cancer/testis genes by database mining and mRNA expression analysis. Int J Cancer. 2002;98(4):485-92.

95. Chen YT, Scanlan MJ, Venditti CA, Chua R, Theiler G, Stevenson BJ, et al. Identification of cancer/testis-antigen genes by massively parallel signature sequencing. Proc Natl Acad Sci U S A. 2005;102(22):7940-5.

96. Almeida LG, Sakabe NJ, deOliveira AR, Silva MC, Mundstein AS, Cohen T, et al. CTdatabase: a knowledge-base of high-throughput and curated data on cancer-testis antigens. Nucleic Acids Res. 2009;37(Database issue):D816-9.

97. Gjerstorff MF, Andersen MH, Ditzel HJ. Oncogenic cancer/testis antigens: prime candidates for immunotherapy. Oncotarget. 2015;6(18):15772-87.

98. Antonia SJ, Vansteenkiste JF, Moon E. Immunotherapy: Beyond Anti-PD-1 and Anti-PD-L1 Therapies. Am Soc Clin Oncol Educ Book. 2016;35:e450-8.

99. Ramos-Vara JA, Beissenherz ME. Optimization of immunohistochemical methods using two different antigen retrieval methods on formalin-fixed paraffin-embedded tissues: experience with 63 markers. J Vet Diagn Invest. 2000;12(4):307-11.

100. Buchwalow IB, Böcker W. Immunohistochemistry: Basics and Methods: Springer; 2010.

101. Torhorst J, Bucher C, Kononen J, Haas P, Zuber M, Kochli OR, et al. Tissue microarrays for rapid linking of molecular changes to clinical endpoints. Am J Pathol. 2001;159(6):2249-56.

51

102. Karlsson C, Bodin L, Piehl-Aulin K, Karlsson MG. Tissue microarray validation: a methodologic study with special reference to lung cancer. Cancer Epidemiol Biomarkers Prev. 2009;18(7):2014-21.

103. Camp RL, Charette LA, Rimm DL. Validation of tissue microarray technology in breast carcinoma. Lab Invest. 2000;80(12):1943-9.

104. Kononen J, Bubendorf L, Kallioniemi A, Barlund M, Schraml P, Leighton S, et al. Tissue microarrays for high-throughput molecular profiling of tumor specimens. Nat Med. 1998;4(7):844-7.

105. Wang Z, Gerstein M, Snyder M. RNA-Seq: a revolutionary tool for transcriptomics. Nat Rev Genet. 2009;10(1):57-63.

106. Maher CA, Kumar-Sinha C, Cao X, Kalyana-Sundaram S, Han B, Jing X, et al. Transcriptome sequencing to detect gene fusions in cancer. Nature. 2009;458(7234):97-101.

107. Wang ET, Sandberg R, Luo S, Khrebtukova I, Zhang L, Mayr C, et al. Alternative isoform regulation in human tissue transcriptomes. Nature. 2008;456(7221):470-6.

108. Conde L, Bracci PM, Richardson R, Montgomery SB, Skibola CF. Integrating GWAS and expression data for functional characterization of disease-associated SNPs: an application to follicular lymphoma. Am J Hum Genet. 2013;92(1):126-30.

109. Dunbar SA. Applications of Luminex xMAP technology for rapid, high-throughput multiplexed nucleic acid detection. Clin Chim Acta. 2006;363(1-2):71-82.

110. Kettman JR, Davies T, Chandler D, Oliver KG, Fulton RJ. Classification and properties of 64 multiplexed microsphere sets. Cytometry. 1998;33(2):234-43.

111. Nilsson P, Paavilainen L, Larsson K, Odling J, Sundberg M, Andersson AC, et al. Towards a human proteome atlas: high-throughput generation of mono-specific antibodies for tissue profiling. Proteomics. 2005;5(17):4327-37.

112. Assarsson E, Lundberg M, Holmquist G, Bjorkesten J, Thorsen SB, Ekman D, et al. Homogenous 96-plex PEA immunoassay exhibiting high sensitivity, specificity, and excellent scalability. PLoS One. 2014;9(4):e95192.

113. Lundberg M, Eriksson A, Tran B, Assarsson E, Fredriksson S. Homogeneous antibody-based proximity extension assays provide sensitive and specific detection of low-abundant proteins in human blood. Nucleic Acids Res. 2011;39(15):e102.

114. Greenwood C, Ruff D, Kirvell S, Johnson G, Dhillon HS, Bustin SA. Proximity assays for sensitive quantification of proteins. Biomol Detect Quantif. 2015;4:10-6.

115. Lohr M, Edlund K, Botling J, Hammad S, Hellwig B, Othman A, 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(2):222-8.

116. Edlund K, Lindskog C, Saito A, Berglund A, Ponten F, Goransson-Kultima H, et al. CD99 is a novel prognostic stromal marker in non-small cell lung cancer. Int J Cancer. 2012;131(10):2264-73.

117. Micke P, Mattsson JS, Djureinovic D, Nodin B, Jirstrom K, Tran L, et al. The Impact of the Fourth Edition of the WHO Classification of Lung Tumours on Histological Classification of Resected Pulmonary NSCCs. J Thorac Oncol. 2016;11(6):862-72.

52

118. Tran L, Mattsson JS, Nodin B, Jonsson P, Planck M, Jirstrom K, 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.

119. Djureinovic D, Hallström BM, Horie M, Mattsson JSM, La Fleur L, Fagerberg L, et al. Profiling cancer testis antigens in non–small-cell lung cancer. JCI Insight. 2016;1(10):e86837.

120. Perez-Gracia JL, Sanmamed MF, Bosch A, Patino-Garcia A, Schalper KA, Segura V, et al. Strategies to design clinical studies to identify predictive biomarkers in cancer research. Cancer Treat Rev. 2017;53:79-97.

121. McShane LM, Altman DG, Sauerbrei W, Taube SE, Gion M, Clark GM, et al. REporting recommendations for tumor MARKer prognostic studies (REMARK). Nat Clin Pract Oncol. 2005;2(8):416-22.

Acta Universitatis UpsaliensisDigital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Medicine 1463

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-348349

ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2018