Cancer Gene Th erapy Group Department of Pathology, Faculty of Medicine Th e National Graduate School of Clinical Investigation Helsinki University Central Hospital & University of Helsinki Finland

CANCER IMMUNOTHERAPY WITH A GENE MODIFIED SEROTYPE 3 ONCOLYTIC ADENOVIRUS

Otto Hemminki

ACADEMIC DISSERTATION

To be presented, with the permission of the Faculty of Medicine of the University of Helsinki, for public examination in lecture hall 2, Haartman Institute on 13th of November 2015, at 12 noon.

Helsinki 2015 Supervised by Professor Akseli Hemminki, MD, PhD Department of Oncology Cancer Gene Th erapy Group Department of Pathology, Faculty of Medicine University of Helsinki Finland

Th esis Committee Professor Ari Harjula, MD, PhD Adjuct Professor Maija Lappalainen, MD, PhD Department of Cardiothoracic Surgery Department. of Virology and Immunology Helsinki University Central Hospital Helsinki University Central Hospital

Reviewers appointed by the Faculty Professor Kari Airenne, PhD Adjunct Professor Timo Muhonen, MD, PhD Professor of Molecular Medicine Department of Oncology University of Eastern Finland University of Helsinki A.I. Virtanen Institute

Offi cial Opponent Professor Ruben Hernandez Alcoceba, MD, PhD Division of Gene Th erapy and Hepatology University of Navarra, Spain

Dissertationes Scholae Doctoralis Ad Sanitatem Investigandam Universitatis Helsinkiensis

No. 95/2015 ISBN 978-951-51-1686-4 (paperback) ISBN 978-951-51-1687-1 (PDF) ISSN 2342-3161 (print) ISSN 2342-317X (online) http://ethesis.helsinki.fi

Layout: Tinde Päivärinta/PSWFOlders Oy Hansaprint, Vantaa 2015

“Just wanted to encourage you to continue the innovative approaches you are trying. N=1 is the way to go to test such therapies, despite what others say. In my view, conventional Phase I trials raise far more concerning ethical issues than those associated with ATAP’s approach.

Best, Bert” A spontaneous e-mail sent to us by the All-time most cited scientist in science Bert Vogelstein Director, Ludwig Center at Johns Hopkins Investigator, Howard Hughes Medical Institute

Parts of this theses text, tables and fi gures are published in a text book chapter (Hem- minki 2014) (Gene Th erapy of Cancer, 3rd Edition, Lattime and Gerson, Elsevier, 2013) written by Hemminki O et A. Permissions to use this material has been obtained from the publishers. To people with cancer TABLE OF CONTENTS

Abstract Tiivistelmä List of original publications Abbreviations Introduction ...... 1 Review of the literature ...... 4 1. Cancer ...... 4 1.1 Classifi cation of cancer ...... 4 1.2 Cancer is a disease of the genome ...... 4 2. Adenoviruses ...... 7 2.1 Adenovirus structure ...... 7 2.2 Adenovirus classifi cation ...... 7 2.3 Adenovirus receptors ...... 10 2.4 Adenovirus replication ...... 13 3. Cancer treatment, History and Future ...... 14 3.1 Current standard treatment of cancer ...... 14 3.2 History of cancer treatments with viruses ...... 16 3.3 Gene modifi ed adenoviruses in cancer therapy ...... 18 3.4 Clinical trials with oncolytic adenoviruses ...... 24 3.5 Cancer immunotherapy in patients ...... 26 3.6 Evaluating the effi cacy of immunotherapies ...... 34 3.7 Immunotherapy in the future ...... 36 Aims of the thesis ...... 41 Materials and methods ...... 42 Results and discussion ...... 50 Summary and conclusions ...... 61 Acknowledgements ...... 65 References ...... 67 Original publications ABSTRACT

In 2012 WHO announced cancer as the leading cause of death. Every day 20 000 people die due to cancer, and the rate is estimated to double before year 2030. While treatments have progressed, there are still few good treatment options for advanced cancer. Th us, there is an urgent need for new treatments. Immunotherapy with gene modifi ed oncolytic adenoviruses provides novel promising means of treating cancer. Th ese treatments incorporate two basic concepts. Firstly, adenoviruses are modifi ed so that they replicate only in cancer cells, which makes the treatments safer. Secondly, the virus induced cancer cell oncolysis elicits a danger signal that awakens the immune system to fi ght the cancer. Viruses can be further armed with diff erent genes that can stimulate the immune system even more. Most of these oncolytic viruses are based on adenovirus serotype 5, as indicated in thousands of publications. However, the primary receptor for serotype 5 is down-regulated in advanced cancer. On the contrary, adenovirus serotype 3 receptor is known to be abundant in advanced cancer making it an interesting subject of research. While a diff erent serotype per see off ers an alternative immune response, serotype 3 incorporates also other interesting features that might further potentiate its utility. Our fi rst goal was to create serotype 3 based oncolytic adenoviruses for the treatment of human cancer. Th e goal was achieved, making this virus, to our knowledge, the fi rst non- adenovirus 5 based oncolytic adenovirus in the world used in humans. Th e publications, study I and II, are now part of this thesis. Th e virus was designed to have a human telomerase reverse transcriptase (hTERT) promoter diverting the replication of the virus into cancer cells. Th is virus, Ad3-hTERT-E1A, was successfully cloned, rescued and produced in large scale, which was followed by rigorous preclinical testing of the virus. Rigorous preclinical testing of the virus followed. Several in vitro and in vivo experiments were performed, including sequencing, qPCR, electron microscopy and neutralizing antibody assays, while the most convincing data was gained from the cell cultures and the animal models. We found the serotype 3 eff ective in all major cancer types in vitro. In vivo, the serotype 3 virus was found at least as potent as serotype 5 based control viruses in several murine models of human cancer. Before clinical treatments, biodistribution and toxicity experiments were performed. In toxicity studies, adenovirus 3 was found less toxic than the serotype 5 based control viruses in an immune competent murine model. Th e histology of all major organs and basic blood values were analyzed. Th e preclinical data suggested strong effi cacy with good safety. In study II, we publish the data of the fi rst 25 patients treated with the Ad3-hTERT-E1A virus. All patients had advanced solid tumors refractory to standard therapies. Th e safety of the treatment was good with up to 4x1012 virus particles given intravenously and/or intratumorally. Overall, all patients experienced mild (grade 1-2) self-limiting fl u-like adverse events. No severe adverse events were noted attributable to the treatments. Aft er treatment, many patients showed signs of effi cacy. Of the 15 patients with elevated tumor markers before the treatment, 73% responded with a decrease or no change in the markers. Even a few complete responses were reported, while some patients also showed a clear decrease in the tumor mass according to imaging. Also the clinical data suggested strong effi cacy with good safety, proposing a need for a randomized study. Our next goal was to evaluate better ways in fi nding treatment responders, as size based computed tomography (CT) is known to be suboptimal in evaluating immunotherapeutics where initial swelling of the tumor due to the immune response is common. In study III, we examined the ability of magnetic resonance imaging (MRI) and spectroscopy (MRS) in immunocompetent Syrian hamsters. T2 weighed MRI seemed encouraging in fi nding responding hamsters as soon as two days aft er treatment. Similar fi ndings were noted with a patient responding to oncolytic treatments. MRS of taurine, choline and unsaturated fatty acids were found to be promising metabolites when evaluating responders aft er oncolytic immunotherapy. Th ese results propose MRI and MRS as potential methods in evaluating responding patients. T2 weighed MRI is already widely used in the clinics, thus a clinical trial should be easy to implement. In study IV, we evaluated the fi rst 16 patients treated with a quadruple modifi ed oncolytic serotype 5 adenovirus. Th e fi ber knob of this virus is from serotype 3, while the virus also produces an immunostimulatory GM-CSF molecule. Th e two other modifi cations restrict the replication to cancer cells. Th e safety profi le of the virus resembled that of the oncolytic serotype 3 virus, and also numerous signs of effi cacy were noted. Immunological studies indicated activation of the immune system in responding patients. Rationale for a randomized study exists also for this virus. TIIVISTELMÄ

WHO arvioi 2012 syövän maailman yleisimmäksi kuolinsyyksi ja saman arvion mukaan syöpäkuolemat kaksinkertaistuvat vuoteen 2030 mennessä. Vaikkakin perinteiset syöpähoidot ovat kehittyneet merkittävästi viime vuosikymmeninä, levinneen taudin ennuste on edelleen huono. Täten uusien syöpähoitojen tarve on ilmeinen. Onkolyyttiset virukset ovat yksi mahdollisuus hoitaa levinnyttä syöpää. Viruksia voidaan muokata siten, että niiden lisääntyminen rajoittuu syöpäkudoksiin tehden onkolyyttisistä viruksista turvallisempia kuin luonnon omat virukset. Halutessa viruksiin voidaan myös lisätä geenejä, jolloin ne saadaan tuottamaan esimerkiksi immuunipuolustusta aktivoivia molekyylejä. Immuunipuolustuksen aktivoituminen vaikuttaisikin olevan ensisijaisen tärkeää hyvän hoitovasteen saavuttamiseksi. Adenovirukset ovat osoittautuneet soveltuvan hyvin tähän käyttöön, ja tuhansia julkaisuja erityisesti adenovirus serotyyppi 5:sta on olemassa. Tässä väitöskirjassa tarkastellaan erityisesti adenovirus serotyyppi 3:n ominaisuuksia syövän immunoterapiassa. Tiettävästi tämä on ensimmäinen ei-serotyypin-5 onkolyyttinen virus, jolla on hoidettu potilaita. Väitöskirjan ensimmäinen osatyö perustuu prekliinisiin töihin serotyypin 3 onkolyyttisella viruksella ja toinen painottuu potilashoitojen raportointiin. Kolmannessa osatyössä tarkastelen magneettikuvantamisen ja magneettispektroskopian mahdollisuuksia onkolyyttisissä immunoterapioissa. Neljäs osatyö käsittelee potilaita, joita on hoidettu onkolyyttisellä viruksella, johon asetettu geeni saa aikaan sen, että virus tuottaa immuunipuolustusta aktivoivaa GM-CSF sytokiinia. Yhteenvetona voidaan todeta, että serotyypin 3 onkolyyttinen adenovirus vaikuttaa lupaavalta hoitomuodolta. Hoidetuista potilaista (N=25), joilla tuumorimarkkerit olivat koholla ennen hoitoa (N=15), markkerit laskivat tai pysyivät ennallaan 73 %:lla. Lisäksi muutamilla potilailla havaittiin selkeitä tuumorimassan pienenemisiä kuvantamisessa. Vastaavia havaintoja tehtiin myös potilaissa, joita hoidettiin GM-CSF sytokiinilla varustetulla viruksella. Prekliiniset kuvantamistutkimukset ja yksittäinen potilas antoivat viitettä siitä, että magneetti tai magneettispektroskopia kuvantamisesta voisi olla hyötyä hoitovastetta arvioitaessa. LIST OF ORIGINAL PUBLICATIONS

Th is thesis is based on the following publications:

I Hemminki O, Bauerschmitz G, Hemmi S, Lavilla-Alonso S, Diaconu I, Koski A, Guse K, Desmond RA, Lappalainen M, Kanerva A, Cerullo V, Pesonen S, Hemminki A. Oncolytic adenovirus based on serotype 3, Cancer Gene Th erapy 2010

II Hemminki O, Diaconu I, Cerullo V, Pesonen S, Kanerva A, Joensuu T, Kairemo K, Laasonen L, Partanen K, Kangasniemi L, Lieber A, Pesonen S, Hemminki A. Ad3-hTERT- E1A, a fully serotype 3 oncolytic adenovirus, in patients with chemotherapy refractory cancer, Molecular Th erapy 2012

III Hemminki O, Immonen R, Närväinen J, Kipari A, Soininen P, Pesonen S, Gröhn O, Hemminki A. In vivo magnetic resonance imaging (MRI) and spectroscopy (MRS) identifi es oncolytic adenovirus responders, Int J Cancer 2014

IV Hemminki O, Parviainen S, Juhila J, Turkki R, Linder N, Lundin L, Kankainen M, Ristimäki A, Koski A, Liikanen I, Oksanen M, Nettelbeck D. M., Kairemo K, Partanen K, Joensuu T, Kanerva A, Hemminki A. Immunological data from cancer patients treated with Ad5/3-E2F-Δ24-GMCSF suggests utility for tumor immunotherapy, Oncotarget 2015

Th e publications are referred to in the text by their roman numerals and can be found at the end of this thesis. Publications are reproduced with permission of the copyright holders. ABBREVIATIONS

Ad adenovirus Ad3 adenovirus serotype 3 Ad5 adenovirus serotype 5 ACT adoptive cell transfer AE adverse event ATAP Advanced Th erapy Access Program BCG Bacillus Calmette–Guérin bp base pair CAR coxsackie‐adenovirus receptor or chimeric antigen receptor CD40L CD40 ligand CMV cytomegalovirus CR complete response CT computed tomography CTL cytotoxic T lymphocyte CTLA‐4 cytotoxic T‐lymphocyte antigen 4 DMEM Dulbecco’s modifi ed Eagle’s medium DSG-2 Desmoglein-2 EMA European Medicines Agency EMT epithelial-to-mesenchymal transition FCS fetal calf serum GFP green fl uorescent protein GM‐CSF granulocyte‐macrophage colony‐stimulating factor hGM‐CSF human GM-CSF HCC hepatocellular carcinoma HSPG heparan sulphate proteoglycans hTERT human telomerase reverse transcriptase HVR hypervariable region IL interleukin IFN interferon i.v. intravenously IRES internal ribosome entry site ITR inverted terminal repeat MHC major histocompatibility complex MR minor response MRI magnetic resonance imaging MRS magnetic resonance spectroscopy MTS 3‐(4,5‐dimethylthiazol‐2‐yl)‐5‐(3‐carboxymethoxyphenyl)‐2‐(4‐ sulfophenyl)‐2H‐ tetrazolium NAb neutralizing antibody NaCl sodium chloride NSCLC non-small cell lung cancer NK natural killer cell ORR overall response rate PBS phosphate buff ered saline PCR polymerase chain reaction PD progressive disease PD-1 programmed cell death protein 1 PET positron emission tomography pfu plaque forming units PR partial response PSA prostate specifi c antigen qPCR quantitative real‐time PCR Rb retinoblastoma RECIST response evaluation criteria in solid tumors RGD arginine‐lysine‐aspartic acid rpm rounds per minute RPMI Roswell Park Memorial Institute medium SCCHN Squamous Cell Carcinoma of the Head and Neck SD stable disease TCID50 tissue culture infectious dose 50% TCR T-cell receptor TNF tumor necrosis factor Treg regulatory T cell TIL tumor infi ltrating lymphocytes VP viral particles WHO World Health Organization

Introduction INTRODUCTION

Published research data suggest that in the near future dramatic changes shall follow in the guidelines of the treatments of most cancers.

In 2012 World Health Organization (WHO) reported the fi rst time in history that cancer causes more deaths (8.2 million) than any other particular disease, bypassing even ischemic hearth disease, stroke and infectious diseases (WHO Global Health Observatory Data Repository, 2012). Both cancer incidence and mortality have been in a steady rise for decades. Despite massive eff orts in prevention, early diagnosis and treatment, advanced cancer remains still without curative treatment options. With many common cancers (e.g. breast, prostate, colon, bladder, melanoma) the awareness of people and doctors and the advances in diagnostic tools (e.g. CT scans, mammography, MRI, blood markers, cytology, scopic procedures etc.) have led to earlier diagnosis, while the advances in treatments have led to safer and more eff ective treatments. However, clinicians are puzzled with the problems caused by detecting smaller and smaller malignant or premalignant tumors. Th is leads to more follow up, more treatments, more complications and a huge psychological stress to the patients. Th us the benefi ts and downsides of many cancer screening programs are under constant debate. For example, it is believed that by simple PSA screening deaths to prostate cancer could be reduced, but when taking into account treatment related complications and the stress caused by diagnosis, it is believed that the screening might fi nally cause more harm to the patients and might not even lead to longer overall survival. Similar debate is ongoing concerning, for example, breast cancer screening (in Finland women at the age of 50-69 are screened at two- years intervals) and colon cancer screening (e.g. performed in some parts of Germany to patients at the age of 60 years). In the USA, cervical, colorectal, and breast cancer screening are currently recommended, and prostate, lung, and ovarian cancer screening are under active review (Wardle et al. 2015). When I started my PhD studies in 2007, cancer treatment relied on three corner stones: surgery, radiation therapy and chemotherapy, exactly aswhen I was born in 1980. In some specifi c tumor types other therapies can be used, such as anti-hormonal therapy, tyrosine kinase inhibitors and small molecular inhibitors. However, these treatments have mainly resulted in only marginal benefi t seen in slight changes in the overall or progression free survival curves. Hundreds or even thousands of vigorously selected patients have been needed to demonstrate the eff ect of the therapy. To me this demonstrates that we are still far from good treatments for the majority of the patients with advanced cancer. While there has been remarkable development in surgery, radiotherapy and medical treatments during the last century, advanced metastasized cancer has still poor prognosis. New treatment modalities are deeply needed. Th is is why I started working with oncolytic gene therapy, a fast growing fi eld due to remarkable technical advances during the last few decades. Gene sequencing and gene modifi cation techniques enable us to rationally modify viruses, for example so that their replication and oncolysis (lysis of tumor cells) can be restricted to cancer cells. Th e theory was that a virus could be injected to tumors creating oncolysis and more viruses would then be ready to infect cancer cells until every single cancer cell would be destroyed. Although the fi rst marketing approval had been recently granted for an oncolytic serotype 5 adenovirus (Xia et al. 2004) by the Chinese regulators in 2005, it seemed that more potent

1 Introduction strategies were needed. However, the safety profi le of the platform was good, as shown by billions of infections of humans by wild type adenoviruses and by thousands of patients treated with modifi ed adenoviruses (Toth et al. 2010). As the main receptor for serotype 5 adenovirus is low in most advanced cancers, we had been making viruses with the serotype 3 fi ber knob (Pesonen et al. 2010; Koski et al. 2013; Hemminki et al. 2015). It was known that the serotype 3 receptor was common in advanced cancers, although the primary receptor itself was not known at this time. My job was to create and investigate the fi rst completely serotype 3 based oncolytic adenovirus. During these years, it has become increasingly clear that oncolysis creates a strong immunoresponse that is crucial in the responses seen (Cerullo et al. 2012; Liikanen et al. 2013; L. 2014). Th us we also made many viruses armed with immunostimulatory molecules (Cerullo et al. 2012; Pesonen et al. 2012; Kanerva et al. 2013; Bramante et al. 2014). It seems these were the correct choices, as demonstrated by the many objective responses seen in more than half of the almost three hundred advanced cancer patients treated in the advanced therapy access program (ATAP, discussed in materials and methods). Although sipuleucel-T was approved for the treatment of hormone-refractory prostate cancer by the FDA in 2010 and shed light to the almost forgotten fi eld of immunotherapy, and Science magazine announced immunotherapy as the breaktrough of the year, the most astonishing results that probably will lead to changes in cancer treatment schemas in most cancers were yet to come. During 2013-2015, cancer and cancer immunotherapy meetings have shown a growing number of truly amazing reports of a broad spectra (SCCHN, NSCLC, gastric, renal, bladder etc.) of advanced tumors responding to treatments. aCTLA4, PD-1 and PD-1 ligand blocking antibodies, commonly known as the, checkpoint inhibitors, were fi rst accepted for the treatment of advanced melanoma but other indications have followed (Redman et al. 2015). Simultaneously, astonishing results have been reported in T-cell therapies: CD19 CAR T-cells in acute lymphatic leukemia (ALL) showed 90% complete response rates, while tumor infi ltrating lymphocytes (TILs) for the treatment of advanced melanoma has shown in multiple clinical trials in centers across the world durable clinical response rates near 50% or more (Turcotte and Rosenberg 2011; Wu et al. 2012). Sometimes the benefi cial results have been so evident that a temporary permit has been granted straight aft er an early phase trial, while typically drug development needs a positive phase III trial before sales permit can be applied (FDA press release, FDA approves Opdivo for advanced melanoma, Dec 22 2014). Also in 2015 FDA stopped two NSCLC phase III trials comparing docetaxel and nivolumab (FDA press release Feb 2015 and April 2015) during interim analysis for ethical reasons, and all patients continued with immunotherapy. Th e oncolytic fi eld has also had good news. A positive trial with an oncolytic virus named T-Vec was completed (Kaufman and Bines 2010) and now FDA voted 22-1 in favor of the approval making T-Vec probably the fi rst oncolytic agent approved in a western country. Th ese and other publications indicate the potency of immunotherapy. In contrary to most other treatments in advanced cancer, immunotherapy responses seem long lasting, and in the case of complete responses patients oft en seem to become cured (Figure 1). At the moment, the problem is that we simply still do not know who to treat, when to treat and how to treat. It seems that one of the most principal problems starts from the classifi cation system. Clinicians rely mostly on the information of how the tissue looks in the microscope, has it spread to neighboring tissues and where in the body according to imaging it is present. So far this information has been most critical in cancer care decision making. However, we must remember that cancer is a disease of the genome, and genes are too small to see in a microscope. Th us in the future we shall be more interested in what mutations or epigenetic alterations are

2 Introduction present in the cancer we are treating and/or how it protects itself from being destroyed by the immune system. When these questions can be addressed in a proper manner, cancer might be mastered. In 2007, it was far from clear that immunotherapy will change the fi eld of cancer treatments. To me this has become clear during my PhD studies. I believe that during the following years this will become evident to the whole medical community and to the patients suff ering from this vicious decease. In brief, more intelligent ways are needed to detect the cancers that need treatments, and better treatments are needed to the ones that need to be treated. Possible solutions are researched in this translational thesis.

Figure 1. Survival curves of typical advanced cancer patients. Long term survivors exist with immunotherapy, a phenomenon that is not oft en seen in advanced cancer patients with other treatment modalities (Hemminki 2015).

3 Review of the Literature REVIEW OF THE LITERATURE 1. Cancer 1.1 Classifi ca on of cancer Today it is well established that cancers arise as a result of the numerous alterations that have occurred in the DNA sequence of cells. However, for decades cancer has been classifi ed according to the site of origin, the grade and the stage of the disease. Typically, a biopsy from a suspected tumor is taken and the site of origin and grade is determined from histological analysis of the tissue. Th e site of origin describes the type of tissue in which the cancer cells begin to develop, for instance, adenocarcinoma (glandular tissue), carcinoma (epithelial tissue), leukemia (blood cells), lymphoma (lymph tissue), myeloma (bone marrow), sarcoma (connective tissue), blastoma (embryonic tissue) and so on. Th e grade of the disease is determined by examining the cells obtained through biopsy under a microscope. Increasing abnormality increases the grade; grade 1 cells have only slight abnormalities while grade 4 cells are immature and undiff erentiated. Th e stage of the disease, on the other hand, is determined typically according to imaging and/or the tissue removed in surgery. Th e most commonly used staging system is the tumor node metastases (TNM) system (Sobin 2001) implemented 60 years ago. It classifi es cancers by size (T0-T4), by amount of cancer positive lymph nodes (N0-N3) and by presence of distant metastases (M0-M1). Most of the common tumors have their own TNM classifi cation. Other parameters of the TNM system include, for instance, serum tumor markers (S0-S3) and completeness of the operation (R0-R2). Commonly a small “p” is present indicating that the stage is given by pathologic examination of a surgical specimen instead of clinical examination that is indicated by “c”. For example, a biopsy from a breast tumor could be taken and classifi ed as breast adenocarcinoma grade I. Aft er surgical resection and lymph node analysis it could further be classifi ed: small (T1), low-grade cancer (G1), no metastasis (M0), no spread to regional lymph nodes (N0), cancer completely removed (R0), and resection material seen by pathologist (p): pT1 pN0 M0 R0 G1. Th is grouping of TNM would be considered stage I. Generally stage I tumors can be cured by operation while higher, for example stage IV, are inoperable. Th is classifi cation system is still the basis of prognosis, although we know that cancer arises from mutations, which today could be sequenced relatively easily. However, at the moment little is known of linking mutations and prognosis. Some cell surface receptors or hormone receptors can be analyzed, and small molecules/monoclonal antibodies/hormone therapy (e.g. trastuzumab) can in some cases be used to target the cancer.

1.2 Cancer is a disease of the genome Sequence variants can be transmitted through the germ line or they can be acquired during life. Germline variants are present in every cell of an individual while acquired mutations are present only in the progeny of a mutated cell. It is approximated that acquired somatic mutations happen at a rate of 10x10-7 cell division in a more or less random fashion (Araten et al. 2005).Th e human species as well as other mammals have developed several mechanisms to protect agaist these mutations. Cells have ways to correct mutated genes, or, if this correction is not possible, they might self-destroy by apoptosis. Normal cells have also a maximum time they can multiply by mitosis. Aft er this so called hayfl inc limit, the cell will enter senescence and die. New replicating

4 Review of the Literature cells will then arise from the unhurriedly replicating stem cells. Finally, also the immune system is effi cient in destroying cells that are abnormal. For a cell to become cancerous, it must fi rst aquire specifi c mutations. Random mutations are seen all the time aft er fertilization, and the number of these mutations grow with time. Random mutations that are not necessary for malignancy are called passenger mutations, while mutations that give advantage for the cells to become tumor (see Table 1 Hallmarks of cancer) are called driver mutations. It is thought that the driver mutations are a prerequisite for malignant tumor growth, while passenger mutations are not. At the moment, it is believed that the great majority of cancers arise when two to eight sequential alterations (usually it takes 20-30 years) have occurred in genes with functions relevant to cancer. To date, about 140 of these driver genes are known (Vogelstein et al. 2013). However, when cancer has developed, the mutational rate grows exponentially and diff erent parts of the tumor become heterogenous also in regard to the DNA. Th us, discovering the critical mutations in a patient and targeting these by drugs is the major objective for personalized medicine. Interestingly, these ca. 140 driver mutation genes function through a dozen signaling pathways (e.g. RAS, MAPK, PI3K) that regulate three core cellular processes: cell fate determination, cell survival and genome maintenance (Vogelstein et al. 2013). Every individual tumor (even from the same histopathology subtype) is distinct in respect to its genetic alterations, but the pathways aff ected are similar. Also genetic heterogeneity among tumor cells of the same patient exist aff ecting treatment results (Vogelstein et al. 2013). Of the cancer genes known to date, approximately 90% have somatic mutations, 20% show germline mutations and 10% both (Futreal et al. 2004). It is estimated that 5-10% of all cancers are inherited, due to high-penetrant germline mutations that cause rare inherited cancer syndromes. Another 15-20% of cancers are known as “familial”, which can be defi ned as cancers clustering in a family more frequently than expected (Nagy et al. 2004; Hemminki et al. 2008). Typical to germline mutations are that patients have many cancer incidents in the family and individuals can have even many diff erent cancers during their life time. Another way of classifying cancer genes is by function. Genes that give growth advantage when mutated are called oncogenes, while other genes control normal cell growth and are called tumor suppressor genes. A mutation in a tumor suppressor gene leads to a loss of function and more uncontrolled behavior of the cell. In 2000, Hanahan and Weinberg published a paper with the title Hallmarks of cancer, which has become the journal Cell’s most sited article (Hanahan and Weinberg 2000). Th e authors suggested that the complexity of cancer can be reduced to a small number of underlying principles, the six hallmarks that a normal tissue must accomplish to become malignant. In 2011, four new hallmarks were suggested. Th ese hallmarks (Table 1) can be imagined as targets when designing new cancer treatments. Oncolytic viruses use many of these hallmarks in their mechanism of action (delta24 deletion/defective p53, hTERT promoter/telomerase activity, oncolysis produces a danger signal for immune system/GM-CSF stimulates this further). Also many of these hallmarks are linked to the immune system, suggesting immunotherapy and oncolytic viruses as potential treatment modalities. To summarize, for a cell to become cancer is not an easy task, but a long evolutionary journey.

5 Review of the Literature

Table 1

Hallmarks of cancer Suggested treatment Avoiding immune destruction Checkpoint inhibitors, T-cell therapy, oncolytic viruses, GM-CSF Enabling replicative immortality Inhibition of telomerase, Myc, Stat3, NF-kappaB, Akt, IL-6 Tumor promoting infl ammation Selective anti-infl ammatory drugs Activation of invasion & metastasis Inhibition of Stat3, NF-kappaB, IL-6, Src Inducing angiogenesis Inhibition of VEGF Genome instability & mutation Upregulation of p53 Resisting cell death Inhibition of Akt, NF-kappaB, Stat3, Bcl2 Deregulating cellular energetics Inhibition of HIF1alpha Sustaining proliferative signaling Inhibition of Myc, Src, Akt, EGF, FGFbeeta, AR, ERalpha, Stat3, Her2/neu Evading growth suppressors Upregulation of p53 Modifi ed from (Hanahan and Weinberg 2000; Hanahan and Weinberg 2011)

6 Review of the Literature 2. Adenoviruses 2.1 Adenovirus structure Adenovirus (Ad) is a non-enveloped, double-stranded DNA virus that is ca. 90 nm in diameter. Viral DNA and associated core proteins are enclosed in an icosahedral capsid, with 20 triangular faces. Th e capsid consists of 240 hexon- and 12 penton proteins. From each of the 12 fi ve- fold capsid vertices projects an elongated fi ber. At its proximal end, the fi ber is bound to the pentameric penton base and from its distal end it forms a globular knob domain (Figure 2). As a general rule, the fi ber knob functions as the major attachment site for cellular receptors, while the penton base is involved in secondary interactions that are required for virus entry into the cell as discussed later.

Figure 2. On the left electron microscope picture of our Ad3-hTERT-E1A virus. Th e appearance is similar to the wild type adenovirus serotype 3. In contrast to serotype 5, it has short straight and rigid fi ber shaft s. On the right, a model of adenovirus showing icosahedral capsid with 20 triangular faces, 12 vertices and 30 edges. Th e main proteins of the capsid are: A=fi ber (knob), B=fi ber (shaft ), C=penton, D=hexon.

2.2 Adenovirus classifi ca on Adenoviridae are a large group of viruses present in many species. Th ey can be divided into fi ve genera: Mastadenovirus (all human adenoviruses and some others), Aviadenovirus (various bird adenoviruses), Atadenovirus (several other vertebrates), Siadenovirus and Ictadenovirus. Classically, 51 human Ad serotypes have been identifi ed. Th is is, however, under debate as for example Field´s Virology (2007) claims that the number is 57. Th e reason for this diff erence is discussed below. Th e gold standard of serotyping has been the virus neutralization test. Th ese serotypes are divided into six subgroups (A to F) based on hemagglutination properties, DNA homology, oncogenicity in rodents and genomic organization (Lukashok and Horwitz 1998);

7 Review of the Literature subgroup B is further divided into B1 and B2 because of obvious striking diff erences in restriction fragment patterns and, in part, tissue tropism.

Classical serotyping Parts of the three capsid proteins (hexon, penton and fi ber) can be used as diagnostically useful antigens. Th e main type-specifi c epitope consists of the loop1 and loop2 of the hexon protein and reacts with serotype specifi c antisera in neutralization tests. Failing theneutralization test (NT) is a common but unresolved issue, and in the case of emerging new serotypes or mutations in old serotypes, it is useless. Th e gamma-determinant epitope of the fi ber knob region has hemmaglutination properties that can be used for hemmaglutination inhibition tests (HI). It is preferred in many laboratories because it is faster and more convenient, but as a downside it cannot diff erentiate all human serotypes because of cross-reactions. Human adenoviruses have a capability of intraspecies and also occasional interspecies recombination, and thus contradictory results in NT and HI tests can be found from clinical isolates.

Future in serotyping To address problems discussed above and to classify emerging serotypes, new methods are needed. As sequencing is becoming more and more aff ordable, the future of adenovirus classifi cation will move from “serotyping” to “typing”, meaning that the classical neutralization test will be replaced with sequencing based methods. Serotyping can already be done by sequencing only the loop2 region of the hexon, whereas additional sequencing data of the hexon loop1 and the gamma-determinant sequence of the fi ber knob should be generated for the identifi cation of new prototype isolates (Madisch et al. 2005).

Subgroups Division of adenovirus serotypes to subgroups has been made through biological capabilities. Th is has been problematic and, for example, group B was further divided into B1 and B2. Roughly, viruses from subgroup B2 can fully inhibit the binding of subgroup B1 viruses; however, viruses from subgroup B1 can only partially inhibit the binding of subgroup B2 viruses (Segerman et al. 2003). A few years later this division was questioned (Madisch et al. 2005). By analyzing the sequence of the hexon, it was found that Ad3 and Ad7 are closely related, but other members of the B1 subgroup clustered with B2 viruses. When looking into the sequences of the fi ber knob (HVR1-HVR6), Madisch and colleagues discovered that Ad3 and Ad7 diff er mainly in HVR5, while other group B viruses (14, 34, 50, 11, 21, 35) have diff erences in other HVRs. Th ese analyses suggest that Ad3 and Ad7 are close relatives in regard of the surface proteins. In addition, phylogenic relations of human Ad group B fi ber knob sequences did not coincide with the hexon sequences, indicating intra subspecies and intersubspecies recombination in the molecular phylogeny of species B. Th e close phylogenetic relationship in the species B assumed by HI cross-reactions between Ad7, Ad11 and Ad14, and between Ad34 and Ad35 was on the other hand confi rmed. An imperfect correlation between tissue tropism and subgroup can be found; for example, viruses in groups B1, C, and E cause mainly respiratory infections; group B2 viruses infect preferably the kidney and urinary tract; group F viruses are known to cause gastroenteritis; and several group D serotypes are associated with epidemic keratoconjunctivitis.

8 Review of the Literature

While most oncolytic adenoviruses are based on serotype 5, we wanted to generate viruses based on a serotype that’s primary receptors are highly expressed in advanced cancer. Group B viruses seem to be such viruses. In the literature group B viruses have been divided historically to B1 and B2. Now also another division (I, II and III) according to the primary receptor has been proposed (Wang et al. 2011). Interestingly, BII and BIII viruses seem to produce also dodecahedral particles (PtDds) with twelve fl at faces.. Th ese “smaller empty capsids” are produced multiple times more than the virus. It has been proposed that they have an important function in opening tight junctions, enabling virus entry and improved access to receptors such as CD46 and Her2/neu. Taken together, emerging data suggests changes in the classical classifi cation. For a long time the virus neutralization test (VN) has served as the gold standard method for typing new adenoviruses. However, this method requires successful isolation and propagation of the virus, as well as the availability of a full panel of hyper-immune sera against every known (approved) serotype (de Jong et al. 2008). In the future, sequencing shall be used more for the classifi cation of adenoviruses. Adenovirus serotypes are presented in table 2.

Table 2. Adenovirus serotypes

Group Adenovirus serotype Main primary Expression of the PtDds receptor receptor in advanced cancers A 12, 18, 31 CAR Low B 3, 7, 11, 14, 16, 21, 34, 35, CD46 / DSG-2 High 50, 55 B1* 3, 7, 16, 21, 50 High B2* 11, 14, 34, 35 High I* 16, 21, 35, 50 CD46 High II* 3, 7p, 14 DSG-2 (& CD46?) High Yes III* 11p CD46 (& DSG-2) High Yes C 1, 2, 5, 6, 57 CAR Low D 8, 9, 10, 13, 15, 17, 19, 20, CAR, CD46 Low * 22, 23, 24, 25, 26, 27, 28, 29, 30, 32, 33, 36, 37, 38, 39, 42, 43, 44, 45, 46, 47, 48, 49, 51, 53, 54, 56 E 4 CAR Low F 40, 41 CAR Low G52 * Group B viruses can be further divedid to B1 and B2 or to I, II and III. BII and BIII viruses produce dodecahedral particles (PtDds) proposed to have an important function in opening tight junctions, enabling virus entry and improved access to receptors such as CD46 and Her2/neu (Wang et al. 2011).

9 Review of the Literature 2.3 Adenovirus receptors In a standard virology textbook, in 1996, it was written that “the identity of the cellular adenovirus receptor remains a mystery” (Shenk 1996). Since then, multiple adenovirus receptors have been identifi ed. To our knowledge, until 2013 all rationally designed oncolytic adenovirus clinical trials have been based on serotype 5 that uses coxsackie-adenovirus receptor (CAR) as a primary binding receptor. In 2013, a phase I trial with Oncos-102 was completed (www.clinicaltrials.org). Th is virus is a serotype 5 oncolytic adenovirus with a fi ber knob from the adenovirus serotype 3. Adenovirus entry into cells, as defi ned by experiments with cultured cells, involves attachment to a primary receptor, followed by interaction with a secondary receptor responsible for internalization. It is thought that the major function of the primary receptor is to hold the virion close to the cell surface, permitting interaction with the secondary receptor. As discussed below, adenoviruses can be quite fl exible in their use of primary receptors. When we started working with the adenovirus serotype 3, the sequence of this virus had been published recently in 2005 by two groups. Th e primary receptor was not known at this point, but reports indicated that it was abundantly present in cancer (Tuve et al. 2006). Among others, CD46 was proposed as the primary receptor, but in 2011 a publication in Nature Medicine suggested desmoglein-2 (DSG-2) as the primary high-affi nity receptor (Wang et al. 2011).

Primary receptors Adenoviruses are known for their ability to use diff erent receptors. If the high avidity receptor is not present, lower avidity receptors can be used. To keep the long story short: Most adenoviruses use CAR as the primary receptor, but adenoviruses from group B do not. Group B binds generally to CD46, except for Ad3, Ad7p and Ad14 that mostly bind to Desmoglein-2. Group B viruses are interesting when it comes to cancer therapy as cancer cells seem to be rich in the receptors they bind to (Tuve et al. 2006). CAR (coxsackie-adenovirus receptor) is the most studied adenovirus receptor. It is a 46-kDa protein that belongs to an immunoglobulin superfamily and has two immunoglobulin- like extracellular domains. CAR is a cell adhesion molecule found mainly in the tight junction of polarized epithelial cells (Cohen et al. 2001). CAR’s distribution in human tissue is not well defi ned, but its mRNA is present in a number of organs (heart, brain, pancreas, intestine (Tomko et al. 1997), lung, liver, and kidney (Fechner et al. 1999). Some adenoviruses (Ad12, Ad2, Ad5, Ad4, Ad41, Ad31, Ad9, Ad19) from all other subgroups (A, C, D, E, F) than subgroup B have been shown to bind to CAR (Roelvink et al. 1998). CAR binds with a site on the outer surface of the trimeric fi ber knob (Roelvink et al. 1999). Simultanious binding to CAR and to the secondary receptor (integrin) imposes geometric constraints on receptor interactions. Th us CAR-binding viruses require fi bers that are both fl exible (Wu et al. 2003) and long (Shayakhmetov and Lieber 2000). CD46 is present in all nucleated cells. Its main function is to protect healthy (noninfected) cells from complement mediated degradation. Th us overexpression of CD46 is also a way for cancer to resist complement activation (Yan et al. 2008). Before any specifi c receptors had been identifi ed, it was observed that Ad3 (group B), did not compete for attachment with Adenoviruses 2 or 5 (both group C), suggesting that it bound to a diff erent receptor. Consistent with this was that none of the group B adenoviruses interacted with CAR. During recent years it has been discovered that some group B viruses bind to CD46 (high or low affi nity) while others bind to DSG-2 (see below) and some to both. Also some group D viruses bind to CD46 with their short

10 Review of the Literature shaft . Interestingly, CD46 functions also as a receptor for a number of other pathogens, including measles virus, Streptococcus pyogenes, human herpes virus 6 and pathogenic Neisseria spp. DSG-2 (desmogein-2) is also a cell adhesion molecule, but it belongs to the cadherin superfamily. Of the group B adenoviruses, Ad3, Ad7, Ad11 and Ad14 use DSG-2 as the main cellular receptor as shown by in vitro experiments. Th ese serotypes represent key human pathogens causing respiratory and urinary tract infections (Wang et al. 2011). It is known that during virus replication, CAR and CD46 binding viruses secrete fi ber proteins in addition to viruses. Unique to DSG-2 binding viruses, however, is their ability to produce and secrete dodecahedral particles consisting of penton base and fi ber proteins but no DNA (Wang et al. 2011). Th ese particles are smaller than the virus and seem to be able to drive the epithelial cells to a mesenchymal state (epithelial-to-mesenchymal transition, EMT). Importantly, EMT mediated by Ad3 results in a mesenchymal phenotype that is more permissive to adenovirus. EMT due to Ad3 dodecahedra has also been shown to sensitize tumors in vitro and in vivo to trastuzumab (Herceptin, targets Her2/neu) and cetuximab (Erbitux). Th e mechanism seems to be the opening of the tight junctions (Wang et al. 2011; Lu et al. 2013). CD80 and CD86 are distantly related immunoglobulin superfamily members that are expressed on antigen-presenting cells and are best known for their important function in T-cell activation. Group B adenoviruses, including Ad3 (or virus with a knob from Ad3), also enters cells aft er binding to either CD80 or CD86. Ad3-mediated transduction of isolated dendritic cells depends on interaction with CD80 and CD86. Viruses that can target receptors expressed on dendritic cells (including CD80/86 and CD46) are interesting in an immunotherapy view point (Short et al. 2004; Short et al. 2006). Th e CD80/CD86 is known to bind to T-cell receptor CTLA-4 with a high specifi city and to CD28 antigen with low specifi city. Th e interaction of CD28 with CD80/CD86 provides a co-stimulatory signal to T-cells, while interaction with CTLA-4 seems to induce peripheral tolerance (Short et al. 2004; Short et al. 2006). Adenovirus has been reported to interact with many other host receptors/molecules not mentioned above. However, with many of the following the data are inconclusive. Several subgroup D viruses (Adenoviruses 37, 8, and 19a) have been shown to infect cells aft er attachment to sialic acid, a common carbohydrate component of glycoproteins and glycolipids (Zhang and Bergelson 2005). It seems that sialic acid binds to a site at the very top of the fi ber knob in most (if not all) group D viruses. eTh alpha-2 domain of the class I major histocompatibility complex (MHC-I) has been reported to promote high-affi nity interaction with Ad5 when expressed on an MHC-defi cient human cell line (Hong et al. 1997), but the same was not observed when hamster cells were used and thus the role of this protein in adenovirus infection remains unclear. Adenovirus serotype 5 has also been shown to attach to vascular cell adhesion molecule 1 (VCAM-1) that is expressed on activated endothelial cells (Chu et al. 2001). VCAM-1 is more highly expressed on atherosclerotic endothelium compared to normal endothelium. Th us it has been suggested that VCAM-mediated infection may be useful for gene therapy of atherosclerosis. Heparan sulfate glycosaminoglycans (HS-GAGs) are heavily sulfated long chains of carbohydrates. Th ey are abundant on the outer surface of cells, within the extracellular matrix and the cellular glycocalyx. Th e glycosaminoglycan is oft en bound to a protein core, forming a proteoglycan. HS-GAGs mediate CAR-independent attachment and infection by some adenoviruses, for example, Adenoviruses 2 and 5. It seems that basic amino acid motifs permit protein recognition of HS-GAG. Perhaps the best known is the KKTK motif within the proximal fi ber shaft . Modifi cation of this area signifi cantly changes Ad5 tropism in vivo (Smith et al. 2003). Dipalmitoyl phosphatidylcholine (DPPC) is known to interact with

11 Review of the Literature

Ad5 hexon. It is a component of pulmonary surfactant. DPPC liposomes enhance virus uptake by a receptor-independent mechanism. It is not known whether interaction with surfactant plays a role in adenovirus infection in vivo (Balakireva et al. 2003).

Secondary receptors Many adenovirus serotypes display an Arg-Gly-Asp (RGD) peptide within the penton base capsid protein. Th is functions as a recognition site for several cellular integrins, members of a large family of heterodimeric adhesion receptors. Integrins consist of alpha and beta units, and they transduce signals to the cells. Th us far 18 alpha and 8 beta integrin subunits have been characterized in humans. Th eir distribution in humans is described in detail by Beaulieu (Beaulieu 1999). Several integrins have been demonstrated to promote adenovirus entry in vivo. Virus attachment may also depend on direct interaction between the penton base and a cell surface integrin, without the need for a primary fi ber receptor. Th e interaction is largely mediated by the RGD motif (Arnberg 2012). Th e engagement of the penton base by integrins induces signals that lead to important rearrangements in the actin cytoskeleton and initiation of virus internalization. Interaction with integrins is also important for virus escape from the endosome. Interestingly, mutation of the penton base RGD sequence slows, but does not prevent, virus internalization and infection. It is unclear whether this suggests that entry occurs by integrin independent routes or whether some virus interactions with integrins are RGD independent. Internalization of RGD-defi cient virus is more rapid in cells that express high levels of fi ber receptor, suggesting that recruitment of multiple fi ber receptors may compensate for the loss of penton-integrin interaction (Zhang and Bergelson 2005).

Liver tropism One problem in the intravenous adenovirus delivery is that a large proportion of the virus ends up in the liver (Arnberg 2012; Koski et al. 2013). Th is has been most evident with serotype 5 viruses in mice. One of the major crevasses with this approach is that mice cells are nonpermissive to human adenovirus. Th us it is rather logical that fi nally the liver clears the virus. Th e mechanism by which hepatocytes are transduced have been studied in great detail, and it seems that coagulation factors, heparin sulfate-containing cell membrane glycoproteins and virus hexon/fi ber shaft proteins seem to be the major components of interaction. In blood circulation the virus seems to bind partly to complement receptor 1, sialic acid and CAR on erythrocytes (Arnberg 2012). It is not known whether these are ways for the virus to avoid clearance or to hide from the immune system. While promising effi cacy has been reported with oncolytic viruses with intratumoral injections, the effi cacy with intravenous approaches has so far not been optimal (Koski et al. 2013). However, humans seems to tolerate large doses of intravenous virus and clinically relevant liver problems are rare (Hemminki 2014).

12 Review of the Literature 2.4 Adenovirus replica on As described above, entry into the cell involves typically two sets of interactions between the virus and the host cell – fi ber binding to the primary receptor and secondary binding of the penton base to the cell integrins. Th e virus is then engulfed by the cell in a clathrin-coated vesicle and is transported to endosomes. Here acidifi cation results in partial disassembly of the capsid and the altered virion escapes into the cytoplasm. It is then transported with the help of microtubules to the nuclear pore complex, whereby the adenovirus particle disassembles. Viral dsDNA (35 000-37 000 bp depending of the serotype) is released, and it enters the nucleus via the nuclear pore. Subsequently, the DNA associates with histone molecules enabling viral gene expression (Meier and Greber 2004; Wolfrum and Greber 2013). Th e adenovirus life cycle is divided by the DNA replication process into two phases: the early and the late phase. Th e early genes are responsible for expressing mainly non-structural, regulatory proteins that prepare the cell for virus production. Th e early genes have three major goals: 1) activating the expression of host proteins necessary for DNA synthesis (driving the cell to S1 simulating phase), 2) activating other virus genes, such as the virus-encoded DNA polymerase, and 3) preventing the infected cell to undergo premature death due to host immune defenses (blockage of interferon and apoptotic activity, and blockage of MHC class I translocation and expression). DNA replication separates the early and late phases. A terminal protein is fi rst attached to the 5´end of the adenovirus genome; this acts as a primer for replication. Th e viral DNA polymerase then uses a strand displacement mechanism to replicate the genome. Th e late phase proteins are focused on producing adequate numbers of structural proteins to pack the produced DNA. Finally, virus DNA is assembled into the protein shells and released from the cells as a result of cell lysis. One cell can produce thousands of viruses that continue to infect when the cell is lysed. Th e typical replication cycle of many adenoviruses takes 24-48 hours (Jogler et al. 2006).

13 Review of the Literature 3. Cancer treatment, History and Future 3.1 Current standard treatment of cancer For decades the treatment of cancer has been divided into surgery, radiotherapy and medical treatment (Figure 3).

Surgery Radiation Medical

Open External Chemo

Scopic Brachy Hormonal

Radioisoto Targeted / Robotic pe Immuno

Figure 3. Treatment options for cancer. Th is thesis concentrates on targeted therapy on immunological oncolytic adenovirus treatments.

While all of these have advanced during the years, the survival for advanced cancer has not changed considerably. Still today aggressive cancer that cannot be surgically removed entirely or effi ciently radiated leads oft en to death. At the same time, general awareness of some cancers (e.g. prostate) leads to diagnosis of less aggressive forms of cancers that might not reduce life- expectancy (Lin et al. 2008). According to autopsy data, most elderly men have malignant tissue in their prostate, although no clinical manifestations were present. Generally, medical treatments, such as diff erent types of chemotherapies and hormonal treatments, can slow down the progression of the disease. However, due to diff erent subtypes of cancer and constant mutations, cancers tend to relapse. In addition, adverse events related to treatments tend to be severe. Th us, there is a great need for better therapies.

Surgery One of the oldest known descriptions for surgical treatment of cancer dates back to approximately 1600 BC in Egypt, where ulcers of breast were treated by cauterization with a toll called the “fi re drill”. Although small surface tumors were treated much like today, great advancements have been made. One of the driving forces was the invention of anesthesia in 1846. Aft er this, surgeons could operate entire tumors together with lymph nodes (Sudhakar 2009). Open surgery is oft en regarded as the gold standard for cancer management. Laparoscopic surgery was fi rst performed in 1901 to the abdominal cavity of pregnant women. Aft er this, laparoscopic procedures were performed mainly for diagnostic purposes. In

14 Review of the Literature the 1980´s, laparoscopic cholecystectomy in humans was performed following fast advancement with several operative procedures in the 1990´s. While some operations might be easier to perform laparoscopically, the main benefi t is the minimally invasive technique off ering minimal pain, low adhesion rate, low hernia rate, short stay at the hospital and quicker return to activities. Today laparoscopic surgery is commonly used in most hospitals. Robotic surgery has advanced during the last years. It off ers some advantages such as higher magnifi cation, higher precision, tremor fi ltration and ergonomics. It also off ers access to places “around the corner”, places that are hard to access with conventional surgery, for instance removal of the prostate. Th us, today it is used especially in urology and gynecology. Problems with robotic surgery include high price and long time for surgery preparation. One downside of robotic surgery is also loosing the “tissue feeling”.

RadiaƟ on therapy Radiation therapy is the use of ionizing radiation to kill cancer cells and shrink tumors. Radiotherapy can be curative if the tumor is located in one place of the body. It can be used aft er surgery as adjuvant therapy to prevent tumor recurrence (e.g. aft er resection of small breast cancer). Synergy with chemotherapy or other medical treatments can be acquired, and thus it is oft en used before, during and aft er chemotherapy in susceptible cancers. In some cases, radiation therapy can be used as palliative treatment to alleviate pain or to aim for local disease control. Th e mechanism of action is based on the ability of ionizing radiation to damage DNA leading to cell death. Th e damage can be direct or indirect. Indirect ionization happens as a result of the ionization of water, forming free radicals, notably hydroxyl radicals, which then damage the DNA. Direct DNA damage can be caused by photons or by charged particles (protons or carbon ions). In photon therapy, most of the eff ect is indirect though free radicals, while the eff ect of charged particles is more direct. In normal cells, damaged DNA is typically repaired, but in cancer these mechanisms are generally mutated, leading to accumulating damage in the DNA and decreased survival of these and the progeny of these cells. While some cancer types are regarded resistant to radiation (e.g. melanoma and kidney cancer), most are regarded susceptible. To spare normal tissue, radiation is today aimed from several angels intersecting at the tumor. In this way the tumor receives much larger absorbed doses than the surrounding tissue. Th e radiation fi elds can also include the draining lymph nodes. Patients are typically tattoo marked and lasers are used to confi rm the exact position of the patient. Sometimes radiation is synchronized with breathing to avoid damage to the surrounding tissues. Typically, fractionation is used in a radiation treatment plan. Here radiation is performed so that patients receive small amounts, 1.8-2 gray (Gy) of radiation for example every weekday for some weeks. In this way, normal cells have time to recover and also cancer cells that were in a resistant phase (cell cycle, hypoxia etc.) at some time point could be aff ected. Total doses in photon therapy vary, but typically ca. 60-80 Gy is used for solid tumors in curative plans and 45-60 Gy in adjuvant therapy. In brachytherapy radiation the source, typically emitting charged particles, is placed inside or next to the area requiring treatment. Th is way a high dose can be locally administered minimizing harm to surrounding tissues. Brachytherapy is commonly used in prostate, cervical, breast and skin cancer. A course of brachytherapy can normally be completed in less time than other radiation techniques.

15 Review of the Literature Medical treatments Some decades ago the medical treatment of cancer was a synonym with chemotherapy. Nowadays, however, the medical treatment can be divided into chemotherapy, hormonal therapy and targeted/immunological therapy, as shown in Figure 4. Chemotherapy has been widely used since the 1940s and works by killing cells that divide rapidly. While this is a known property of cancer cells, also normal rapidly dividing cells, such as cells in the bone marrow, digestive tract and hair follicles, suff er from collateral damage. Th is leads to the common and sometimes lethal side-eff ects: myelosupression (low blood cells), immunosuppression, mucositis (infl ammation of the digestive tract mucosa) and alopesia (hair loss). Chemotherapy can be used as single-agent chemotherapy (one drug at a time) or as combination therapy (several drugs at once). It can be combined with radiation (chemo radiotherapy), or sometimes light is used to convert the inactive agent to active chemotherapeutic (photo chemotherapy or photodynamic therapy). Hormonal therapy, fi rst in the form of androgen deprivation for treatment of prostate cancer, started raising interest as treatment for cancer as early as in the 1940s (Crawford 2004). At the moment, new targeted treatment modalities are entering the clinics. Th ere are over ten monoclonal antibodies that can be used to treat diff erent cancers. Also various immunotherapy based treatment modalities have been introduced. Immunotherapy has resulted in advances not seen in decades. Th ese will be discussed in chapters 3.5 - 3.7 in more detail.

3.2 History of cancer treatments with viruses Th e end of the 19th century is traditionally considered as the beginning of modern medicine. Also in cancer therapy, new treatment modalities were introduced, prior to which predominantly the only cancer therapy was excision of the tumor by surgery. Anesthesia was becoming accessible, but surgery was still very rudimentary. Tumors that were superfi cial and easy to operate could be removed, but relapses were common. As primitive chemotherapy, castrol oil and arsenic was used. In 1895, radiotherapy was discovered and quickly adapted to the treatment of cancer. In spite of these advances, cancer was rarely cured. However, already at this time it was occasionally observed that cancer patients who contacted an infectious disease went to brief periods of clinical remission. William Coley, a surgical oncologist, developed a mixture of bacteria in the late 19th century as a treatment for cancer (Coley 1891; Nauts et al. 1946). Some years later viruses were “discovered” as some agents that could pass through fi lters that bacteria could not pass (Kelly and Russell 2007). Plaque assay was discovered in the 1915 giving some hints of the nature of viruses. Since the 1920s, viruses have been used for oncolysis (Sinkovics and Horvath 1993). Th e fi rst reports of electron microscopy are from 1939, initiating a period of better understanding. Ten years later, cell culture as a propagation method of viruses was becoming possible and many viruses were shown to have an eff ect on tumors in rodent models. Aft er this, viruses in general and adenovirus in particular have been studied with an excessive intensity, resulting in their biology being now understood more thoroughly than most other organisms in nature (Hemminki 2014). During the last 150 years there has been a steady fl ow of case reports describing tumor regression of patients aft er natural virus infections (e.g. Varicella, measles, infl uenza, hepatitis, glandular fever). Reports describe rare but sometimes dramatic responses in cancer patients recovering from viral diseases (Kelly and Russell 2007; Eager and Nemunaitis 2011). Most remarks

16 Review of the Literature are from patients suff ering of lymphoma or leukemia, known to be associated with substantial immune suppression. Until recently, these were also the easiest tumors to be diagnosed and basically the only tumors where treatment eff ects could be measured. Th us these are probably artifi cially overrepresented in historical case reports. Typically, these responding patients have been young, and the remissions have been short-lived, lasting for one or two months (Toth and Wold 2010). Th ese observations can be interpreted so that in the right conditions certain viruses can destroy tumors without causing harm to the patient, but in advanced disease resistance can develop quickly. On the basis of clinical observations, several viruses with low pathogenicity to normal tissue and high oncolytic capacity have been selected for clinical investigation (Eager and Nemunaitis 2011). First attempts to treat cancer patients with viruses date back a hundred years. At this time, the virus was collected from heterogeneous and impure specimens from diff erent plants, humans or animals. During the fi rst half of the twentieth century little was known of the biological nature of viruses, but still many viruses were used to treat cancer patients. Success and side eff ects of the treatment varied, and the purity, quantity and scientifi c standards were diff erent from today. Interest in the fi eld has oscillated, reaching an early peak in the 1950s and 1960s when clinical testing became more common and many diff erent wild type viruses (e.g. Hepatitis, Epstein-Barr, West Nile, Uganda, dengue, yellow fewer) were used to treat diff erent cancers. Intratumoral virus replication was oft en confi rmed, but clear curative or survival increasing responses were rare. Aft er hundreds of patient series on diff erent cancers (e.g. with Egypt 101 virus, a type of West Nile virus, more than 150 “trials” against diff erent cancer types (Southam and Moore 1952)) with diff erent wild type viruses, it was becoming clear that most wild type viruses lacked effi cacy or safety. Some of the most promising results and acceptable side-eff ects already at this time were associated with adenoidal-pharyngeal-conjunctival (APC) virus (Huebner et al. 1955; Huebner et al. 1956; Georgiades et al. 1959; Zielinski and Jordan 1969), which is nowadays known as the adenovirus. In 1956, for example, 30 women with advanced epidermoid carcinoma of the cervix were treated with adenovirus. Th e virus was administered intra-arterial, intravenous or by intratumoral routes. Within ten days in two-thirds of the cases necrosis was observed, and most remarkably, it appeared to be restricted to the cancerous tissue. No safety problems were reported in the use of this wild type virus. In the 1970s and 1980s, the regulatory aspects of clinical trials with living pathogens became stricter. More importantly, the arsenal of drugs based on chemistry was expanding exponentially resulting in a belief that cancer would soon be cured (Hemminki 2015). Th us, there seemed to be little need for diffi cult-to-handle and tricky-to- produce substances, which might be one of the reasons why the fi eld was nearly abandoned. For the past two decades, oncolytic viruses have again gained increasing interest as advances in molecular biology, tumor biology, immunology and virology seem to suggest that oncolytic viruses might be an unexploited resource in cancer therapy. Th e fi rst marketing approval for an oncolytic virus in the western countries is likely to be granted in late 2015, as FDA has voted 22-1 in favor of approving T-Vec (discussed later in chapter 3.5).

17 Review of the Literature 3.3 Gene modifi ed adenoviruses in cancer therapy

Classical hypothesis Emerging hypothesis

18 Review of the Literature

Figure 4. Left panel. Classical hypothesis of the function of oncolytic virus in patients. Oncolytic viruses reach the tumor by direct injection or blood stream. Th ey infect the tumor cells and start rep- licating. Oncolysis of a tumor cell causes millions of new virions to be released. New virions spread to neighboring tumor cells and to distant metastasis via the blood stream. Eventually all tumor has been destroyed. Right panel. Emerging hypothesis of the function of immunostimulatory oncolytic virus in patients. Oncolytic viruses reach the tumor by direct injection or blood stream. Th ey infect the tu- mor cells and start replicating. Oncolysis of the tumor cells cause infl ammation that activates Antigen Presenting Cells (APC) to phagocytize lysed tumor cell remnants. Th e APCs present tumor associated antigens to Cytotoxic T Lymphocytes (CTL) and T Helper (TH) cells. All tumor cells are not lysed by the Adenoviruses as the replication and spread is inhibited by the immune system and other mechanisms. APC stimulated TH cells activate B-cells, macrophages and CTLs. Activated B-cells produce antibodies against the tumor while macrophages and CTLs have a direct anti-tumor activity. During this complex process multiple immunomodulatory -cytokines, -co-receptors and -cells are also crucial (not shown for clarity). Aft er tumor destruction, T-memory and B-memory cells are generated as a sign of anti-tumor immunity (Hemminki 2014).

Th ere are several advantages with adenoviruses in cancer therapy: 1) Adenoviruses are found naturally oncolytic to some degree. 2) Safety: even the wild type replicating adenovirus is found safe in several human trials. More than 400 modern time clinical adenovirus trials have been performed (Ginn et al. 2013). During the years, large doses of wild type or modifi ed adenoviruses have been administered to patients or volunteers, including multiple injection strategies, suggesting safety. Likewise, with mounting clinical experience, virus shedding or latent infections have not been a concern (Hemminki 2012). 3) Adenoviruses have been examined in great detail. Th eir replication and eff ect to host cells, as well as the eff ect to host immune system, has been reported in thousands of publications. Adenovirus has been used as a model organism to study DNA replication, apoptosis, oncogenic transformation and mRNA processing. Th e biology of natural adenovirus infections in humans are also reasonably well understood (Toth and Wold 2010). 4) Adenovirus can be easily produced in large quantities. 5) Th e genome is easy to manipulate and it is genetically stable.

First generaƟ on, replicaƟ on defi cient adenoviruses Th e fi rst application of adenoviruses in cancer gene therapy was their use as replication defi cient vectors (Toth et al. 2010) to transfer specifi c transgenes to the malignant cells. Th is was found to be one of the most effi cient ways to express transgenes compared to any other viral or non- viral-vector. Adenovirus could transduce a wide variety of cells, both replicating (such as cancer cells) and nonreplicating (e.g. normal tissues and cancer stem cells). Transgene expression was, however, found transient and oft en insuffi cient to generate a signifi cant therapeutic eff ect (Toth et al. 2010). Th us, there was a need for more potent vectors to treat cancer.

Second generaƟ on, oncolyƟ c adenoviruses Once tumor transduction was recognized as the critical obstacle to cancer gene therapy, the oncolytic platform gained in popularity due to its ability for intratumoral amplifi cation and

19 Review of the Literature penetration. For safety reasons, oncolytic viruses are modifi ed (or biologically selected) so that they preferentially replicate and lyse cancer cells (Toth et al. 2010). In theory, the virus would replicate and lyse the tumor cells and the progeny virions would fi nd new cancer cell hosts until no tumor is left (see fi gure 4, left panel). Th ere are several ways of making a virus oncolytic: 1) Delete parts of the viral genome (e.g. 24bp deletion in the E1A) that are responsible for freeing E2F. For virus replication in normal cells, the E2F is needed to drive the cell to S-phase. Th is is not necessary in tumor cells defective in the Rb/p16 pathway, including most if not all tumors. 2) Replace essential virus promoters with promoters that are active mainly in tumor cells (e.g. hTERT, E2F1). 3) Remove viral genes that prevent premature apoptosis of infected cells, as these genes are already commonly inactive in cancer cells. Oncolytic adenoviruses show generally effi cient cell killing in vitro and good potency in vivo. In cancer patients they have shown some signs of effi cacy but complete cures are rare (Hemminki et al. 2012).

Third generaƟ on, oncolyƟ c and armed adenoviruses As the tolerability and safety of oncolytic adenoviruses appears good compared to traditional cancer treatments in advanced cancer, but the effi cacy has left room for improvement, a need for more eff ective viruses has become evident. Globally, scientists have produced multiple approaches in arming the viruses with anti-cancer transgenes. Most frequently these transgenes have been inserted either in E1 or in a deleted E3 region of the virus. An appealing variant of the latter strategy is placing the transgene under the native E3 virus promoter, resulting in transgene production in the early or late phase of Ad replication, as defi ned by the alterative splicing of this area (Jin et al. 2005). With this approach, most of the E3 gene can be retained, which may be appealing from the point of view of immunological aspects, optimal oncolysis and spreading of the virus. As an alternative, the transgenes can be placed under cancer specifi c promoters (e.g. E2F, hTERT), a strong constitutive promoter (e.g. CMV, active in most cells) or some inducible foreign promoters (eg. tetracycline). Th e several therapeutic molecules can be grouped by function as done below. While many of these aim to potentiate the classical approach (direct cell killing), the immunomodulatory armed viruses aim at triggering the anti-tumor immunity.

Immunomodulatory oncolytic adenoviruses. Evading the immune system is one of the crucial tasks cancer must accomplish before it can thrive (Hanahan and Weinberg 2011). While the immune system is in a major role in abolishing cancer cells, it also helps to pick the most suitable cancer cells to survive and helps establishing conditions within the tumor microenvironment that facilitate tumor outgrowth (Schreiber et al. 2011). In this immunoediting process cancer cells use multiple mechanisms. Th ey, for example, secrete immunosuppressive cytokines, downregulate Major Histocompability Class I molecules, express Fas ligand to kill reactive cytotoxic lymphocytes and recruit regulatory T-cells. Th us, in advanced cancer the tumor microenvironment is highly immunosuppressive and unaff ected by the immune system, although the tumor presents tumor associated antigens recognized as nonself (de Souza and Bonorino 2009). During recent years, there is growing evidence that arming oncolytic adenoviruses with immunomodulatory molecules helps breaking the tolerance acquired by the tumor (Cerullo et al. 2010). It seems that the infection and immunologic cell death elicit a strong infl ammatory immune response – “a danger signal” - which may overcome the immunosuppression (see fi gure 4, right panel). Immunomodulatory molecules might further enhance the anti-tumor immune reaction.

20 Review of the Literature

At the moment, granulocyte macrophage colony-stimulating factor (GMCSF) seems to be one of the most promising cytokines. As an example, Ad5-D24-GMCSF (serotype 5 Ad based oncolytic adenovirus coding for GMCSF) was used to treat immunocompetent Syrian hamsters with subcutaneous hamster pancreatic cancer. Tumors were eradicated and anti-tumor immunity raised, as these hamsters did not grow new pancreatic tumors even aft er new cancer cell injections (Cerullo et al. 2010). Later this virus was taken to patients in the ATAP program (described in materials and methods) showing good safety and promising results. Another immunostimulatory molecule armed oncolytic virus used in patients codes for CD40L (Pesonen et al. 2012). Immunomodulatory viruses with promising in vitro and in vivo data include the following: aCTLA-4 for activating cytotoxic T-cells (Dias et al. 2011), interferon alpha (Shashkova et al. 2008), beta (He et al. 2008) and gamma (Su et al. 2006) to induce pro- infl ammatory eff ect, RANTES for recruiting immunological cells (Lapteva et al. 2009), IL-24, IL-12 and B7-1 to enhance the immune response (Lee et al. 2006; Liu et al. 2012) and Fas ligand, IL-27 and hsp70 chaperone to assist antigen presentation (Haviv et al. 2001; Tagawa et al. 2006). While many of these still wait to be tested in the clinic, aCTLA-4 as a monoclonal antibody (ipilimumab) has been approved by the FDA for the treatment of unresectable stage III or IV melanoma (Hodi et al. 2010).

Adenoviruses that produce prodrug-converting enzymes. Th ese viruses produce enzymes (thymidine kinase, cytosine deaminase, nitroreductase) (Freytag et al. 1998; Chen et al. 2004) that convert non-toxic prodrugs to active forms. In this way the oncolytic activity is further enhanced by a bystander eff ect caused by the toxic product when the systemically given prodrug is administered. Th ymidine kinase has been of special interest as its eff ect can also be seen in PET or SPECT (Barton et al. 2008). Timing of the administration of the prodrug is critical, as this will typically destroy also the virus and terminate the treatment. Th is “suicide gene” might be regarded also as a safety procedure. Ad5-CD/Tkrep is the fi rst from this group that has been tested in humans (Freytag et al. 2002), and it is being evaluated in a randomized phase 3 trial in fi rst line treatment of prostate cancer in combination with radiation therapy (Lu et al. 2011).

Vector spread enhancing oncolytic adenoviruses. Clinical tumors are complex, and especially the stromal areas of the tumor pose several obstacles for eff ective virus penetration. To make the tumor microenvironment more permeable for adenoviruses, a relaxin producing adenovirus has been generated. Relaxin increases the production of certain matrix metalloproteases and decreases the production of collagen, therefore helping the virus spread (Ganesh et al. 2007). Also adenoviruses expressing ADP (Barton et al. 2006) or adenoviruses without the E3 gp19k gene (Gros et al. 2008) have been made. Th ese were shown to accelerate the virus release from the cells in vitro and in vivo. One interesting experiment engaged a fusiogenic protein from an enveloped virus, which caused the infected cell to fuse with neighboring cells (Galanis et al. 2001).

Direct anti-tumor eff ect producing oncolytic adenoviruses. Many oncolytic adenoviruses coding for apoptosis inducing genes such as TRAIL (Liu et al. 2012), TNF-alpha and Fas ligand (Tagawa et al. 2006) have been made. Th e secreted protein can kill uninfected cells causing a bystander eff ect, but this might hinder the spread of the virus further.

21 Review of the Literature

Anti-angiogenic oncolytic adenoviruses. Tumor neo-vasculature is a hallmark of cancer (Hanahan and Weinberg 2011). Adenoviruses are naturally targeted to newly formed blood vessels. Alpha-v-beta3 integrins are found abundantly in the neo-vasculature (Hood et al. 2002), and these are one of the co-receptors for adenoviruses. Some of the viruses used in ATAP (described in materials and methods) target also integrins. Th e anti-tumor eff ect has been further enhanced by creating oncolytic anti-angiogenic adenoviruses expressing antiVEGF (Yoo et al. 2007) or endostatin (Lin et al. 2007).

Combination therapy. Combining armed oncolytic adenoviruses with existing and new treatment modalities seems to be a feasible approach. Multiple preclinical studies have been conducted to evaluate the best combinations of adenoviruses and chemotherapeutics, radiation therapies, surgery and antibody based therapies (Kumar et al. 2008; Dias et al. 2010; Young et al. 2013). Especially appealing is the combination of checkpoint inhibitors and oncolytic virus. One of the newer ideas incorporate T-cell therapies to oncolytic viruses (Melcher et al. 2011). Also several clinical trials have been started or are being planned to assess these important questions.

Using diff erent cell surface receptors. Targeting adenoviruses to diff erent cell surface receptors is called transductional targeting. All species C (serotypes 1, 2, 5, 6) adenoviruses and most other adenoviruses use the Coxsackie-Adenovirus Receptor (CAR) as a primary attachment receptor. Th e majority of research has been performed with serotype 5 and 2 adenoviruses that both use CAR as the primary receptor. CAR is expressed in a wide variety of cells, but in many advanced cancers it seems to be downregulated (Kuster et al. 2010). Another conceptual issue with CAR is that it seems to be located below the tight junction, facilitating lateral spread. So far it is not clear how adenoviruses initially reach CAR from the epithelial surface (Strauss and Lieber 2009) although recent hypothesis has been proposed (Kotha et al. 2015; Luisoni et al. 2015). In vitro many other receptors (most of which have lower affi nity) for adenoviruses has been found (Arnberg 2009). Some species B adenoviruses have been of growing interest as they use other high affi nity receptors than CAR, and these receptors seem to be widely expressed in advanced human tumors (Tuve et al. 2006). Research from the past few years seem to indicate that these receptors are CD46 and DSG2 (Wang et al. 2011; Lu et al. 2013) (although also CD80/86 have been suggested (Ulasov et al. 2007)). For this reason, a chimeric Ad5/3 virus was made. Here the serotype 5 adenovirus knob has been changed with serotype 3 knob (Kanerva et al. 2002). Many oncolytic forms of this construct have been then used successfully in the ATAP (described in materials and methods) patents (Raki et al. 2011). Inspired by the good results associated with this approach, an oncolytic adenovirus completely based on serotype 3 was made in my stydy (study I and II). Another approach, also used safely in patients, features arginine-glysine-aspartate (RGD) modifi cation of the fi ber to target alpha-v-beta integrins highly expressed in advanced tumors (Nokisalmi et al. 2010; Pesonen et al. 2012). Other targeting modifi cations have also been used in preclinical models (Kreppel and Kochanek 2008) (van Geer et al. 2009) (Alba et al. 2009). Chemical modifi cations of adenoviruses can result in altered tropism as well. Here the virus is coated with a hydrophilic polymer, most oft en with polyetheneglycol (PEG) or poly-N- (2hydroxypropyl)methacrylamine (pHPMA) (Kreppel and Kochanek 2008). Th is shield protects against antibodies and can be further coupled with ligands for tumor associated receptors (e.g. VEGF). Progeny virions, nonetheless, are left unshielded.

22 Review of the Literature

Using alternative physical delivery methods. Th e complex architecture of the tumor prevents spreading of the virus through the tumor, and therefore multiple means to circumvent this problem have been tried. Performing multiple injections of the virus around the tumor is one way of helping the virus spread (Barton et al. 2004). Another way is to inject the virus to the artery leading to the tumor (Reid et al. 2001). Intravenous injection of adenovirus seems to be feasible as well, especially if the amount of neutralizing antibodies is low or the amount of virus is high (Hemminki et al. 2012). Intracavitary injection to treat, for instance, bladder or ovarian cancer is also one good way of increasing the local consentration of the virus (Malmstrom et al. 2010). Using mesenchymal stem cells (MSC) as delivery vehicles is another means of transferring the virus to the cancer. Th ese cells can be effi ciently infected with adenoviruses, and they are shown to home in to tumors aft er intravenous injection (Komarova et al. 2006), although not all studies agree (Hakkarainen et al. 2007).

Preventing the anti-viral immune response. While it has become increasingly clear that the immune response is very important in oncolytic adenovirus treatments, Syrian hamster is becoming a progressively popular model in the fi eld (Diaconu et al. 2010). For unknown reasons, the human adenovirus serotype 5 can replicate in some degree in this immunocompetent model. Oncolytic eff ect can be potentiated by using immunosuppressed Syrian hamster models, but at the same time the ability to raise anti-tumor immunity is disabled and the net result might be negative. In general, there is lack of agreement if the immune response is a friend or a foe in the context of oncolytic adenovirus therapy, and it could be both, or either, depending on the model, virus, transgene and outcome variable. With regard to human data from ATAP (see materials and methods) immunity certainly seems to be an important player in regard to effi cacy (Cerullo et al. 2010), but there are also case examples where lack of neutralizing antibody induction coincided with disappearance of tumors (Nokisalmi et al. 2010). Th us, the consequences of immune response could vary from patient to patient.

Bio-selection of new oncolytic adenoviruses in vitro. Gain-of-function mutations in the viral genome can lead to better anti-cancer agents. One way to do this is to amplify the virus in the presence of mutagens (Gros et al. 2008), or to pool diff erent serotype viruses together and grow them at low multiplicity on diff erent cancer cells. Aft er serial passages, viruses with better anti- cancer abilities (in this in vitro cell line) outgrow other variants. One such experiment resulted in an Ad3/Ad11 chimera with promising oncolytic features (Kuhn et al. 2008; Di et al. 2014), supporting the potential of subgroup B viruses. Th is ColoAd1 construct is now under phase I/ II clinical evaluation (www.clinicaltrials.org). In a similar way as the ColoAd1 was made, any oncolytic adenovirus could be incubated at low multiplicity on cancer cell lines, and viruses with attractive oncolytic abilities on the cells could be harvested. Wheather such viruses correlate with clinical safety or potency, obviously needs to be addressed and may be obtainable only in humans.

Avoiding over-attenuation. Drastic modifi cations of the virus genome may seem appealing for enhancement of safety, but could compromise potency. Building on the good safety of wild type adenovirus injections performed in the 1950s (Huebner et al. 1955; Huebner et al. 1956; Georgiades et al. 1959; Zielinski and Jordan 1969), it has been proposed also now that using wild type adenovirus (which has some innate tumor selectivity) or an armed virus without tumor

23 Review of the Literature selectivity mechanisms, might be safe. Th is hypothesis was allowed by the FDA to enter a Phase I clinical trial for intratumoral injection into solid tumors of any indication (Ying et al. 2009).

3.4 Clinical trials with oncoly c adenoviruses In spite of remarkable advances in prevention, diagnosis and treatments of cancer, its incidence has been increasing constantly, and cancer is now the most frequent cause of mortality in the world according to WHO. With some exceptions, curative treatments are not available for most metastasized solid tumors. Many treatments result in only a minor increase in the time of survival, and many cause severe side-eff ects. Every day tens of thousands of people – some of whom are of working age - die of this cruel disease. Th us it is obvious that there is an urgent need for new treatments. Oncolytic viruses present a strong rationale and an untapped approach for cancer therapy. Also synergy with other treatment modalities have been suggested in numerous preclinical studies (Bauzon and Hermiston 2008), and even some clinical data indicating synergy is available (Khuri et al. 2000; Reid et al. 2005; Kirn 2006; Hemminki et al. 2012). Th e vast majority of clinical trials conducted with oncolytic adenoviruses have been Phase I dose escalation studies designed to assess the safety of the treatment (Kumar et al. 2008). Several oncolytic adenoviruses have appeared safe and no dose limiting toxicities or maximum tolerated doses have been reached in most clinical trials. Instead, the highest dose used has typically been set as the maximum feasible dose for further studies. Adverse events have typically also been mild and self-limiting, although the non-linear toxicity of adenoviruses is given due respect in the fi eld (Raper et al. 2003). Generally it has been thought that the effi cacy of the unarmed oncolytic adenoviruses have been modest (Cody and Douglas 2009) when used as a single modality. However, traditional size-based surrogate endpoints might not be good in evaluating the effi cacy due to the swelling caused by infl ammation resulting from virus replication and/or transgene expression (Reid et al. 2005). Furthermore, the primary endpoint in most trials has not been effi cacy, as only few higher than phase I/II trials have been performed. Below, some interesting clinical trials are described, most of which have been performed with serotype 5 adenoviruses. All of them suggest safety, and also signs of effi cacy can be noted in most.

ONYX-015 (also known as dl1520 or CI-1042, similar constructs include H101, i.e. OncorineR). Originally described as dl1520 (Barker and Berk 1987), a patient isolate, Onyx-015 was the fi rst adenovirus in the second emergence of cancer treatments with viruses occurring during the 1990s. It was tested in a dozen phase 1 and 2 trials but unfortunately never made it into a randomized study. It is a serotype 5 based adenovirus for which several possible selectivity mechanisms have been proposed, including p53/p14ARF defects and aberrant late mRNA transport, both of which may relate to the absence of the E1B-55 K gene. While the deletion of this gene provides attenuation and some specifi city to tumors, it is by default also able to replicate in normal tissues since it was isolated from a non-tumor bearing host. Onyx-015 has been through many Phase I and Phase II clinical studies in a variety of indications (Lu et al. 2004; Crompton and Kirn 2007). Th e use of this adenovirus vector has been proven safe, even when high (3x10e11 PFU) intravenous injections were used. Th e effi cacy data of these trials are moderate when evaluated by classical evaluation criteria. In part this could be due to suboptimal criteria used in effi cacy evaluation, but the attenuation of the virus could also contribute to low potency. Already in the original publication describing dl1520, the strain was reported as severely crippled in comparison to wild type virus (Barker and Berk 1987).

24 Review of the Literature

Some of the most interesting data comes from patients with colorectal liver metastases treated with an injection of ONYX-015 into the hepatic artery with the combination of intravenous 5-FU/leucovorin. Th is resulted in tumor regression in 46% of the patients, and the median survival time increased to 19 months from the expected 4-6 months. Interestingly, some patients initially showed enlargement of the tumors, while later regression was seen (Reid et al. 2005). Th is initial swelling is reported as progressive disease using classical evaluation criteria, which might underestimate underlying therapeutic eff ects, as subsequently described for ipilimumab (anti-CTLA4 antibody) (Wolchok et al. 2009). In this particular ONYX-015 trial, the inside of these swollen tumors were reported to be necrotic, and this could be observed with PET-CT. Th is and other observations suggest that multimodal follow-up is useful for oncolytic viruses. In regard to immunologically armed oncolytic viruses, this eff ect is probably even more prominent than with unarmed viruses or “passive” immunotherapy with anti-CTLA4 antibodies. Tumor size, however, is only a surrogate endpoint (used in lieu of survival), and interpretation of non-randomized data is very diffi cult. Th us, it was satisfying to see Shanghai Sunway Biotech taking H101 (a closely related although distinct virus from dl1520) to a randomized phase 3 trial (Xia et al. 2004), which resulted in approval of Oncorine by the Chinese FDA in 2005 for the treatment of head and neck squamous cell carcinoma. According to the companys home page (http://www.sunwaybio.com.cn/en/product.html) Oncorine is recommended nowadays also in advanced lung cancer, liver cancer, pleural and peritoneal eff ucions and pancreatic cancer.

Ad5-CD/Tkrep. Ad5-CD/Tkrep is the fi rst replication-competent adenovirus vector that expresses a therapeutic gene that was tested in humans in a randomized setting (Freytag et al. 2002). Ad5-CD/Tkrep is replication selective since it contains the same E1B-55K gene deletion as ONYX-015. Th e virus expresses a cytosine deaminase (CD) and herpes simplex virus thymidine kinase (TK) fusion gene to allow for therapeutic treatment with both 5-FC and ganciclovir. In a fi ve-year follow up report of a phase I trial in human prostate cancer, the prostate-specifi c antigen doubling time increased following the gene therapy from a mean of 17 to 31 months (P=0.014). Th is postponed the salvage androgen suppression therapy (that is associated with high morbidity) by an average two years with good safety. Th us, this approach could provide an attractive treatment option for patients experiencing PSA relapse following defi nitive therapy (Freytag et al. 2007). Th is virus is now being tested in a randomized phase 3 study in fi rst line therapy of prostate cancer in combination with radiation (Lu et al. 2011).

CG7060 and CG7870 (also called CV706 and CV787). CG7060 is a serotype 5 based prostate- specifi c antigen selective replication-competent adenovirus indicated for use in prostate cancer. A single Phase I clinical trial with CG7060 has been reported. Here 20 patients in fi ve dose cohorts (1 x 1011 to 1 x 1013 viral particles, VP) were treated with an intra-prostatic injection. All fi ve patients who achieved 50% reduction in PSA were treated with the highest two doses of the vector. Th is treatment was found safe and effi cacy was suggested (DeWeese et al. 2001). CG7870 is another serotype 5 based vector that is intended for use in prostate cancer. Previously the vector was found safe and well tolerated in a phase I/II clinical trial of intra-prostatic delivery for locally recurrent prostate cancer. Next, the CG7870 was administered as a single intravenous infusion in a dose escalation design (1 x 1010 to 6 x 1012 VP) to 23 patients with hormone- refractory metastatic prostate cancer. Flu-like symptoms were commonly observed. No partial or complete prostate-specifi c antigen responses were observed; however, vefi patients had a

25 Review of the Literature decrease in serum PSA of 25% to 49% following a single treatment, including three out of eight patients at the highest dose levels (Small et al. 2006).

CG0070. An Integrated Phase II/III, open label, randomized and controlled study to assess the safety and effi cacy of CG0070 is now recruiting (NTC01438112). Th is adenovirus serotype 5 vector expresses GM-CSF. Patients with non-muscle invasive bladder cancer with carcinoma in situ who have failed BCG (Bacillus Calmette–Guérin) are included.

Telomelysin (also called OBP-301). Telomelysin is a human telomerase reverse transcriptase (hTERT) promoter driven serotype 5 oncolytic adenovirus. One phase I clinical trial was conducted in patients with advanced variable solid tumors. A single intratumoral injection of Telomelysin was administered to three cohorts of patients (1 x 1010, 1 x 1011, 1 x 1012 VP). All doses were well tolerated. One patient had a partial response of the injected malignant lesion. Seven patients (of the sixteen) fulfi lled Response Evaluation Criteria in Solid Tumors (RECIST) defi nition for stable disease at two months aft er treatment. Post-injection biopsies performed at day 28 on four of the patients with stable disease revealed necrosis of the tumor. One patient had 33% reduction of injected lesion at day 28 and 56.7% reduction of injected lesion at day 56. Hence also evidence of antitumor activity was suggested (Nemunaitis et al. 2010).

H103 is a serotype 5 based oncolytic adenovirus similar to H101 and Onyx-015, but it expresses heat shock protein (HSP)70, which is suggested to be important in antigen presentation and activating the immune system. A phase I clinical trial of intratumoral injection of H103 was conducted in 27 patients with advanced solid tumors. Injections were performed in a dose- escalation manner from a dose of 2.5 x 107 to 3.0 x 1012 VP. Th e treatment was well tolerated. Th e objective response (complete response and partial response) to H103-injected tumors was 11.1% (3/27), and the clinical benefi t rate (complete response, partial response, minor response and stable disease) was 48.1%. Partial transient regression of distant, uninjected tumors was also noted in three patients (Li et al. 2009).

ONCOS-102 (also called CGTG-102 and Ad5/3-d24-GMCSF). A CGTG spin off company Oncos Th erapeutics performed a phase I clinical trial in 2013 with the Ad5/3-D24-GMCSF virus (www.clinicaltrials.org) used before in the ATAP. Virus was suggested safe in patients with injectable refractory solid tumors with concominant low dose cyclophosphamide. In January 2015 the virus (Oncos-102) was granted US lung cancer orphan drug designation. According to the press release other indications might include soft tissue sarcoma, mesothelioma and ovarian cancer. In June 2015 the company joined with a Norwegian immunotherapy company (Targovax) becoming the largest company in immunotherapy in northen Europe.

3.5 Cancer immunotherapy in pa ents It has been known for decades that cancer cells have antigens, diff erent to normal cells, on their cell surface. Th ese should be recognized by the immune system and destroyed. However, during the evolution of cancer growth, the immune system has been inhibited by multiple methods. Cancer can become invisible to the immune system by two principal ways, the antigen presenting machinery can be altered (e.g. downregualtion of MHCI) or tumors can have

26 Review of the Literature variants with reduced sensitivity to immune eff ectors. Other ways for the tumor to hide from immunodestruction is to suppress the antitumor immune response by infi ltration of suppressor myeloid cells, infi ltration of regulatory T-cells, or production of immunosuppressive factors. If these points can be eff ectively addressed, then most, also metastatic, cancers might be treated. For decades multiple diff erent approaches have been experimented with in order to boost the immune response (Mellman et al. 2011). However, complete cures have been rare, as advanced cancer has usually evolved to overcome these treatments. Furthermore, cross reactivity with normal cells have lead at occasions to unpleasant side eff ects. Th ese days the understanding of the immune system seems to have reached a limit where therapeutic eff ects are becoming apparent. Th is is clarifi ed with a handful of newly approved immunological drugs discussed below and partly shown in table 3. Before, the most evident problem seemed to be that it was not known how to “release the brakes” of the immune system. During the last few years a dozen phase III trials have shown the effi cacy of checkpoint inhibitors in releasing the brakes, thus re-opening the stage for many therapies not found eff ective before.

Table 3. History timeline of modern medicine cancer treatments

Cancer Immunotherapy Timeline Other cancer treatments 1840s Surgery under ether anesthesia (Long) 1850s 1860s 1870s 1880s Coley´s toxins (bacteria were noted to 1890s Radiotherapy decrease tumors) Case reports of tumor regression aft er 1900s natural viral infections Hormonal therapy (estrogen in 1910s Prostate CA) Chemotherapy (Nitrogen mustard) 1920s 1930s 1940s Hundreds of case series treating cancer 1950s Linear accelerator for RT with viruses (e.g. Varicella, Measels, Stereotactic RT, Combination Vaccinia, West Nile, Adenovirus, Mumps) chemotherapy, 1960s Tamoxifen synthesized BCG in Bladder CA 1970s -85 1st study adoptive T cell transfer 1980s Tamoxifen approved -86 IFN-alpha approved -Oncolytic viruses are re-invented and 1990s Scopic techniques become standard in rationally designed many operations -92 IL-2 approved -97 1st mAB approved -05 Oncolytic adenovirus (H101) 2000s Da Vinci robotic system approved by approved in China the FDA

27 Review of the Literature

Table 3 cont. Cancer Immunotherapy Timeline Other cancer treatments 1st cellular immunotherapy approved 2010 (sipuleucel-T) 1st checkpoint inhibitor approved 2011 (ipilimumab, CTLA4 mAB) T cell therapies show evidence of effi cacy 2012 in multiple centers and many cancer types 2013 2014 More checkpoint inhibitors approved 2015 (nivolumab and pembrolizumab, PD1 mAB). FDA votes 22-1 for approving T-Vec oncolytic virus Th e rapid development of cancer immunotherapy is regarded to be due to fast development in fi elds such as gene therapy and immunology during the last decades

Historically, cancer immunotherapy has been most oft en divided into cytokine-, antibody-, cancer vaccine- and cell based therapies. Nowadays also oncolytic viruses should be added to the list. Although the mechanism of action is diff erent between the treatment approaches, the uniting factor is that these therapies stimulate the immune system of the patient to fi ght cancer.

Cytokines BCG. In the early 20th century BCG (Bilié/Bacillus de Calmette et Guérin) tuberculosis vaccine was invented. It contains live bovine mycobacterium tuberculosis that is almost nonvirulent to humans. While BCG is not a cytokine and the mechanism of action is still somewhat unclear, the eff ect in cancer therapy is probably mediated via the cytokine storm it produces (Redelman- Sidi et al. 2014). In 1970, it was shown that BCG inhibits the growth of melanoma when injected locally. However, the results systemically were not as good as treatment with chemotherapy. Likewise, the combinations did not result in increased survival. Instead, BCG has been used since the 1970s successfully in medium to high risk noninvasive bladder cancer as the only treatment that has proven to reduce the risk of recurrence. It is regarded as the best treatment in this patient group (Perez-Jacoiste Asin et al. 2014). Side-eff ects usually last some days aft er the inoculation in the bladder. Th ese include bladder irritation, mild fewer and ufl like symptoms. Serious adverse events are rare. IFN-alpha. Another example of immunotherapy that has been used for decades is the interferons (IFN). In 1957, it was discovered that a destroyed infl uenza virus could interfere with the replication of a wild type infl uenza virus when implanted in an egg. Th is was explained to be due to a protein that was named interferon. Interferons are divided to alpha-, beta- and gamma interferons. Of these, alpha interferons have been used the most. Kari Cantell, a Finnish scientist, has been a pioneer one the fi eld and his Finnferon product have been widely used for decades. Best responses are acquired in an atypical leukemia, hairy-cell-leukemia, where basically all patients achieved remission. Also other leukemias, lymphoma, multiple myeloma

28 Review of the Literature and Kaposi´s sarcoma have shown good response to interferon treatments. In regard to solid tumors, only melanoma and kidney cancer have shown objective benefi ts. IFN remains the only currently available adjuvant option for melanoma. It was approved 1986 as adjuvant therapy in patients with melanoma at high-risk of recurrence aft er surgical resection (Ascierto et al. 2014). Interferon is usually given into the muscle as a daily dose or as three doses a week. Clear survival benefi ts have been hard to prove with interferons. Th e side eff ects have, however, been mild compared to chemotherapy. In 2011, FDA approved a new peg interferon alfa-2b drug that is used for the treatment of Hepatitis C. IL-2. Interleukin-2 is a protein produced by activated T-cells. It also auto activates T-cells and natural killer cells. It has been called the T cell growth factor. In 1992, FDA approved Proleukin (recombinant interleukin-2) for the treatment of metastatic renal cell carcinoma. In clinical studies, about 1 in 15 patients had no evidence of disease ranging from 7 months to 10+ years (Rosenberg 2007). Later it received also indication for metastatic melanoma. About 1 in 17 patients with metastatic melanoma become free of all evidence of disease ranging from 3 months to 10+ years (Amin and White 2013). Also other cytokines have shown some promise, for instance tumor necrosis factor (TNF) that has shown effi cacy in colorectal and soft tissue sarcomas in phase I-II trials. Th e same substance has shown some effi cacy in liver metastases when given to the feeding arteria.

AnƟ bodies Antibodies are the key components of the adaptive immune system, playing a fundamental role in both the recognition of foreign antigens and the stimulation of an immune response to them. Most antibodies used in cancer immunotherapy are naked; some, however, are conjugated with toxic or radioactive molecules. Antibodies are also referred as murine, chimeric, humanized and human. Antibody binding to the cancer cell can lead to NK cell mediated cell death, complement activation and cell death, or the antibody binding can interfere with cell signaling leading to reduced cell survival. Tumor specifi c antibodies have become part of the therapeutic repertoire in treating colorectal, breast, head and neck cancers and leukemia aft er improving the overall survival and progression-free survival in randomized phase 3 trials. Th e response rates have been 8-10% when used as single agents. Combining traditional chemotherapy and/ or radiotherapy have increased response rates to 30% (Kirkwood et al. 2012). Table 4 shows 14 approved antibodies used in cancer treatment. Two are conjugated with toxic compounds and two are radiolabeled, while the rest are naked.

29 Review of the Literature

Table 4. Approved antibodies for the treatment of cancer Brand Antibody Antibody Year Generic name Approved treatment(s) name Type Target Approved B-cell Chronic Alemtuzumab Campath humanized CD52 2001 lymphocytic leukemia (CLL) metastatic colorectal 2004 cancer non-small cell lung Bevacizumab Avastin humanized VEGF 2006 cancer 2009 renal cell carcinoma 2009 glioblastoma multiforme relapsed Hodgkin 2011 Brentuximab lymphoma Adcetris chimeric CD30 vedotin* relapsed anaplastic large- 2011 cell lymphoma 2004 colorectal cancer advanced squamous cell 2006 carcinoma of the head and neck (SCCHN) recurrent locoregional or Cetuximab Erbitux chimeric EGFR 2011 metastatic squamous cell head and neck cancer EGFR-expressing 2012 metastatic colorectal cancer Gemtuzumab acute myelogenous Mylotarg humanized CD33 2000 ozogamicin* leukemia Ibritumomab Zevalin murine CD20 2002 non-Hodgkin lymphoma tiuxetan** Ipilimumab Yervoy human CTLA4 2011 metastatic melanoma 2014 metastatic melanoma PD-1 Nivolumab Opdivo human squamous non-small receptor 2015 cell lung cancer Ofatumumab Arzerra human CD20 2009 refractory CLL metastatic colorectal Panitumumab Vectibix human EGFR 2006 cancer PD-1 melanoma aft er Pembrolizumab Keytruda human 2014 pathway ipilimumab Rituxan, 1997 non-Hodgkin lymphoma Rituximab chimeric CD20 Mabthera 2010 CLL Non-Hodgkin Tositumomab** Bexxar murine CD20 2003 lymphoma Trastuzumab Herceptin humanized ErbB2 1998 breast cancer *with toxic compound ** radiolabelled Checkpoint inhibitors in bold. Modifi ed from (Scott et al. 2012)

30 Review of the Literature

Checkpoint inhibitors. Of these antibodies, I would like to point out checkpoint inhibitors ipilimumab, nivolumab and pembrolizumab. Th ese three antibodies have gained recent approval and have shown remarkable promise in a large repertoire of advanced cancers. CTLA4 antibody ipilimumab (Yervoy, also known as MDX-010 and MDX-101) is a human IgG1 antibody that binds to the cell surface CTLA4. CTLA4 is expressed on T-cells, and activation of the protein switches off the T-cell activity. Ipilimumab binds to this protein so that cytotoxic T-cells are not inactivated by the immunosuppressive cancer microenvironment. While ipilimumab seems to aff ect more T-cells in the lymph nodes, another antibody called nivolumab (Opdivo, ONO-4538, BMS-936558, MDX1106) seems to work better in the tumor microenvironment. Nivolumab is also a fully human IgG4 monoclonal antibody. When the PD-1 inhibitor was tested in the Checkmate 66 phase III randomized trial for the treatment of melanoma, the study was stopped early in 2014 due to analysis conducted by the independent data monitoring committee, as the survival of the PD-1 arm was superior to the control arm. Patients were unblinded and allowed to cross over to PD-1 treatment. Nivolumab was approved by the FDA for the treatment of unresectable or metastatic melanoma (Dec 2014) and for the treatment of squamous non- small cell lung cancer (Mar 2015). Some months later (Jun 2015) also EMA granted marketing authorisation in Europe. It is thought that tumors exploit the PD-1 checkpoint pathway to turn off an immune response at the tumor site by inactivating T-cells. Th ese T-cells cannot mount eff ective immune responses. Nivolumab prevents the PD-1/PD-L1 interaction, thus reinvigorating the immune system. Effi cacy has also been seen aft er ipilimumab (CTLA-4) failure (Hamid O 2014). Th e third checkpoint inhibitor recently (2014) approved by FDA is called pembrolizumab (Keytruda). It is intended for use in melanoma, following treatment with ipilimumab. Some months later (May 2015) also EMA recommended granting a marketing authorisation for pembrolizumab. It is recommended as monotherapy for the treatment of advanced melanoma. Its mechanism of action is similar to nivolumab, but, instead of acting at the cell surface receptor, it blocks the PD-1 intracellular pathway. Th e adverse event profi le is told to be slightly favorable, having only 12% grade 3-4 adverse events (Antoni Ribas 2014).

Cancer vaccines Th ere are two types of cancer vaccines - preventive and therapeutic. Preventive vaccines against two viruses that cause cancer are in wide global use: vaccine against hepatitis B and human papillomavirus (HPV). In 2010, FDA approved the fi rst therapeutic vaccine. In 2013, marketing authorization for the drug was granted in Europe. Th is vaccine, sipuleucel-T (Provenge) was approved for men with castration resistant metastatic prostate cancer. It stimulates an immune response to prostatic acid phospahatase (PAP). Th is antigen is found on most prostate cancer cells. Before, it was used as a tumor marker in prostate cancer, but nowadays it has been replaced almost completely by the more sensitive prostate specifi c antigen (PSA). In a clinical trial, a survival benefi t of about four months was noted. Adverse events that were more frequently reported in the sipuleucel-T group than in the placebo group included chills, fever and headache (Kantoff et al. 2010). While so far no clinical benefi t has been observed with vaccines containing only antigen that have been injected strait to the patient, the sipuleucel-T vaccine is customized to each patient. First, antigen presenting cells are isolated from patient blood. Th ese are then cultured with a protein called PAP-GM-CSF. Aft er culture the cells are returned to the patient. Patients receive three treatments usually two weeks apart. Each cycle of treatment involves the same in vitro culture. Th e mechanism of action seems to be that the antigen presenting cells that

31 Review of the Literature have taken up PAP-GM-CSF stimulate T-cells to kill tumor cells that express PAP. High cost and complicated treatement (Sweden is the closest location to Finland for the treatment) has been obstacles for common practice. Prostavac is a new vaccine used against metastatic prostate cancer as subcutaneous injections. It utilizes recombinant poxviruses to express PSA, along with 3 immune-enhancing costimulatory molecules (LFA-3, ICAM-1, and B7.1). Treatment is initiated by subcutaneous administration of a priming dose of vaccinia encoding PSA and the costimulatory molecules, followed by 6 subsequent boosting doses of fowlpox encoding the same PSA-costimulatory molecule cassette. Using this heterologous prime-boost dosing regimen, the immune system becomes focused on inducing PSA-specifi c T cell responses, designed to kill tumor cells expressing PSA. A phase 2 trial with 125 castration resistant prostate cancer patients was completed showing an overall survival increase of 8.5 months (25.1 versus 16.6 months) and a 44% reduction in the risk of death (stratifi ed log-rank P=.0061). Prostvac was generally well tolerated, with the most common side eff ects including injection site reactions, fever, fatigue, and nausea (Kantoff et al. 2010). A phase 3 study (Prospect) is ongoing.

Cell-based therapies At the moment there are three types of adoptive cell transfer (ACT) therapies advancing towards regulatory approval: 1) tumor infi ltrating lymphocytes (TIL), 2) chimeric antigen receptor (CAR) engineered T cells, and 3) T-cell receptor (TCR) engineered T-cells (June et al. 2015). TILs do not need genetic engineering. Th e theory is that the patients own T-cells recognize the tumor, but the tumor has disabled these important T-cells. Results are gained when these cells are activated. Th e commonly used method is to extract the T-cells from excised tumor, expand and activate them in vitro (IL-2 is commonly used) before infusing the cells back to the patient. Usually the patient is lymphodepleated by chemotherapy before the cells are returned. It is anticipated that cancers with several mutations are prone to be also heavily immunosuppressed and thus benefi t from checkpoint inhibitors (CTLA4 and PD1) as well as TIL therapies. So far TIL therapies have been developed mainly by the National Cancer Institute, and recently an international phase 3 randomized metastatic melanoma trial (NCT02278887) was started (June et al. 2015). Engineered CAR and TCR cells, on the other hand, are harvested from the peripheral blood. T-cells are then transduced by viral or nonviral methods so that the desired receptor is introduced to the T cell surface. By this method a large pool of cells targeted to a known receptor can be achieved. At the moment, there are at least 20 companies developing CAR and TCR therapies, while only few companies concentrate on TIL therapies. ACT has been generally safe, but on-target and off -target recognition of normal tissue can occur with engineered T-cells. To date, most patients have been treated with mouse antibody modifi ed T-cells. It is though that using fully human modifi ed T-cells might decrease toxicity. However, as observed with the checkpoint inhibitors (CTLA-4 and PD-1), autoimmune and infl ammatory side eff ects can be expected also with cell based therapies. As elucidated with the sipuleucel-T (Provenge) dendritic cell vaccine: Cell therapies can be manufactured and delivered to physicians, but the effi cacy must justify the cost and complexity of these new treatments compared to more easily administered pharmaceuticals. A recent publication highlights the potential of cell based therapies (Maude et al. 2014). Relapsed acute lymphoblastic leukemia (ALL) is diffi cult to treat regardless of the aggressive treatments. CAR cells targeting CD19 may overcome many limitations of the conventional

32 Review of the Literature therapies. A total of 25 pediatric and 5 adult patients were treated with autologous T-cells transduced with lentiviral vector. Half of the patients had undergone stem-cell transplantation. Of the 30 patients, 27 (90%) received complete response and three received partial response. Th e six month overall survival was 78% and the responses seem durable (Maude et al. 2014). Modifi ed T-cells have shown to be potent in leukemia and lymphoma. New surface molecules are currently being targeted: epidermal growth factor (EGFR) for glioblastoma, prostate specifi c membrane antigen (PSMA) for prostate cancer, mesothelin for ovarian, pancreatic and mesothelioma, and C-met for triple negative breast cancer.

OncolyƟ c viruses Oncolytic adenoviral trials have been discussed in more detail in chapter 3.4. In summary, accumulating data suggest that the potential effi cacy of oncolytic viruses is due to immunological mechanisms (Figure 4). Viruses can also be armed with immunomodulatory genes as discussed in chapter 3.3. In table 5, some interesting oncolytic viruses with clinical data are shown.

T-Vec (Talimogene laherparepvec) oncolytic virus is discussed here, as FDA voted 22-1 in favor of its approval 4/2015, making it probably the fi rst oncolytic virus to be approved in a western country. It is an oncolytic herpes simplex virus type 1 (HSV1) modifi ed so that it produces human GM-CSF. It is given intratumorally to injectable tumors but systemic responses in non- injectable tumors are also commonly seen. It was compared in a randomized phase III trial (OPTiM) with GM-CSF in the treatment of stage III-IV melanoma. In IIIB/C, IV M1a patients the median survival was 41,1 months in the T-Vec arm, while 21,5 months was observed with the GM-CSF arm (P<0.001). Also progression free survival increased with the T-Vec arm (Robert Hans Ingemar Andtbacka 2013). No diff erence was observed in the stage IV M1b/c subgroup. A combination phase Ib trial with T-Vec + ipilimumab (CTLA-4) has been started. In preliminary analysis of 17 patients with investigator assessed response, ORR was 41% (24% CR, 18% PR); 35% had SD. Median time to response was 2.9 months (Puzanov 2014). Also according to the pharmaceutical company’s poster (shown in ESMO, Madrid, 2014), four other phase II trials with melanoma have been started with the T-Vec: 1) combination with pembrolizumab (MK- 3475), 2) correlation between objective response rate and baseline CD8+ cells, 3) as neoadjuvant before surgery, and 4) biodistribution trial to evaluate shedding of the virus.

33 Review of the Literature

Table 5. Some promising oncolytic viruses and their clinical status

Clinical Product Company Virus Lead indication status Head-and-neck cancer; Adenovirus modifi ed lung cancer; liver Shanghai by E1B-55KD deletion Oncorine cancer; malignant Approved in Sunway Biotech for conditional (H101) pleural and peritoneal China, 2005 Co replication in P53- eff usions; pancreatic defi cient cancer cells cancer Modifi ed HSV-1 carrying ICP34.5 Phase 3, T-Vec & ICP47 deletions, FDA voted (Talimogene Amgen expressing US11 as an Metastatic melanoma 22-1 in favor laherparepvec) immediate early gene of approval and encoding GM- 4/2015 CSF Reolysin (respiratory Oncolytics Reovirus serotype 3 Head-and-neck cancer Phase 3 enteric orphan Biotech (Dearing strain) virus) Cold Genesys Conditionally CG0070 (Irvine, replicating adenovirus Bladder cancer Phase 2/3 California) encoding GM-CSF Pexastimogene Th ymidine kinase- devacirepvec Jennerex deleted vaccinia virus Liver cancer Phase 2b (Pexa-Vec, JX- Biotherapeutics encoding GM-CSF 594) Viralytics Unmodifi ed Cavatak (Sydney, Melanoma Phase 2 Coxsackievirus A21 Australia) Conditionally Virttu Biologics replicating HSV1716 Seprehvir High-grade glioma Phase 2 (London) carrying ICP34.5 deletion Oncolytic virus derived by directed PsiOxus evolution of chimeric ColoAD1 Th erapeutics Metastatic cancer Phase 1/2 virus based on (Abingdon, UK) adenovirus serotypes 3 and 11p Conditionally DNAtrix replicating adenovirus DNX-2401 Glioma Phase 1 (Houston) encoding an integrin- binding peptide Oncos Conditionally Th erapeutics CGTG-102 replicating adenovirus Solid tumors Phase 1 (Helsinki, encoding GM-CSF Finland) Modifi ed from (Sheridan 2013)

34 Review of the Literature 3.6 Evalua ng the effi cacy of immunotherapies At the moment, treatment effi cacy of cancer patients is evaluated mainly by imaging, performance status and sometimes tumor markers (e.g. PSA). In regard to immunotherapeutics, there are certain weaknesses in all of these.

In vitro and in vivo models. A clear discrepancy in effi cacy between tissue culture, animal experiments and human patients has been observed as long as adenoviruses or immunotherapies have been used to treat cancer. For example, if a particular oncolytic virus kills eff ectively a given cell line, the results observed in vivo can already be contradictory. Multiple factors are thought to lie behind this phenomenon (Strauss and Lieber 2009): 1) lack of a primary adenovirus attachment receptor on tumor cells in situ, 2) partly unknown innate host defense mechanisms that recognize adenoviruses as pathogens and shut down adenovirus replication in vivo making tumors resistant to further virus injections (Liikanen et al. 2011), 3) Virus spread in the tumor is limited due to the 3D structure containing fi brotic stroma, necrotic-, hypoxic- and low pH- areas of real life tumors as well as the epithelial phenotype containing tight and adherens junctions of cancer cells. Even the mere size of virus particles, as compared to many chemical drugs, could present a problem for intratumoral spread. 4) When it comes to immunocompetent animals or patients, the adaptive immune system could also play a role in elimination of virus within weeks. 5) Clinical treatments have been usually applied to terminal patients with advanced disease and history of multiple cycles of cancer treatments. Th ese cancers have evolved for years, developing the crucial mutations needed to avoid the multiple ways of the body to eliminate cancerous tissues. Th is is in great contrast with the cancers implanted to murine models that typically grown for a few weeks. In immunocompetent mice, it is possible to collect extensive biological materials for laboratory analysis. However, while the quality of the data can be good, its relevance in the context of humans can vary. Even the most sophisticated mouse models develop tumors over weeks or months at best, while human tumors grow over years and decades, and thus the same level of immunological complexity cannot be reproduced. Moreover, if tumor cell line tumors are transplanted into mice, they do not resemble natural carcinogenesis. Transgenic models can mimic better development of tumors in situ, which is an improvement, but a caveat is that cancer in humans does not occur subsequent to engineered mutations in tumor suppressors and oncogenes, and thus several diff erences in phenotype are likely. It has thus become clear that immunocompromised (e.g. SCID or nude mice) animal models with injected human cancer cell lines or immunocompetent animals (eg. Syrian hamsters) with hamster cancer, or any other animal model have poor prognostic capabilities when it comes to advanced human cancers.

Evaluating patient responses with imaging. In the clinics, CT imaging is the most widely used imaging method, while PET-CT and MRI have gained popularity during the years. Th e problem with immunotherapeutics is that sometimes tumor swelling due to infl ammation makes the tumor larger, and this is categorized as progression of the disease while in reality it might be a sign of effi cacy (Wolchok et al. 2009; Hoos et al. 2010; Prieto et al. 2012). Same is true with lymph nodes (Focosi et al. 2008). If these become enlarged, it might be interpreted as metastases to the lymph nodes, while it might just suggest that the immune system has started to recognize malignant tissue and troops are multiplying in the lymphatic tissue (Koski et al. 2012). MRI has the same weakness if only the total size of the tumor is evaluated. However, MRI is better

35 Review of the Literature in evaluating changes inside the tissue, as discussed in study III. Similar problems can be seen with increased sugar intake with PET-CT (Koski et al. 2012), while other labeling might provide better insight into what is truly happening. Analogous problems in evaluating responses have been evident with the checkpoint inhibitors (Saenger and Wolchok 2008). In one report with CTLA-4 blocking ipilimumab, it took a median of 30 months to achieve response with the patients who fi nally achieved complete response (Prieto et al. 2012). In one long-term analysis of melanoma patients who lived at least 4 years aft er ipilimumab, 25% had never achieved an outcome better than progressive disease, as defi ned by WHO criteria (Wolchok et al. 2013). Th us, it is clear that standard radiographic response criteria do not necessary capture possible benefi cial eff ects of immunotherapies (Hoos et al. 2010). Some new Immune-Related Response Criteria have been evaluated (Wolchok et al. 2009) and they have now been used in some of the checkpoint inhibitor trials (Wolchok et al. 2013; Wolchok et al. 2013). It is known that with immunotherapeutics one pattern of response includes initial swelling/growing of the tumor and only later the response is noted. With RECIST criteria (developed for chemotherapy) this initial increase is graded as progressive disese and other treatments are considered. With the Immune-Related Response Criteria this pattern of response is is not neglected and patients are followed longer. Similarly, appearance of new lesions are graded progressive disease with RECIST while the Immune-Related Response Criteria try to take this in account with immunotherapeutics (Wolchok et al. 2009).

Survival and progression free survival have been the traditional endpoints in clinical trials. Both of these pose problems when it comes to immunological evaluation. It is common with many cancer treatments that some patients benefi t, while some do not, and some might suff er from the treatment. Generally the total benefi t can be evaluated, for example, how many months the median patient survives. With immunotherapeutics it is, however, not unusual that some patients survive very long and seem even to be cured from advanced cancer (see Figure 1). Th is clear benefi t is not seen with the median overall survival. A better way might be to evaluate patients that are alive aft er 3 or 5 years from treatment or to use (restricted) mean survival (Zhao et al. 2015). When analyzing progression free survival, many patients might be progressing fi rst while later a response is seen. As discussed before, these patients might thus be deemed progressors while the benefi t is only seen later.

Evaluating patient responses with tumor markers. Tumor markers have proven to be relatively good indicators with certain tumors, to certain levels. Th e biggest problem associated with the markers is that tumor destruction can release large amounts of tumor markers to the blood stream. Th us we (Topi M. O. Turunen 2009) and others (Hoos et al. 2010) have reported that an initial tumor marker growth in blood might not necessary be a sign of poor effi cacy.

Emerging ways to evaluate response. In our studies, some new possible immunological markers have emerged. Amounts of anti-cancer antibodies in serum seemed to correlate with treatment effi cacy (study IV). Also other immunological markers, such as T-cell types and amounts, MRI, MRS, calprotectin (study III) and serum HMGB1 (Liikanen et al. 2015) have shown some potential.

36 Review of the Literature 3.7 Immunotherapy in the future According to National Cancer Institute web pages (4/2015), there are 142 active (currently accepting patients) phase III/IV immunotherapy treatment trials recruiting in the USA. If we include also phase I-II, the number of active immunotherapy trials increases to 1703. Of all cancer trials, this is a massive proportion, as 372 phase III/IV and 3882 phase I-IV treatment trials in cancer were active in this registry. Th us, approximately 40% of all active cancer trials at the moment are on immunotherapy! While immunotherapy has been around as long as modern medicine has existed (Coley´s toxins, virus experiments, BCG etc.), the true potential has just started to reveal itself as the checkpoint inhibitors seem to show benefi t in a large repertoire of cancers. Some of the data are published in peer-review papers, while a large amount of the most recent data is gathered from meeting presentations and abstracts, or from FDA publications. Below some interesting recent publications are discussed with conclusions and future visions. Checkpoint inhibitors in melanoma have certainly proven to be one of the better treatment options in advanced disease. At the moment, it seems that BRAF staining status does not correlate with response, while some evidence shows that PD-L1 staining status aff ects the results, but checkpoint inhibitors seem to work also in PD-L1 staining negative patients. Th is was confi rmed in a recently published report (Long et al. 2015). PD-L1 positive patients survived longer with both nivolumab and dacarbazine (chemotherapy) treatments, while PD-L1 negative patients survived longer when treated with nivolumab compared to the chemotherapy. Th us, it seems that PD-L1 staining suggests good survival, but also PD-L1 negative patients seem to benefi t from nivolumab. Drug-related grade 3–4 AEs were reported in 12% of patients receiving nivolumab, in contrast to 18% receiving dacarbazine. Th e authors concluded that compared to dacarbazine, nivolumab signifi cantly improved OS and PFS in previously untreated patients with BRAF wild- type metastatic melanoma with an acceptable safety profi le. In a recent publication, CTLA-4 antibody ipilimumab was combined with a PD-1 blocking antibody nivolumab in the treatment of metastatic melanoma without previous treatment (Postow et al. 2015). Objective response was seen in 61% (44 out of 72 patients) in the combination group versus 11% (4 of 37 patients) in the group that received ipilimumab and placebo. Drug-related adverse events of grade 3 or 4 were reported in 54% of the patients who received the combination therapy compared to 24% of the patients who received ipilimumab monotherapy. Most of these events resolved with immune- modulating medication and the authors concluded that the safety profi le was acceptable. Another trial (phase I with 53 patients) used this same combination of ipilimumab and nivolumab and showed that 53% of patients had at least 80% tumor shrinkage with manageable safety profi le (Wolchok et al. 2013). Sometimes the results have been so dramatic and rapid, even aft er a single dose of the combination to a bulky melanoma, that clinicians have been worried, especially if the tumor is present in the myocardium or transmural metastasis in the small bowel (both common sites of metastatic melanoma) (Chapman et al. 2015). Checkpoint inhibitors have also started to show potential in other cancers. FDA approved recently nivolumab (Opdivo) for intravenous use, for the treatment of patients with metastatic squamous non-small cell lung cancer (NSCLC) with progression on or aft er platinum-based chemotherapy (FDA press release 3/2015). Nivolumab demonstrated signifi cantly superior overall survival (OS) compared to docetaxel, with a 41% reduction in the risk of death (hazard ratio: 0.59 [95% CI: 0.44, 0.79; p=0.00025]), in a prespecifi ed interim analysis of a phase III clinical trial. Approval was based on the results of CheckMate -017 and CheckMate -063 trials. In Hodgins lymphoma, 23 patients were treated with nivolumab (Ansell et al. 2015). An

37 Review of the Literature objective response was reported in 20 patients (87%), including 17% with a complete response and 70% with a partial response; the remaining 3 patients (13%) had stable disease. Th e rate of progression-free survival at 24 weeks was 86%. Drug-related adverse events of any grade and of grade 3 occurred in 78% and 22% of patients, respectively. In renal cell carcinoma a phase II trial was performed. Patients with clear-cell mRCC previously treated with agents targeting the vascular endothelial growth factor pathway were randomly assigned to nivolumab 0.3, 2, or 10 mg/kg intravenously once every three weeks (Motzer et al. 2015). Th e primary objective, progression-free survival, was 2.7, 4.0 and 4.2 months, respectively. Secondary end point was overall survival: 18.2 months, 25.5 months (80% CI, 19.8 to 28.8 months), and 24.7 months, respectively. Th e most common treatment-related adverse event was fatigue (24%, 22% and 35%, respectively). Nineteen patients (11%) experienced grade 3 to 4 treatment-related AEs. Authors concluded that nivolumab demonstrated antitumor activity with a manageable safety profi le, and a phase III trial was started. Recently (25.9.2015) results from the trial were published and nivolumab was found superior to everolimus in the median overall survival (25.0 versus 19.6 months) and also grade 3 or 4 treatment related adverse events were favorable for nivolumab (19% versus. 37%) (Motzer et al. 2015). According to data presented in the ESMO 2014 Madrid and CIMT 2015 Mainz meetings also gastric, head and neck, and bladder cancers have shown good results in ongoing trials. Some cancers, however, seem not to respond so well. For instance hepatocellular, colon and prostate cancer have shown poor responses. To conclude the data on checkpoint inhibitors we can state that immunologic checkpoint blockade with antibodies has resulted in long-term responses with minimal side eff ects in signifi cant numbers of patients with melanoma, lung, kidney, bladder and triple-negative breast cancer, as well as in chemotherapy-refractory Hodgkin disease (Homet Moreno and Ribas 2015). In general, moderate numbers (ca. 10-20%) of grade 3-4 adverse events are reported. Th e effi cacy and safety seems to be somewhat better with PD-1 or PD-L1 blocking antibodies than with the CTLA-4 antibody. However, as CTLA-4 has shown additive effi cacy (and side eff ects) when combined to a PD-L1 blocking antibody, it is likely to become part of the treatments, especially with patients in good general status. No safety signals are available, but most side eff ects have attenuated by time. Common side eff ects include skin reactions, colitis, endocrine disorders, hepatitis and neurological adverse events. Adverse events have appeared usually 3-7 weeks aft er treatment and can last for many weeks. Sometimes high dose corticosteroids have been used, but so far drug related mortality has been low. It is anticipated that approvals by drug regulatory bodies will be forthcoming in several cancers in the next months. Th e potential is enormous. For example, as presented in the ESMO meeting, Madrid 2014, one pharmaceutical company´s (Bristol-Mayer-Squibb) pipeline included the following: CNS, gastric, mCRC, advanced HCC, advanced RCC, prostate, ovarian, NSCLC, SCLC, melanoma, follicular lymphoma and other hematologic tumors, as well as advanced solid tumors. While in the future checkpoint inhibitors will be used in many diff erent cancers, it is also likely that the use will start earlier (at least if cost is not an issue). Even at present, treatment before surgery is considered. While two T-cell inhibitory receptors (CTLA-4 and PD-1) have now been targeted in positive randomized trials, new promising targets are under research (especially TIM-3 and LAG-3). Now that the “brakes” can be deactivated, there are also large numbers of T-cell receptors (including CD26, OX40 and CD137) to be tested for activating the “gas pedal”. Immunotherapies have generally worked especially well with melanoma. Th e reason for this is explained (at the moment) mainly with the immunological nature of melanoma and the large number of mutations

38 Review of the Literature seen in melanoma compared to many other cancers (Vogelstein et al. 2013). More mutations means that the malignant tissue is more likely to be recognized by the immune system, and thus unleashing the brakes from the T-cells leads more likely to tumor regression than with tumors that might not be so well recognized by the immune system. As shown in study IV, biopsies from the tumor present varying amounts of immunological cells. Analysis of these cells, or even just their quantity, might give a hint of what kind of immune treatment this patient might need. In the absence of T-cells, oncolytic virus might be the correct tool, or combining oncolytic virus with checkpoint inhibitors might be rational. Durable complete cancer regressions observed in patients with metastatic melanoma receiving adoptive cell transfer (ACT) has demonstrated the effi cacy of this approach for the treatment of cancer (Rosenberg 2011). For non-necessary organs (e.g. prostate, breast, thyroid, ovary), it might be easier to just target and destroy both normal and cancer cells. Th e feasibility of this approach is demonstrated with lymphoma/leukemia (Kochenderfer and Rosenberg 2013). Another option is to target cancer-testis antigens that are a group of proteins expressed during fetal development and epigenetically silenced aft er this. However, in 20-80% of common epithelial cancers (e.g. bladder, lung, ovary, and liver), they are re-expressed. As an example, NY-ESO-1 seems to be expressed only in cancers and not in normal tissues. Th us, T-cells targeting this antigen have shown dramatic regressions in patients with metastatic synovial cell sarcoma and metastatic melanoma, while other cancer types are been investigated (Robbins et al. 2011). While there are over 100 cancer-testis antigens, some might have low expression in normal tissues or cross reactivity might be present. Th us careful analysis to avoid toxicity is necessary. Th ese antigens could be also used as possible targets in future cancer treatments. Many viruses can be used in cancer immunotherapy and some of the more promising have been discussed (chapters 3.4 and 3.5). To me adenovirus in general and serotype 3 in particular seems promising. While all adenoviruses are effi cient in causing infl ammation and can generate a potent danger signal, stimulating the immune system, serotype 3 might also have some other advantages not present with the most commonly used serotype 5. It has been shown that adenovirus serotype 3 can use CD80 (also known as B7.1) and CD86 (also known as B7.2) as cellular attachment receptors (Short et al. 2004). Th ese receptors are molecules that are present on mature dendritic cells and B lymphocytes, and are involved in stimulating (via CD28) or deactivating (via CTLA4) T-cells (Pardoll 2012). CTLA4 seems to have a much higher overall affi nity for both ligands (CD80 and CD86). It has been suggested that its expression on the surface of T-cells reduces the activation of T-cells by outcompeting CD28 in binding CD80 and CD86, as well as actively delivering inhibitory signals to the T-cell. Maybe this is one reason why Ad3 has shown promising results in cancer patients (study II). However, the exact mechanisms of CTLA4 action are still under considerable debate (Pardoll 2012). Th is hypothesis and its immunological grounds need more study. Another explanation for the good results might be that tumors seem to upregulate CD80/CD86 to bind to CTLA4 and to downregulate T-cell activation (Pardoll 2012). As serotype 3 adenovirus can use these as entry receptors, higher effi cacy might be expected. Th is was demonstrated with preclinical glioma experiments. Here, anti-CD80/CD86 monoclonal antibody was used effi ciently to block Ad5/3 binding (Ulasov et al. 2007).

39 Review of the Literature

To conclude, all cancers contain multiple unique mutations, and future progress in cancer gene therapy will likely result from the immunologic targeting of these mutated proteins (Vogelstein et al. 2013). As known with traditional cancer therapies, best results will probably result aft er fi nding the right combinations. At the moment is seems that checkpoint inhibitors will form the solid base of this new evolving platform.

Below are some promising combinations/new targets that might become part of conventional cancer treatments: INF + checkpoint blockade (PD-1 and/or CTLA-4) Oncolytic virus + checkpoint blockade Oncolytic virus + T cell based therapy Oncolytic virus + checkpoint blockade + T cell based therapy Double checkpoint blockade: PD-1 + CTLA-4 Checkpoint blockade + CD137 co-stimulation Targeting CD40, ICOS, CD160 receptors PD-1 k.o. TILs or engineered CAR/TCR T-cells Modulation of Treg or MDSCs Targeting cancer-testis antigens Other possible targets on T-cells (Mellman et al. 2011): Inhibitory T-cell receptors: CTLA-4, PD-1, TIM-3, BTLA, VISTA, LAG-3. Activating T-cell receptors: CD28, OX40, GITR, CD137, CD27, HVEM.

40 Aims of the Th esis AIMS OF THE THESIS

Th e fi rst objective for the thesis was to create and preclinically test a serotype 3 based oncolytic adenovirus for the treatment of human cancer. Th e next objective was to retrospectively evaluate patients treated with this virus and a quadruple modifi ed virus producing GM-CSF. Th e fi nal objective was to evaluate magnetic resonance imaging and spectroscopy as a method for evaluating oncolytic immunotherapy treatments.

Study I PCR, sequencing and restriction analyses to verify that the cloning and the rescuing of the virus were performed as anticipated. To evaluate the virus in vitro and in vivo, cell killing effi cacy in cancer cell lines and lack of killing in non-cancer cells. Various animal models for anti-tumor activity. Biodistribution and toxicity assays.

Study II Analysis of patients treated with the virus and case series type retrospective epidemiological evaluation. Adverse events, neutralizing antibodies, qPCR for presence of virus. Signs of effi cacy: markers, imaging, survival, case examples, anti-tumor antibodies. Immunological activity, T-cell activation/accumulation.

Study III Evaluation of the potential of magnetic resonance imaging (MRI) and spectroscopy (MRS) in the context of oncolytic immunotherapy. Studying whether MRI or MRS could be used as a biomarker.

Study IV Evaluation of patients treated with an immunostimulatory cytokine (GMCSF) armed quadruple modifi ed oncolytic adenovirus: treatment safety, tolerability, signs of effi cacy. Case series type retrospective epidemiological evaluation.

41 Materials and Methods MATERIALS AND METHODS

Th is section presents a summary of materials and methods used in the studies. For more detailed descriptions, please refer to the original publications.

Cell lines. A repertoire of cell lines has been used in this study, as shown in Table 6. In study I, cell lines representing several major cancer types were used. Most cell lines are commonly used human cancer cell lines. HUVECs (Human Umbilical Vein Endothelial Cells) and fi broblasts were used as cell lines representing normal cells. Some hamster cancer cell lines were also used (Study III and IV), as with the Syrian hamster model we were better able to study immunological eff ects. Th e cell lines were mostly acquired commercially (ATCC, Manassas, VA) or received as kind gift s from other groups. Also primary cells from humans were used.

Table 6. Cell lines used in this thesis

Cell line Study Description/reference/other 911-1C11 I Special cells for Ad3 rescue and culture of non-replicating viruses. Keep in 1mg/ml G418 to select the neomycin resistant cells. (Fleischli et al. 2007) 293-2v6.11 I Special cells for Ad3 rescue and culture of non-replicating viruses. Harbors a ponansterone inducable E4orf6 (Sirena et al. 2005)(Brough et al. , 1996) FSH173WE I Human fi broblasts (non-malignant) HUVEC I Human umbilical vein endothelial cells (non-malignant) Skov3-luc I Human ovarian CA. Firefl y luciferin, emit photons aft er giving luciferase and can be imaged in vitro and in vivo Skov3.ip1 I, IV Human ovarian adeno CA PC3-MM2 I, IV Human hormone refractory metastatic prostate CA A549 I., IV Human lung CA CAMA-1 I Human breast adeno CA 293 I Human embryonic kidney cells LNM-35/EGFP I Human lymphagenous metastatic subline of lung large cell CA PANC-1 I Human pancreas CA HTC116 I Human colorectal CA ACHN I Human kidney cell CA TF-1 IV GMCSF dependent erythroleukemia cells HapT1 III, IV Hamster pancreatic cancer DDT1-MF-2 III Hamster leiomyosarcoma

Adenovirus constructs. Adenovirus constructs are shown in Table 7. We used the wild type adenovirus serotype 3 and 5. We also used non-replicative and replication competent versions of these viruses that had a transluminant transgene (luciferase, luc or green fl uorescence protein, GFP) cloned into the virus. Th ese viruses could be imaged in vitro and/or in vivo. Viruses that

42 Materials and Methods were serotype 5 but the fi ber knob (the main primary attachment receptor) was from serotype 3 (Ad5/3 constructs) were also used. Most oncolytic viruses were cloned in our lab, while some viruses were received from other groups.

Table 7. Viruses used in this theses

Serotype Knob Targeting Tumor specifi city Arming serotype Ad5 wt 5 5 CAR none none Ad5-D24-GMCSF 5 5 CAR 24 bp deletion in GMCSF (Cerullo et al.) E1A2) Ad5/3-luc1 53Partly replication luciferase (Bilbao et al. 2000) DSG-21) defi cient virus for imaging Ad5/3-D24 53Partly 24 bp deletion in none (Kanerva et al. 2003) DSG-21) E1A2) Ad5/3-D24-GMCSF 53Partly 24 bp deletion in GMCSF (Koski et al.) DSG-21) E1A2) Ad5/3-hTERT-E1A 53Partly hTERT none (Bauerschmitz et al. DSG-21) 2008)4) Ad5/3-E2F-D24-GMCSF 53Partly E2F1 promoter and GMCSF (Hemminki et al. 2014) DSG-21) 24 bp deletion in E1A 2) Ad3 wt (Fleischli et al. 3 3 DSG-2 none none 2007) Ad3-hTERT-E1A 3 3 DSG-2 hTERT promoter3) none (Hemminki et al. 2012) Ad3-CMV-eGFP 3 3 DSG-2 replication eGFP for (Fleischli et al. 2007) defi cient virus imaging Ad3-CMV-luciferase 3 3 DSG-2 replication luciferase (Fleischli et al. 2007) defi cient virus for imaging Ad3-GFP (Wang et al. 3 3 DSG-2 none GFP for 2011) imaging wtAd3 (Wang et al. 2011) 3 3 DSG-2 none none 1) In vitro studies suggest that changing only the serotype knob is not enough to get high affi nity binding to DSG-2 (Wang et al. 2011) (Rosewell 2012) 2) Replication in cells with a defi cient Rb/p16 pathway (a hallmark of cancer) 3) Replication in cells with active telomerase (a hallmark of cancer) 4) Named here as Ad5/3-hTERT-gp

Virus cloning. While cloning of a modifi ed virus can certainly be problematic, the basic principle is relatively simple, and sometimes successful cloning and viruses can be done in just a couple of months by only one person. Everything starts from rationale planning, and for this the DNA sequence needs to be known. Today sequencing is relatively cheap enabling endless possibilities. Commonly, a BAC (bacterial artifi cial chromosome) is used. Here, the adenovirus DNA is stored

43 Materials and Methods to a large plasmid, and it can be multiplied in bacteria, and DNA can be then purifi ed. Typically, a large amount of pure DNA is essential for successful cloning. Th ere are some hundred or so diff erent enzymes that cut DNA from certain locations, for instance, the sequence GAATTC is recognized and cut by EcoRI restriction enzyme if present. Nowadays it is relatively common to order sequences of DNA (usually provided in a smaller plasmid) from diff erent companies, as custom made DNA pieces tend to be relatively cheap. Th us, traditional PCR based making of inserts is becoming more uncommon. While other enzymes are used to cut, others are used to “glue” pieces together. In all simplicity we need the starting viral DNA and a piece that shall be inserted and some enzymes. Finally, when the DNA is ready, it is delivered to cells, where the DNA changes the infected cell into a virus factory.

In vitro cytotoxicity assay is a commonly used assay when virus cell killing needs to be assessed. We used this assay especially in study I, but also in studies II-IV. Here approximately 10 000 cells/well are plated to a 96-well plate. Th e following day the cells are infected with diff erent concentrations of virus (e.g. 0.1-100 virus particles per cell). Th e infection is commonly done in triplicates/quadruplicates. Plates are then observed daily and growth media changed if needed. Depending on the cell line and virus, some days later cell viability is analyzed with a MTS assay. A substance is added to the cells and, if still alive, the mitochondria of the cells will uptake the added substance, and this can be visualized some minutes later with an automated plate reader.

Progressive infectivity assay. Aft er making a large scale adenovirus stock, the quality and quantity of these viruses is typically assessed with two methods. First, an optical reading of the diluted stock is gained and as DNA absorbs light at 260nm wave length, an approximation of the virus quantity can be counted. Aft er this, a PFU assay is typically performed. Diff erent dilutions of the virus stock are incubated with 293 cell line cells, and fi ve days later the plates are read and an approximation of virus quantity is gained. When the PFU value is compared to the optical reading, an approximation of the quality is gained. Th is system has been validated for serotype 5 viruses. However, our studies showed that serotype 3 seems to enter cells, replicate, and kill the cells in a somewhat diff erent way. Th us, we generated an assay we named progressive infectivity assay. We believe this assay might be better in evaluating the quality and quantity of serotype 3 viruses than the PFU assay validated for serotype 5 viruses. In this assay, there is no termination on day 5. Th e virus dilutions on the plates are observed at a few day intervals until no more cell death is seen. Growth media is changed or added when necessary. Th e lowest concentration that kills cells represents a dilution where functional virus is present.

Animal experiments were used in all studies (I-IV). In most studies, various human cancer cells were injected (s.c. or i.p.) to immunocompromised mice. Th e cells grow to measurable tumors in some days or in a few weeks. Th en virus treatments were performed and tumor growth was measured by various methods (manual measurement, luciferase imaging, MRI/MRS imaging). Immunocompetent mice were also used in experiments when human tumor cell lines were not needed (e.g. when virus biodistribution and toxicity or neutralizing antibodies were evaluated). Finally, in the latter two studies (III-IV), the Syrian hamster model was used. Adenovirus serotype 5 replicates in some degree in this immunocompetent model as opposite to mice. Th us, hamster cancer can be treated with human oncolytic viruses in an immunocompetent model.

44 Materials and Methods

Statistical analyses. By far the students T-test was the most common statistical method used (study I-IV), while also Mann-Whitney’s non-parametric test and F-test was used in study I with the help of a statistician.

Neutralizing antibody experiments were used in various forms in studies I-II and IV. Th e basic idea, as defi ned by the classical serotyping of viruses, is that when an immunocompetent human/ animal is acquainted with a virus of a certain serotype, it produces neutralizing antibodies some weeks later against this serotype. Th ese antibodies can be collected by venesection, also heating the serum antibodies can be experimented with in vitro and in vivo. When antibodies are present at a suffi cient quantity, they neutralize the virus, and infection (at least in vitro) is blocked. Th is blocking is serotype specifi c.

Human interferon-gamma enzyme-linked immunosorbent spot (ELISPOT) assay was used when we assessed tumor T-cell responses from the peripheral blood (study II, IV). PBMCs (peripheral blood mononuclear cells) were isolated by gradient centrifugation (Percoll, Sigma). Th en Mabtech´s protocol was followed. In brief, the PBMCs were plated and stimulated with various peptides. Typically, the cells were stimulated with viral or tumor associated peptides, and reactive T-cells (the ones that produced interferon-gamma) formed spots on the plate aft er staining and could be counted.

Histology and immunohistology was used in all studies. Normal organs were sometimes analyzed, but more typically tumors were collected, parafi nized, cryotomised, stained and analyzed. While hematoxylin-eosin staining was the routine, many diff erent stains were used to gain information, especially from immunological changes. Calprotectin was successfully used to indicate immunological activity, and various CD stains shined light on the processes at the tumor site. While the staining was sometimes done by us, a pathologist analyzed the results. qPCR and PCR were used in all studies (I-IV), conditions and primers are described in the articles.

Sequencing was used to control the results of cloning. Primers with the virus DNA were taken to the sequencing unit.

Advanced Th erapy Access Program (ATAP). Patients described in this thesis (study II-IV) have been treated in the ATAP program. In study II-IV we discuss treatments of 42 patients and concentrate in two viruses, while the total amount of patients treated in the ATAP program was 290 and 10 diff erent viruses were used. Th e diff erence between a trial and a treatment (Hemminki 2015) is shown in table 8.

45 Materials and Methods

Table 8

Trial Treatment Predetermined protocol Patients treated case by case Strict inclusion criteria No absolute inclusion or exlusion criteria Sometimes placebo included No placebo May involve interventions without benefi t to Only procedures directly relevant for the patient the patient, for example biopsies are performed May have a sponsor with commercial interests Cost paid by patient, community, insurance Clinical trials are tighly regulated and very Few regulations apply (559/1994, 15§ in expensive Finland), exept for “advanced therapies” (EU 1394/2007) May benefi t the society and facilitate products Goal to help the patient eventually available to millions May or may not benefi t the patient Limited benefi t for society

ATAP was set up to off er oncolytic virus treatments to patients that did not have access to clinical trials (Hemminki 2012). It is based on the European Commission Advance Th erapies Regulation (EC/1394/2007), which determine rules for patient-by-patient use of gene therapy and cell products. On one hand, the goal was to apply regulation in an area where it had been absent previously. On the other hand, scientifi c and medical progress was encouraged. In ATAP, each patient is monitored for safety, effi cacy and survival. All data is reported in peer reviewed journals and to the Finnish regulatory authority (FIMEA). Although new drugs always entail a certain risk factor, ATAP attempts to balance this against the risk of death posed by incurable and progressing tumors (Hemminki 2012). Overall, 10 diff erent viruses were used in a total of 290 patients who had disease incurable with current therapies. Th e typical patient had a tumor progressing aft er all routine therapies had been exhausted. Each of the 821 treatments was individually designed, typically employing intratumoral injection guided by ultrasound or computer tomography. Intrapleural, intraperitoneal and intravenous injection were also used depending on the location of the patients’ tumors. Th e patient population in ATAP resembles a typical Phase 1 population in that patients have incurable advanced solid tumors progressing aft er routine treatments, and in fact most patients have gone through multiple regimens of chemotherapy. Patients sign informed consent. Aft er treatment, patients are monitored 24 h in the hospital and thereaft er as outpatients. As required by the philosophy of individualized therapy, each patient was treated according to our best knowledge, taking into account what we knew about their disease, what we had learned about our viruses in the laboratory and in animals, and – probably most importantly – from previous patients. Each patient taught us something and sometimes one patient taught us more than a thousand mice. ATAP had only one goal: trying to help the patient. However, as medical professionals, we were interested in the results and in ways to improve them, and thus we tried to learn as much as possible. Th e non-binding inclusion criteria for ATAP are described in table 9.

46 Materials and Methods

Table 9. Non-binding guidelines for inclusion of patients in ATAP

Inclusion criteria Exclusion criteria - Solid tumor - Confi rmed brain met. or glioma - Refractory disease = failed treatments for - Organ transplant, HIV which there is strong scientifi c evidence - Severe cardiovascular, metabolic or pulmo- - Good performance status: WHO 0-2 (3-4 nary disease were also safe but less effi cacy seemed to re- - Elevated serum bilirubin sult) - Serum AST or ALT > 3x normal - Written informed consent - Th rombocytes <75

All viruses (see table 10) used in ATAP are designed to work in most tumors, and they are second or third generation viruses – hence replication competent and some viruses are also armed. Each virus was carefully tested preclinically before patient treatment. All of these viruses were designed so that replication takes place primarily in tumor cells (hTERT, Cox2 and E2F promoters or D24 deletion, or combination thereof). Th ese modifi cations make the viruses more safe (Toth and Wold 2010). However, it was rapidly discovered that the safety in patients with all constructs was good, and it became also relatively quickly clear that oncolysis alone was unlikely to cure patients with advanced tumors. Th us improving selectivity seemed less critical than improving effi cacy. Th erefore, we rapidly moved to armed viruses and utilization of drugs such as low-dose cyclophosphamide – useful for counteracting regulatory T-cells - to enhance effi cacy. Various modifi cations to the virus capsid were done to achieve better transduction of tumor cells, as downregulation of CAR is suggested as a problem in oncolytic virus treatments (Hemminki et al. 2011). Some of the viruses were targeted to integrins and some to the Ad3 receptor by changing the knob of the Ad5 virus fi ber to an Ad3 knob (Ad5/3 viruses). Integrins (Pesonen et al. 2012) and serotype 3 receptors (Wang et al. 2011; Hemminki et al. 2012) (DSG-2) are suggested to be abundant in patient tumor tissues, thus better effi cacy with these viruses was anticipated. Finally, a completely serotype 3 oncolytic adenovirus (Ad3-hTERT-E1A, study I, II) was made to avoid anti-Ad5 immunity and to achieve stronger binding to DSG-2. Good results associated with this unarmed virus seem to suggest that Ad5 might not be the only feasible serotype for cancer therapy. A particularly attractive aspect of DSG-2 binding is the synergy with monoclonal antibodies (Wang et al. 2011). Our preliminary patient data seems to corroborate the notion that this would be interesting for formal testing (see study II table 4, patients treated with concomitant trastuzumab).

47 Materials and Methods

Table 10. Viruses used in ATAP

Serotype Targeting Tumor specifi city Arming Ad5-D24-GMCSF 5 CAR 24 bp deletion in E1A2) GMCSF (Cerullo et al.) Ad5-RGD-D24 (Pesonen 5 CAR and 24 bp deletion in E1A2) No et al. 2012) Integrins Ad5-RGD-D24-GMCSF 5 CAR and 24 bp deletion in E1A2) GMCSF (Pesonen et al. 2012) Integrins ICOVIR-7 (Nokisalmi et 5 CAR and E2F1 promoter & 24 No al. 2010) Integrins bp deletion in E1A2) Ad5/3-Cox2L-D24 5 Partly DSG-21) Cox2L promoter & 24 No (Bauerschmitz et al. bp deletion in E1A2) 2006) Ad5/3-D24-GMCSF 5 Partly DSG-21) 24 bp deletion in E1A2) GMCSF (Koski et al.) Ad5/3-hTERT-hCD40L 5 Partly DSG-21) hTERT promoter3) CD40L (Bauerschmitz et al.) Ad5/3-E2F1-D24- 5 Partly DSG-21) E2F1 promoter& 24 bp GMCSF GMCSF (Ranki 2012) deletion in E1A 2) Ad5/3-D24-hNIS 5 Partly DSG-21) 24 bp deletion in E1A2) hNIS (Rajecki et al. 2011) Ad3-hTERT-E1A 3 DSG-2 hTERT promoter3) No (Hemminki et al. 2012) 1) In vitro studies suggest that changing only the serotype knob is not enough to get high affi nity binding to DSG-2 (Wang et al. 2011) (Rosewell 2012) 2) Replication in cells with a defi cient Rb/p16 pathway (a hallmark of cancer) 3) Replication in cells with active telomerase (a hallmark of cancer)

Adverse reactions of all treated patients were collected. Typically, the patients received fl u- like symptoms that alleviated by themselves in some days. Injection site pain and leukocytopenia were also commonly observed. Th e latter, and in particular “lymphopenia”, may in fact refl ect redistribution of white blood cells from the blood to target organs, including tumors, and is thus not an adverse event but in fact part of the mechanism of the therapy. Mild decrease in hemoglobin observed on the next day was thought to relate to the fl uids the patients received aft er treatment. Another possible explanation is that the virus binds to erythrocytes, and some of them are subsequently cleared by the reticular endothelial system (Seiradake et al. 2009). In contrast to high dose administration in animal models, liver enzymes were seldom elevated. In general, treatments were found safe and well tolerated. Each patient treatment in ATAP was designed using all the information accumulated so far. In this way, the bench-to-bedside-and-back cycle became extremely swift , as basically every patient represented one cycle of the process. Indeed, the cycle was frequently shortcut from bedside to the next bedside. On the contrary, with phase I trials (for those who have the money to perform them), each cycle takes typically several years per cycle. If the treatment plan is found suboptimal, it can be hard to correct during the trial, leading to unsatisfying results and in some cases exposure of patients to ineff ective drugs. Th us the faster learning process is in the interest

48 Materials and Methods of patients who do not have the years to wait for the cutting edge treatments to become available in pharmacies. Th e 4.5 years of ATAP were an exciting time, since we saw the approach improve rapidly. At the same time, we believe that we have helped many individual patients fi ght their deadly disease. In the future, we believe that we are more likely to be able to plan successful trials and thus minimize patient exposure to ineff ective interventions. A key aspect of ATAP, however, is the focus on each patient. Formal trials are needed to corroborate the results, and in particular randomized studies would be critical to assess the magnitude of the benefi t, if present.

49 Results and Discussion RESULTS AND DISCUSSION Study I Background of study I. Publications describing attempts to treat cancer with viruses date back hundred years. Only half of this time have pioneers of this fi eld been able to purify, grow and fi nally also visualize the viruses. Advances in molecular technology during the last decade have enabled scientists to manipulate the genome of viruses and other organisms. In the fi eld of gene therapy, adenovirus 5, the best known adenovirus, has been rationally manipulated in many diff erent ways. Genes have been deleted and inserted, some viruses could replicate while others could not. Some viruses were armed with cytokines and other genes that could be therapeutic. Growing amount of data suggested, however, that there was a major problem in using serotype 5 adenovirus in cancer gene therapy, as the primary receptor of this virus was generally low in advanced cancer (Kanerva et al. 2002; Volk et al. 2003). First solution to this problem was to use chimeric viruses. Th ese viruses had a fi ber knob region from another virus. One potent modifi cation was the Ad5/3 chimeric virus, where the knob was from serotype 3 adenovirus. Th is virus was shown to be more potent in killing cancer cells in many publications (Kanerva et al. 2002; Volk et al. 2003; Bauerschmitz et al. 2006; Koski et al. 2010). However, it seemed that a fully serotype 3 oncolytic virus might be even better, and also the immunological response should be diff erent, as the whole virus is of another serotype. In 2005, the sequence of this virus was published, and this enabled the rational design of the new virus. Serotype 3 contains only 63% of identical genome with serotype 5 virus, therefor many other diff erences were also expected. Th e primary receptor was not known at this time, but a mass of data suggested that it should be highly expressed in advanced cancer (Tuve et al. 2006). In study I, we constructed probably the fi rst non-serotype 5 based oncolytic adenovirus, Ad3-hTERT-E1A. Th e replication of this virus is controlled by a human telomerase reverse transcriptase (hTERT) promoter. Telomerase activity is considered a critical step in carcinogenesis and most tumors are believed to have high telomerase activity (Fujiwara et al. 2007). Th e replication of Ad3-hTERT-E1A should thus be restricted to cancer cells. In vitro experiments. Aft er cloning the Ad3-hTERT-E1A construct backbone, the virus was prepared. Th e DNA was transfected to 911-1c11 cells and the subsequent amplifi cation was done in 293-2v6-11 cells. Final large scale amplifi cation was then perforemed in commonly used A549 (lung cancer) cells. In our hands, the kinetics of the virus replication seemed slower than with the commonly used serotype 5 virus. While Ad5 viruses caused CPE (cytopatic eff ect, a typical adenovirus induced cell morphology change seen in a microscope) in a couple of days, it seemed that serotype 3 virus needed a much longer time. Slower replication cycle might be due to the genetic alterations done to the virus (insertion of hTERT promoter in the place of the naive promoter of the E1A region). However, the wild type serotype 3 had the same kinetics in our experiments. Th e slower kinetics was not noted by our colleagues in Zurich or Seattle also working with adenovirus 3. Still today the reason to this fi nding is somewhat unclear. However, the reason might be due to the dodecahedral particles produced in a massive excess (Lu et al. 2013) by the serotype 3 virus. It might be that the cesium chlorite gradient purifi cation we performed resulted in collection of mainly the Ad3 virus or mainly the dodecahedra, while purifi cation in other groups might lead to collection of other compartments. Also the quantifi cation of the purifi ed virus might aff ect the results. We decided to use optical quantifi cation (based on light and DNA interactions, dodecahedra do not contain DNA), while other ways to quantify the virus amount

50 Results and Discussion per volume are also available; for example, plaque forming units. We analyzed the virus by many diff erent ways including PCR, qPCR, restriction analyses, electron microscopy, and even a new assay called the progressive infectivity assay was created for the virus. We also took the virus to the University Hospital lab to be quantifi ed by their qPCR system that should recognize serotype 3 adenovirus. For some reason the results were negative. Later it became evident that the Hospital qPCR was unsuitable to detect serotype 3 and needed recalibration. Finally, aft er a few learning years and a large number of experiments, we were confi dent that we really had the right product. Oncolytic potency of the Ad3-hTERT-E1A. Monolayers of diff erent cancer cell lines were used to evaluate the cancer cell killing potency of the serotype 3 viruses. Ad3 wild type and the modifi ed Ad3-hTERT-E1A virus were able to kill all cancer cell lines in vitro that were tested (prostate, lung, colon, ovarian, lung, breast, pancreas). No signifi cant diff erence was noted between the wild type and the oncolytic virus, suggesting that the engineering of the virus had not slowed down or deactivated the virus. Cell killing of non-malignant cell lines (representing normal cells) was signifi cantly reduced with the oncolytic Ad3-hTERT-E1A virus compared to the wild type virus, suggesting safety of the modifi cation. Th e in vivo work thus indicated that the virus seems to work as designed. It retains its effi cacy in cancer cell killing compared to the wild type, while in normal cells the cell killing ability is reduced, thus making the virus safer that the wild type. Th is might be an important feature, although adenovirus wild type 3 is not known to cause mortality in healthy adults. Th e quantity of the virus used and the immunological imbalance in cancer patients might, however, pose a greater risk for complications. In vivo experiments. Th e oncolytic potency of the Ad3-hTERT-E1A was tested in mice bearing prostate cancer (PC-3MM2) or lung cancer (A549) xenograft s subcutaneously. A clear reduction of tumor growth was seen compared to PBS treated animals. Interestingly, although in vitro the virus seemed somewhat slower that the serotype 5 based controls, in vivo experiments indicated that serotype 3 was somewhat better than the controls. As subcutaneous tumors might not be optimal surrogates of human cancer, we utilized also an orthotropic ovarian cancer (SKOV3-luc) model. Here, the luciferase producing tumor cells were injected in the peritoneum to induce intraperitoneal disseminated carsinomatosis which could be imaged. Similar results as with the subcutaneous models were observed. Interestingly, one of the six mice treated with the Ad3-hTERT-E1A seemed completely cured as no signal could be detected at 120 days. Th e autopsy was also found normal. One interesting feature of the intraperitoneal model was that virus injections to the peritoneum were given on the fi rst three imaging time points (days 3, 7 and 14). A clear reduction of signal was discovered aft er the fi rst injection, some further benefi t was seen aft er the second injection, while the third injection seemed to lose its eff ect. We believe this is due to intracellular resistance mechanisms not well understood and might be explained with changes in interferon signaling (Liikanen et al. 2011). As the last experiment in study I, we wanted to demonstrate that neutralizing antibodies against serotype 5 would block the infection of serotype 5 but not serotype 3. We made antibodies in immunocompetent mice against serotype 5 virus and collected the serum aft er a month. Later the heated serum was used in the intraperitoneal model described above. Th e serum blocked the anti-tumor eff ect of serotype 5 virus and partly also the eff ect of the chimeric 5/3 virus. As expected, no blocking of the Ad3- hTERT-E1A was seen with the anti-Ad5 serum.

51 Results and Discussion Study II Background of study II. Preclinical testing suggested that the Ad3-hTERT-E1A is a potent virus with replication controlled by a telomerase promoter, providing added safety. A totally new serotype in oncolytic virotherapy could be of benefi t as anti-viral immunity might restrict the effi cacy of repeated serotype 5 usage. Other publications had suggested that serotype 3 virus has also other potentially important capabilities, such as inducing epithelial-to-mesenchymal transition, opening of the tight junctions and revealing important receptors that might potentiate the usage of, for example, Herceptin and Erbitux (Beyer et al. 2011). While in theory epithelial- to-mesenchymal transition might pose a risk of generating metastases, no evidence or signals to this regard has to our knowledge been provided by the scientifi c community. One publication also indicated that the virus could potentiate the eff ect of chemotherapy (Beyer et al. 2012). Th e receptor of the Ad3 was discovered, and it was known to be high in advanced tumors (Tuve et al. 2006; Wang et al. 2011). Th e Ad3 is also known to bind to CD80/CD86 in antigen presenting cells. Th e CD80/CD86 on the other hand bind to CTLA-4 (high specifi city) and CD28 (low specifi city) having an important role in both activating T cells and inducing tolerance. Th us, in theory, Ad3 might aff ect the immune system also directly via the CD80/86 mediated binding.

Preclinical part of study II Biodistribution and toxicity experiments in study II. Before starting patient treatments, we wanted to gain more information of the virus in an immunocompetent model. In biodistribution experiments, mice were intravenously injected with the serotype 3 virus and six hours later the animals were sacrifi ced and organs collected for qPCR analysis. Virus seemed to accumulate mainly in the liver, spleen, lungs, blood clot and bone marrow. Low quantities of virus were detected in other organs. Most viruses seemed to bind to platelets or PBMCs in blood. Th e biodistribution in rodents resembled results gained with Ad5/3 and Ad5 viruses. Th e signifi cance of these results for humans is not known, as it seems that mice lack the relevant high affi nity adenovirus receptors present in humans. Acute immunological toxicity. To evaluate acute immunological toxicity, mice were injected with Ad5, Ad5/3 and Ad3 virus. Blood samples were collected six hours post injection and subsequently analyzed for cytokines. Ad3 was shown to elicit higher Rantes, interleukin-6 and tumor necrosis factor alpha than control viruses, indicating serotype 3 as a potent activator of mouse macrophages and T-cells. Cytokine values were well below values associated with toxicity. Toxicity assay was performed by comparing PBS, Ad3wt, Ad3-hTERT-E1A, Ad5/3-hTERT- E1A, Ad5wt and Ad5/3-D24. Mice in the two latter groups appeared ill 72 h aft er the intravenous high dose virus injection and all mice were sacrifi ced. Livers of the animals were macroscopically yellow, while all other organs were normal. All organ samples were analyzed by a pathologist and no toxicity was noted in other organs than the livers of the Ad5 and Ad5/3 viruses. Liver enzymes were 36-176 fold elevated in these mice, while only minor elevations were seen with Ad3wt and Ad3-hTERT-E1A (three-fold) and Ad5/3-hTERT-E1A (fi ve-fold) compared to the mock group. According to this immunocompetent rodent model, Ad3 seems not to pose signifi cant safety issues. However, the caveat is that the important adenovirus receptors are not present in murine models.

52 Results and Discussion PaƟ ent treatments in study II Using the epidemiological method of retrospective registry research, the fi rst 25 patients treated in the ATAP program (see materials and methods) with Ad3-hTERT-E1A were included in this case series. Patients were heavily pretreated with a median of three chemotherapeutic regimens and represented nine diff erent cancer types. Median age of the patients was 60 years. About half of the patients had been pretreated with serotype 5 based oncolytic viruses. Treatments were started with a relatively low dose of 1010VP and the dose was gradually increased to over 1012VP. Since the quantity of neutralizing antibodies (NAbs) against serotype 3 was believed to be small, many patients received most (or all) of the virus intravenously. Rest of the virus was injected into the tumor. Adverse reactions were similar to the ones seen with Ad5 or Ad5/3 patients. Flu like symptoms were present in most patients and no serious adverse events leading to patient hospitalization due to treatment were encountered. Many patients received a decrease in lymphocytes aft er treatment. Interestingly, the patients that had been pretreated with a serotype 5 oncolytic adenovirus showed a prolonged lymphodepletion, up to three weeks, in the blood, suggesting a diff erent immune response compared to the oncolytic treatment naive patients. In an attempt to optimize the treatment for each patient, we assessed antiviral and antitumoral T-cells in patient blood. Clear changes in the T-cell reactivity when stimulated by Ad3 hexon (antiviral) or various tumor associated antigens (antitumoral) were seen before and aft er Ad3-hTERT-E1A treatment. As we were analyzing cells from blood, not from the tumor, defi nite conclusions, other than that something is happening, were diffi cult to draw. However, we hypothesized that traffi cking of white blood cells to tumors might be a possible explanation for the fi ndings (Kanerva et al. 2013). Prolonged virus appearance in blood was detected with the Ad3-hTERT-E1A. Especially the patients who had received high dose (over 1012VP) of virus, eight out of nine of them had still measurable virus in blood at three and six weeks post therapy. From low dose (under 1012VP) patients, two out of six showed virus in blood at three and six weeks post treatment. Commonly, Ad5 based virus treated patients have been negative for virus some days aft er the treatment (Escutenaire et al. 2011). At this point we do not know if the reason for the diff erence is technical (qPCR more sensitive) or if it has true biological background. Virus in blood compartments. We detected usually ten times more virus in the clot compared to the serum, but there was immense variance sometimes also within the same patient. From the two fi rst treated patients, blood samples were taken at 10 min, 2 h, 6 h and 20 h time points with the rationale that this would help optimize the next treatment of the patient. However, most of the virus was cleared from the blood in minutes. No virus was detected in red blood cells and most of the virus was seen in PBMCs and plasma. Neutralizing antibody (Nab) assay. Most patients had low titer (median 256) of NAbs against Ad3 virus, indicating a previous wild type infection. At baseline no diff erence was detected between Ad5 pretreated and non-pretreated patients in regard to Ad3 NAbs. In contrary, Ad5 NAbs were signifi cantly increased in serotype 5 pretreated patients. Aft er Ad3-hTERT-E1A treatment, Ad3 NAbs increased signifi cantly while no change was seen with serotype 5 NAbs. Th e results were as expected. Cross neutralization between the serotypes was not seen. Evidence of antitumor activity. Most patients were treated in a serial manner, including three treatments before a new imaging. While some evidence of an anti-tumor eff ect was seen with 15/23 (65%) patients, it cannot be concluded that Ad3-hTERT-E1A was the reason for the results since the patients had also received other viruses. Keeping with the philosophy of ATAP, patient benefi t was the only goal, and scientifi c dissection of the data was not a factor in treatment 53 Results and Discussion design. One patient (S171) was imaged before and one month aft er Ad3-hTERT-E1A treatments without other treatments in between. Here, a 30% reduction in the injected tumor volume was seen. In regard to tumor markers, 15 patients had elevated tumor markers before treatment, and this was compared to a value measured three weeks aft er the Ad3-hTERT-E1A treatment. Of these 15 patients, 11 (75%) showed signs of antitumor activity (decreasing or stable markers). Two patients (R217 and R263) were noted via normalization of tumor markers (CEA and CA15- 3, respectively) that were elevated before Ad3-hTERT-E1A treatment. Patient K260 had a high CEA level of 854 before Ad3-hTERT-E1A treatment, and this was reduced to 460, suggesting response. Two more patients showed decreasing tumor markers while fi ve patients showed stable marker values aft er treatment. Four patients showed increase in markers suggesting progressive disease. Th us, of the 16 patients that could be evaluated with CT or tumor markers (one by CT, rest by tumor markers), 12 (75%) indicated disease control (stable markers or better), suggesting potential of the serotype 3 unarmed virus. Interestingly, fi ve out of the six patients that received Ad3-hTERT-E1A only intravenously (no intratumoral injection) showed stable markers or better in the evaluations. Th is might suggest that the Ad3-hTERT-E1A could be successfully delivered as an intravenous injection. Survival. Patients that showed stable markers or better (N=15) survived 295 days (median), while other patients (N=8) survived signifi cantly shorter (P<0.001) for 108 days. If this is due to the virus treatment can not be assessed with this retrospective patient series. Patient examples. Of the 25 treated patients some interesting fi ndings were noted. Altogether fi ve breast cancer patients were treated; all showing decreasing or stable markers aft er treatment. According to preclinical work, serotype 3 adenovirus should open the tight junction and reveal the receptor for trastuzumab Her2/neu (Beyer et al. 2011; Wang et al. 2011). Two of the patients were on trastuzumab, but progression was noted before the Ad3-hTERT-E1A treatment. Both patients showed a decrease in tumor markers (CEA 11->5 and Ca12-5 15->10) and both experienced a long survival (alive at 310 and at 630 days). At the time of writing (19.10.2015) the latter patient was alive at 2135 days while the survival of the other patients is not known as the patient is not Finnish. Although little can be concluded from two patients, the results indicate that there might be synergy with these treatments. Durable CT response with patient S171. Patient S171 with a malignant fi brous histocytoma had been pretreatment by multiple operations and chemo regimens, but the disease was progressing before initiation of virus therapy. Th e patient was fi rst treated with fi ve serotype 5 based adenovirus treatments over a six-month period. Aft er a single cycle of Ad3-hTERT- E1A virus, the patient showed a tumor reduction of 30% aft er one month and 44% reduction aft er four months of the injected largest tumor. While at four months the injected tumors were stable (N=5, -0.4%), non-injected tumors had grown and progressive situation was diagnosed. However, there seemed to be anti-tumor eff ect (graded stable disease according to RECIST) aft er the single Ad3-hTERT-E1A injection. During this four mouth period the patient did not receive any additional treatments and was feeling physically well (over 10km daily cross country skiing in Lappland). High virus replication of the Ad3-hTERT-E1A virus was noted with patient N227. Th e patient was a 3-year-old with neuroblastoma. Before the Ad3-hTERT-E1A treatment she had gone through radiation therapy and seven cycles of chemotherapy, but the NSE tumor marker was still rising and bone marrow biopsies indicated progression in the immunofl uorescence analysis. Th e patient was rstfi treated with a serotype 5 based virus and then with Ad3-hTERT- E1A. High virus titers were noted the following days (suggesting virus replication), and some

54 Results and Discussion virus was detected weeks (at three and six weeks) aft er the treatment. NSE decreased from 25 to 21 at three weeks and at six weeks the bone marrow biopsy was found tumor free. Long term virus presence in the blood was detected in many patients. One interesting patient was a 58-year-old man with pancreatic cancer (H305). He was operated a year before and multiple chemotherapeutic regimens had been tried, but progression of the cancer was noted. He received Ad3-hTERT-E1A as the fi rst and second virus treatment. Virus was detected three weeks aft er the fi rst treatment, but at six weeks, before the serotype 5 based virus treatment, he was negative for the Ad3-hTERT-E1A. However, at 3 and 19 days aft er the serotype 5 virus treatment, Ad3-hTERT-E1A re-emerged to the blood according to the qPCR. Inspired by this fi nding we analyzed all the samples we had and noted that with 6/9 patients the Ad3-hTERT- E1A was detected from blood aft er a later serotype 5 virus treatment. Next, we analyzed patients that were fi rst treated with a serotype 5 virus and then with Ad3-hTERT-E1A and found that 5/7 patients showed re-emerging serotype 5 virus aft er the treatment with a new serotype. Th e most logical explanation to me is that aft er initial oncolysis the virus gets turned off by the cancer cells somehow, for instance by interferons produced by the tumor stroma (Liikanen et al. 2011). Th en, when the serotype is changed, it also activates the pre-used virus that is waiting in the tumors cells, or the virus/virus DNA is released due to cell lysis. Th ese fi ndings suggest that it might be rational to change the serotype or the virus aft er some cycles of therapy (Sarkioja et al. 2008). Along with the results seen with the prolonged lymphocytopenia (traffi cking to the tumor? (Kanerva et al. 2013)) and other T-cell changes, we hypothesize that priming with serotype 5 virus and then boosting with a serotype 3 virus might be a potent immunological treatment modality especially when combined with trastuzumab, EGFR-targeting cetuximab, tumor infi ltrating lymphocyte therapy and/or checkpoint inhibitors.

Study III Background of study III. We and others have seen that some patients seem to benefi t from oncolytic virus treatments as well as some other immunological treatments. However, at the moment there are no good methods of evaluating who benefi ts and who does not. Traditional imaging based methods, such as CT might not be optimal for immunological treatments as tumor swelling might happen before regression (Wolchok et al. 2009; Hoos et al. 2010; Prieto et al. 2012). PET-CT has also been reported to give false positive results (Kuruppu et al. 2007; Focosi et al. 2008; Koski et al. 2012) as immunological activity increases glucose intake in, for example, lymph nodes. In study III, we wanted to assess magnetic resonance imaging (MRI) and spectroscopy (MRS) in evaluating oncolytic treatments. Syrian hamster was used as the model, as it is known to be semi permissive to human Ad5 and GM-CSF. Hamster pancreatic carcinoma (HaP-T1) and hamster leiomyosarcoma (DDT1-MF-2) were injected subcutaneously to Syrian hamsters, which were later treated with an oncolytic virus or with a similar virus that also expresses GM-CSF. Th e hamsters were followed with MRIS. As expected, tumors treated with the GM-CSF expressing virus slowed down the tumor growth the most. With the unarmed virus, the eff ect was not as evident and only a nonsignifi cant trend of tumor growth inhibition compared to the PBS was seen (Study III, Figure 1). Similar results were observed with the carcinoma and the sarcoma model. Interestingly, it seemed that some of the hamsters responded better to the treatment than others. In the carcinoma hamsters this was more evident; aft er initial growth some tumors started to shrink and others continued to grow. Th e tumors that started to shrink (N=5) were termed as responders while others were termed 55 Results and Discussion nonresponders (N=5). Most of the responders were from the armed GM-CSF producing virus group. Similar results were seen with the sarcoma model, although not as evident tumor growth reduction was observed. T2 weighted MRI of the carcinomas showed interesting results. Tumors that later decreased in size showed a distinct dark core in the imaging already from day two onwards. Th is was not seen in the nonresponding or the PBS treated tumors. Instead, hyper intensive areas (white) were commonly seen, indicating rapid tumor growth. Th ese fi ndings are visualized nicely in Figure 2 of study III. Further study and other publications suggested that the hypo intense (dark) core visualized in the responders consists of coagulative necrosis, while the hyper intense (white) areas of the nonresponders are due to liquefactive necrosis due to fast tumor growth. In the sarcoma tumors, MRI detected acute hemorrhages, leading to signal loss and preventing similar T2 quantifi cation and spectroscopy. Small tumor size (from some mm to two cm) induced technical problems leading to low quality imaging and long imaging times. As tumors of patients are typically larger, this might not be a problem in a clinical setting. MRS of the tumors revealed also fascinating fi ndings. Th is technique consisted of stimulating the hydrogen protons of the in vivo hamster tumors by the magnetic fi eld and then analyzing the spectra emitted from the tumor. With this technique we could analyze what substances are present at the tumor. Most of the spectra come from water. We noted that the T2 relaxation time decreased with the responding animals (starting from day two), while this was not seen with the nonresponding or the PBS treated hamsters. We also noted that the amounts of taurine, unsaturated fatty acids and choline were signifi cantly lower in the responders compared to nonresponders or PBS. Th us we concluded that both MRI and MRS could be used at early timepoints to evaluate how the treatments work. T-cell and heterophil infi ltration and calprotectin positivity indicated high immunological activity in the armed oncolytic virus treated hamster tumors that responded to the treatment. Heterophils are equivalent to human neutrophils and calprotectin is known to be produced from activated neutrophils/heterophils. A fecal test based on calprotectin is already in clinical use to screen patients that might need colonoscopy (Vestergaard et al. 2008). Calprotectin use as a biomarker in oncolytic immunotherapy is an interesting subject of further study. Here calprotectin staining was performed to tumor samples, but analyses from the blood or other body fl uids should also be feasible. Patient N21. We had seen that MRI and MRS could be used in evaluating responding immunocompetent hamsters. Next, we wanted to see if this could help optimize the treatment of patients. We found one patient, treated in ATAP, that was imagined with MRI before and aft er virus treatment. Th e patient was a 6-year-old boy with advanced neuroblastoma that was previously treated with three diff erent chemotherapy regimens and autologous stem cell transplant, but the disease was progressing. Aft er the oncolytic virus therapy, a partial response in injected tumors was seen and a complete response in the non-injected bone marrow, indicating that this patient seemed to respond to the virus treatment. Interestingly, in the T2 weighted MRI scans the tumor became darker. Less than a year later the tumor started growing again and also a distinctly lighter contrast was seen in the tumor. While this one retrospectively analyzed patient certainly does not prove anything, it shows that the technique is promising and a clinical study is warranted. In conclusion, our studies indicate that T2 contrast chance and the T2 water relaxation time could be adapted relatively easily to the clinics, as they can be measured with widely available MRI equipment. Also MRS of choline, taurine and unsaturated fatty acids might provide interesting information, although this technique is more time consuming and challenging. However, it has already been evaluated for example in prostate cancer (Sarkar et al. 2014). 56 Results and Discussion Study IV Background of study IV. Immunotherapy has matured to a stage where its clinical potential starts to fl ourish. A handful of new drugs have already become available. Some of these seem to produce complete responses in a portion of metastatic patients. However, most patients are still incurable and thus there is room for improvement. So far two randomized trials have been completed with oncolytic immunotherapy, both with positive results. An oncolytic serotype 5 adenovirus improved the effi cacy of chemotherapy for treatment of metastatic head and neck cancer (Xia et al. 2004) and gained license in China. In the west, a herpes virus armed with GM-CSF cytokine (T Vec) was eff ective in metastatic melanoma (Robert Hans Ingemar Andtbacka 2013) and recently (29.4.2015) FDA voted in favor of the approval. For study IV, an Ad5/3-E2F-d24-GMCSF (also called CGTG-602) virus was used. Th e design of this virus includes three concepts: 1) Serotype 3 knob is used for increased entry into tumor cells. Th e Ad5/3 construct has shown enhanced killing of cancer cells, clinical specimens and xenograft tumors in mice while retaining safety in humans (Koski et al. 2010; Kanerva et al. 2013). 2) Th e E2F promoter and the d24 deletion of the E1A region of the virus are designed for specifi c replication of the virus in cancer cells and not in normal cells. Th e E2F promoter is active in cell lines that have the mutated pRb pathway, common in most tumors (Fueyo et al. 2000; Alonso et al. 2008). Th us, while in normal cells the translation of the virus E1A (critical for virus replication) does not happen, in cancer cells it does. Another control mechanism of virus replication is the deletion of the 24 base pairs from the E1A region. In normal cells this deletion obstructs the replication of the virus. However, the 24bp deletion leads to a mutant E1A protein translation that can result in toxicity or anti-viral immunity. In most malignant cells the 24 bp deletion does not hinder virus replication (Cerullo et al. 2010). 3) GM-CSF is a cytokine that is used widely as an immunostimulatory molecule. It has been proven eff ective in a randomized phase III trial when combined with a herpes virus (Robert Hans Ingemar Andtbacka 2013). It has also been used successfully in a vaccine concept with metastatic prostate cancer. In 2010, FDA granted marketing authorization to this fi rst therapeutic vaccine (sipuleucel-T, Provenge) and in 2013 it was approved also in Europe. A survival benefi t of about four months was noted (Kantoff et al. 2010). GM-CSF is a potent cytokine that induces systemic anti-tumor immunity. It recruits and maturates antigen presenting cells, NK-cells and neutrophils. We expect the local production of the GM-CSF at the tumor site to be useful in avoiding adverse events while retaining effi cacy.

Preclinical part of study IV Some of the most important preclinical results are shown in the Figure 1 of study IV. Th e viability of TF-1 cells is dependent on GM-CSF. Increased viability of these cells was seen with the supernatant that was collected and fi ltered from Ad5/3-E2F-d24-GMCSF infected cells. Commercial human GM-CSF was used as a positive control. Filtered supernatant from cells infected by a similar, but not a GM-CSF producing, virus was used as a negative control. As expected, this supernatant did not enhance the survival compared to wells where no supernatant was used. As a result, we concluded that functional human GM-CSF is produced by the virus. Next, we demonstrated that the replication of Ad5/3-E2F-d24-GMCSF virus was low in normal cells. For this experiment, primary human hepatocytes were used. Low quantities of infective virus particles were found compared to the wild type adenovirus and to a 24 base pair deleted oncolytic virus, indicating that the double control of replication (E2F and d24) was eff ective. To

57 Results and Discussion strengthen the tumor selectivity data, a Syrian hamster model was used. Tumor free hamsters received an injection of the Ad5/3-E2F-d24-GMCSF virus to the livers. Tumor bearing hamsters received similar injections. Later, livers and tumors were collected and qPCR was used to determine virus quantities in livers and tumors. Minimal virus was detected in the livers while high amounts of virus were seen in the tumors. Finally, the Ad5/3-E2F-d24-GMCSF virus was tested in diff erent cancer cell lines (including lung, ovarian and prostate) in vitro and in vivo. Th e E2F promoter modifi cation did not seem to slow down the virus replication in human cancer cell lines compared to a virus with a naïve promoter. Th e virus seemed to kill effi ciently all tested cancer lines in vitro, and potency was also demonstrated in the immunocompetent Syrian hamster cancer model in vivo. In this in vivo model, a low dose cyclophosphamide did not signifi cantly improve the anti-tumor eff ect of the Ad5/3-E2F-d24-GMCSF. Th e in vivo data of the Syrian hamster model was somewhat disappointing. A signifi cant reduction in tumor growth was seen, but the results had been better with similar constructs that were fully serotype 5 (Cerullo et al. 2010). Th e explanation to this fi nding seems to be that the Ad5/3 construct (or a fully serotype 3 construct) is less permissive in many (if not all) hamster cells lines (Bramante et al. 2013). At the moment, it seems that the Syrian hamster model is a good rodent model with the serotype 5 adenovirus when immunological eff ects are evaluated, while it has limited value with other serotypes (including serotype 3 or Ad5/3). To our knowledge, no good non-primate animal models exist for serotype 3 viruses.

Clinical part of study IV Patients were treated with the Ad5/3-E2F-d24-GMCSF virus in the ATAP program (see materials and methods). Using the epidemiological method of retrospective registry research (case series), data was collected from the fi rst 13 patients treated with the virus. Patients had advanced metastatic tumors refractory to and progressing aft er standard therapy. Th ey were treated with 2-4 cycles of the virus adding up to 39 treatment cycles. One patient was treated fi rst with the virus and then with hyperthermic intraperitoneal chemotherapy (HIPEC). Five patients had ovarian, three breast, two pancreatic, one rectum, one colon, one melanoma, one sarcoma, and one fi brosarcoma cancer. Patient age varied from 40 to 74 years. Patients were treated in a personalized manner. One to ten tumor sites were injected. Variation between tumors and patients represents well “real life” advanced cancer patients compared to a classical clinical trial population. However, the vast patient variation must be kept in mind and comparison between treatment results is diffi cult. As required by FIMEA (EU/1394/2007, Dnro 608/03.01.01/2009), we were mandated to collect safety data. We collected all adverse events from the 39 treatment cycles. Grade 1-2 fl u like symptoms including fever, fatigue, and pain were experienced in more than half of the treatments. Most grade 3 events were self-limiting and treated as outpatient. No grade 4-5 adverse events attributable to the treatments were observed. We concluded that the treatments were well tolerated. Next, we studied virus replication. Prior to therapy, all patient serum samples were negative for Ad5/3-E2F-d24-GMCSF qPCR. One day aft er the treatment 8/13 patient serums were positive for the virus. Th e highest titer was 1141 VP/ml. On days 3-8 we had serum from four patients; of these, two were positive and increased from day 1, suggesting virus replication. Th e highest titer on days 3-8 was 11523 VP/ml. Aft er the 2nd and 3rd treatment cycles, no virus was detected aft er day 1 post treatment. Evidence of virus replication was seen aft er the fi rst treatment in the serum, while subsequent treatments did not lead to virus detection aft er day 1. Th e reason for this is 58 Results and Discussion unknown, but some evidence of internal resistance against virus replication, mediated perhaps by the interferons, has been suggested (Liikanen et al. 2011). However, as we have not seen here or in the other patient series that the virus amount in blood would correlate with effi cacy, it is unknown if this fi nding has relevance. Nevertheless, we have preclinically investigated substances that block the resistance (data not published). Another way of avoiding resistance would be to use diff erent viruses. However, the signifi cance of aggressive virus replication is unknown, as at the moment we believe that the more important aspect of the treatments involve the activation of the body’s own immune system and the role of the virus is more to produce a danger signal. We evaluated the neutralizing antibodies (NAb) during the virus treatment. As expected, a signifi cant increase in the NAb titer was observed aft er treatment. For an alternative view on the anti-viral antibodies, we also analyzed anti-hexon IgG from the patient serums. At baseline a low titer (20-300U/ml) was observed. while at three weeks all patients showed an increase (titers between 200-3000). From some patients also ascites samples were analyzed and an increase of titers was noted also here. Th e antibody results did not off er any surprising findings. No correlation with NAb titers or IgG with effi cacy (or other parameters) was observed here or in analyses done with other ATAP patients. PET-CT responses. All patients had progressing tumors prior to treatments. Six patients were assessable with PET-CT. Th e response was generally assessed 3-4 weeks aft er the last virus injection. Typically, three injections were given three weeks apart. Of the six evaluable patients, R319 had a 49% reduction in the metabolic activity in the injected liver tumor and a complete response in a non-injected mediastinal tumor. S354 had a complete metabolic response, O340 had a minor metabolic response, S352 and C312 had stable metabolic disease, and H344 had a progressive metabolic disease. Th us, 83% (5/6) had stable metabolic disease or better and 50% had a decrease in the metabolic activity in PET-CT. While PET-CT has been proven good in cancer diagnostics, there is a problem with false positive fi ndings. It has been shown that immunological activity due to viral infection increases metabolic activity in the lymph nodes giving false positive results (Kuruppu et al. 2007; Focosi et al. 2008; Koski et al. 2012). While the possibility of an increase in the metabolic activity aft er adenovirus treatments is possible, we did not seem to have evident problems of it in this setting. Tumor marker responses. Ten of the patients had elevated tumor markers before treatment. Th ree had reduction of marker levels during the treatments. Two had an initial decrease followed by an increase, and one patient had fi rst an elevation and then a subsequent reduction in the tumor marker levels. Four patients had elevation in the marker levels. Th ere seemed to be quite a good correlation with PET-CT results and tumor markers. Survival data. If we combine the results gained from the PET-CT and tumor marker data, we can conclude that 9 out of the 12 evaluable patients (75%) had some kind of a positive response (stable metabolic disease/stable markers or better) to the treatments. Th e survival of these patients were 135 days, while the survival of patients without any signs of response was 80 days, suggesting that there is a positive correlation with the objective responses (in PET-CT and/or markers) and survival. However, from this retrospective analysis we can not conclude that this correlation is due to the virus treatment. As seen now with many recent immunological treatments, it seems that some patients respond to the treatments while others seem not to benefi t. Similar fi ndings were seen with evaluated patients - some patients show no signs of response while others display strong and clear responses (e.g. patient S354 complete metabolic response, survival over 1000 days). Unfortunately, at the moment there is a lack of defi nite markers indicating patients that will benefi t from the treatment.

59 Results and Discussion

Peripheral blood T-cell activity. Cancer immunotherapy has noted the importance of T-cells in cancer development and treatment. For example, the recently approved checkpoint inhibitors activate T-cells that are in a passive state, leading sometimes to dramatic curative responses. We wanted to optimize the treatment for each patient, and one way to do this was to gain information on the T-cells of the patients. While the peripheral blood is not the optimal place to investigate T-cells, it is the most noninvasive and accessible source. As individual patient treatment is the priority, unnecessary biopsies were not an option. Th e blood is the “highway for T-cells” traveling from their normal habituate (lymph tissue or places with non-self-material such as bacteria, malignant cells etc.). Although some T-cells are known to patrol the body, only a minority is found from the peripheral blood. Patient blood, before and aft er treatments was used, the T-cells were collected and pulsed with adenovirus peptide or with tumor associated peptide pools. Th en the INF-gamma production was analyzed with ELISPOT. With this method we could evaluate anti-viral and anti-tumor T-cells circulating in the blood. We noted a concordance in 9/11 patients so that the patients who had an increase in anti-tumor T-cell activity had also an increase in anti-viral T-cell activity and vice versa. However, the results are very diffi cult to interpret, as the patients and data are very heterogeneous. Nevertheless, we can conclude that aft er the treatment with Ad5/3-E2F-d24-GMCSF changes in peripheral T-cell activity were commonly seen. Whether this means that T-cells traffi c to the tumor, multiply or that that the general immunity is increased, cannot be answered at this state, although there is increasing evidence in this direction (Kanerva et al. 2013; Tahtinen et al. 2015). Antibodies against tumor associated antigens. Many publications indicate that tumor associated antigens are elevated in cancer patients, while some also suggest that decrease of these antigens might indicate treatment effi cacy. With our patients we noted that antibodies against CEA, survivin, MUC-1 or NY-ESO-1 frequently decreased in patients with signs of anti-tumor effi cacy. Signifi cant correlation between anti-tumor antibody decrease and positive clinical signs of benefi t were noted. Th e hypothesis for this fi nding is that the immune systems anti-tumor activity is reclaimed aft er virus injection leading to clearance of tumor associated antigens. Biopsies. We had the possibility to evaluate biopsies of two patients, before and aft er therapy. One patient had ovarian cancer (O340) and the other breast cancer (R356). Interestingly, O340 seemed to have only few immunological cells present at the tumor before treatment. Aft er treatment, the quantity of immune cells in and around the tumor was found to increase multiple folds. Th e patient seemed to respond well, as response was seen both in PET-CT and markers. A relatively long survival of 890 days followed. Also a decrease in anti-tumor antibodies and a clear decrease in anti-tumor T-cells in blood (suggestive to traffi cking to the tumor) were observed. On the other hand, patient R356 did not show drastic changes in the quantity of immunological cells in the tumor, and only a partial response in markers was noted while the survival was only 102 days. We hypothesized that patient O340 was susceptible for oncolytic immunotherapy while patient R356 was immunologically resistant to the therapy. It would have been very attractive to try the checkpoint inhibitors on patient R356, as the tumor seemed to contain immunological cells that were in anenergy or otherwise not capable of destroying the target it had already recognized. On the other hand, we could hypothesize that maybe patient O340 would not have benefi ted of the checkpoint inhibitors as low quantities of T-cells were present in the tumor. Maybe a good trial design would take a biopsy, analyze immunological cells, and divide the patients to checkpoint inhibitor group or to a group which would fi rst receive an oncolytic virus for creating oncolysis and in this way epitope presentation for T-cells.

60 Summary and Conclusions SUMMARY AND CONCLUSIONS

“Now this is not the end. It is not even the beginning of the end. But it is, perhaps, the end of the beginning.”

– Winston Churchill, aft er Britain’s fi rst victorious battle of World War II

Cancer immunotherapies have provided several success stories during the last years. While BCG and interferon-alpha used to be the only immunotherapies for decades, now the range is increasing. Checkpoint inhibitors and T-cell therapies have shown their capability in a few cancer types and new indications are being approved. In China, an oncolytic virus (H101) has been approved years ago, while latest reports indicate that in the near future also the fi rst western approval is expected with the T-Vec oncolytic virus (Voted 22-1 in favor of approval by the FDA 29.4.2015, (Robert Hans Ingemar Andtbacka 2013)). Th e limitations of these therapies at the moment seem to be the cost, while the effi cacy in the responding patients is oft en dramatic and long lived. At the moment, it seems almost inevitable that in the following years many, if not all, cancers shall have their own form of immunotherapy (Homet Moreno and Ribas 2015). Th ese might fi nally even take the place of traditional treatments which cause more side-eff ects (especially chemotherapy). Also treatments at early stage cancer are considered. Th e biggest obstacle for cancer therapy, however, seems to be the same old: who to treat, when to treat and how to treat. Hopefully better imaging, sequencing and immunological knowledge will tackle these questions. In study I, we showed that cloning of a functional oncolytic adenovirus of other serotype than 5 is possible. We show that the Ad3-hTERT-E1A is eff ective in cell lines representing seven major cancer types, while low toxicity was seen in cell lines representing normal cells. In vivo, serotype 3 seemed to be at least as potent as the serotype 5 based oncolytic control viruses in three independent human cancer models. As the serotype is diff erent and the receptor for serotype 3 adenovirus is known to be highly present in advanced cancer, Ad3-hTERT-E1A is a promising new oncolytic virus. In study II, we fi rst continued the work started in study I. We performed biodistribution and toxicity experiments with the Ad3-hTERT-E1A and found that in mice higher amounts of cytokines were released but less liver damage was induced compared to the serotype 5 or 5/3 control viruses. Next, the results of the fi rst 25 patients treated with this virus were published. Th e safety profi le of the virus was good, resembling mild adverse events seen with serotype 5 viruses. Patients that had been previously treated with a serotype 5 based virus showed deeper and longer lasting lymphocytopenia than patients that received Ad3-hTERT-E1A as the fi rst treatment. We also noted frequent alterations in the T-cells collected from blood, indicating immunological activity. Interestingly, patients that were pretreated with serotype 5 viruses showed re-emerging of this virus (or its DNA) to blood, and vice versa, indicating that perhaps the old virus is lurking passively inside malignant cells. Signs of effi cacy was seen in many patients, and decrease in tumors markers was observed in 11/15 (73%) patients. Four of these responding patients were treated only intravenously, indicating that at least the fi rst oncolytic virus of a certain serotype could be given by intravenous route only. Interestingly, good results were seen with breast cancer patients, especially the ones on trastuzumab, indicating that the synergy suggested in preclinical

61 Summary and Conclusions models might be true. As the safety and effi cacy seemes good, further testing or development of armed serotype 3 viruses seems rational. During studies I-II, we noted that an oncolytic adenovirus based on serotype 3 seems to have many benefi ts compared to the serotype 5 (or some other coxsackie-adenovirus receptor binding adenovirus), which has been the most used adenovirus on the fi eld. While the main receptor for serotype 3 (DSG-2) is widely present in advanced cancer, it seems to have other potentially important aspects, such as the suggested opening of the tight junction, production of dodecahedral particles (PtDds), transforming the cells to an epithelial-to-mesenchymal transition (EMT), binding to the CD80/CD86 present at antigen-presenting cells, and might thus have a role in activating T-cells. In study III, we wanted to study if magnetic resonance imaging (MRI) and/or spectroscopy (MRS) could be used in the evaluation of oncolytic immunotherapy, as little previous data existed. Currently used methods are proven to be unreliable due to infl ammation and initial tumor swelling. A GM-CSF armed and a non-armed control adenovirus was used. We show that in T2-weighed MRI a hypodense core area develops in the responding tumors in an immunocompetent Syrian hamster carcinoma model. Th is core area is consistent with coagulative necrosis. Similar fi ndings were observed in a neuroblastoma patient who responded to the treatment. We found that MRS of choline, taurine and unsaturated fatty acids can be useful indicators of response at early days aft er treatment. A high amount of calprotectin positive infl ammatory cells and T-cells were found surrounding necrotic areas, suggesting a possible link between necrosis, oncolysis and/or immune response in the GM-CSF treated cells. MRI and MRS could thus be used in the estimation of effi cacy in oncolytic immunotherapy at early time points. Patients benefi tting from the therapy could be selected from nonresponding ones treatable with other modalities. In study IV, a quadruple modifi ed CM-CSF producing oncolytic adenovirus was evaluated. First the virus was shown to function as designed. Th e modifi cations made it prone in entering and replicating in cancer cells while replication in normal human cells was low. Th en we demonstrate that the virus produces functional GM-CSF. At the clinical part of the study, we evaluated the 13 fi rst patients treated with the virus. Treatments were found well tolerated and effi cacy was suggested. Signs of potential anti-tumor effi cacy (PET-CT and markers) was observed in 9/12 (75%) evaluable patients. Radiological disease control rate (stable metabolic disease or better) with PET-CT was 83% while the response rate (minor metabolic response or better) was 50%. Accumulation of immunological cells aft er treatment was seen in tumor biopsies. RNA expression analysis of these biopsies indicated immunological activity and metabolic changes aft er virus replication. Conclusion from studies I-IV. A fully functional serotype 3 oncolytic adenovirus was constructed. Th e treatments showed promising safety and effi cacy data. Even though this virus was not armed with any immunomodulatory transgene, it seemed to be as good, or even better, as a GM-CSF armed Ad5/3 (serotype 5 virus with a serotype 3 fi ber knob) oncolytic virus. A side by side comparison of two independent studies is hard, but if we hypothesize that the patients reported in study II and study IV were similar as well as all other major error factors, the non-armed serotype 3 virus seems promising. With serotype 3, 11/15 patients (73%) showed possible signs of effi cacy according to tumor markers while this was seen with 6/10 (60%) of the armed Ad5/3 virus. However, we must keep in mind that the number of patients is low for any comparison and there was signifi cant heterogeneity among the patients. Th e goal was to treat patients, not to evaluate effi cacy. Accordingly, no defi nite conclusions can be made of

62 Summary and Conclusions effi cacy. Th is data is, however, provocative as the GM-CSF armed Ad5/3 construct seems to be one of the best oncolytic viruses we have used. Th erefore, the further development of a fully serotype 3 armed oncolytic virus is highly suggested. Preclinical work seems to propose that combining adenovirus serotype 3 with approved therapies, such as chemotherapy, trastuzumab or cetuximab (Beyer et al. 2011) (Beyer et al. 2012), may off er further potency. Results have been promising and adenovirus 3 dodecahedral particles are entering into clinical testing (Beyer et al. 2012). Other combinations that will probably be seen in the near future include combinations with checkpoint inhibition and T-cell based therapies (see table 11). Th ere even seems to be rationale for combining them all.

Table 11. Hypothetical new cancer treatment schema

Biopsy fi nding Problem Treatment Eff ect High amount of T-cells are being Checkpoint inhibitor Eff ector T-cell activation eff ector T-cells inactivated by the tumor Th e ratio of Tregs/ Too many Checkpoint inhibitor/ Eff ector T-cell activation MDSC to T eff ector regulatory cells Collect T-cells, grow cells unfavorable and modulte in vitro, return to patient Low amount of Tumor not Oncolytic (armed) Oncolysis, danger signal, eff ector T-cells recognized by virus immunostimulation, T-cells presentation of tumor associated antigens High serotype Intravenous Use serotype 3 virus / Oncolysis, danger signal, 5 NAbs in serotype 5 not inject serotype 5 virus immunostimulation, blood (previous eff ective intratumorally presentation of tumor treatment) associated antigens and activation of previous virus treatment High serotype Intravenous Use serotype 5 virus / Oncolysis, danger signal, 3 NAbs in serotype 3 not inject serotype 3 virus immunostimulation, blood (previous eff ective intratumorally presentation of tumor treatment) associated antigens and activation of previous virus treatment Low Her2/neu trastuzumab not Use adenovirus Improved access to the receptor eff ective serotype 3 as a tight Her2/neu receptor junction opener Low EGFR receptor cetuximab not Use adenovirus Improved access to the eff ective serotype 3 as a tight EGFR receptor junction opener If patients is in a good general status, checkpoint combinations (CTLA-4 and PD-1 antibody) preferable in stead of monotherapy. For non-necessary organs (e.g. prostate, breast, thyroid, ovary) targeting these cells with transfected T-cells might be an option. Feasibility of this approach is demonstrated with lymphoma/leukemia. Similar therapy might be preferable if tumor presents cancer-testis antigens.

63 Summary and Conclusions

My favorite imaginary treatment design at the moment for advanced cancer:  Operation of the tumor and growing the tumor infi ltrating lymphocytes from the sample.  Priming the residual tumor with an immunostimulatory serotype 5 adenovirus and then boosting the treatment later with a serotype 3 adenovirus.  Th ese treatments would be combined with checkpoint inhibition and with trastuzumab and/or cetuximab when rational.  Finally returning the tumor infi ltrating lymphocytes to the patient.

64 Acknowledgements ACKNOWLEDGEMENTS

Th is work was conducted during 2007-2015 in the Cancer Gene Th erapy Group (CGTG), University of Helsinki, Finland. CGTG was formerly part of the Molecular Cancer Biology Program in Biomedicum Helsinki and currently affi liated to Medicum, Haartman Institute. In brief, I am deeply grateful to all people working in these institutes and their help and assistance during these years. Without the cleaning man or the head of the department the journey would have been a lot harder, thank you all. First I want to thank all the 290 patients that participated in the 821 ATAP treatments. Th eir bravery and faith in these new treatments enabled us to gain experience and to develop the treatments further. While not all of the patients showed benefi t, I believe there were many that received advantage themselves. A special thanks for the Dean of Faculty of Medicine, Professor Risto Renkonen for his continuous support over these years. I wish to thank also the head of Haartman Institute, Professor Tom Böhling and the current and former directors of Molecular Cancer Biology Program for the good research facilities. I am grateful for professor Ruben Hernandez Alcoceba who accepted (in minutes aft er sending the email) the request to be my opponent. Professor Kari Airenne and adjunct professor Timo Muhonen are entitled to special recognition for their helpful comments during the pre- examination. I am very grateful for my supervisor, group leader, professor and brother Akseli for giving me the opportunity to get acquainted into this truly fascinating world of gene therapy, rational cloning, viruses, immune cells, oncolysis, cancer, cancer treatments and oncolytic immunotherapy. I wish also to thank all former and present group members of the CGTG. Without Akseli´s driving force and the CGTG group’s team work, the accomplishments would not have been possible. During my time more than a hundred publications were published in highly ranked journals and also many patients were treated with the oncolytic viruses that were cloned and preclinically evaluated by the group. I saw the evolution from preclinical promising oncolytic agents to clinical treatments. Akseli and the CGTG made the leap from the bench to the bedside, something that is rarely accomplished with preclinical groups. I saw the encouraging responses and the attempts to fi nd money for clinical testing. Now the group has evolved back to the bench with the information gained from these treatments and is designing even more potent treatments. Hopefully someday these attempts will lead to late stage clinical trials so that the effi cacy of these promising treatments can be evaluated. Th is enables the possibility for global large scale patient benefi t. I would like to thank Sergio, my hands on mentor, for introducing me to the cell culture and mouse work, and Sari for helping me out to plan and execute many of the experiments that were done and helping me for preparing the articles. I am grateful for the time Iulia helped me with the experiments and animal work. I would like to thank Gerd for passing on the information about the serotype 3 viruses; he was the one in our group who had tried to clone these viruses before me. For teaching of cloning, I would like to thank Eerika and our collaborators from Zurich and Seattle. Silvio kindly invited me to Zurich to learn new tricks with cloning, and Andre and Hongjie from Seattle were extremely helpful regarding the cloning of the armed viruses. Simona and Suvi, you have helped me fi nalize critical experiments. Saila, thank you especially for the help with the qPCR and Kikka for help with the ELISPOT, you both emit good atmosphere around you. Minna and Aila, thank you for the million little things you have done for me. Essential was

65 Acknowledgements also the help and knowledge that more experienced researchers such as Iulia, Lotta, Tuuli, Kilian, Markus, Vince and Anna were able to give. Younger GCGT PhD students, Mikko, Siri, Riikka, Kristian, Noora, you are good company and seem enthusiastic, keep up the good work and spirit. You are the heart of CGTG. Th ank you for the fruitful collaboration in the MRIS project with the University of Eastern Finland. Riikka, Johanna and Olli, this would have not been possible without your welcoming attitude. Th e scientifi c conferences we joined were memorable not only because of the scientifi c background and the many connections with collaborators we made but because we were able to project the work we are doing with work done by others. Some trips we made with friends from the lab were unforgettable, including kitesurfi ng in the Long Island, powder skiing in Utah, touring the Alps, road tripping Baja California or hanging out at diff erent summer cottages in Finland. Ilkka, Marko, Kilian, Petri, Joao, Iulia, Sari, Akseli, Anniina, Maria and others, thanks also for the time outside the lab. Equally great were the many scientifi c and non-scientifi c interaction we had with diff erent groups in Meilahti. Especially regular fl oor ball with the Lauri Aaltonen group and beach volley with a heterogeneous team have oft en been good balance in the weekly cycle. Especial thanks to Alexandra, Heli, Erkka, Boris for keeping things active outside the lab. Th ank you Ari Harjula and Maija Lappalainen for giving your time and advice during the yearly follow-up group meetings. Your comments were valuable and the meetings helped to concretely see the progress as sometimes it felt that no progress was made in ages. Special thanks also to Maija for helping with the validation of the Ad3-hTERT-E1A virus; electron microscopy, culture, serotyping and PCR. Special thanks for Ari Harjula and Mika Matikainen for accepting to become the reviewers of this thesis. Even greater thanks to Kari Airenne and Timo Muhonen who also accepted and were then appointed as reviewers by the Faculty. I wish to thank the National Graduate School of Clinical Investigation for educational and fi nancial support during these years. In addition, I would like to thank the other financial supporters including our CGTG and the following foundations: Duodecim, Biomedicum Helsinki, Irja Karvonen, Finnish Cancer Organisations, Sohlberg, Research on Viral Diseases, Maud Kuistila, Irja Karvonen, Eemil Aaltonen and Ida Montin. I guess, I could not change my destiny. Th e genes had it encrypted or then it just was the environment. Having a research professor as an mother, as a father and nowadays also as a brother did not help. I feel I was newer pressurized or even lured, but sometime in mid-med school science started to interest me more and more. For few years I searched passively for something interesting and aft er graduating I ended up starting with the most interesting group I had come up with. Coincidence or not, it was the one of my brother. Although combining research with clinical work seems to be quite demanding, I have, however, enjoyed the experience and, if feasible, shall try to continue on this road. Th us, I would like to thank people involved in my upbringing, Inna, Ossi, Kari, Asta, Aksu, Liina. I don’t know how you did, it but I think you did it quite well. Finally, I want to thank my beloved Anni for understanding, caring and standing by my side through this sometimes stressful journey. Last but not least I thank all my friends. Without you my life would be somewhat dull.

Helsinki, October 2015

66 References REFERENCES

Alba, R., Bradshaw, A. C., Parker, A. L., Bhella, D., Waddington, S. N., Nicklin, S. A., van Rooijen, N., Custers, J., Goudsmit, J., Barouch, D. H., McVey, J. H. and Baker, A. H. (2009). ”Identifi cation of coagulation factor (f)x binding sites on the adenovirus serotype 5 hexon: Eff ect of mutagenesis on fx interactions and gene transfer.” Blood 114(5): 965-971. Alonso, M. M., Alemany, R., Fueyo, J. and Gomez-Manzano, C. (2008). ”E2f1 in gliomas: A paradigm of oncogene addiction.” Cancer Lett 263(2): 157-163. Amin, A. and White, R. L., Jr. (2013). ”High-dose interleukin-2: Is it still indicated for melanoma and rcc in an era of targeted therapies?” Oncology (Williston Park) 27(7): 680-691. Ansell, S. M., Lesokhin, A. M., Borrello, I., Halwani, A., Scott, E. C., Gutierrez, M., Schuster, S. J., Millenson, M. M., Cattry, D., Freeman, G. J., Rodig, S. J., Chapuy, B., Ligon, A. H., Zhu, L., Grosso, J. F., Kim, S. Y., Timmerman, J. M., Shipp, M. A. and Armand, P. (2015). ”Pd-1 blockade with nivolumab in relapsed or refractory hodgkin’s lymphoma.” N Engl J Med 372(4): 311-319. Araten, D. J., Golde, D. W., Zhang, R. H., Th aler, H. T., Gargiulo, L., Notaro, R. and Luzzatto, L. (2005). ”A quantitative measurement of the human somatic mutation rate.” Cancer Res 65(18): 8111-8117. Arnberg, N. (2009). ”Adenovirus receptors: Implications for tropism, treatment and targeting.” Rev Med Virol 19(3): 165-178. Arnberg, N. (2012). ”Adenovirus receptors: Implications for targeting of viral vectors.” Trends Pharmacol Sci 33(8): 442-448. Ascierto, P. A., Chiarion-Sileni, V., Muggiano, A., Mandala, M., Pimpinelli, N., Del Vecchio, M., Rinaldi, G., Simeone, E. and Queirolo, P. (2014). ”Interferon alpha for the adjuvant treatment of melanoma: Review of international literature and practical recommendations from an expert panel on the use of interferon.” J Chemother 26(4): 193-201. Balakireva, L., Schoehn, G., Th ouvenin, E. and Chroboczek, J. (2003). ”Binding of adenovirus capsid to dipalmitoyl phosphatidylcholine provides a novel pathway for virus entry.” J Virol 77(8): 4858- 4866. Barker, D. D. and Berk, A. J. (1987). ”Adenovirus proteins from both e1b reading frames are required for transformation of rodent cells by viral infection and DNA transfection.” Virology 156(1): 107- 121. Barton, K. N., Paielli, D., Zhang, Y., Koul, S., Brown, S. L., Lu, M., Seely, J., Kim, J. H. and Freytag, S. O. (2006). ”Second-generation replication-competent oncolytic adenovirus armed with improved suicide genes and adp gene demonstrates greater effi cacy without increased toxicity.” Mol Th er 13(2): 347-356. Barton, K. N., Stricker, H., Brown, S. L., Elshaikh, M., Aref, I., Lu, M., Pegg, J., Zhang, Y., Karvelis, K. C., Siddiqui, F., Kim, J. H., Freytag, S. O. and Movsas, B. (2008). ”Phase i study of noninvasive imaging of adenovirus-mediated gene expression in the human prostate.” Mol Th er 16(10): 1761- 1769. Barton, K. N., Xia, X., Yan, H., Stricker, H., Heisey, G., Yin, F. F., Nagaraja, T. N., Zhu, G., Kolozsvary, A., Fenstermacher, J. D., Lu, M., Kim, J. H., Freytag, S. O. and Brown, S. L. (2004). ”A quantitative method for measuring gene expression magnitude and volume delivered by gene therapy vectors.” Mol Th er 9(4): 625-631. Bauerschmitz, G. J., Guse, K., Kanerva, A., Menzel, A., Herrmann, I., Desmond, R. A., Yamamoto, M., Nettelbeck, D. M., Hakkarainen, T., Dall, P., Curiel, D. T. and Hemminki, A. (2006). ”Triple- targeted oncolytic adenoviruses featuring the cox2 promoter, e1a transcomplementation, and serotype chimerism for enhanced selectivity for ovarian cancer cells.” Mol Th er 14(2): 164-174.

67 References

Bauerschmitz, G. J., Ranki, T., Kangasniemi, L., Ribacka, C., Eriksson, M., Porten, M., Herrmann, I., Ristimaki, A., Virkkunen, P., Tarkkanen, M., Hakkarainen, T., Kanerva, A., Rein, D., Pesonen, S. and Hemminki, A. (2008). ”Tissue-specifi c promoters active in cd44+cd24-/low breast cancer cells.” Cancer Res 68(14): 5533-5539. Bauzon, M. and Hermiston, T. W. (2008). ”Exploiting diversity: Genetic approaches to creating highly potent and effi cacious oncolytic viruses.” Curr Opin Mol Th er 10(4): 350-355. Beaulieu, J. F. (1999). ”Integrins and human intestinal cell functions.” Front Biosci 4: D310-321. Beyer, I., Cao, H., Persson, J., Song, H., Richter, M., Feng, Q., Yumul, R., van Rensburg, R., Li, Z., Berenson, R., Carter, D., Roffl er, S., Drescher, C. and Lieber, A. (2012). ”Coadministration of epithelial junction opener jo-1 improves the effi cacy and safety of chemotherapeutic drugs.” Clin Cancer Res 18(12): 3340-3351. Beyer, I., van Rensburg, R., Strauss, R., Li, Z., Wang, H., Persson, J., Yumul, R., Feng, Q., Song, H., Bartek, J., Fender, P. and Lieber, A. (2011). ”Epithelial junction opener jo-1 improves monoclonal antibody therapy of cancer.” Cancer Res 71(22): 7080-7090. Bilbao, R., Bustos, M., Alzuguren, P., Pajares, M. J., Drozdzik, M., Qian, C. and Prieto, J. (2000). ”A blood-tumor barrier limits gene transfer to experimental liver cancer: Th e eff ect of vasoactive compounds.” Gene Th er 7(21): 1824-1832. Bramante, S., Koski, A., Kipar, A., Diaconu, I., Liikanen, I., Hemminki, O., Vassilev, L., Parviainen, S., Cerullo, V., Pesonen, S. K., Oksanen, M., Heiskanen, R., Rouvinen-Lagerstrom, N., Merisalo- Soikkeli, M., Hakonen, T., Joensuu, T., Kanerva, A., Pesonen, S. and Hemminki, A. (2013). ”Serotype chimeric oncolytic adenovirus coding for gm-csf for treatment of sarcoma in rodents and humans.” Int J Cancer. Bramante, S., Koski, A., Kipar, A., Diaconu, I., Liikanen, I., Hemminki, O., Vassilev, L., Parviainen, S., Cerullo, V., Pesonen, S. K., Oksanen, M., Heiskanen, R., Rouvinen-Lagerstrom, N., Merisalo- Soikkeli, M., Hakonen, T., Joensuu, T., Kanerva, A., Pesonen, S. and Hemminki, A. (2014). ”Serotype chimeric oncolytic adenovirus coding for gm-csf for treatment of sarcoma in rodents and humans.” Int J Cancer 135(3): 720-730. Cerullo, V., Koski, A., Vaha-Koskela, M. and Hemminki, A. (2012). ”Chapter eight--oncolytic adenoviruses for cancer immunotherapy: Data from mice, hamsters, and humans.” Adv Cancer Res 115: 265-318. Cerullo, V., Pesonen, S., Diaconu, I., Escutenaire, S., Arstila, P. T., Ugolini, M., Nokisalmi, P., Raki, M., Laasonen, L., Sarkioja, M., Rajecki, M., Kangasniemi, L., Guse, K., Helminen, A., Ahtiainen, L., Ristimaki, A., Raisanen-Sokolowski, A., Haavisto, E., Oksanen, M., Karli, E., Karioja-Kallio, A., Holm, S. L., Kouri, M., Joensuu, T., Kanerva, A. and Hemminki, A. ”Oncolytic adenovirus coding for granulocyte macrophage colony-stimulating factor induces antitumoral immunity in cancer patients.” Cancer Res 70(11): 4297-4309. Cerullo, V., Pesonen, S., Diaconu, I., Escutenaire, S., Arstila, P. T., Ugolini, M., Nokisalmi, P., Raki, M., Laasonen, L., Sarkioja, M., Rajecki, M., Kangasniemi, L., Guse, K., Helminen, A., Ahtiainen, L., Ristimaki, A., Raisanen-Sokolowski, A., Haavisto, E., Oksanen, M., Karli, E., Karioja-Kallio, A., Holm, S. L., Kouri, M., Joensuu, T., Kanerva, A. and Hemminki, A. (2010). ”Oncolytic adenovirus coding for granulocyte macrophage colony-stimulating factor induces antitumoral immunity in cancer patients.” Cancer Res 70(11): 4297-4309. Chapman, P. B., D’Angelo, S. P. and Wolchok, J. D. (2015). ”Rapid eradication of a bulky melanoma mass with one dose of immunotherapy.” N Engl J Med. Chen, M. J., Green, N. K., Reynolds, G. M., Flavell, J. R., Mautner, V., Kerr, D. J., Young, L. S. and Searle, P. F. (2004). ”Enhanced effi cacy of escherichia coli nitroreductase/cb1954 prodrug activation gene therapy using an e1b-55k-deleted oncolytic adenovirus vector.” Gene Th er 11(14): 1126-1136.

68 References

Chu, Y., Heistad, D., Cybulsky, M. I. and Davidson, B. L. (2001). ”Vascular cell adhesion molecule-1 augments adenovirus-mediated gene transfer.” Arterioscler Th romb Vasc Biol 21(2): 238-242. Cody, J. J. and Douglas, J. T. (2009). ”Armed replicating adenoviruses for cancer virotherapy.” Cancer Gene Th er 16(6): 473-488. Cohen, C. J., Shieh, J. T., Pickles, R. J., Okegawa, T., Hsieh, J. T. and Bergelson, J. M. (2001). ”Th e coxsackievirus and adenovirus receptor is a transmembrane component of the tight junction.” Proc Natl Acad Sci U S A 98(26): 15191-15196. Coley, W. B. (1891). ”Ii. Contribution to the knowledge of sarcoma.” Ann Surg 14(3): 199-220. Crawford, M. H. (2004). ”History of human biology (1929-2004).” Hum Biol 76(6): 805-815. Crompton, A. M. and Kirn, D. H. (2007). ”From onyx-015 to armed vaccinia viruses: Th e education and evolution of oncolytic virus development.” Curr Cancer Drug Targets 7(2): 133-139. de Jong, J. C., Osterhaus, A. D., Jones, M. S. and Harrach, B. (2008). ”Human adenovirus type 52: A type 41 in disguise?” J Virol 82(7): 3809; author reply 3809-3810. de Souza, A. P. and Bonorino, C. (2009). ”Tumor immunosuppressive environment: Eff ects on tumor- specifi c and nontumor antigen immune responses.” Expert Rev Anticancer Th er 9(9): 1317-1332. DeWeese, T. L., van der Poel, H., Li, S., Mikhak, B., Drew, R., Goemann, M., Hamper, U., DeJong, R., Detorie, N., Rodriguez, R., Haulk, T., DeMarzo, A. M., Piantadosi, S., Yu, D. C., Chen, Y., Henderson, D. R., Carducci, M. A., Nelson, W. G. and Simons, J. W. (2001). ”A phase i trial of cv706, a replication-competent, psa selective oncolytic adenovirus, for the treatment of locally recurrent prostate cancer following radiation therapy.” Cancer Res 61(20): 7464-7472. Di, Y., Seymour, L. and Fisher, K. (2014). ”Activity of a group b oncolytic adenovirus (coload1) in whole human blood.” Gene Th er 21(4): 440-443. Diaconu, I., Cerullo, V., Escutenaire, S., Kanerva, A., Bauerschmitz, G. J., Hernandez-Alcoceba, R., Pesonen, S. and Hemminki, A. (2010). ”Human adenovirus replication in immunocompetent syrian hamsters can be attenuated with chlorpromazine or cidofovir.” J Gene Med 12(5): 435-445. Dias, J. D., Guse, K., Nokisalmi, P., Eriksson, M., Chen, D. T., Diaconu, I., Tenhunen, M., Liikanen, I., Grenman, R., Savontaus, M., Pesonen, S., Cerullo, V. and Hemminki, A. (2010). ”Multimodal approach using oncolytic adenovirus, cetuximab, chemotherapy and radiotherapy in hnscc low passage tumour cell cultures.” Eur J Cancer 46(3): 625-635. Dias, J. D., Hemminki, O., Diaconu, I., Hirvinen, M., Bonetti, A., Guse, K., Escutenaire, S., Kanerva, A., Pesonen, S., Loskog, A., Cerullo, V. and Hemminki, A. (2011). ”Targeted cancer immunotherapy with oncolytic adenovirus coding for a fully human monoclonal antibody specifi c for ctla-4.” Gene Th er. Eager, R. M. and Nemunaitis, J. (2011). ”Clinical development directions in oncolytic viral therapy.” Cancer Gene Th er 18(5): 305-317. Escutenaire, S., Cerullo, V., Diaconu, I., Ahtiainen, L., Hannuksela, P., Oksanen, M., Haavisto, E., Karioja-Kallio, A., Holm, S. L., Kangasniemi, L., Ribacka, C., Kauppinen, S., Joensuu, T., Arstila, T. P., Pesonen, S., Kanerva, A. and Hemminki, A. (2011). ”In vivo and in vitro distribution of type 5 and fi ber-modifi ed oncolytic adenoviruses in human blood compartments.” Ann Med 43(2): 151- 163. Fechner, H., Haack, A., Wang, H., Wang, X., Eizema, K., Pauschinger, M., Schoemaker, R., Veghel, R., Houtsmuller, A., Schultheiss, H. P., Lamers, J. and Poller, W. (1999). ”Expression of coxsackie adenovirus receptor and alphav-integrin does not correlate with adenovector targeting in vivo indicating anatomical vector barriers.” Gene Th er 6(9): 1520-1535.

69 References

Fleischli, C., Sirena, D., Lesage, G., Havenga, M. J., Cattaneo, R., Greber, U. F. and Hemmi, S. (2007). ”Species b adenovirus serotypes 3, 7, 11 and 35 share similar binding sites on the membrane cofactor protein cd46 receptor.” J Gen Virol 88(Pt 11): 2925-2934. Focosi, D., Caracciolo, F., Galimberti, S., Papineschi, F. and Petrini, M. (2008). ”False positive pet scanning caused by inactivated infl uenza virus vaccination during complete remission from anaplastic t-cell lymphoma.” Ann Hematol 87(4): 343-344. Freytag, S. O., Khil, M., Stricker, H., Peabody, J., Menon, M., DePeralta-Venturina, M., Nafziger, D., Pegg, J., Paielli, D., Brown, S., Barton, K., Lu, M., Aguilar-Cordova, E. and Kim, J. H. (2002). ”Phase i study of replication-competent adenovirus-mediated double suicide gene therapy for the treatment of locally recurrent prostate cancer.” Cancer Res 62(17): 4968-4976. Freytag, S. O., Rogulski, K. R., Paielli, D. L., Gilbert, J. D. and Kim, J. H. (1998). ”A novel three-pronged approach to kill cancer cells selectively: Concomitant viral, double suicide gene, and radiotherapy.” Hum Gene Th er 9(9): 1323-1333. Freytag, S. O., Stricker, H., Peabody, J., Pegg, J., Paielli, D., Movsas, B., Barton, K. N., Brown, S. L., Lu, M. and Kim, J. H. (2007). ”Five-year follow-up of trial of replication-competent adenovirus- mediated suicide gene therapy for treatment of prostate cancer.” Mol Th er 15(3): 636-642. Fueyo, J., Gomez-Manzano, C., Alemany, R., Lee, P. S., McDonnell, T. J., Mitlianga, P., Shi, Y. X., Levin, V. A., Yung, W. K. and Kyritsis, A. P. (2000). ”A mutant oncolytic adenovirus targeting the rb pathway produces anti-glioma eff ect in vivo.” Oncogene 19(1): 2-12. Fujiwara, T., Urata, Y. and Tanaka, N. (2007). ”Telomerase-specifi c oncolytic virotherapy for human cancer with the htert promoter.” Curr Cancer Drug Targets 7(2): 191-201. Futreal, P. A., Coin, L., Marshall, M., Down, T., Hubbard, T., Wooster, R., Rahman, N. and Stratton, M. R. (2004). ”A census of human cancer genes.” Nat Rev Cancer 4(3): 177-183. Galanis, E., Bateman, A., Johnson, K., Diaz, R. M., James, C. D., Vile, R. and Russell, S. J. (2001). ”Use of viral fusogenic membrane glycoproteins as novel therapeutic transgenes in gliomas.” Hum Gene Th er 12(7): 811-821. Ganesh, S., Gonzalez Edick, M., Idamakanti, N., Abramova, M., Vanroey, M., Robinson, M., Yun, C. O. and Jooss, K. (2007). ”Relaxin-expressing, fi ber chimeric oncolytic adenovirus prolongs survival of tumor-bearing mice.” Cancer Res 67(9): 4399-4407. Georgiades, J., Zielinski, T., Cicholska, A. and Jordan, E. (1959). ”Research on the oncolytic eff ect of apc viruses in cancer of the cervix uteri; preliminary report.” Biul Inst Med Morsk Gdansk 10: 49-57. Ginn, S. L., Alexander, I. E., Edelstein, M. L., Abedi, M. R. and Wixon, J. (2013). ”Gene therapy clinical trials worldwide to 2012 - an update.” J Gene Med 15(2): 65-77. Gros, A., Martinez-Quintanilla, J., Puig, C., Guedan, S., Mollevi, D. G., Alemany, R. and Cascallo, M. (2008). ”Bioselection of a gain of function mutation that enhances adenovirus 5 release and improves its antitumoral potency.” Cancer Res 68(21): 8928-8937. Hakkarainen, T., Sarkioja, M., Lehenkari, P., Miettinen, S., Ylikomi, T., Suuronen, R., Desmond, R. A., Kanerva, A. and Hemminki, A. (2007). ”Human mesenchymal stem cells lack tumor tropism but enhance the antitumor activity of oncolytic adenoviruses in orthotopic lung and breast tumors.” Hum Gene Th er 18(7): 627-641. Hanahan, D. and Weinberg, R. A. (2000). ”Th e hallmarks of cancer.” Cell 100(1): 57-70. Hanahan, D. and Weinberg, R. A. (2011). ”Hallmarks of cancer: Th e next generation.” Cell 144(5): 646-674.

70 References

Haviv, Y. S., Blackwell, J. L., Li, H., Wang, M., Lei, X. and Curiel, D. T. (2001). ”Heat shock and heat shock protein 70i enhance the oncolytic eff ect of replicative adenovirus.” Cancer Res 61(23): 8361- 8365. He, L. F., Gu, J. F., Tang, W. H., Fan, J. K., Wei, N., Zou, W. G., Zhang, Y. H., Zhao, L. L. and Liu, X. Y. (2008). ”Signifi cant antitumor activity of oncolytic adenovirus expressing human interferon-beta for hepatocellular carcinoma.” J Gene Med 10(9): 983-992. Hemminki, A. (2012). Mol Th er: 1654–1655. Hemminki, A. (2012). ”Treatment of chemotherapy-refractory cancer in the advanced therapy access program.” Mol Th er 20(9): 1654-1655. Hemminki, A. (2015). Across the valley of death with advanced therapy. Turku, Finland, Nomerta Hemminki, K., Li, X., Sundquist, K. and Sundquist, J. (2008). ”Familial risks for common diseases: Etiologic clues and guidance to gene identifi cation.” Mutat Res 658(3): 247-258. Hemminki, O., Bauerschmitz, G., Hemmi, S., Lavilla-Alonso, S., Diaconu, I., Guse, K., Koski, A., Desmond, R. A., Lappalainen, M., Kanerva, A., Cerullo, V., Pesonen, S. and Hemminki, A. (2011). ”Oncolytic adenovirus based on serotype 3.” Cancer Gene Th er 18(4): 288-296. Hemminki, O., Diaconu, I., Cerullo, V., Pesonen, S. K., Kanerva, A., Joensuu, T., Kairemo, K., Laasonen, L., Partanen, K., Kangasniemi, L., Lieber, A., Pesonen, S. and Hemminki, A. (2012). ”Ad3-htert-e1a, a fully serotype 3 oncolytic adenovirus, in patients with chemotherapy refractory cancer.” Mol Th er. Hemminki, O., Immonen, R., Narvainen, J., Kipar, A., Paasonen, J., Jokivarsi, K. T., Yli-Ollila, H., Soininen, P., Partanen, K., Joensuu, T., Parvianen, S., Pesonen, S. K., Koski, A., Vaha-Koskela, M., Cerullo, V., Pesonen, S., Grohn, O. H. and Hemminki, A. (2014). ”In vivo magnetic resonance imaging and spectroscopy identifi es oncolytic adenovirus responders.” Int J Cancer 134(12): 2878- 2890. Hemminki, O., Parviainen, S., Juhila, J., Turkki, R., Linder, N., Lundin, J., Kankainen, M., Ristimaki, A., Koski, A., Liikanen, I., Oksanen, M., Nettelbeck, D. M., Kairemo, K., Partanen, K., Joensuu, T., Kanerva, A. and Hemminki, A. (2015). ”Immunological data from cancer patients treated with ad5/3-e2f-delta24-gmcsf suggests utility for tumor immunotherapy.” Oncotarget 6(6): 4467-4481. Hemminki, O. H., A (2014). Oncolytic adenoviruses in the treatment of cancer in humans Gene therapy of cancer. S. G. a. E. C. Lattime, Elsevier/Academic Press. 3rd edition. Hodi, F. S., O’Day, S. J., McDermott, D. F., Weber, R. W., Sosman, J. A., Haanen, J. B., Gonzalez, R., Robert, C., Schadendorf, D., Hassel, J. C., Akerley, W., van den Eertwegh, A. J., Lutzky, J., Lorigan, P., Vaubel, J. M., Linette, G. P., Hogg, D., Ottensmeier, C. H., Lebbe, C., Peschel, C., Quirt, I., Clark, J. I., Wolchok, J. D., Weber, J. S., Tian, J., Yellin, M. J., Nichol, G. M., Hoos, A. and Urba, W. J. (2010). ”Improved survival with ipilimumab in patients with metastatic melanoma.” N Engl J Med 363(8): 711-723. Homet Moreno, B. and Ribas, A. (2015). ”Anti-programmed cell death protein-1/ligand-1 therapy in diff erent cancers.” Br J Cancer 112(9): 1421-1427. Hong, S. S., Karayan, L., Tournier, J., Curiel, D. T. and Boulanger, P. A. (1997). ”Adenovirus type 5 fi ber knob binds to mhc class i alpha2 domain at the surface of human epithelial and b lymphoblastoid cells.” EMBO J 16(9): 2294-2306. Hood, J. D., Bednarski, M., Frausto, R., Guccione, S., Reisfeld, R. A., Xiang, R. and Cheresh, D. A. (2002). ”Tumor regression by targeted gene delivery to the neovasculature.” Science 296(5577): 2404-2407.

71 References

Hoos, A., Eggermont, A. M., Janetzki, S., Hodi, F. S., Ibrahim, R., Anderson, A., Humphrey, R., Blumenstein, B., Old, L. and Wolchok, J. (2010). ”Improved endpoints for cancer immunotherapy trials.” J Natl Cancer Inst 102(18): 1388-1397. Huebner, R. J., Bell, J. A., Rowe, W. P., Ward, T. G., Suskind, R. G., Hartley, J. W. and Paff enbarger, R. S., Jr. (1955). ”Studies of adenoidal-pharyngeal-conjunctival vaccines in volunteers.” J Am Med Assoc 159(10): 986-989. Huebner, R. J., Rowe, W. P., Schatten, W. E., Smith, R. R. and Th omas, L. B. (1956). ”Studies on the use of viruses in the treatment of carcinoma of the cervix.” Cancer 9(6): 1211-1218. Jin, F., Kretschmer, P. J. and Hermiston, T. W. (2005). ”Identifi cation of novel insertion sites in the ad5 genome that utilize the ad splicing machinery for therapeutic gene expression.” Mol Th er 12(6): 1052-1063. Jogler, C., Hoff mann, D., Th eegarten, D., Grunwald, T., Uberla, K. and Wildner, O. (2006). ”Replication properties of human adenovirus in vivo and in cultures of primary cells from diff erent animal species.” J Virol 80(7): 3549-3558. June, C. H., Riddell, S. R. and Schumacher, T. N. (2015). ”Adoptive cellular therapy: A race to the fi nish line.” Sci Transl Med 7(280): 280ps287. Kanerva, A., Mikheeva, G. V., Krasnykh, V., Coolidge, C. J., Lam, J. T., Mahasreshti, P. J., Barker, S. D., Straughn, M., Barnes, M. N., Alvarez, R. D., Hemminki, A. and Curiel, D. T. (2002). ”Targeting adenovirus to the serotype 3 receptor increases gene transfer effi ciency to ovarian cancer cells.” Clin Cancer Res 8(1): 275-280. Kanerva, A., Nokisalmi, P., Diaconu, I., Koski, A., Cerullo, V., Liikanen, I., Tahtinen, S., Oksanen, M., Heiskanen, R., Pesonen, S., Joensuu, T., Alanko, T., Partanen, K., Laasonen, L., Kairemo, K., Kangasniemi, L. and Hemminki, A. (2013). ”Antiviral and antitumor t-cell immunity in patients treated with gm-csf-coding oncolytic adenovirus.” Clin Cancer Res 19(10): 2734-2744. Kanerva, A., Zinn, K. R., Chaudhuri, T. R., Lam, J. T., Suzuki, K., Uil, T. G., Hakkarainen, T., Bauerschmitz, G. J., Wang, M., Liu, B., Cao, Z., Alvarez, R. D., Curiel, D. T. and Hemminki, A. (2003). ”Enhanced therapeutic effi cacy for ovarian cancer with a serotype 3 receptor-targeted oncolytic adenovirus.” Mol Th er 8(3): 449-458. Kantoff , P. W., Higano, C. S., Shore, N. D., Berger, E. R., Small, E. J., Penson, D. F., Redfern, C. H., Ferrari, A. C., Dreicer, R., Sims, R. B., Xu, Y., Frohlich, M. W. and Schellhammer, P. F. (2010). ”Sipuleucel-t immunotherapy for castration-resistant prostate cancer.” N Engl J Med 363(5): 411- 422. Kantoff , P. W., Schuetz, T. J., Blumenstein, B. A., Glode, L. M., Bilhartz, D. L., Wyand, M., Manson, K., Panicali, D. L., Laus, R., Schlom, J., Dahut, W. L., Arlen, P. M., Gulley, J. L. and Godfrey, W. R. (2010). ”Overall survival analysis of a phase ii randomized controlled trial of a poxviral-based psa-targeted immunotherapy in metastatic castration-resistant prostate cancer.” J Clin Oncol 28(7): 1099-1105. Kaufman, H. L. and Bines, S. D. (2010). ”Optim trial: A phase iii trial of an oncolytic herpes virus encoding gm-csf for unresectable stage iii or iv melanoma.” Future Oncol 6(6): 941-949. Kelly, E. and Russell, S. J. (2007). ”History of oncolytic viruses: Genesis to genetic engineering.” Mol Th er 15(4): 651-659. Khuri, F. R., Nemunaitis, J., Ganly, I., Arseneau, J., Tannock, I. F., Romel, L., Gore, M., Ironside, J., MacDougall, R. H., Heise, C., Randlev, B., Gillenwater, A. M., Bruso, P., Kaye, S. B., Hong, W. K. and Kirn, D. H. (2000). ”A controlled trial of intratumoral onyx-015, a selectively-replicating adenovirus, in combination with cisplatin and 5-fl uorouracil in patients with recurrent head and neck cancer.” Nat Med 6(8): 879-885.

72 References

Kirkwood, J. M., Butterfi eld, L. H., Tarhini, A. A., Zarour, H., Kalinski, P. and Ferrone, S. (2012). ”Immunotherapy of cancer in 2012.” CA Cancer J Clin 62(5): 309-335. Kirn, D. (2006). ”Th e end of the beginning: Oncolytic virotherapy achieves clinical proof-of-concept.” Mol Th er 13(2): 237-238. Kochenderfer, J. N. and Rosenberg, S. A. (2013). ”Treating b-cell cancer with t cells expressing anti-cd19 chimeric antigen receptors.” Nat Rev Clin Oncol 10(5): 267-276. Komarova, S., Kawakami, Y., Stoff -Khalili, M. A., Curiel, D. T. and Pereboeva, L. (2006). ”Mesenchymal progenitor cells as cellular vehicles for delivery of oncolytic adenoviruses.” Mol Cancer Th er 5(3): 755-766. Koski, A., Kangasniemi, L., Escutenaire, S., Pesonen, S., Cerullo, V., Diaconu, I., Nokisalmi, P., Raki, M., Rajecki, M., Guse, K., Ranki, T., Oksanen, M., Holm, S. L., Haavisto, E., Karioja-Kallio, A., Laasonen, L., Partanen, K., Ugolini, M., Helminen, A., Karli, E., Hannuksela, P., Joensuu, T., Kanerva, A. and Hemminki, A. ”Treatment of cancer patients with a serotype 5/3 chimeric oncolytic adenovirus expressing gmcsf.” Mol Th er 18(10): 1874-1884. Koski, A., Kangasniemi, L., Escutenaire, S., Pesonen, S., Cerullo, V., Diaconu, I., Nokisalmi, P., Raki, M., Rajecki, M., Guse, K., Ranki, T., Oksanen, M., Holm, S. L., Haavisto, E., Karioja-Kallio, A., Laasonen, L., Partanen, K., Ugolini, M., Helminen, A., Karli, E., Hannuksela, P., Joensuu, T., Kanerva, A. and Hemminki, A. (2010). ”Treatment of cancer patients with a serotype 5/3 chimeric oncolytic adenovirus expressing gmcsf.” Mol Th er 18(10): 1874-1884. Koski, A., Karli, E., Kipar, A., Escutenaire, S., Kanerva, A. and Hemminki, A. (2013). ”Mutation of the fi ber shaft heparan sulphate binding site of a 5/3 chimeric adenovirus reduces liver tropism.” PLoS One 8(4): e60032. Koski et al., A. K., Helena Ahtinen, Anne Roivainen, Heidi Liljenbäck, Minna Oksanen, Kaarina Partanen, Leena Laasonen, Kalevi Kairemo, Akseli Hemminki (2012). [18f]-fdg-pet versus ct for evaluation of oncolytic virus treatment in advanced cancer patients. ASGCT. Philadelphia, USA. Kotha, P. L., Sharma, P., Kolawole, A. O., Yan, R., Alghamri, M. S., Brockman, T. L., Gomez- Cambronero, J. and Excoff on, K. J. (2015). ”Adenovirus entry from the apical surface of polarized epithelia is facilitated by the host innate immune response.” PLoS Pathog 11(3): e1004696. Kreppel, F. and Kochanek, S. (2008). ”Modifi cation of adenovirus gene transfer vectors with synthetic polymers: A scientifi c review and technical guide.” Mol Th er 16(1): 16-29. Kuhn, I., Harden, P., Bauzon, M., Chartier, C., Nye, J., Th orne, S., Reid, T., Ni, S., Lieber, A., Fisher, K., Seymour, L., Rubanyi, G. M., Harkins, R. N. and Hermiston, T. W. (2008). ”Directed evolution generates a novel oncolytic virus for the treatment of colon cancer.” PLoS One 3(6): e2409. Kumar, S., Gao, L., Yeagy, B. and Reid, T. (2008). ”Virus combinations and chemotherapy for the treatment of human cancers.” Curr Opin Mol Th er 10(4): 371-379. Kuruppu, D., Brownell, A. L., Zhu, A., Yu, M., Wang, X., Kulu, Y., Fuchs, B. C., Kawasaki, H. and Tanabe, K. K. (2007). ”Positron emission tomography of herpes simplex virus 1 oncolysis.” Cancer Res 67(7): 3295-3300. Kuster, K., Koschel, A., Rohwer, N., Fischer, A., Wiedenmann, B. and Anders, M. (2010). ”Downregulation of the coxsackie and adenovirus receptor in cancer cells by hypoxia depends on hif-1alpha.” Cancer Gene Th er 17(2): 141-146. L., V. (2014). Repeated intratumoral administration of oncos-102 leads to robust cellular and transcriptional immune activation at tumor site in an ovarian cancer patient. CIMT. Mainz, Germany). Lapteva, N., Aldrich, M., Weksberg, D., Rollins, L., Goltsova, T., Chen, S. Y. and Huang, X. F. (2009). ”Targeting the intratumoral dendritic cells by the oncolytic adenoviral vaccine expressing rantes elicits potent antitumor immunity.” J Immunother 32(2): 145-156. 73 References

Lee, Y. S., Kim, J. H., Choi, K. J., Choi, I. K., Kim, H., Cho, S., Cho, B. C. and Yun, C. O. (2006). ”Enhanced antitumor eff ect of oncolytic adenovirus expressing interleukin-12 and b7-1 in an immunocompetent murine model.” Clin Cancer Res 12(19): 5859-5868. Li, J. L., Liu, H. L., Zhang, X. R., Xu, J. P., Hu, W. K., Liang, M., Chen, S. Y., Hu, F. and Chu, D. T. (2009). ”A phase i trial of intratumoral administration of recombinant oncolytic adenovirus overexpressing hsp70 in advanced solid tumor patients.” Gene Th er 16(3): 376-382. Liikanen, I., Ahtiainen, L., Hirvinen, M. L., Bramante, S., Cerullo, V., Nokisalmi, P., Hemminki, O., Diaconu, I., Pesonen, S., Koski, A., Kangasniemi, L., Pesonen, S. K., Oksanen, M., Laasonen, L., Partanen, K., Joensuu, T., Zhao, F., Kanerva, A. and Hemminki, A. (2013). ”Oncolytic adenovirus with temozolomide induces autophagy and antitumor immune responses in cancer patients.” Mol Th er 21(6): 1212-1223. Liikanen, I., Koski, A., Merisalo-Soikkeli, M., Hemminki, O., Oksanen, M., Kairemo, K., Joensuu, T., Kanerva, A. and Hemminki, A. (2015). ”Serum hmgb1 is a predictive and prognostic biomarker for oncolytic immunotherapy.” Oncoimmunology 4(3): e989771. Liikanen, I., Monsurro, V., Ahtiainen, L., Raki, M., Hakkarainen, T., Diaconu, I., Escutenaire, S., Hemminki, O., Dias, J. D., Cerullo, V., Kanerva, A., Pesonen, S., Marzioni, D., Colombatti, M. and Hemminki, A. (2011). ”Induction of interferon pathways mediates in vivo resistance to oncolytic adenovirus.” Mol Th er 19(10): 1858-1866. Lin, K., Lipsitz, R., Miller, T. and Janakiraman, S. (2008). ”Benefi ts and harms of prostate-specifi c antigen screening for prostate cancer: An evidence update for the u.S. Preventive services task force.” Ann Intern Med 149(3): 192-199. Lin, X., Huang, H., Li, S., Li, H., Li, Y., Cao, Y., Zhang, D., Xia, Y., Guo, Y., Huang, W. and Jiang, W. (2007). ”A phase i clinical trial of an adenovirus-mediated endostatin gene (e10a) in patients with solid tumors.” Cancer Biol Th er 6(5): 648-653. Liu, X., Cao, X., Wei, R., Cai, Y., Li, H., Gui, J., Zhong, D., Liu, X. Y. and Huang, K. (2012). ”Gene-viro- therapy targeting liver cancer by a dual-regulated oncolytic adenoviral vector harboring il-24 and trail.” Cancer Gene Th er 19(1): 49-57. Long, G. V., Atkinson, V., Ascierto, P. A., Brady, B., Dutriaux, C., Maio, M., Mortier, L., Hassel, J. C., Rutkowski, P., McNeil, C., Kalinka-Warzocha, E., Savage, K. J., Hernberg, M., Lebbe, C., Charles, J., Mihalcioiu, C., Chiarion-Sileni, V., Mauch, C., Schmidt, H., Schadendorf, D., Gogas, H., Horak, C., Sharkey, B., Waxman, I. M. and Robert, C. (2015). ”Nivolumab improved survival vs dacarbazine in patients with untreated advanced melanoma.” J Transl Med 13: 2063. Lu, M., Freytag, S. O., Stricker, H., Kim, J. H., Barton, K. and Movsas, B. (2011). ”Adaptive seamless design for an effi cacy trial of replication-competent adenovirus-mediated suicide gene therapy and radiation in newly-diagnosed prostate cancer (recap trial).” Contemp Clin Trials 32(3): 453-460. Lu, W., Zheng, S., Li, X. F., Huang, J. J., Zheng, X. and Li, Z. (2004). ”Intra-tumor injection of h101, a recombinant adenovirus, in combination with chemotherapy in patients with advanced cancers: A pilot phase ii clinical trial.” World J Gastroenterol 10(24): 3634-3638. Lu, Z. Z., Wang, H., Zhang, Y., Cao, H., Li, Z., Fender, P. and Lieber, A. (2013). ”Penton-dodecahedral particles trigger opening of intercellular junctions and facilitate viral spread during adenovirus serotype 3 infection of epithelial cells.” PLoS Pathog 9(10): e1003718. Luisoni, S., Suomalainen, M., Boucke, K., Tanner, L. B., Wenk, M. R., Guan, X. L., Grzybek, M., Coskun, U. and Greber, U. F. (2015). ”Co-option of membrane wounding enables virus penetration into cells.” Cell Host Microbe 18(1): 75-85. Lukashok, S. A. and Horwitz, M. S. (1998). ”New perspectives in adenoviruses.” Curr Clin Top Infect Dis 18: 286-305.

74 References

Madisch, I., Harste, G., Pommer, H. and Heim, A. (2005). ”Phylogenetic analysis of the main neutralization and hemagglutination determinants of all human adenovirus prototypes as a basis for molecular classifi cation and taxonomy.” J Virol 79(24): 15265-15276. Malmstrom, P. U., Loskog, A. S., Lindqvist, C. A., Mangsbo, S. M., Fransson, M., Wanders, A., Gardmark, T. and Totterman, T. H. (2010). ”Adcd40l immunogene therapy for bladder carcinoma-- the fi rst phase i/iia trial.” Clin Cancer Res 16(12): 3279-3287. Maude, S. L., Frey, N., Shaw, P. A., Aplenc, R., Barrett, D. M., Bunin, N. J., Chew, A., Gonzalez, V. E., Zheng, Z., Lacey, S. F., Mahnke, Y. D., Melenhorst, J. J., Rheingold, S. R., Shen, A., Teachey, D. T., Levine, B. L., June, C. H., Porter, D. L. and Grupp, S. A. (2014). ”Chimeric antigen receptor t cells for sustained remissions in leukemia.” N Engl J Med 371(16): 1507-1517. Meier, O. and Greber, U. F. (2004). ”Adenovirus endocytosis.” J Gene Med 6 Suppl 1: S152-163. Melcher, A., Parato, K., Rooney, C. M. and Bell, J. C. (2011). ”Th under and lightning: Immunotherapy and oncolytic viruses collide.” Mol Th er 19(6): 1008-1016. Mellman, I., Coukos, G. and Dranoff , G. (2011). ”Cancer immunotherapy comes of age.” Nature 480(7378): 480-489. Motzer, R. J., Escudier, B., McDermott, D. F., George, S., Hammers, H. J., Srinivas, S., Tykodi, S. S., Sosman, J. A., Procopio, G., Plimack, E. R., Castellano, D., Choueiri, T. K., Gurney, H., Donskov, F., Bono, P., Wagstaff , J., Gauler, T. C., Ueda, T., Tomita, Y., Schutz, F. A., Kollmannsberger, C., Larkin, J., Ravaud, A., Simon, J. S., Xu, L. A., Waxman, I. M. and Sharma, P. (2015). ”Nivolumab versus everolimus in advanced renal-cell carcinoma.” N Engl J Med. Motzer, R. J., Rini, B. I., McDermott, D. F., Redman, B. G., Kuzel, T. M., Harrison, M. R., Vaishampayan, U. N., Drabkin, H. A., George, S., Logan, T. F., Margolin, K. A., Plimack, E. R., Lambert, A. M., Waxman, I. M. and Hammers, H. J. (2015). ”Nivolumab for metastatic renal cell carcinoma: Results of a randomized phase ii trial.” J Clin Oncol 33(13): 1430-1437. Nagy, R., Sweet, K. and Eng, C. (2004). ”Highly penetrant hereditary cancer syndromes.” Oncogene 23(38): 6445-6470. Nauts, H. C., Swift , W. E. and Coley, B. L. (1946). ”Th e treatment of malignant tumors by bacterial toxins as developed by the late william b. Coley, m.D., reviewed in the light of modern research.” Cancer Res 6: 205-216. Nemunaitis, J., Tong, A. W., Nemunaitis, M., Senzer, N., Phadke, A. P., Bedell, C., Adams, N., Zhang, Y. A., Maples, P. B., Chen, S., Pappen, B., Burke, J., Ichimaru, D., Urata, Y. and Fujiwara, T. (2010). ”A phase i study of telomerase-specifi c replication competent oncolytic adenovirus (telomelysin) for various solid tumors.” Mol Th er 18(2): 429-434. Nokisalmi, P., Pesonen, S., Escutenaire, S., Sarkioja, M., Raki, M., Cerullo, V., Laasonen, L., Alemany, R., Rojas, J., Cascallo, M., Guse, K., Rajecki, M., Kangasniemi, L., Haavisto, E., Karioja-Kallio, A., Hannuksela, P., Oksanen, M., Kanerva, A., Joensuu, T., Ahtiainen, L. and Hemminki, A. (2010). ”Oncolytic adenovirus icovir-7 in patients with advanced and refractory solid tumors.” Clin Cancer Res 16(11): 3035-3043. Pardoll, D. M. (2012). ”Th e blockade of immune checkpoints in cancer immunotherapy.”Nat Rev Cancer 12(4): 252-264. Perez-Jacoiste Asin, M. A., Fernandez-Ruiz, M., Lopez-Medrano, F., Lumbreras, C., Tejido, A., San Juan, R., Arrebola-Pajares, A., Lizasoain, M., Prieto, S. and Aguado, J. M. (2014). ”Bacillus calmette-guerin (bcg) infection following intravesical bcg administration as adjunctive therapy for bladder cancer: Incidence, risk factors, and outcome in a single-institution series and review of the literature.” Medicine (Baltimore) 93(17): 236-254.

75 References

Pesonen, S., Diaconu, I., Cerullo, V., Escutenaire, S., Raki, M., Kangasniemi, L., Nokisalmi, P., Dotti, G., Guse, K., Laasonen, L., Partanen, K., Karli, E., Haavisto, E., Oksanen, M., Karioja-Kallio, A., Hannuksela, P., Holm, S. L., Kauppinen, S., Joensuu, T., Kanerva, A. and Hemminki, A. (2012). ”Integrin targeted oncolytic adenoviruses ad5-d24-rgd and ad5-rgd-d24-gmcsf for treatment of patients with advanced chemotherapy refractory solid tumors.” Int J Cancer 130(8): 1937-1947. Pesonen, S., Diaconu, I., Kangasniemi, L., Ranki, T., Kanerva, A., Pesonen, S. K., Gerdemann, U., Leen, A. M., Kairemo, K., Oksanen, M., Haavisto, E., Holm, S. L., Karioja-Kallio, A., Kauppinen, S., Partanen, K. P., Laasonen, L., Joensuu, T., Alanko, T., Cerullo, V. and Hemminki, A. (2012). ”Oncolytic immunotherapy of advanced solid tumors with a cd40l-expressing replicating adenovirus: Assessment of safety and immunologic responses in patients.” Cancer Res 72(7): 1621- 1631. Pesonen, S., Nokisalmi, P., Escutenaire, S., Sarkioja, M., Raki, M., Cerullo, V., Kangasniemi, L., Laasonen, L., Ribacka, C., Guse, K., Haavisto, E., Oksanen, M., Rajecki, M., Helminen, A., Ristimaki, A., Karioja-Kallio, A., Karli, E., Kantola, T., Bauerschmitz, G., Kanerva, A., Joensuu, T. and Hemminki, A. (2010). ”Prolonged systemic circulation of chimeric oncolytic adenovirus ad5/3- cox2l-d24 in patients with metastatic and refractory solid tumors.” Gene Th er 17(7): 892-904. Postow, M. A., Chesney, J., Pavlick, A. C., Robert, C., Grossmann, K., McDermott, D., Linette, G. P., Meyer, N., Giguere, J. K., Agarwala, S. S., Shaheen, M., Ernstoff , M. S., Minor, D., Salama, A. K., Taylor, M., Ott, P. A., Rollin, L. M., Horak, C., Gagnier, P., Wolchok, J. D. and Hodi, F. S. (2015). ”Nivolumab and ipilimumab versus ipilimumab in untreated melanoma.” N Engl J Med. Prieto, P. A., Yang, J. C., Sherry, R. M., Hughes, M. S., Kammula, U. S., White, D. E., Levy, C. L., Rosenberg, S. A. and Phan, G. Q. (2012). ”Ctla-4 blockade with ipilimumab: Long-term follow-up of 177 patients with metastatic melanoma.” Clin Cancer Res 18(7): 2039-2047. Puzanov, I. e. a. (2014). Primary analysis of a phase 1b multicenter trial to evaluate safety and effi cacy of talimogene laherparepvec (t-vec) and ipilimumab (ipi) in previously untreated, unresected stage iiib-iv melanoma. ASCO. Rajecki, M., Kangasmaki, A., Laasonen, L., Escutenaire, S., Hakkarainen, T., Haukka, J., Ristimaki, A., Kairemo, K., Kangasniemi, L., Kiljunen, T., Joensuu, T., Pesonen, S. and Hemminki, A. (2011). ”Sodium iodide symporter spect imaging of a patient treated with oncolytic adenovirus ad5/3- delta24-hnis.” Mol Th er 19(4): 629-631. Raki, M., Sarkioja, M., Escutenaire, S., Kangasniemi, L., Haavisto, E., Kanerva, A., Cerullo, V., Joensuu, T., Oksanen, M., Pesonen, S. and Hemminki, A. (2011). ”Switching the fi ber knob of oncolytic adenoviruses to avoid neutralizing antibodies in human cancer patients.” J Gene Med 13(5): 253-261. Ranki, T. (2012). Preclinical and clinical evaluation of oncolytic immunotherapy with ad5/3-e2f1- δ24-gmcsf (cgtg-602), a gm-csf producing adenovirus targeted to tumors on four levels ASGCT 2012. Philadelphia. Raper, S. E., Chirmule, N., Lee, F. S., Wivel, N. A., Bagg, A., Gao, G. P., Wilson, J. M. and Batshaw, M. L. (2003). ”Fatal systemic infl ammatory response syndrome in a ornithine transcarbamylase defi cient patient following adenoviral gene transfer.” Mol Genet Metab 80(1-2): 148-158. Redelman-Sidi, G., Glickman, M. S. and Bochner, B. H. (2014). ”Th e mechanism of action of bcg therapy for bladder cancer--a current perspective.” Nat Rev Urol 11(3): 153-162. Redman, J. M., Hill, E. M., AlDeghaither, D. and Weiner, L. M. (2015). ”Mechanisms of action of therapeutic antibodies for cancer.” Mol Immunol. Reid, T., Galanis, E., Abbruzzese, J., Sze, D., Andrews, J., Romel, L., Hatfi eld, M., Rubin, J. and Kirn, D. (2001). ”Intra-arterial administration of a replication-selective adenovirus (dl1520) in patients with colorectal carcinoma metastatic to the liver: A phase i trial.” Gene Th er 8(21): 1618-1626.

76 References

Reid, T. R., Freeman, S., Post, L., McCormick, F. and Sze, D. Y. (2005). ”Eff ects of onyx-015 among metastatic colorectal cancer patients that have failed prior treatment with 5-fu/leucovorin.” Cancer Gene Th er 12(8): 673-681. Robbins, P. F., Morgan, R. A., Feldman, S. A., Yang, J. C., Sherry, R. M., Dudley, M. E., Wunderlich, J. R., Nahvi, A. V., Helman, L. J., Mackall, C. L., Kammula, U. S., Hughes, M. S., Restifo, N. P., Raff eld, M., Lee, C. C., Levy, C. L., Li, Y. F., El-Gamil, M., Schwarz, S. L., Laurencot, C. and Rosenberg, S. A. (2011). ”Tumor regression in patients with metastatic synovial cell sarcoma and melanoma using genetically engineered lymphocytes reactive with ny-eso-1.” J Clin Oncol 29(7): 917-924. Robert Hans Ingemar Andtbacka, F. A. C., Th omas Amatruda, Neil N. Senzer, Jason Chesney, Keith A. Delman, Lynn E. Spitler, Igor Puzanov, Susan Doleman, Yining Ye, Ari M. Vanderwalde, Robert Coffi n and Howard Kaufman (2013). ”Optim: A randomized phase iii trial of talimogene laherparepvec (t-vec) versus subcutaneous (sc) granulocyte-macrophage colony-stimulating factor (gm-csf) for the treatment (tx) of unresected stage iiib/c and iv melanoma.” Journal of Clinical Oncology, 2013 ASCO Annual Meeting Abstracts. Vol 31, No 18_suppl (June 20 Supplement), 2013: LBA9008. Roelvink, P. W., Lizonova, A., Lee, J. G., Li, Y., Bergelson, J. M., Finberg, R. W., Brough, D. E., Kovesdi, I. and Wickham, T. J. (1998). ”Th e coxsackievirus-adenovirus receptor protein can function as a cellular attachment protein for adenovirus serotypes from subgroups a, c, d, e, and f.” J Virol 72(10): 7909-7915. Roelvink, P. W., Mi Lee, G., Einfeld, D. A., Kovesdi, I. and Wickham, T. J. (1999). ”Identifi cation of a conserved receptor-binding site on the fi ber proteins of car-recognizing adenoviridae.” Science 286(5444): 1568-1571. Rosenberg, S. A. (2007). ”Interleukin 2 for patients with renal cancer.” Nat Clin Pract Oncol 4(9): 497. Rosenberg, S. A. (2011). ”Cell transfer immunotherapy for metastatic solid cancer--what clinicians need to know.” Nat Rev Clin Oncol 8(10): 577-585. Rosewell, A. (2012). Effi cient apical transduction of polarized human airway epithelium by adenovirus serotype 3 ASGCT 2012. Philadelphia. Saenger, Y. M. and Wolchok, J. D. (2008). ”Th e heterogeneity of the kinetics of response to ipilimumab in metastatic melanoma: Patient cases.” Cancer Immun 8: 1. Sarkar, B. K., Chakraborty, C., Sharma, A. R., Bae, K. J., Sharma, G., Doss, G. P., Dutta, D., Ding, S., Ganbold, B., Nam, J. S. and Lee, S. S. (2014). ”Novel biomarker for prostate cancer diagnosis by mrs.” Front Biosci (Landmark Ed) 19: 1186-1201. Sarkioja, M., Pesonen, S., Raki, M., Hakkarainen, T., Salo, J., Ahonen, M. T., Kanerva, A. and Hemminki, A. (2008). ”Changing the adenovirus fi ber for retaining gene delivery effi cacy in the presence of neutralizing antibodies.” Gene Th er 15(12): 921-929. Schreiber, R. D., Old, L. J. and Smyth, M. J. (2011). ”Cancer immunoediting: Integrating immunity’s roles in cancer suppression and promotion.” Science 331(6024): 1565-1570. Scott, A. M., Wolchok, J. D. and Old, L. J. (2012). ”Antibody therapy of cancer.” Nat Rev Cancer 12(4): 278-287. Segerman, A., Arnberg, N., Erikson, A., Lindman, K. and Wadell, G. (2003). ”Th ere are two diff erent species b adenovirus receptors: Sbar, common to species b1 and b2 adenoviruses, and sb2ar, exclusively used by species b2 adenoviruses.” J Virol 77(2): 1157-1162. Seiradake, E., Henaff , D., Wodrich, H., Billet, O., Perreau, M., Hippert, C., Mennechet, F., Schoehn, G., Lortat-Jacob, H., Dreja, H., Ibanes, S., Kalatzis, V., Wang, J. P., Finberg, R. W., Cusack, S. and Kremer, E. J. (2009). ”Th e cell adhesion molecule ”car” and sialic acid on human erythrocytes infl uence adenovirus in vivo biodistribution.” PLoS Pathog 5(1): e1000277. 77 References

Shashkova, E. V., Kuppuswamy, M. N., Wold, W. S. and Doronin, K. (2008). ”Anticancer activity of oncolytic adenovirus vector armed with ifn-alpha and adp is enhanced by pharmacologically controlled expression of trail.” Cancer Gene Th er 15(2): 61-72. Shayakhmetov, D. M. and Lieber, A. (2000). ”Dependence of adenovirus infectivity on length of the fi ber shaft domain.” J Virol 74(22): 10274-10286. Shenk, T. (1996). Fields virology, 3rd ed. Adenoviridae: Th e viruses and their replication, Lippincott-Raven Publishers, Philadelphia, Pa. Sheridan, C. (2013). ”Amgen announces oncolytic virus shrinks tumors.” Nat Biotechnol 31(6): 471- 472. Short, J. J., Pereboev, A. V., Kawakami, Y., Vasu, C., Holterman, M. J. and Curiel, D. T. (2004). ”Adenovirus serotype 3 utilizes cd80 (b7.1) and cd86 (b7.2) as cellular attachment receptors.” Virology 322(2): 349-359. Short, J. J., Vasu, C., Holterman, M. J., Curiel, D. T. and Pereboev, A. (2006). ”Members of adenovirus species b utilize cd80 and cd86 as cellular attachment receptors.” Virus Res 122(1-2): 144-153. Sinkovics, J. and Horvath, J. (1993). ”New developments in the virus therapy of cancer: A historical review.” Intervirology 36(4): 193-214. Sirena, D., Ruzsics, Z., Schaff ner, W., Greber, U. F. and Hemmi, S. (2005). ”Th e nucleotide sequence and a fi rst generation gene transfer vector of species b human adenovirus serotype 3.” Virology 343(2): 283-298. Small, E. J., Carducci, M. A., Burke, J. M., Rodriguez, R., Fong, L., van Ummersen, L., Yu, D. C., Aimi, J., Ando, D., Working, P., Kirn, D. and Wilding, G. (2006). ”A phase i trial of intravenous cg7870, a replication-selective, prostate-specifi c antigen-targeted oncolytic adenovirus, for the treatment of hormone-refractory, metastatic prostate cancer.” Mol Th er 14(1): 107-117. Smith, T. A., Idamakanti, N., Rollence, M. L., Marshall-Neff , J., Kim, J., Mulgrew, K., Nemerow, G. R., Kaleko, M. and Stevenson, S. C. (2003). ”Adenovirus serotype 5 fi ber shaft infl uences in vivo gene transfer in mice.” Hum Gene Th er 14(8): 777-787. Sobin, L. H. (2001). ”Tnm: Principles, history, and relation to other prognostic factors.” Cancer 91(8 Suppl): 1589-1592. Southam, C. M. and Moore, A. E. (1952). ”Clinical studies of viruses as antineoplastic agents with particular reference to egypt 101 virus.” Cancer 5(5): 1025-1034. Strauss, R. and Lieber, A. (2009). ”Anatomical and physical barriers to tumor targeting with oncolytic adenoviruses in vivo.” Curr Opin Mol Th er 11(5): 513-522. Su, C., Peng, L., Sham, J., Wang, X., Zhang, Q., Chua, D., Liu, C., Cui, Z., Xue, H., Wu, H., Yang, Q., Zhang, B., Liu, X., Wu, M. and Qian, Q. (2006). ”Immune gene-viral therapy with triplex effi cacy mediated by oncolytic adenovirus carrying an interferon-gamma gene yields effi cient antitumor activity in immunodefi cient and immunocompetent mice.” Mol Th er 13(5): 918-927. Sudhakar, A. (2009). ”History of cancer, ancient and modern treatment methods.” J Cancer Sci Th er 1(2): 1-4. Tagawa, M., Kawamura, K., Shimozato, O., Ma, G., Li, Q., Suzuki, N., Shimada, H. and Ochiai, T. (2006). ”Virology- and immunology-based gene therapy for cancer.” Cancer Immunol Immunother 55(11): 1420-1425. Tahtinen, S., Gronberg-Vaha-Koskela, S., Lumen, D., Merisalo-Soikkeli, M., Siurala, M., Airaksinen, A. J., Vaha-Koskela, M. and Hemminki, A. (2015). ”Adenovirus improves the effi cacy of adoptive t-cell therapy by recruiting immune cells to and promoting their activity at the tumor.” Cancer Immunol Res.

78 References

Tomko, R. P., Xu, R. and Philipson, L. (1997). ”Hcar and mcar: Th e human and mouse cellular receptors for subgroup c adenoviruses and group b coxsackieviruses.” Proc Natl Acad Sci U S A 94(7): 3352-3356. Topi M. O. Turunen, P. N., Vincenzo Cerullo, Sari Pesonen, Minna Oksanen, Sophie Escutenaire, Akseli Hemminki (2009). Eff ect of oncolytic adenovirus on tumor marker levels in cancer patients and in preclinical test systems. ASGCT. Toth, K., Dhar, D. and Wold, W. S. (2010). ”Oncolytic (replication-competent) adenoviruses as anticancer agents.” Expert Opin Biol Th er 10(3): 353-368. Toth, K. and Wold, W. S. (2010). ”Increasing the effi cacy of oncolytic adenovirus vectors.” Viruses 2(9): 1844-1866. Turcotte, S. and Rosenberg, S. A. (2011). ”Immunotherapy for metastatic solid cancers.” Adv Surg 45: 341-360. Tuve, S., Wang, H., Ware, C., Liu, Y., Gaggar, A., Bernt, K., Shayakhmetov, D., Li, Z., Strauss, R., Stone, D. and Lieber, A. (2006). ”A new group b adenovirus receptor is expressed at high levels on human stem and tumor cells.” J Virol 80(24): 12109-12120. Ulasov, I. V., Rivera, A. A., Han, Y., Curiel, D. T., Zhu, Z. B. and Lesniak, M. S. (2007). ”Targeting adenovirus to cd80 and cd86 receptors increases gene transfer effi ciency to malignant glioma cells.” J Neurosurg 107(3): 617-627. van Geer, M. A., Bakker, C. T., Koizumi, N., Mizuguchi, H., Wesseling, J. G., Oude Elferink, R. P. and Bosma, P. J. (2009). ”Ephrin a2 receptor targeting does not increase adenoviral pancreatic cancer transduction in vivo.” World J Gastroenterol 15(22): 2754-2762. Wang, H., Li, Z., Yumul, R., Lara, S., Hemminki, A., Fender, P. and Lieber, A. (2011). ”Multimerization of adenovirus serotype 3 fi ber knob domains is required for effi cient binding of virus to desmoglein 2 and subsequent opening of epithelial junctions.” J Virol 85(13): 6390-6402. Wang, H., Li, Z. Y., Liu, Y., Persson, J., Beyer, I., Moller, T., Koyuncu, D., Drescher, M. R., Strauss, R., Zhang, X. B., Wahl, J. K., 3rd, Urban, N., Drescher, C., Hemminki, A., Fender, P. and Lieber, A. (2011). ”Desmoglein 2 is a receptor for adenovirus serotypes 3, 7, 11 and 14.” Nat Med 17(1): 96-104. Wardle, J., Robb, K., Vernon, S. and Waller, J. (2015). ”Screening for prevention and early diagnosis of cancer.” Am Psychol 70(2): 119-133. Vestergaard, T. A., Nielsen, S. L., Dahlerup, J. F. and Hornung, N. (2008). ”Fecal calprotectin: Assessment of a rapid test.” Scand J Clin Lab Invest 68(4): 343-347. Vogelstein, B., Papadopoulos, N., Velculescu, V. E., Zhou, S., Diaz, L. A., Jr. and Kinzler, K. W. (2013). ”Cancer genome landscapes.” Science 339(6127): 1546-1558. Wolchok, J. D., Hoos, A., O’Day, S., Weber, J. S., Hamid, O., Lebbe, C., Maio, M., Binder, M., Bohnsack, O., Nichol, G., Humphrey, R. and Hodi, F. S. (2009). ”Guidelines for the evaluation of immune therapy activity in solid tumors: Immune-related response criteria.” Clin Cancer Res 15(23): 7412- 7420. Wolchok, J. D., Kluger, H., Callahan, M. K., Postow, M. A., Rizvi, N. A., Lesokhin, A. M., Segal, N. H., Ariyan, C. E., Gordon, R. A., Reed, K., Burke, M. M., Caldwell, A., Kronenberg, S. A., Agunwamba, B. U., Zhang, X., Lowy, I., Inzunza, H. D., Feely, W., Horak, C. E., Hong, Q., Korman, A. J., Wigginton, J. M., Gupta, A. and Sznol, M. (2013). ”Nivolumab plus ipilimumab in advanced melanoma.” N Engl J Med 369(2): 122-133. Wolchok, J. D., Weber, J. S., Maio, M., Neyns, B., Harmankaya, K., Chin, K., Cykowski, L., de Pril, V., Humphrey, R. and Lebbe, C. (2013). ”Four-year survival rates for patients with metastatic melanoma who received ipilimumab in phase ii clinical trials.” Ann Oncol 24(8): 2174-2180.

79 References

Wolfrum, N. and Greber, U. F. (2013). ”Adenovirus signalling in entry.” Cell Microbiol 15(1): 53-62. Volk, A. L., Rivera, A. A., Kanerva, A., Bauerschmitz, G., Dmitriev, I., Nettelbeck, D. M. and Curiel, D. T. (2003). ”Enhanced adenovirus infection of melanoma cells by fi ber-modifi cation: Incorporation of rgd peptide or ad5/3 chimerism.” Cancer Biol Th er 2(5): 511-515. Wu, E., Pache, L., Von Seggern, D. J., Mullen, T. M., Mikyas, Y., Stewart, P. L. and Nemerow, G. R. (2003). ”Flexibility of the adenovirus fi ber is required for effi cient receptor interaction.” J Virol 77(13): 7225-7235. Wu, R., Forget, M. A., Chacon, J., Bernatchez, C., Haymaker, C., Chen, J. Q., Hwu, P. and Radvanyi, L. G. (2012). ”Adoptive t-cell therapy using autologous tumor-infi ltrating lymphocytes for metastatic melanoma: Current status and future outlook.” Cancer J 18(2): 160-175. Xia, Z. J., Chang, J. H., Zhang, L., Jiang, W. Q., Guan, Z. Z., Liu, J. W., Zhang, Y., Hu, X. H., Wu, G. H., Wang, H. Q., Chen, Z. C., Chen, J. C., Zhou, Q. H., Lu, J. W., Fan, Q. X., Huang, J. J. and Zheng, X. (2004). ”[phase iii randomized clinical trial of intratumoral injection of e1b gene-deleted adenovirus (h101) combined with cisplatin-based chemotherapy in treating squamous cell cancer of head and neck or esophagus].” Ai Zheng 23(12): 1666-1670. Yan, J., Allendorf, D. J., Li, B., Yan, R., Hansen, R. and Donev, R. (2008). ”Th e role of membrane complement regulatory proteins in cancer immunotherapy.” Adv Exp Med Biol 632: 159-174. Ying, B., Toth, K., Spencer, J. F., Meyer, J., Tollefson, A. E., Patra, D., Dhar, D., Shashkova, E. V., Kuppuswamy, M., Doronin, K., Th omas, M. A., Zumstein, L. A., Wold, W. S. and Lichtenstein, D. L. (2009). ”Ingn 007, an oncolytic adenovirus vector, replicates in syrian hamsters but not mice: Comparison of biodistribution studies.” Cancer Gene Th er 16(8): 625-637. Yoo, J. Y., Kim, J. H., Kwon, Y. G., Kim, E. C., Kim, N. K., Choi, H. J. and Yun, C. O. (2007). ”Vegf- specifi c short hairpin rna-expressing oncolytic adenovirus elicits potent inhibition of angiogenesis and tumor growth.” Mol Th er 15(2): 295-302. Young, B. A., Spencer, J. F., Ying, B., Toth, K. and Wold, W. S. (2013). ”Th e eff ects of radiation on antitumor effi cacy of an oncolytic adenovirus vector in the syrian hamster model.” Cancer Gene Th er 20(9): 531-537. Zhang, Y. and Bergelson, J. M. (2005). ”Adenovirus receptors.” J Virol 79(19): 12125-12131. Zhao, L., Claggett, B., Tian, L., Uno, H., Pfeff er, M. A., Solomon, S. D., Trippa, L. and Wei, L. J. (2015). ”On the restricted mean survival time curve in survival analysis.” Biometrics. Zielinski, T. and Jordan, E. (1969). ”[remote results of clinical observation of the oncolytic action of adenoviruses on cervix cancer].” Nowotwory 19(3): 217-221.

80