Radiolabeled HER-2 Binding Affibody Molecules for Tumor Targeting

Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Medicine 127

Preclinical Studies

ANN-CHARLOTT STEFFEN

ACTA UNIVERSITATIS UPSALIENSIS ISSN 1651-6206 UPPSALA ISBN 91-554-6503-X 2006 urn:nbn:se:uu:diva-6678 ! "#$% & ' ( & & ' ' )* & + ,- .' / /'-

&& 01- !- 230 4 ( && + & . . ( (- - - $ 5- !6 - - 748 "$0%%90!% 60:-

1 ' ' ' && ( - 4 ( ' ' & ' - . ( ( ( ' '- 7 ' ' && / ' '(' && & ' 230 ' ( / ' (( & & / - & ( / ' $ %7 ' && & / 230 ( - 7 / ' / ' ' 230 ( ( && ( & ' - .' ; & ' && & ' (' ' - 0 0 & $ (' ' < / ' ' ; ( - .' ( && & ' && && / / ' ' && & ' ; - 7 / & ' ' (' ( ' ' && && ( - .' && / / ' ' ' 0 ( $$ - & / 230 ( ( ' $$ && - .' ' && / & && / ' ; ;- .' $$ && / /' & ' $$ / - . ' <( ' / / ( & $$ ( - .' ' ' ; & && & & ' & -

230 && $$ ( ( ( ' ( (

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

= 01' && !

78 $!%$0! ! 748 "$0%%90!% 60: # ### 0!!5> )' #?? --? @ A # ### 0!!5>, List of papers

I Wikman M, Steffen AC, Gunneriusson E, Tolmachev V, Adams GP, Carlsson J, Ståhl S: Selection and characterization of HER2/neu-binding affibody ligands. Protein Engineering Design & Selection 2004, 17(5):455-462.

II Steffen AC, Wikman M, Tolmachev V, Adams GP, Nilsson FY, Stahl S, Carlsson J: In Vitro Characterization of a Bivalent Anti- HER-2 Affibody with Potential for Radionuclide-Based Diagnos- tics. Cancer Biotherapy and Radiopharmaceuticals 2005, 20(3):239-248.

III Steffen AC, Orlova A, Wikman M, Nilsson FY, Ståhl S, Adams GP, Tolmachev V, Carlsson J: Affibody mediated tumour targeting of HER-2 expressing xenografts in mice. European Journal of Nuclear Medicine and Molecular Imaging 2006, In press

IV Ekerljung L, Steffen AC, Carlsson J, Lennartsson J: Effects of HER2-binding Affibody Molecules on Intracellular Signaling Pathways. Tumor Biology 2006, vol 27, In press

V Steffen AC, Göstring L, Tolmachev V, Carlsson J: Differences in 211 sensitivity for At-(ZHER2:4)2 treatment between three HER-2 overexpressing cell lines. Submitted to Radiation Research 2006

VI Steffen AC, Almqvist Y, Chyan MK, Lundqvist H, Tolmachev V, Wilbur DS, Carlsson J: Biodistribution and dose calculations for 211At labeled HER-2 binding affibody molecules. Manuscript, 2006

Reprints were made with permission from Oxford Journals (I), Mary Ann Liebert, Inc. Publishers (II), Springer Science and Business Media (III) and S. Karger AG, Basel (IV).

Contents

Introduction...... 9 Cancer ...... 9 Cancer treatments today ...... 9 Experimental therapies ...... 9 Cancer detection ...... 10 Tumor targeting...... 10 Molecular targets ...... 11 The EGF receptor family ...... 11 The EGF receptor /HER-1...... 11 HER-2 / neu...... 12 HER-3 and HER-4...... 13 Signaling...... 13 Targeting agents ...... 14 Antibodies...... 15 Antibody fragments...... 17 Peptides...... 19 Affinity proteins...... 19 Affibody molecules ...... 20 Selection systems ...... 21 Phage display...... 21 Other systems ...... 23 Therapeutic/visualizing moieties ...... 23 Therapy ...... 23 Radionuclides ...... 23 Toxins...... 25 Cancer drugs...... 25 Visualization ...... 25 The present study ...... 27 Selection of HER-2 binding affibody molecules (I & II)...... 27 Biacore testing (I & II) ...... 28 125I labeled HER-2 binding affibody molecules ...... 29 Cellular testing (I & II)...... 29 Specific binding / epitope comparison...... 29 Cellular internalization and retention...... 30 Cellular retention...... 31 Biodistribution in mice (III)...... 32 Gamma camera imaging (III) ...... 34 Autoradiography/immunohistochemistry (III) ...... 35 Cellular effects of the HER-2 binding affibody molecules (IV) ...... 35 211At labeled HER-2 binding affibody molecules ...... 37 In vitro therapy (V)...... 38 Biodistribution (VI) ...... 43 Other ongoing studies with tumor targeting affibody molecules...... 47 Summary...... 48 Future studies...... 50 Acknowledgments...... 51 References...... 54 Abbreviations

AHNP Anti-HER-2/neu peptide CDR Complementarity determining region CPM Counts per minute DAG Diacylglycerol DNA Deoxyribonucleic nucleic acid ECD Extracellular domain EGF Epidermal growth factor EGFR Epidermal growth factor receptor Erk Extracellular regulated kinase Fab Antigen-binding fragment FACS Fluorescence activated cell sorting Fv Variable fragment GDP Guanosine diphosphate GTP Guanosine triphosphate Gy Gray HAMA Human anti-mouse antibody HER Human EGF receptor i.p. Intraperitoneally i.v. Intravenously kBq Kilobequerel

KD Dissociation equilibrium constant kDa Kilodalton

kon Association rate constant

koff Dissociation rate constant LET Linear energy transfer MAb Monoclonal antibody MAPK Mitogen-activated protein kinase p.i. Post-injection PCR Polymerase chain reaction PDK Phosphatidylinositol-dependent protein kinase PET Positron emission tomography PI(3)K Phosphatidylinositol-3-OH kinase

PIP2 Phosphatidylinositol 4,5-biphosphate

PIP3 Phosphatidylinositol 3,4,5-triphosphate PKC Protein kinase C PLC phospholipase C PAB N-succinimidyl-para-astatobenzoate PIB N-succinimidyl-para-iodobenzoate PSA Prostate specific antigen RBE Relative biological effectiveness RNA Ribonucleic acid s.c. Subcutaneously scFv Single-chain Fv SD Standard deviation SELEX Systematic evolution of ligands by exponential enrichment SPECT Single photon emission computed tomography SPMB N-succinimidyl-para-(tri-methylstannyl)benzoate TGF Transforming growth factor Introduction

Cancer Cancer is one of the most common diseases in the western world today. Ap- proximately one out of three people will receive a cancer diagnose in their life time, and cancer is responsible for about 23% of all deaths. That makes cancer the second most common cause of death, after cardiovascular disease, in the western world. The number of diagnosed cases has dramatically increased during the last decades. The aging population is believed to be responsible for a large part of the increase, but taking this into account, there has still been an increase in diagnosed cases [1]. The mortality has, during the same time, decreased somewhat, and the expectation of life has markedly increased [2]. This indicates improvements in cancer diagnosis and treatment.

Cancer treatments today Modern cancer treatments are mainly divided into two categories based on the properties of the tumor. 1) Localized primary tumors and metastasis, where surgery and external radiation therapies can be applied, and 2) sub- clinical tumors and disseminated tumor cells, where chemotherapy and hormone therapy are used.

Experimental therapies While there is a constant improvement in current treatment modalities, much research is now focusing on new approaches in cancer therapy. Many of these modalities are based on basic tumor biology and often a combination of approaches is used in one treatment setting. 1. Anti-angiogenesis [3], where the formation of new blood vessels is targeted, thereby shutting off the supply of nutrients to the tumor and preventing tumor growth. 2. Apoptosis regulation [4]. Most tumors are defective in their apoptosis ability - their ability to self-destruct when too severely damaged. Also, apoptosis dysregulation is often responsible for resistance to chemotherapy and radiotherapy. By different ap-

9 proaches against this characteristic, tumor formation can be prevented and the applicability of current treatments can be enlarged. 3. Differentiation therapy [5], where biological approaches are used to turn a tumor cell off or reverting to a more benign phenotype. 4. Gene therapy [6], where therapeutic or cytotoxic genes are trans- ferred to the patient, and specifically expressed in the tumor, where tumor-associated transcription factors are present. 5. Immunotherapy [7], where the patients own immune system is stimulated to attack the tumor. 6. Radionuclide therapy [8], where therapeutic radionuclides are specifically targeted to tumor cells resulting in deposition of high doses of radioactivity locally to the tumor. This will be discussed in more detail later. 7. Signal transduction modification [9], for down-regulating systems associated with uncontrolled growth, and up-regulating control systems, such as the apoptosis machinery.

Cancer detection An early cancer diagnose often determines the possibility of curing the disease and there has been a massive debate about whether to screen for various cancer forms in the population or not. In Sweden, mammography has been offered to detect early breast cancer since the 1980’s. Cell sample analysis to screen for cervix cancer has been offered since the 1970’s. There is also a need to screen for the most common form of cancer – prostate cancer, and there is a screening method available – the PSA screen. This method has, however, been widely debated [10] and is not used in routine today. Tumor visualization techniques are widely used to determine the exact lo- cation of the tumors before surgery, and also to determine if metastases are present and to monitor the tumor burden during therapy. Apart from the x- ray technique, which is used in mammography, gamma-camera techniques and PET-technology are used for visualization. A tumor tracing molecule, coupled to a suitable radionuclide (gamma-emitting for use in the gamma- camera or positron-emitting for use in PET) is administrated, and when the tracer has reached the tumor, images are taken. The most used PET-tracer is FDG - a 18F labeled glucose molecule, that visualizes areas with high me- tabolism, such as tumor and inflammation areas.

Tumor targeting Most cancer treatments (e.g. radiation therapy and chemotherapy) affect all dividing cells to some degree. This is often the dose-limiting factor for those

10 treatment modalities, since e.g. stem cells in the bone marrow and proliferat- ing cells in the intestine are also affected. If the dose to the normal tissues could be reduced, a higher total dose could be administrated, resulting in a higher chance of curing the cancer. This idea was stated over 100 years ago by the German scientist Paul Ehrlich in his “magic bullet” concept [11].

Molecular targets The perfect target structure for use in tumor targeting should be present only on tumor cells, and be easily accessible for targeting agents located in the blood. This makes cell surface structures, or structures located in the extracellular matrix, suitable. It is also important that the target is available during the whole treatment time, so that the cancer cells do not down- regulate its expression during treatment. This can be achieved if the target structure is needed for the cancer cell to grow and divide.

The EGF receptor family The Epidermal Growth Factor (EGF) receptor family, also referred to as the ErbB family, consists of four members – the EGF receptor, or Human EGF receptor 1 (HER-1), HER-2, HER-3 and HER-4, see Figure 1 [12]. These growth factor receptors are often overexpressed by tumor cells and the overexpression gives the cells a growth advantage over normal cells. However, the members in the EGF receptor family also play an important role in the growth of normal cells.

Figure 1. The EGF receptor family (HER-1 to HER-4)

The EGF receptor /HER-1 The growth factor EGF was first described by Cohen in 1962 [13] and EGF and the EGF receptor has since then been thoroughly investigated [14, 15]. EGF receptors are overexpressed in many different carcinomas, such as glioma [16, 17], urinary bladder carcinomas [18], lung cancers and gastrointestinal cancers [19]. The expression in normal tissues is most often low compared with the expression in tumor cells. One important exception is the expression of EGF receptors in the hepatocytes in the liver, that express

11 about 105-106 EGF receptors per cell [20, 21], which is at approximately the same level as most tumor cells. This limits the use of the EGF receptor as target in systemic treatments. Instead, local therapies, such as urinary bladder installation for bladder carcinomas [22] and injection into the surgical cavity in glioma patients are of interest.

Figure 2. EGF receptor activation. I-IV: domains of the extracellular part of the receptor, TM: transmembrane domain, IC: Intracellular domain.

The EGF receptor has many ligands [23], such as EGF and the Transforming Growth Factor-D (TGF-D). As seen in Figure 2, ligand binding results in a conformational change, exposing the dimerization arm of domain II. As an effect of the dimerization event, the tyrosine kinase domains located on the intracellular domain of the receptor are phosphorylated.

HER-2 / neu HER-2 was first described as a rat oncogene called neu, which was found in the late 1970’s in rat neuroblastoma. Later, the homology to the EGF receptor was discovered, and the human variant of the protein was found. Around 1990, the search for ligands for the orphan receptor resulted in many candidate peptides, all sharing a common structure [24] and the ability to phosphorylate HER-2 [25-32]. It was later discovered that the newly found peptides, or neuregulins, did not bind directly to HER-2, but were ligands to HER-3 and HER-4 [33, 34] and activated HER-2 only through heterodimerization. HER-2 is the preferred dimerization partner for all the other members in the EGF receptor family [35], but is now again considered to be an orphan receptor. However, HER-2 is not dependent on ligand activation for activation. As seen in Figure 3, the dimerization loop of the extracellular domain II is constitutively exposed, allowing for dimerization with any member of the EGF receptor family. When overexpressed, HER-2 homodimers are often found (Figure 3).

12 Figure 3. Homodimerization and activation of HER-2. I-IV: domains of the extracellular part of the receptor, TM: transmembrane domain, IC: Intracellular domain.

HER-2 is overexpressed in many malignances, such as urinary bladder carcinomas [36, 37] and carcinomas in the gastrointestinal tract [38]. However, the disease most associated with HER-2 is breast cancer. About 30% of all breast cancer patients have overexpression of HER-2, and the overexpression is highly correlated with the aggressive forms of the disease [39, 40]. Importantly, the HER-2 expression is similar between the primary tumor and the corresponding metastases [41].

HER-3 and HER-4 The final two members of the EGF receptor family, HER-3 and HER-4 were discovered when the search for HER-2 ligands was the most intense [42, 43]. HER-3 has a deficient tyrosine kinase domain [44], and can thus not trans- duct signals as a homodimer, but the heterodimer HER-2/HER-3 is considered to have the strongest oncogenic potential in the family [45]. HER-3 and HER-4 are normally described as membrane associated proteins, but immunohistochemical staining of tissues and cells expressing HER-3 and HER-4 show a mainly cytoplasmic staining [46, 47]. Thus, the use of HER-3 and HER-4 as molecular targets seems difficult. HER-3 and HER-4 are dependent on ligand binding in order to dimerize, in a similar way as the EGF receptor (Figure 2).

Signaling The members of the EGF receptor family serve as components in a complex communication network between the different ligands and the intracellular signaling proteins [48]. Different ligands have affinity for certain receptors. The receptors can form different homo- or heterodimers, and the subsequent intracellular signal depends on all those (and probably more) parameters.

13 Many different signaling pathways have been implicated for the EGF receptor family. The most commonly mentioned pathways are shown in Fig- ure 4, and include:

Figure 4. Intracellular signaling pathways involved in the EGF receptor family.

1. the Ras- and Shc-activated Mitogen-Activated Protein Kinase (MAPK) pathway, including Erk1 and Erk2, which promotes proliferation [49] 2. the phospholipase C J (PLCJ) pathway, that is associated with migration [50, 51] 3. the Phosphatidylinositol-3-OH Kinase (PI(3)K) activated Akt pathway, which activates cell-survival pathways [52].

Targeting agents The choice of targeting agent influences many aspects of tumor targeting. The size of the targeting agent is one factor. It affects primarily the residence time in the circulation (i.e. the blood). Small molecules are filtrated in glomerulus in the kidneys, whereas larger molecules are not, and will instead circulate longer and eventually be degraded in the liver and excreted in the feces. Molecules smaller than ~6 kDa are freely filterable in the kidneys, whereas molecules larger than 70 kDa are not filterable at all. Molecules with a molecular weight in between those limits are filterable to some degree, dependent on size and charge [53]. Long residence time in the blood will lead to high background in imaging settings and high normal tissue exposure in therapeutic settings. Too short residence time in the blood, will, however, lead to elimination of the targeting agent before sufficient levels in

14 the tumor are reached. Another aspect of the size of the targeting agent is that smaller agents generally penetrates the tumor tissue better [54-56].

Antibodies The first-used targeting agents for tumor targeting were nature’s own targeting agents – antibodies. The invention of the hybridoma technology in 1975 [57] brought new light on the field, making large scale production of highly tumor specific antibodies possible. Still today, antibodies are the most used targeting agents in the clinic. Two types of monoclonal antibodies are used in the clinic: 1. Non-conjugated antibodies, without any toxins or radioactive nuclide attached to them. 2. Conjugated antibodies, with a radionuclide or chemotherapy agent attached.

Table 1: Monoclonal antibodies approved in the United States (US) or Europe (EU). Modified from [58]. Trade namn Conjug- Year of Description Indication (Generic name) ated to approval Orthoclone OKT3 Murine mAb against Reversal of acute kidney - 1986 (US) (Muromomab CD3) CD3 transplant rejection

OncoScint Murine mAb against Detection of colorectal and 111In 1992 (US) (Satumomab Pendetide) TAG-72 ovarian cancers

MyoScint Murine mAb against Detection of myocardial 111In 1996 (US) (Imiciromab Pentetate) cardiac myosin infarction

ProstaScint Murine mAb against Detection of prostate 111In 1996 (US) (Capromab Pendetide) PSMA adenocarcinoma

Rituxan (US), Chimeric mAb Non-Hodkin’s lymphoma - 1997 (US) Mabthera (EU) against CD20 1998 (EU) (Rituximab)

Simulect Chimeric mAb Prophylaxis of acute organ - 1998 (EU) (Basiliximab) against IL-2 rejection in allogenic renal transplantation

HumaSPECT Human mAb against Detection of carcinoma of 99mTc 1998 (EU) (Votumumab) cytokeratin tunor the colon and rectum associated antigen

Zenapax Humanized mAb Prevention of acute kidney - 1997 (US) (Daclizumab) against IL-2 transplant rejection 1999(EU)

15 Synagis Humanized mAb Prophylaxis of lower tract - 1998 (US) (Palivizumab) against respiratory respiratory disease 1999 (EU) syncytical virus

Remicade Chimeric mAb Crown’s disease - 1998 (US) (Infliximab) against TNF-D 1999 (EU)

Herceptin Humanized mAb Treatment of metastatic - 1998 (US) (Trastuzumab) against HER-2 breast cancer 2000 (EU)

Mylotarg Humanized mAb Acute myeloid leukemia Cytotoxin 2000 (US) (Gemtuzumab against CD33 zogamicin zogamicin)

Campath (US), Humanized mAb Chronic lymphocytic - 2001 Mabcampath (EU) against CD52 leukemia (US, EU) (Alemtuzumab)

Zevalin Murine mAb against Non-Hodgkin’s lymphoma 111In 2002 (US) (Ibritumomab tiuxetan) CD20 90Y 2004 (EU)

Humira Human mAb against Rheumatoid arthritis - 2002 (US) (Adalimumab) TNF 2003 (EU)

Xolair Humanized mAb Asthma - 2003 (US) (Omalizumab) against IgE 2005 (EU)

Bexxar Murine mAb against Non-Hodgkin’s lymphoma 131I 2003 (US) (Tositumomab) CD20

Erbitux Chimeric mAb Carcinoma of the colon - 2004 (Cetuximab) against EGFR and rectum (US, EU)

Avastin Humanized mAb Metastatic carcinoma of - 2004 (US) (Bevacizumab) against VEGF the colon and rectum 2005 (EU)

Tysabri Humanized mAb Multiple Sclerosis - 2004 (US) (Natlizumab) against D4-integrin

NeutroSpec Murine mAb against Diagnosis of appendicitis 99mTc 2004 (US) (fanolesomab) CD15

As seen in Table 1, most antibodies approved for human use today, are non- conjugated antibodies. They act most often as antagonists or by recruiting immune cells. One of the most known non-conjugated antibodies is trastuzumab, the active component in the breast cancer drug Herceptin. It is a humanized version of the murine mAb 4D5. Many different mechanisms of

16 action have been postulated for trastuzumab [59, 60]. The major effects of trastuzumab are: 1. Down-regulation of HER-2 from the cell membrane [60, 61] 2. Blockage of cell-cycle progression, decreasing the S-phase fraction [60] 3. Antibody Dependent Cellular Cytotoxicity (ADCC) [62, 63] 4. Inhibition of the constitutive growth-signaling properties of HER-2 [60, 64] 5. Angiogenesis inhibition [65, 66] 6. Sensitizing the tumor against chemotherapeutic agents, such as doxorubicin, paclitaxel, docataxel, etoposide and cisplatin [67, 68], and radiotherapy [69, 70].

Another HER-2 binding antibody is currently investigated in a phase II study – the humanized version of the murine mAb 2C4; pertuzumab (Omnitarg). Pertuzumab has similar anti-proliferate effects as trastuzumab [71], but binds to the dimerization loop of the HER-2 ECD and blocks both HER-2 homo- and hetero-dimerization. A problem that arises when using antibodies of murine origin, which is the case for most monoclonal antibodies, is the immune response in the human patient that the species-specific Fc region causes. This problem, often called HAMA response, can, at least partially, be overcome using humaniza- tion. This is a procedure, in which the constant region of the antibody is replaced by a human version using antibody engineering. Another problem with antibodies is the large size. A molecular weight of ~155 kDa limits tumor penetration and prevents filtration in the kidneys, leading to long circulation times, as discussed above.

Antibody fragments The binding regions of antibodies, the Complementarity Determining Re- gions (CDR’s), constitute a very small portion of the total antibody size (Figure 5). It is thus possible to reduce the size of the antibody without de- stroying the binding capacity. The easiest way of reducing the antibody size, is to use enzymatic degradation for generating smaller antibody fragments, as seen in Figure 5. The fragments obtained by enzymatic degradation is, however, not small enough to be readily filtrated in the kidneys. Even smaller constructs, such as scFv’s and different kinds of multivalent variants (Figure 6), can be produced using genetic engineering. In those cases, a covalent, non-reducible, linker is introduced between the heavy and light chain. The only antibody fragments approved for human use today are Fab’s, but several scFv’s, diabodies and minibodies are on the way [72]. For some applications, such as tumor imaging, it is important to achieve high contrast between the tumor and normal tissue within only a few hours. In those cases, even a scFv can be too large to be optimal. Also, antibody fragments are dependent on disulphide bonds to retain their structure. This

17 results in sensitivity against denaturing conditions such as heat and reducing agents, demanding expensive and complicated production routes.

Figure 5. Antibody fragments produced by enzymatic degradation. Dark gray segments, corresponds to the heavy chains, whereas the light gray segments correspond to the light chains. The black stripes on the upper parts of the antibody represent the CDR’s.

Figure 6. Antibody fragments produced by antibody engineering

18 Peptides Small peptides, consisting of only a few amino acids, are often considered ideal agents for tumor imaging. Most peptides used for tumor targeting today are based on natural ligands, and thereby target peptide receptors. The, by far, most used peptides for tumor imaging, are derivatives of the hormone somatostatin [73], that binds to the somatostatin receptor, which often is overexpressed in neuroendocrine tumors. Endogenous somatostatin has a plasma half-life of only 3 minutes, and is thereby unsuitable as a drug. The development of the somatostatin analog octreotide, with a plasma half-life of about 100 minutes, started a whole new field of research. Somatostatin analogs are primarily used as imaging agents, radiolabeled with 111In [74], but have also been tested for their therapeutic potential, labeled with 111In, 90Y or 177Lu [75]. The therapeutic potential, especially when 111In is applied, can probably be dramatically improved now, when new nuclear localizing versions of somatostatin analogs have been developed [76]. Peptides for tumor targeting can also be developed from larger proteins, such as antibodies. One example of this is the peptidomimetic AHNP, that mimics on of the CDR’s of trastuzumab [77]. The apparent affinity was, however, 300 times lower, compared with the original antibody.

Affinity proteins In the last decades, the immunoglobulin framework, or scaffold, has faced increasing competition from new, non-immunoglobulin scaffolds. Some of these scaffolds are listed in Table 2.

Table 2: Examples of protein scaffolds of non-immunoglobulin origin for development of novel affinity proteins (modified from [78] and [79]). No. of resi- Name Scaffold Cross-links etc. dues Affibody Protein A 58 - Anticalin Lipocalin (BBP) 160-180 (174) 0-3 (2) S-S FliTrx / Peptide aptamer Thioredoxin 108 1 S-S Immunity protein ImmE7 87 - Knottin / microprotein CBD / EETI-II 36 / 28 2 / 3 S-S Kunitz domain BPTI / APPI 58 3 S-S Peptamer Staphylococcal nuclease 149 - Repeat-motif protein Ankyrin repeat 33 - Tendamistat D-Amylase inhibitor 74 2 S-S Trinectin Fibronectin III 94 - Zinc-finger Zinc-finger 26 Bound Zn2+

19 Most of these non-immunoglobulin affinity proteins are derived from proteins with intrinsic binding properties and are extraordinary stable and solu- ble. Their structure allows for introduction of new binding regions, or re- placement of the original. Large libraries, comprising 108-1012 possible clones can then be screened for the desired binding characteristics. In general, the library complexity determines the strength of the binding, with larger libraries resulting in higher affinity [80]. Some of the affinity proteins in Table 2 are schematically drawn in Figure 7.

Figure 7. Affinity proteins representative of different non-immunoglobulin scaffolds. (a) Affibody molecule, (b) Immunity protein ImmE7, (c) Cytochrom b562, (d) Ankyrin Repeat Protein, (e) Kunitz-domain Inhibitor, (f) Fibronectin, (g) Knottin, (h) Carbohydrate Binding Module and (i) anticalin FluA. Red regions correspond to an D-helical fold, blue regions to a ȕ-sheet fold and yellow regions represent the binding regions. Modified from [79].

Affibody molecules Protein A is a protein localized on the surface of staphylococcus aureus and protects the bacterium from the immune system of the host. The protein consists of five homologous domains (Figure 8a), each of which binds both the Fc [81] and Fab [82] region of IgG. A slightly modified version of domain B, called domain Z [83], has been extensively used as affinity purification reagent [84-86]. Of the 58 amino acids in the domain, thirteen solvent exposed amino acids localized on helix 1 and helix 2 (Figure 8b) were randomized to construct a library [87]. This so-called affibody library has been used to find affinity ligands against a number of proteins [88-90]. The scaffold contains no cysteine residues. This allows for recombinant production both in the cytoplasm and in the periplasm, and the introduction of a terminal cysteine residue makes site-specific immobilization and labeling possible [91]. The small size (about 6 kDa) and the absence of disulfide bonds also

20 enables synthetic production of affibody molecules [92], with opportunity for introduction of suitable groups e.g. chelating molecules for radiometal labeling. Affibody molecules have also been found to be stable against both high temperatures, proteolytic degradation and high pH [93]. The high stability in combination with its small size makes affibody molecules highly interesting for tumor targeting applications.

Figure 8. Schematic representation of affibody molecules. a) The five Fc binding domains of protein A. b) The three-helix-bundle structure of the affibody molecule. The thirteen randomized positions located on helix 1 and 2 are indicated (black), as well as their position in the amino acid sequence.

Selection systems A popular and highly efficient strategy to find a molecule that binds with high affinity to a desired protein, is to construct a library of millions or more variants. From this library, variants that bind to the target are selected, and amplified, and the obtained sub-library is the selected for binding again, perhaps using more harsh conditions. This cycle is repeated a number of times until one variant dominates the pool. Most often, the desired affinity molecule is a protein or a peptide, and the library is constructed in nucleic acid (for easy amplification and sequencing). This requires that a link between the genotype (the library) and the phenotype (the resulting protein or peptide) must be established.

Phage display A bacterial virus, called phage, is used in the most common selection system today – phage display. In phage display, the genotype and the phenotype is linked by genetic fusion of the gene coding for desired protein or peptide to

21 the gene goding for one of the proteins that constitute the coat of the phage. Amplification is performed by infection of E. coli bacteria. The most commonly used phage for use in phage display is the filamentous phage M13. It contains a single stranded circular DNA that codes for eleven proteins, of which five are capsid proteins. Of these, pIII (present in about 5 copies – drawn black in Figure 9) and pVIII (present in about 2700 copies – drawn in gray in Figure 9) are the most commonly used display partners [94].

Figure 9. One round of biopanning. a) Phages containing the protein/peptide library fused to the pIII gene are panned against the solid-phase immobilized target protein. b) Bound phages are, after washing, released from the solid-phase using e.g. low pH elution. c) Released phages are allowed to infect E. coli. Co-infection with a helper phage allows for production of new phages, which again can be panned against the target molecule.

Phage display libraries are now commercially available in kit formulations, where both linear and disulfide constrained peptides of various sizes can be selected, as well as different antibody fragments.

22 Other systems While phage display has been the golden standard for protein display, many other methods have emerged over the years. Cell surface display, expressing protein libraries on the surface of e.g. bacteria and yeast, has the advantage of quick and efficient screening using FACS [95]. Eukaryotic systems, such as the yeast surface display also has the advantage of post-translational modifications [96]. Another approach is used in the cell-free selection techniques [97], where in vitro transcription / translation systems are used. There, ribosome display is the golden standard, in which a lack of stop codons prevents ribosomal release, and thereby establishes a physical linkage between the mRNA and the translated protein [98]. Another cell-free method is called mRNA display, with the largest difference from ribososome display being a covalent linkage to between the mRNA and the protein [99]. The advantages with cell-free systems are that they are not dependent on transfection efficiency and libraries of larger size can generally be obtained. Also, the possibility to introduce PCR-based mutagenesis to these systems [97] makes the cell-free systems interesting for protein selection. When it comes to selecting binders composed of nucleic acid, SELEX is the method of choice. The method can be used for selecting DNA or RNA aptamers with affinities comparable to monoclonal antibodies [100]. How- ever, in order to prevent in vivo degradation of the aptamers, “post-SELEX modifications” or modified nucleotides must be used.

Therapeutic/visualizing moieties When a tumor associated structure and a targeting agent have been selected, it must determined whether the agent should be used for therapy or imaging (visualization).

Therapy Two of the most common cancer treatment modalities, radiotherapy and chemotherapy, can be used also in a tumor targeting setting, in order to reduce the normal tissue toxicity.

Radionuclides Targeted radiotherapy is achieved by linking a tumor targeting molecule, such as an antibody, to a moiety containing a radionuclide suitable for therapy. Three types of radionuclides can than be used: D-emitters, ȕ-emitters and auger electron emitters, see Table 3. Importantly, to kill a tumor cell, the radiosensitive target – the DNA – must be hit [101]. A ȕ-particle has a much longer range than a D-particle of the same energy as a result of the differences is mass between the ȕ-particle (electron) and the D-particle (helium atom). That means that the energy deposition from the D-

23 particle is localized to a much smaller volume, thus resulting in more severe damages in the surrounding tissue. Thereby, the D-particle has a higher RBE (Relative Biological Effectiveness) compared with the ȕ-particle. The longer range of the ȕ-particle however, will lead to cross-fire irradiation. That means that radiation from one targeted cell will also irradiate surrounding cells. The benefit with cross-fire irradiation is that there is no need to target each and every tumor cell. From the negative side, surrounding normal tissues will also be damaged, and some of the advantages with targeted radiotherapy are lost.

Table 3: Radionuclides of interest for targeted radiotherapy Main therapeutic Radionuclide Daughter nuclide Physical half-life Energy mode of decay 211At 7.2 h Alpha 5.9 MeV 211Po 516 ms Alpha 7.4 MeV 212Bi 60.6 m Alpha 6.1 MeV 212Po 298 ns Alpha 8.8 MeV 213Bi 45.6 m Alpha 5.9 MeV 213Po 4.2 µs Alpha 8.4 MeV 123I 13.3 h Auger cascade 125I 59.4 d Auger cascade 131I 8.0 d Beta 606 keV 177Lu 6.7 d Beta 498 keV 111In 2.8 d Auger cascade 90Y 64 h Beta 2.3 MeV

Auger electrons are released from an atom’s outer shell as a result of excita- tion, leaving a vacancy in internal electron shell [102]. They have very short range in tissue (a few nm) as a result of their low energy, but the fact that they are released in a cascade (with about 20 electrons per decay for 125I [103]) results in many decays within a small volume. Auger emitting radionuclides are thus highly cytotoxic if localized very close to the DNA. By introducing a nuclear localization signal on a tumor targeting agent, this phenomenon can be utilized [76]. Another aspect involving the cellular localization of the targeting agent is internalization. In general, internalization results in lysosomal degradation, where the radionuclide is removed from the targeting agent and excreted from the cell. Several strategies have been tested to overcome this problem [104], including attaching a non-degradable molecule, such as dextran, to the targeting molecule, and locating the radiolabel to this moiety [105]. A second approach has been to use radiometals. Radiometals, such as 177Lu, and 111In are positively charged, and thus highly hydrophilic, and will remain within the cytoplasm for a long time [106, 107]. A third approach utilizes the same principle, but allows for residualization of non-metallic nuclides. In

24 this approach, linker molecules that are charged at lysosomal pH, are used [108].

Toxins Toxins originating from plants or bacteria are often used in the formulation of “immunotoxins”, conjugated to antibodies, or antibody fragments. The most used toxins are the bacterial toxins Diphtheria toxin and Pseudomonas exotoxin [109, 110]. The toxins inhibit different steps in the translation machinery and are dependent on internalization for their effect.

Cancer drugs Cytotoxic drugs, such as doxorubicin and daunorubicin, have been used for conjugation to antibodies or antibody fragments for targeted delivery, even in clinical trials [111]. Moreover, targeted liposomes – immunoliposomes – filled with cytotoxic drugs have been used, in order to further increase the delivered concentration of the drugs [112].

Visualization Tumor specific imaging is an excellent way of detecting small metastases not visible using standard techniques. Gamma- and positron-emitting radionuclides are by far the most common labels used [113]. Gamma camera techniques, such as SPECT is frequently used in nuclear medicine today. Most gamma cameras are optimized for 99mTc, but e.g. 111In is also readily used, especially in the Octreoscan formulation [74]. Radionuclides used in tumor imaging are listed in Table 4.

Table 4: Radionuclides of interest for tumor imaging Radionuclide Physical half-life Emission 11C 20 m Positron 76Br 16 h Positron, gamma 18F 110 m Positron 68Ga 68 m Positron, gamma 123I 13 h Gamma 124I 4.2 d Positron, gamma 111In 2.8 d Gamma 99mTc 6 h Gamma

The PET technique is superior to the available gamma camera techniques when it comes to sensitivity, resolution and quantification [114], but is much more expensive and is not as available as gamma camera techniques. The best positron emitting nuclides are 18F and 11C since they are pure positron emitters. However, due to the time-consuming and inefficient labeling chemistry of 18F, and short physical half-life of 11C, other nuclides are increasing in interest [114, 115].

25 The physical half-life of the radionuclide should match (no more than three physical half-lives should pass during the examination) the time needed for obtaining good tumor-to-background for the targeting agent used. This means that antibodies which, due to their large size, circulate in the blood stream for days, must be labeled with a radionuclide like 111In or 124I to be used for visualization. However, shorter time between injection and measurement is normally preferred, in order to reduce the hospital time for the patient. It is also important in order to limit the problems with radiation protection around the patient as well as the needed measurement time. This makes small peptides in combination with short-lived radionuclides optimal for tumor visualization.

26 The present study

Selection of HER-2 binding affibody molecules (I & II) Affibody molecules with affinity for HER-2 were selected using phage display. The filamentous phage M13 and a phagemid library containing 8.7×108 variants, representing affibody molecules with 13 of the 58 amino acids randomized, was used. The target protein - HER-2 ECD was biotinylated and attached to streptavidin coated magnetic beads. The phages (expressing about one affibody molecule per phage as a pIII fusion) were subjected to four rounds of panning against HER-2 ECD. In the panning process, the amount of HER-2 ECD was decreased and the number of washes increased, between every round (Table 5).

Table 5: Conditions used in each round of panning and fraction of the added phages that remained after each round. Cycle Amount of HER-2 ECD Number of washes Eluted fraction (×10-6) 1 40 µg 1 25 2 20 µg 3 1 3 20 µg 6 23 4 5 µg 12 800

After each round of panning, the binding phages were eluted from the magnetic beads using low pH and were than allowed to infect E. coli for amplification. After four rounds of panning, 49 colonies were randomly picked and the phagemid was sequenced. As seen in Figure 10, one sequence, corresponding to ZHER2:4 was dominating (33/49 clones) and sequence homology at certain positions was found. There is often a correlation between the clone abundance and the binding strength [89] of the protein. Thus, all sequences that were found more than once were cloned into expression vectors. The corresponding proteins (ZHER2:4, ZHER2:7, ZHER2:8 and ZHER2:2) were produced in E. coli and purified using a His-tag introduced by the vector.

27 Figure 10. Deduced amino acid sequence in the 49 sequenced clones after four rounds of panning against HER-2 ECD. The wild-type (wt) sequence is shown at the top and the frequency of each sequence is found to the right.

Previous experiments [116] have shown that the apparent binding strength between a ligand and its target can be increased by the construction of multivalent ligands. Since affibody molecules are based on the multivalent protein A (Figure 6a), construction of multivalent affibody molecules is quite straightforward. A second gene fragment, encoding the most abundant sequence – corresponding to ZHER2:4, was inserted into the expression vector corresponding to ZHER2:4. The resulting protein, (ZHER2:4)2, was then produced in E. coli and purified using the His-tag.

Biacore testing (I & II) The produced affibody molecules were tested for binding to HER-2 ECD immobilized to a Biacore sensorchip. The association rate constant kon and the dissociation rate constant koff were measured and the dissociation equilibrium constant KD was then calculated as koff/kon. ZHER2:4, that was the by far most abundant clone according to the DNA sequencing, was also shown to be the strongest HER-2 binder with a KD of about 50 nM. The second strongest clone was ZHER2:7 with a KD of about 140 nM. ZHER2:2 did also bind specifically to HER-2 ECD, but weaker than the clone 4 and 7. ZHER2:8 did not show any specific HER-2 binding. The bivalent version of clone 4 - (ZHER2:4)2 - showed higher apparent affinity (KD = 3 nM) against HER-2 ECD compared with the monovalent

28 form. As seen in Table 6, the difference in KD is mainly originating from differences in koff.

Table 6. Kinetic data obtained with Biacore technology

-1 -1 -1 Affibody clone kon (M s ) koff (s ) KD (nM)

5 -3 ZHER2:4 1.8×10 9.9×10 50 5 -4 (ZHER2:4)2 2.5×10 7.6×10 3

125I labeled HER-2 binding affibody molecules Direct radioiodination is directed against tyrosine residues in the protein, and as seen in Figure 10, the only two tyrosine residues in ZHER2:4 are located in the two helices responsible for binding. It was thus not surprising that this method destroyed the binding capacity of ZHER2:4. However, indirect radioiodination using the linker molecule PIB retained the binding capacity of the affibody molecule, as determined using Biacore technology (data not shown). As shown in Figure 11, the precursor SPMB is first radiolabeled with 125I and is then coupled to amine groups, mainly lysine residues, located on the protein. One of the lysine residues available on ZHER2:4 is located in close proximity to one of the helices responsible for binding (Figure 10) and labeling directed against this group sometimes destroy the binding capacity (data not shown).

Figure 11. Labeling of a protein with 125I using the PIB linker

Cellular testing (I & II) Specific binding / epitope comparison To prove that the HER-2 specific affibody molecules could not just bind to 125 the recombinant HER-2 ECD, but also native HER-2, [ I]PIB-ZHER2:4 was tested for binding to the HER-2 overexpressing breast cancer cell line SKBR-3. To prove that the binding was specific, the uptake of 125I in cells

29 125 receiving only [ I]PIB-ZHER2:4 was compared with cells receiving excess amount of unlabeled ZHER2:4 to block the binding. It was also investigated if the monoclonal antibody trastuzumab could interfere with the binding. As seen fin Figure 12a, the uptake after administration of ZHER2:4 (white bar) could be significantly reduced by addition of excess amounts of unlabeled ZHER2:4 (black bar). Addition of unlabeled trastuzumab did not reduce the binding of ZHER2:4, implicating that trastuzumab and ZHER2:4 do not bind to the same epitope on HER-2. To confirm the data about the non-overlapping epitopes, trastuzumab was labeled with 125I and the uptake of 125I in cells receiving only 125I-trastuzumab was compared with cells receiving excess amount of unlabeled trastuzumab or ZHER2:4 to block the binding. As seen in Figure 12b, also 125I-trastuzumab showed specific binding to SKBR-3 cells, and ZHER2:4 could not displace the binding.

125 Figure 12. Test for specific binding and epitope comparison for [ I]PIB-ZHER2:4 and 125I-trastuzumab in HER-2 overexpressing SKBR-3 cells. a) Radioactivity up- 125 take per cell after exposure to [ I]PIB-ZHER2:4. b) Radioactivity uptake per cell after 125 exposure to I-trastuzumab. A 500-fold excess of unlabeled ZHER2:4 or trastuzumab to determine specific binding and overlapping of epitopes. Values are an average of three dishes and the error bars show the standard deviation.

Cellular internalization and retention The internalization kinetics and cellular retention was compared between 125 125 [ I]PIB-ZHER2:4 and [ I]PIB-(ZHER2:4)2 in SKBR-3 cells. To measure internalization, an acid wash protocol, originally described for internalization of EGF [117], was used. In this method, 0.1 M Glycine-HCl, pH 2.5 is added to the cells after removal of unbound radioactivity. During six minutes in 4 ºC, the ligands located on the cellular membrane are allowed to dissociate in the low-pH buffer. The remaining radioactivity is considered to be internalized. Figure 13 shows that both ZHER2:4 and (ZHER2:4)2 are internalizing in SKBR-3 cells. The internalized fraction is almost constant between 0.5 h and 24 h. However, the internalized fraction is higher for the bivalent version (about 50%) compared with the monovalent version (about 30%).

30 125 Figure 13. Cellular uptake of radioactivity in SKBR-3 cells. a) [ I]PIB-ZHER2:4 and 125 b) [ I]PIB-(ZHER2:4)2. Values are an average of three dishes and the error bars show the standard deviation.

Cellular retention

As previously discussed, the main difference between ZHER2:4 and (ZHER2:4)2 in terms of binding, is the rate of dissociation from HER-2. To test if this was true also for native HER-2, SKBR-3 cells were exposed for 2 h with 125 125 [ I]PIB-ZHER2:4 or 4 h with [ I]PIB-(ZHER2:4)2 (times corresponding to maximum uptake). The cells were then washed to remove unbound radioactivity and the remaining radioactivity was followed as it was released from the cells.

125 Figure 14. Cellular retention of radioactivity in SKBR-3 cells. a) [ I]PIB-ZHER2:4 125 and b) [ I]PIB- (ZHER2:4)2. Values are an average of three dishes and the error bars show the standard deviation. Note the difference in scale between the x-axes.

In Figure 14, the cellular retention of radioactivity is shown as a function of time. Discrimination between membrane-bound and internalized radioactivity was done using the acid wash method, described above. As seen in Fig- ure 14a, almost all membrane bound radioactivity dissociated during the first 125 15 minutes when delivered as the monovalent [ I]PIB-ZHER2:4. Comparing

31 the total radioactivity retention, 50% of the initial radioactivity was removed from the cells after 3 minutes when delivered with ZHER2:4 whereas 2.2 h was needed when delivered with (ZHER2:4)2.

Biodistribution in mice (III) When specific binding to HER-2 expressing cells was proved in vitro, the next step was to explore the in vivo targeting properties. The monovalent affibody molecule ZHER2:4, was abandoned, and all focus was set on the bivalent version - (ZHER2:4)2.

An animal model was chosen, where immune-deficient mice (BALB/c nu/nu) were xenografted with human HER-2 overexpressing cells. The previously used breast cancer cell line, SKBR-3, did not form tumors in vivo and was thus replaced by the ovarian cancer cell line SKOV-3, that also overexpress HER-2 and has been reported to form tumors in immune- deficient mice [118, 119]. About 10 million cells were injected subcutaneously in the right hind leg. The tumors were then allowed to establish for about one to two months. 125 The radioiodinated HER-2 binding affibody molecule [ I]PIB-(ZHER2:4)2 (~1 µg, 100 kBq), was injected into the blood stream of the mice via the tail vein. At different time-points between 1 and 24 hours, groups of four mice were euthanized and organs and tissues of interest were collected. First, the specificity in the uptake in various organs, 4 h after the injection, was determined. This was done by comparing the uptake with the up- 125 take in a group of mice receiving the same amount of [ I]PIB-(ZHER2:4)2 but also receiving an excess amount if unlabeled (ZHER2:4)2. The uptake was also compared with the uptake of a 125I labeled non-specific affibody molecule - 125 [ I]PIB-ZTaq4:5. As seen in Figure 15a, the tumor uptake was significantly reduced when an excess of unlabeled (ZHER2:4)2 was administrated. No significant difference was seen in any other organ. Figure 15b shows that the uptake of the 125 unspecific affibody molecule [ I]PIB-ZTaq4:5 was significantly lower in the tumor, compared with the specific affibody molecule, whereas the uptake in all normal organs, except for the kidney, was unaffected. The lower kidney uptake was believed to be a result of differences in the isoelectric point of the two affibody molecules. The theoretical isoelectric point of the two affibody molecules predicts that the HER-2 binding affibody molecule should carry a net positive charge at physiological pH, whereas the unspecific affibody molecule should carry a net negative charge. Renal reabsorbtion is known to be dependent on the charge of the molecule to be reabsorbed [120, 121], and positively charged molecules are more efficiently reabsorbed in the proximal tubule.

32 Figure 15. In vivo specificity. a) comparison between a group receiving only 125 [ I]PIB-(ZHER2:4)2 and a “blocked” group that also received an excess of unlabeled (ZHER2:4)2 b) comparison between a group receiving the radiolabeled HER-2 binding affibody molecule and a group receiving a non-specific affibody molecule. Values are averages from four mice. Error bars are standard deviations. For the thyroid, %ID/organ is shown.

In conclusion, the two independent specificity tests both showed that there was a specific uptake of radioactivity in the tumor, but not in any other organ.

125 Figure 16. Biodistribution of [ I]PIB-(ZHER2:4)2 in tumor bearing mice. Values are averages from four mice. Error bars are standard deviations. For the thyroid, %ID/organ is shown.

33 125 The biodistribution of [ I]PIB-(ZHER2:4)2 was also followed over time (from 1 to 24 hours) to investigate the biokinetic properties of the compound. It was found that (ZHER2:4)2 was excreted through the kidneys, like most proteins of this small size [122]. As seen in Figure 16, this resulted in rapidly decreasing blood levels, which, in combination with high tumor uptake, resulted in impressive tumor-to-blood ratios. Ten times higher concentration of radioactivity was found in the tumor compared with the blood 8 h after the injection. Already at the 4 h time-point, the only organs with higher concentration of radioactivity compared with the tumor, was the kidneys. Thus, the radioiodinated (ZHER2:4)2 has the potential to be a good tumor imaging agent.

Gamma camera imaging (III) 125 The biodistribution of [ I]PIB-(ZHER2:4)2 clearly showed potential for tumor imaging. To confirm this potential, tumor bearing mice were injected with 125 about 3 MBq (~3 µg) [ I]PIB-(ZHER2:4)2. The mice were euthanized 6 or 8 hours later and the radioactivity distribution was visualized in a gamma camera. Figure 17 shows that the tumor could clearly be visualized already 6 h after injection. At 8 h p.i., the background radioactivity was lower and the tumor was more clearly visible. The imaging potential of the affibody molecule investigated could thus be confirmed.

Figure 17. Gamma camera image of the radioactivity distribution 6 h (right mouse) 125 and 8 h (left mouse) after injection of [ I]PIB-(ZHER2:4)2. The mice were injected i.v. in the tail. Xenografted tumors were placed in the right flank.

34 Autoradiography/immunohistochemistry (III) To determine the intratumoral localization of the radioactivity when admin- 125 istrated as [ I]PIB-(ZHER2:4)2 as well as the HER-2 expression in the xenografts, tumors from the gamma camera study was analyzed. The tumors were fixed in formaldehyde, embedded in paraffin and cut in 4-µm-thick sections. Consecutive sections were stained with hematoxylin, immunohistochemi- cally stained for HER-2 expression and subjected to autoradiography. As seen in Figure 18b, HER-2 is homogenously expressed in the tumor. The autoradiography (Figure 18c) however, showed that the 125I in the tumor is mainly localized to the cell layers closest to the blood vessels, 6 h p.i. of 125 125 [ I]PIB-(ZHER2:4)2. Further studies about the intratumoral localization of I 125 after delivery of [ I]PIB-(ZHER2:4)2 are needed to determine the importance of timing and protein amount. Until then, no definite conclusions, regarding potential for targeted radiotherapy, can be draw from this experiment.

125 Figure 18. Serial tumor sections, 6 h after the injection of [ I]PIB-(ZHER2:4)2. a) Hematoxylin, b) immunohistochemical staining showing HER-2 expression and c) autoradiography showing the intratumoral distribution of 125I. The arrows indicate blood vessels. The inserted bar is 100 µm.

Cellular effects of the HER-2 binding affibody molecules (IV) As previously discussed, the monoclonal antibody trastuzumab has an intrinsic anti-tumor activity involving down-regulation of HER-2, ADCC and sensitizing against chemo- and radiotherapy. The molecular background of many of the intrinsic anti-tumor activities of trastuzumab is believed to be derived from effects in the intracellular signaling network, where a decreased phosphorylation of Akt and Erk are believed to be responsible [70, 123, 124]. To determine if the two affibody molecules (ZHER2:4)2 and its monovalent affinity maturated variant ZHER2:342 (described in Orlova et al. [125]) had similar anti-tumor properties as trastuzumab, all three substances were used to investigate their effect on intracellular signaling. SKOV-3 cells were treated with the substances for between 15 minutes and 8 h. The cells were then lysed and immunoprecipitated with antibodies against HER-2 and

35 PLCJ1. Cell lysate was also separated on SDS-PAGE gels and thereafter immunoblotted with phosphor-specific antibodies against Akt and Erk1/2. The membranes were then re-probed with antibodies against total Akt and Erk2. The immonoprecipitates were separated on a SDS-PAGE gel and immunoblotted with antibodies specific for phosphotyrosine. For each of the four analyzed proteins, a ratio between the phosphorylated form and the total amount of that protein was calculated. The signals were then compared to the signals obtained without stimulation with trastuzumab or affibody molecules. Table 7 summarizes the phosphorylation data obtained.

Table 7. Summary of protein phosphorylation data in SKOV-3 cells. +++ = More than 10-fold increase, ++ = 3- to 10-fold increase, + = 1.5- to 3-fold increase, --- = More than 45% reduction, -- = 25-45% reduction, - = 15-24 % reduction

Protein Trastuzumab (ZHER2:4)2 ZHER2:342

HER-2 +++ ++ --

PLC-J1 (migration) +++ + --- Akt (anti-apoptosis) -- + + Erk1/2 (proliferation) -- + -

The effects of trastuzumab on phosphorylation of HER-2, Akt and Erk are consistent with previous data [70, 123, 124]. Data about the effects of trastuzumab on PLCJ1 have not been found in the literature. Interestingly, the effects of the bivalent affibody molecule were much more similar to the effects of trastuzumab than the effects of the monovalent affinity maturated affibody molecule, possibly as a result of bivalent binding. To determine if the differences in intracellular signaling had any biological effect, migration (related to the activity of PLCJ1 [50, 51]), apoptosis (related to Akt activity [52]) and proliferation (related to Erk activity [49]) was measured with and without the three HER-2 targeting substances.

Table 8. Summary of biological effects of the three substances in SKOV-3 cells. + = increase, - = decrease, +/- = no difference, -- = large decrease, (-) = tendency (not significant) of decrease.

Protein Trastuzumab (ZHER2:4)2 ZHER2:342

Migration -- + - Anti-apoptosis +/- +/- +/- Proliferation (-) + (-)

Migration was studied by introducing a wound in a layer of confluent, serum-starved SKOV-3 cells. The wound was then allowed to repair under the influence of the substances, and photographs were taken after 24, 32, 48, 55

36 and 72 h. As seen in Table 8, addition of (ZHER2:4)2 closed the wound faster than without the addition. Both trastuzumab and ZHER2:342 decreased the rate of wound healing. Trastuzumab decreased the wound healing rate the most. These results agree with the signaling study, except for the trastuzumab case, where PLCJ1 was extensively phosphorylated, but a significant decrease in migration was seen. Apoptosis was studied using the annexin V technique [126]. The apoptosis levels were quantified by flow cytometry using annexin V/propidium iodide. The basal apoptosis fraction in untreated cells was about 0.3% and was not affected by addition of the substances. The cells were also treated with camptothecin, which works as a type-I DNA topoisomerase inhibitor and is known to induce apoptosis [127]. This caused the apoptosis level to increase to about 5.7% but it was still not affected by addition of any of the three substances. According to the signaling study, phosphorylation of Akt, that is associated with anti-apoptosis, was increased when the cells were treated with the two affibody molecules, whereas a decrease was seen after treatment with trastuzumab. The fact that no difference in apoptosis was found in this study can be related to the absence of expression of the apoptosis protein p53 in the SKOV-3 cell line [128]. The final biological test was a proliferation test. SKOV-3 cells were cultured in the presence of trastuzumab, (ZHER2:4)2, ZHER2:342 or without the substances. The substances were replaced three times per week, and once every week, the cells were counted and a suitable number of cells were seeded in new bottles. Cell numbers were then counted back as if all cells were kept. As seen in Table 8, (ZHER2:4)2 significantly increased the proliferation, whereas trastuzumab and ZHER2:342 reduced the proliferation rate (not significant) in SKOV-3 cells. These results are in correlation with the signaling study, where (ZHER2:4)2 increased the phosphorylation of the proliferation- associated Erk, and trastuzumab and ZHER2:342 reduced the phosphorylation.

211At labeled HER-2 binding affibody molecules As earlier discussed, targeted radiotherapy with alpha emitters is a promising concept due to the high-LET radiation from the alpha decay. One problem associated with alpha-emitters is their short physical half-life (Table 3). The physical half-life of the radionuclide must be in correlation with the biological half-life of the targeting agent so that as many of the decays as possible will take place at the target site. With antibodies, it usually takes days before the tumor-to-normal tissue ratio is beneficial for therapy. This is a result of the large size of antibodies, eliminating the possibility of filtration through the kidneys. The pharmacokinetic properties of the much smaller HER-2 targeting affibody molecules can thus be an attractive approach.

37 125 The biodistribution of [ I]PIB-(ZHER2:4)2 in mice was encouraging and the pharmacokinetic properties, with a half-life in the elimination phase of 45 minutes, was beneficial for targeted radiotherapy with short-lived radionuclides. It was thus decided to analyze the potential for targeted radiotherapy using 211At labeled HER-2 binding affibody molecules.

In vitro therapy (V) 211 The possibility to radiolabel (ZHER2:4)2 with At using the same chemistry as for 125I (Figure 11) was first investigated. It was also tested if [211At]PAB- (ZHER2:4)2 could specifically bind to three different HER-2 overexpressing cell lines and kill the cells in a dose-dependent manner. 211 Two different amounts of [ At]PAB-(ZHER2:4)2 was used in each experiment - one corresponding to approximately one affibody molecule per cellular HER-2 (1:1), and one corresponding to an approximate 5-fold excess of affibody molecules (5:1). To test the binding specificity, a 500-fold excess of 211 unlabeled (ZHER2:4)2 compared with the 5:1 concentration of [ At]PAB- (ZHER2:4)2 was added to three dishes (blocked group). To test the influence of this high concentration of (ZHER2:4)2 (about 500 nM) on the growth, three dishes were treated with only (ZHER2:4)2. The cells were exposed to the substances for 24 h, and were then washed and the growth of the cells was then monitored for about two months. To measure how many radioactive decays each cell was exposed to, the cell associated radioactivity in parallel dishes was measured. The values were corrected for the decays occurring from the time the sample was taken to the time of measurement. Accumulated dose was then calculated as the area under the curve (Figure 19).

211 Figure 19. Uptake of [ At]PAB-(ZHER2:4)2 in a) SKOV-3 cells, b) SKBR-3 cells and c) BT-474 cells. Values are an average of three dishes and the error bars show the standard deviation.

The cells were counted once every week and a suitable amount of cells were kept each week. Growth curves were then constructed as if all cells would have been saved at each sub-cultivation. When all groups had reached expo-

38 Figure 20. Growth of a) SKOV-3, b) SKBR-3 and c) BT-474 after exposure to 211 [ At]PAB-(ZHER2:4)2. Plotted values are averages from three bottles. Error bars are standard deviations. nential growth, the exponential curve was extrapolated back to the time of exposure, to determine what number of cells was responsible for the re- growth (Figure 20). From this, an approximate estimate of the surviving fraction can be calculated [129]. Table 9 summarizes the accumulated dose, survival data and obtained fi- 211 nal growth rate. The sensitivity against both At-(ZHER2:4)2 and unlabeled (ZHER2:4)2 differed to a large extent. In both cases, SKOV-3 was the most resistant, SKRB-3 intermediately sensitive and BT-474 the most sensitive. The sensitivity of especially BT-474 against unlabeled (ZHER2:4)2 is highly remarkable. We showed in paper IV that continuous incubation with (ZHER2:4)2 influences the growth rate of SKBR-3. In that case, it increased the growth rate. Here, only 24 h of incubation with (ZHER2:4)2 results in a growth delay of several days. It is also remarkable if the sensitivity against high- LET, such as 211At radiation, differs between cell lines. Cell type dependent sensitivity to high-LET is normally not expected [130].

39 Table 9. Survival data extracted from Figure 20 and accumulated dose calculation according to Figure 19. Dose (DPC) Surviving fraction (%) Growth rate (h) SKOV-3 Control 100 39 Blocked control 99 39 Blocked 5:1 10 51 38 1:1 22 81 38 5:1 60 18 38 SKBR-3 Control 100 61 Blocked control 10 60 Blocked 5:1 19 6 62 1:1 34 18 62 5:1 137 0.3 60 BT-474 Control 100 121 Blocked control 61 120 Blocked 5:1 17 36 129 1:1 24 30 133 5:1 83 0 -

One plausible explanation to why the cell lines responded differently to the 211At irradiation could be that the radioactive decay takes place at different distances from the radiosensitive target. Thus, the cellular geometries as well 211 as the degree of internalization of [ At]PAB-(ZHER2:4)2 in all cell lines was measured. Table 10 summarizes the geometry data of the three cell lines.

Table 10. Summary of cellular geometry factor. The cytoplasmic distance corresponds to the average distance between the cellular membrane and the nucleus. Cell line Cellular diameter Nuclear diameted Cytoplasmic distance

SKOV-3 15.6 ± 3.1 µm 8.4 ± 1.6 µm 3.6 µm SKBR-3 18.1 ± 3.0 µm 10.3 ± 2.2 µm 3.9 µm BT-474 17.5 ± 2.0 µm 11.8 ± 1.8 µm 2.9 µm

Internalization was measured by conjugating the fluorophor CypHer5E to (ZHER2:4)2. CypHer5E is only fluorescent at low pH, and will indicate presence of (ZHER2:4)2 in endosomes and lysosomes. Cells were treated with Cy- pHer5E-(ZHER2:4)2 for one hour in 37 ºC and the internalization was then quantified using flow cytometry and visualized in confocal microscopy. As negative control, a 100-fold excess of unconjugated (ZHER2:4)2 was used in the flow cytometry assay and incubation on ice (where no internalization

40 should occur) was used in the confocal microscopy experiment. As seen in Figure 21, specific internalization was found in all cell lines. However, much less internalization was found in SKOV-3 cells compared with SKBR- 3 and BT-474 cells. The geometrical average of the fluorescence signal (found on the x-axis in Figure 21) was 14.8 for SKOV-3 cells, 56.7 for SKBR-3 cells and 68.3 for BT-474 cells. This means that SKOV-3 cells internalized CypHer5E-(ZHER2:4)2 to a much less degree compared with the other cell lines, and BT-474 was the cell line with highest amount of internalization.

Figure 21. Internalization of CypHer-5E-(ZHER2:4)2 measured with confocal microscopy (CM) and flow cytometry (FC). In the CM pictures, the nucleus is colored in blue, whereas internalized (ZHER2:4)2 is colored in red. In the FC, red curves represent cells exposed to only CypHer-5E-(ZHER2:4)2, whereas green curves represent cells exposed to CypHer-5E-(ZHER2:4)2 in combination with a 100-fold excess of unlabeled (ZHER2:4)2. a) CM: SKOV-3, 37 °C b) CM: SKOV-3, 0 °C c) FC: SKOV-3 d) CM: SKBR-3, 37 °C e) CM: SKBR-3, 0 °C f) FC: SKBR-3 g) CM: BT-474, 37 °C h) CM: BT-474, 0 °C i) FC: BT-474

When the internalization was investigated with confocal microscopy (Figure 21), it was difficult to find any internalization in SKOV-3 cells, whereas internalized vesicles was clearly found in SKBR-3 and BT-474. The most

41 pronounced internalization was found in BT-474 cells also in the confocal microscopy study. Incubation on ice resulted in no visible internalization in any of the cell lines. Since no information about the intrinsic radiosensitivity of the investigated cell lines was found in the literature, a low-LET irradiation study was performed. The three cell lines were irradiated with 137Cs J-ray photons at doses between 2 and 8 Gy and the survival was measured both using the conventional clonogenic survival assay and using the cell growth assay described above. As seen in Figure 22, the sensitivity order between the cell lines is the same with low-LET as seen with high-LET.

Figure 22. Survival curves derived from a) the cell growth assay, and b) the clonogenic survival assay. The data from the clonogenic survival assay are mean values from 3-12 cultures while the data from the growth assay are from three cultures. Mean values and standard deviations are shown. The curves were obtained form curve fitting using a linear quadratic model.

Although the sensitivity against high-LET is normally not cell line dependent, such differences can be attributed to differences in intrinsic radiosensitivity, dependent on e.g. nuclear size and p53 status. Also, differences in internalization of the targeting agent are important, since the distance between the decay and the nucleus depicts the probability of hitting the radiosensitive target - the DNA. When 211At is used, internalization is even more important, since its daughter nuclide, the alpha emitter 211Po, which is responsible for about 60% of the energy deposition [131], easily diffuses away from the targeted cell if not internalized [132]. The in vitro effects of 211At

42 has been thoroughly been investigated before and high potency for 211At in tumor therapy has been shown [133, 134].

Biodistribution (VI) 125 The biodistribution of the I labeled (ZHER2:4)2 showed that if the same biodistribution could be obtained using the alpha-emitter 211At, this was a prom- 211 ising therapeutic setting. Also, it was shown that the At labeled (ZHER2:4)2 in a specific manner could bind to and kill HER-2 overexpressing cells. To 211 evaluate the possibility for targeted radiotherapy with [ At]PAB-(ZHER2:4)2, a biodistribution study in tumor bearing mice was performed. The biodistribution was then compared with the previous biodistribution of [125I]PIB- (ZHER2:4)2.

125 Figure 23. Comparison between the concentrations of [ I]PIB-(ZHER2:4)2 (open 211 squares) and [ At]PAB-(ZHER2:4)2 (filled triangles) in selected tissues. The tissues in the graph are: a) blood, b) tumor, c) lungs, d) liver, e) spleen, f) kidneys, g) stomach, and h) salivary glands. Values are mean values obtained from four mice. Error bars are standard deviations.

In most cases, when changing one radiohalogen for another, the biodistribution remains unaffected [135-137]. However, the are some reported cases, especially when smaller proteins, like antibody fragments, have been used, where the biodistribution has changed significantly when the radiohalogen has been changed [138, 139]. Figure 23 is a comparison between the biodistribution of [125I]PIB- 211 (ZHER2:4)2 and [ At]PAB-(ZHER2:4)2 in the organs with the greatest difference between 125I and 211At, and some other interesting organs. Remarkably higher concentration of radioactivity was found in lung, spleen, stomach and salivary glands when the 211At labeled affibody molecule was used.

43 Since lung, spleen, stomach and salivary glands are organs that are known to accumulate free 211At, it was hypothesized that the 211At label was not 211 stable when administrated as [ At]PAB-(ZHER2:4)2. By changing the linker molecule from the benzamide derivative PAB to a decaborate based new linker molecule, it was further hypothesized that the binding strength of 211At would increase and hence the in vivo stability, as previously had been found [140]. The decaborate based linker molecule was site-specifically coupled to the c-terminal cysteine on the affinity maturated HER-2 binding affibody mole- 211 cule, ZHER2:342-cys [125], using maleimide chemistry. The At labeled compound was then injected into normal mice, and the biodistribution was measured 4 h p.i.. As seen in Figure 24, introduction of the decaborate based linker molecule reduced the radioactivity concentration in lung, stomach, thyroid and salivary glands, which indicated higher stability of the conjugate. However, much higher concentration of radioactivity was instead found in kidneys and liver, probably as a result of increased residualization of the double-negatively charged linker molecule.

211 Figure 24. Biodistribution of [ At]B10-ZHER2:342 in NMRI mice 4 h p.i. (gray bars) 211 compared with [ At]PAB-(ZHER2:4)2 (white bars). Values are mean values obtained from four mice. Error bars are standard deviations. A standard thyroid weight of 5 mg/20 g mouse was used.

According to a previous study, the uptake of free 211At in normal tissues can be prevented by co-administration of Na-thiocyanate [141]. To test this hy- pothesis, and also to test if we could block the release of free 211At by administration of L-lysine, that would block kidney uptake and thereby degradation there, a 4-h biodistribution study in normal mice was performed. One group received 2 mg L-lysine per gram body weight i.p. 30 min before ra- 211 dioactive injection, and 1, 2 and 3 h after injection of [ At]PAB-(ZHER2:4)2.

44 Another group received 97 µg Na-thiocyanate per gram body weight i.p. 24 and 1 h before the radioactive injection. The last group received L-lysine and Na-thiocyanate as described above. Comparison was made with a group 211 receiving only the injection of [ At]PAB-(ZHER2:4)2. As seen in the decaborate study, a reduction in the concentration of radioactivity was found in stomach, thyroid and salivary glands, especially in the group receiving both L-lysine and Na-thiocyanate (Figure 25). In the combination group, significant reduction was found also in heart, lung, pancreas and the intestine.

211 Figure 25. Biodistribution of [ At]PAB-(ZHER2:4)2 in normal mice four hours after injection. White bars are controls, gray bars represent mice that received L-lysine, dashed bars represent mice that received Na-thiocyanate and gray dashed bars represent mice that received both these substances. Values are mean values from four mice. Error bars are standard deviations. The inserted image shows the thyroid uptake (standard thyroid weight – 5 mg/20 g mouse) represented by a box that extends from the 25:th percentile to the 75:th percentile with a line at the median. The bars in the inserted figure show the highest and lowest value.

To be able to calculate if this would be enough to perform a therapy study, a kinetic biodistribution from 1 to 16 hours was done in tumor bearing mice. 211 About 160 kBq [ At]PAB-ZHER2:342-cys injected i.v. in all mice. L-lysine and Na-thiocyanate was administrated as above.

45 Based on the kinetic biodistribution obtained, absorbed dose for all organs was calculated (Table 11). The dosimetry calculation showed that the thyroid received the highest radiation dose from 211At – about 8 times more than the tumor. Also the lungs and spleen received higher dose than the tumor – about twice that of the tumor. Liver and kidneys were exposed to about the same dose as the tumor. The lung is assumed to be a critical organ, making it imperative that the dose to the lungs must be further decreased before this approach can be applied in a therapeutic setting. In conclusion, changing from one radiohalogen to another (i.e. from 125I to 211At) and also by changing linker molecule in the radiolabeling process, can result in dramatic changes in the biodistribution. The best solution found so far, when applying 211At labeled affibody molecules for therapy, seems to be to use the PAB linker and to administrate the conjugate in combination with L-lysine and Na-thiocyanate. However, the remaining problem with high concentrations of the radionuclide in e.g. lungs and spleen limits the use for this setting in radionuclide therapy, and optimization regarding decaborate-based linker molecules will be further explored.

Table 11. Calculation of absorbed dose in different organs and tissues after expo- 211 sure to [ At]PAB-ZHER2:342-cys after co-administration of L-lysine and Na- thiocyanate. About 160 kBq 211At per mouse was injected. Absorbed dose (Gy) ± SD Blood 1.39 ± 0.13 Heart 1.18 ± 0.15 Lung 3.78 ± 0.71 Liver 1.66 ± 0.18 Spleen 4.09 ± 1.03 Pancreas 0.45 ± 0.07 Kidney 2.18 ± 0.37 Stomach 1.20 ± 1.59 S. Intestine 0.75 ± 0.22 L. Intestine 0.85 ± 0.16 Salivary glands 1.04 ± 0.15 Thyroid 17.2 ± 51.4 Skin 1.13 ± 0.14 Muscle 0.28 ± 0.07 Bone 0.72 ± 0.16 Tumor 2.05 ± 0.19 Brain 0.25 ± 0.06

46 Other ongoing studies with tumor targeting affibody molecules

From the original group of four people (Maria Wikman, Stefan Ståhl, Jörgen Carlsson and myself), the number of people working with tumor targeting affibody molecules has grown tremendously. Writing about what everyone is doing would result in at least one more thesis. Some of the ongoing activities are only mentioned shortly here. At the Royal Institute of Technology (KTH), a number of EGFR binding affibody molecules has been selected in a similar manner as used for the selection of the HER-2 binding versions. Competitive binding with EGF and the monoclonal antibody cetuximab was tested and confirmed at our depart- ment in Uppsala. The EGFR binders had lower affinity than the corresponding HER-2 binders. KD for the first generation of EGFR binding affibody molecules, obtained after four rounds of panning against EGFR ECD, was about 0.5-1 µM and for the bivalent constructs 25-50 nM. Affinity maturation has now been initiated and will probably increase the affinity further. Epitope mapping of the HER-2 binding affibody molecules has been discussed for a long time, and finally this work has been initiated by researchers at KTH. In the method used, bacterial display and FACS technology is used for rapid screening. The company Affibody AB was responsible for the successful affinity maturation of the HER-2 binding affibody molecule, where the affinity increased more than 2000 times. The company is also responsible for many interesting animal studies. In one of them, where excellent therapeutic responses were obtained, the ȕ-emitter 177Lu was coupled to an affibody construct, consisting of two domains of ZHER2:342 and one albumin binding domain. The albumin binding domain prolongs the serum half-life and prevents kidney excretion, thereby optimizing for 177Lu therapy and sparing the kidneys. Affibody AB has also achieved impressive tumor imaging using 111In- DOTA- ZHER2:342, where the DOTA chelator was introduced in the synthesis. This product will enter clinical trials in short. Also, labeling of affibody molecules with 99mTc for tumor imaging has been investigated [142], with the hope to simplify the use at conventional nuclear medicine departments and reduce the cost.

47 Summary

The work presented in this thesis deals with selection, development and testing of HER-2 binding affibody molecules. For the first time, affibody molecules for medical use were developed. Phage display technology was used to select an affibody molecule with high affinity (KD = 50 nM) for the tumor marker HER-2. To improve the binding characteristics, a bivalent version - (ZHER2:4)2 - was constructed, thereby decreasing the dissociation equilibrium 125 constant, KD, to 3 nM. The affibody molecules were radiolabeled with I and tested for binding to HER-2 overexpressing cells in vitro. Specific binding, at a site different from trastuzumab, was found. The bivalent version showed much better cellular retention compared with the monovalent form and was thus chosen for further studies. The 125I labeled bivalent affibody molecule was also investigated in vivo in tumor bearing mice. We showed that the tumor could be specifically targeted and that most normal organs received low uptake of 125I. Tumor-to blood ratios of about 10 could be achieved 8 h post injection. We further showed that it was possible to visualize the tumor uptake in a gamma camera. The biologic effects of the bivalent affibody molecule and a monovalent affinity maturated version - ZHER2:342 - was measured and compared with the effects of trastuzumab. It was found that although all molecules target the same protein, the effects differed greatly. The effects found in the signaling pathways correlated to some degree with changes if proliferation and migration, e.g. (ZHER2:4)2 stimulated phosphorylation of Erk1/2 and PLCJ1, as well as growth and migration, while ZHER2:342 did not. ZHER2:342 even inhibited phosphorylation of PLCJ1 and migration. Small proteins, such as affibody molecules, are suitable for targeted radiotherapy with alpha emitting radionuclides since their fast biokinetics match the short physical half-lives of the alpha-emitters. (ZHER2:4)2 was therefore radiolabeled with the alpha emitter 211At and the effect was tested in three different HER-2 expressing cell lines in vitro. We showed that all three cell lines could be specifically targeted and killed by the 211At labeled affibody molecule and the sensitivity against the radioactivity and unlabeled (ZHER2:4)2 differed among the cell lines. This was surprising, since differences in radiosensitivity are generally not seen for high-LET radiation. Further investiga- tions showed that at least some of the differences in sensitivity could be ex-

48 plained by differences in internalization in the different cell lines and geometrical differences of the cells. The 211At labeled bivalent affibody molecule was then tested in vivo, where it was found that the 211At label was not stable, causing release of free radiolabel and uptake in normal organs. By changing linker molecule to a decaborate based linker, the stability was markedly increased. However, new problems, such as high concentration of radioactivity in liver and kidneys, originating from the residualizing properties of the charged linker molecule were found. Instead, the first linker molecule was tried again, but degradation in kidneys was blocked with L-lysine and uptake of free 211At in normal organs was blocked by Na-thiocyanate. The results were promising and a kinetic biodistribution and dosimetry study was initiated. However, the reduction of 211At uptake in lungs and spleen was not enough, causing the dose to those organs to be twice that of the tumor. Liver and kidneys received about the same dose as the tumor. Especially the dose to the lungs was considered a problem, which has to be solved before applying this approach in a therapeutic setting. The best approach might be to optimize the decaborate based linker molecule so that the uptake in liver and kidneys can be reduced while retaining the stabilizing property of the decaborate linker.

49 Future studies

There are always new exciting studies to do, but in terms of a Ph.D. project, the line has to be drawn somewhere. Hopefully, this does not mean that the project will not be continued. Some interesting studies appear as a direct consequence of the work presented in this thesis. x The intratumoral localization study in paper III should be expanded to increase the understanding of how the radiolabeled affibodies penetrate the tumor. Perhaps should also the affibody molecules be stained using IHC in order to discriminate between free radioactivity and radioactivity bound to the affibody molecules. Similar studies in the kidneys could increase the understanding of the kidney uptake. x As shown in paper IV, the bivalent affibody molecule induced a signaling pattern similar to trastuzumab, whereas the monovalent affibody molecule induced a completely different signaling pattern. It would be interesting to investigate if this is a result of dimerization of receptors in the EGFR family and in that case - which receptor dimers are responsible for which signal. x The affibody molecule-based delivery of 211At to tumors in vivo should be optimized. As previously discussed, optimization of the decaborate based linker molecules is probably the best way to go. Alternatively, the alpha-emitting Bi isotopes could be tested. A second alternative could be to introduce a nuclear localization signal in the affibody sequence and test therapy with Auger electrons from e.g. 111In or 123I.

In a wider perspective, it would be interesting to select affibody molecules directed against different epitopes on HER-2 and study the biological effects depending on where the molecule binds. Bi-specific affibody molecules directed against both HER-2 and EGFR is an obvious approach in order to increase the tumor specificity, but also bi-specific affibody molecules directed against different domains of HER-2 would be interesting to study. Selection of an affibody molecule that prevents dimerization, like the monoclonal antibody pertuzumab, has been discussed. In conclusion, the concept of using affibody molecules in tumor targeting has been very successful, and this field will most likely be further exploited, with new molecular targets and effector moieties as a result.

50 Acknowledgments

Arbetet i denna avhandling har huvudsakligen utförts på avdelningen för Biomedicinsk strålningsvetenskap (BMS) vid Uppsala universitet med finan- siering från Cancerfonden. Massor av människor har hjälp till på en mängd olika sätt, och jag skulle speciellt vilja tacka några av dem här.

Först av allt vill jag tacka min huvudhandledare Jörgen Carlsson för att du trott på mig och stått på min sida och för att du bidrar till den extremt trevliga stämningen på avdelningen. Tack för alla tappra försök att skaka fram ett fungerande projekt när AHNP-projektet havererade! Ett helt gäng projekt blev det... Tur att åtminstone ett av dem gav resultat.

Tack även till min biträdande handledare Vladimir Tolmachev för ditt oänd- liga kunnande i radiokemi och för att du alltid kommer med nya idéer.

Jag vill även tacka min andra biträdande handledare Lars Gedda för att du alltid tar dig tid och förklarar saker och ting.

Greg Adams, for supply of the extracellular domains of all receptors in the EGFR family. Without the HER-2 ECD there wouldn’t be an affibody project… But most of all, thanks for being such a nice guy!

Scott Wilbur, for patiently supplying me with linker molecules for 211At labeling and for answering all my questions.

Maria & Stefan, för att ni nappade på vår briljanta idé att selektera affibodymolekyler mot HER-2. Vad hade hänt om vi inte träffat på er? Tack för roliga möten och för att ni alltid är så positiva!

Anna Orlova, för att du introducerade mig för djurförsök och lärde mig handskas med möss (och hur man får dem att kissa på beställning).

Lovisa, för att du ställde upp och gjorde den snygga internaliseringsstudien på så kort varsel.

51 Johan Lennartsson, för att du sökte upp samarbetet med oss och för all ex- pertis i signaleringsbranchen. Vi behöver verkligen en expert som kan över- blicka det extremt trassliga signaleringsnätverket.

Stort tack även till Lina Ekerljung som tog sig an uppgiften att analysera den påverkan på signaleringsvägarna som våra affibodymolekyler ger och för tappra försök att inducera apoptos i SKOV-3...

Veronika, för all hjälp med absolut allting! Framförallt på sistone, då jag varit upptagen med avhandlingen. Hur skulle vi på BMS klara oss utan dig?

Fredrik, Rickard och Anders och resten av gänget på Affibody AB för anga- gemang, trevliga möten och kontinuelig leverans av diverse affibodymolekyler.

Åsa, för att du övertalade mig att göra exjobb på BMS (det ångrar jag inte för en sekund), och för att du lärt mig det mesta av vad det innebär att vara en BMS-are. Tack dessutom för att du drog med oss till Spanien på en helt underbar vecka med god mat och trevligt sällskap i en underbar miljö.

Ylva, för att du stått ut med fullspäckade tidsscheman och underhållit mig alla sena kvällar på djuravdelningen. Att göra djurförsök med dig är ju fak- tiskt roligt! Tack även för du delat rollen som reseledare på skidresorna med mig och för att du är rummets telefonist!

Per Nilsson, för att du introducerade mig för fjällvandring. Bättre avstress- ning finns inte! Nu blir du så illa tvungen att följa med till Norge i höst...

Marika och Micke för att ni lät er bli övertalade att komma till BMS. Där- med säkerställdes molekylärarnas dominans på avdelningen. Marika - dra inte ner på koffeinet. Det är värt en och annan kängru att få höra dina histo- rier! Jag tror den blinda hunden håller med.

Tack dessutom till alla gamla, nya och tillfälliga BMS:are för att ni gör det roligt att gå till jobbet varje dag. Tänk så gamla vi kommer bli om vi fortsät- ter och skratta så mycket som vi gör... Tack även till alla BMS-respektivar för att ni involverar er i BMS-aktiviteterna och jämnar ut könsfördelningen.

Johan och alla andra i ”lördagsfikagänget”, för att ni drar ut oss på lördags- fika (och badutflykter sommartid) lite då och då.

Mamma och pappa för att ni alltid stöttat och uppmuntrat mig och hjälpt mig på alla tänkbara sätt. Tack även Björn för att du som storebror banat väg för mig...

52 Till sist vill jag rikta ett stort tack till min älskade Jesper som ger mig så mycket kärlek och trygghet.

53 References

1. Cancer i siffror: Cancerfonden and Socialstryrelsen; 2005. 2. Talback M, Stenbeck M, Rosen M, Barlow L, Glimelius B: Cancer survival in Sweden 1960-1998--developments across four decades. Acta Oncol 2003, 42(7):637-659. 3. Christofferson R, Claesson-Welsh L, Muhr C: [Anti-angiogenic drugs probable complement in cancer therapy]. Lakartidningen 2002, 99(42):4138-4139, 4142-4138. 4. Reed JC: Apoptosis mechanisms: implications for cancer drug discovery. Oncology (Williston Park) 2004, 18(13 Suppl 10):11-20. 5. Spira AI, Carducci MA: Differentiation therapy. Curr Opin Pharmacol 2003, 3(4):338-343. 6. Lo HW, Day CP, Hung MC: Cancer-specific gene therapy. Adv Genet 2005, 54:235-255. 7. Blattman JN, Greenberg PD: Cancer immunotherapy: a treatment for the masses. Science 2004, 305(5681):200-205. 8. Carlsson J, Forssell Aronsson E, Hietala SO, Stigbrand T, Tennvall J: Tu- mour therapy with radionuclides: assessment of progress and problems. Radiother Oncol 2003, 66(2):107-117. 9. Klein S, McCormick F, Levitzki A: Killing time for cancer cells. Nat Rev Cancer 2005, 5(7):573-580. 10. Gambert SR: Screening for prostate cancer. Int Urol Nephrol 2001, 33(2):249-257. 11. Winau F, Westphal O, Winau R: Paul Ehrlich--in search of the magic bullet. Microbes Infect 2004, 6(8):786-789. 12. Bazley LA, Gullick WJ: The epidermal growth factor receptor family. Endocr Relat Cancer 2005, 12 Suppl 1:S17-27. 13. Cohen S: Isolation of a mouse submaxillary gland protein accelerating incisor eruption and eyelid opening in the new-born animal. J Biol Chem 1962, 237:1555-1562. 14. Carpenter G, Cohen S: Epidermal growth factor. Annu Rev Biochem 1979, 48:193-216. 15. Sorkin A: Endocytosis and intracellular sorting of receptor tyrosine kinases. Front Biosci 1998, 3:d729-738. 16. Shinojima N, Tada K, Shiraishi S, Kamiryo T, Kochi M, Nakamura H, Makino K, Saya H, Hirano H, Kuratsu J et al: Prognostic value of epidermal growth factor receptor in patients with glioblastoma multiforme. Cancer Res 2003, 63(20):6962-6970.

54 17. Agosti RM, Leuthold M, Gullick WJ, Yasargil MG, Wiestler OD: Expres- sion of the epidermal growth factor receptor in astrocytic tumours is specifically associated with glioblastoma multiforme. Virchows Arch A Pathol Anat Histopathol 1992, 420(4):321-325. 18. Colquhoun AJ, Mellon JK: Epidermal growth factor receptor and bladder cancer. Postgrad Med J 2002, 78(924):584-589. 19. Salomon DS, Brandt R, Ciardiello F, Normanno N: Epidermal growth factor-related peptides and their receptors in human malignancies. Crit Rev Oncol Hematol 1995, 19(3):183-232. 20. Dunn WA, Connolly TP, Hubbard AL: Receptor-mediated endocytosis of epidermal growth factor by rat hepatocytes: receptor pathway. J Cell Biol 1986, 102(1):24-36. 21. Gladhaug IP, Refsnes M, Christoffersen T: Regulation of surface expression of high-affinity receptors for epidermal growth factor (EGF) in hepatocytes by hormones, differentiating agents, and phorbol ester. Dig Dis Sci 1992, 37(2):233-239. 22. Bue P, Holmberg AR, Marquez M, Westlin JE, Nilsson S, Malmstrom PU: Intravesical administration of EGF- dextran conjugates in patients with superficial bladder cancer. Eur Urol 2000, 38(5):584-589. 23. Harris RC, Chung E, Coffey RJ: EGF receptor ligands. Exp Cell Res 2003, 284(1):2-13. 24. Peles E, Yarden Y: Neu and its ligands: from an oncogene to neural factors. Bioessays 1993, 15(12):815-824. 25. Tarakhovsky A, Zaichuk T, Prassolov V, Butenko ZA: A 25 kDa polypep- tide is the ligand for p185neu and is secreted by activated macro- phages. Oncogene 1991, 6(12):2187-2196. 26. Holmes WE, Sliwkowski MX, Akita RW, Henzel WJ, Lee J, Park JW, Yansura D, Abadi N, Raab H, Lewis GD et al: Identification of heregulin, a specific activator of p185erbB2. Science 1992, 256(5060):1205-1210. 27. Lupu R, Colomer R, Kannan B, Lippman ME: Characterization of a growth factor that binds exclusively to the erbB-2 receptor and induces cellular responses. Proc Natl Acad Sci U S A 1992, 89(6):2287-2291. 28. Dobashi K, Davis JG, Mikami Y, Freeman JK, Hamuro J, Greene MI: Characterization of a neu/c-erbB-2 protein-specific activating factor. Proc Natl Acad Sci U S A 1991, 88(19):8582-8586. 29. Lupu R, Colomer R, Zugmaier G, Sarup J, Shepard M, Slamon D, Lippman ME: Direct interaction of a ligand for the erbB2 oncogene product with the EGF receptor and p185erbB2. Science 1990, 249(4976):1552-1555. 30. Yarden Y, Weinberg RA: Experimental approaches to hypothetical hormones: detection of a candidate ligand of the neu protooncogene. Proc Natl Acad Sci U S A 1989, 86(9):3179-3183. 31. Bacus SS, Huberman E, Chin D, Kiguchi K, Simpson S, Lippman M, Lupu R: A ligand for the erbB-2 oncogene product (gp30) induces differentiation of human breast cancer cells. Cell Growth Differ 1992, 3(7):401- 411. 32. Peles E, Bacus SS, Koski RA, Lu HS, Wen D, Ogden SG, Levy RB, Yarden Y: Isolation of the neu/HER-2 stimulatory ligand: a 44 kd gly-

55 coprotein that induces differentiation of mammary tumor cells. Cell 1992, 69(1):205-216. 33. Carraway KL, 3rd, Sliwkowski MX, Akita R, Platko JV, Guy PM, Nuijens A, Diamonti AJ, Vandlen RL, Cantley LC, Cerione RA: The erbB3 gene product is a receptor for heregulin. J Biol Chem 1994, 269(19):14303- 14306. 34. Plowman GD, Green JM, Culouscou JM, Carlton GW, Rothwell VM, Buckley S: Heregulin induces tyrosine phosphorylation of HER4/p180erbB4. Nature 1993, 366(6454):473-475. 35. Graus-Porta D, Beerli RR, Daly JM, Hynes NE: ErbB-2, the preferred heterodimerization partner of all ErbB receptors, is a mediator of lat- eral signaling. Embo J 1997, 16(7):1647-1655. 36. Kruger S, Weitsch G, Buttner H, Matthiensen A, Bohmer T, Marquardt T, Sayk F, Feller AC, Bohle A: HER2 overexpression in muscle-invasive urothelial carcinoma of the bladder: prognostic implications. Int J Can- cer 2002, 102(5):514-518. 37. Coogan CL, Estrada CR, Kapur S, Bloom KJ: HER-2/neu protein overexpression and gene amplification in human transitional cell carcinoma of the bladder. Urology 2004, 63(4):786-790. 38. Ross JS, McKenna BJ: The HER-2/neu oncogene in tumors of the gastrointestinal tract. Cancer Invest 2001, 19(5):554-568. 39. Slamon DJ, Clark GM, Wong SG, Levin WJ, Ullrich A, McGuire WL: Human breast cancer: correlation of relapse and survival with amplification of the HER-2/neu oncogene. Science 1987, 235(4785):177-182. 40. Berchuck A, Kamel A, Whitaker R, Kerns B, Olt G, Kinney R, Soper J, Dodge R, Clarke-Pearson D, Marks P: Overexpression of HER-2/neu is associated with poor survival in advanced epithelial ovarian cancer. Cancer Res 1990, 50(13):4087-4091. 41. Carlsson J, Nordgren H, Sjostrom J, Wester K, Villman K, Bengtsson NO, Ostenstad B, Lundqvist H, Blomqvist C: HER2 expression in breast cancer primary tumours and corresponding metastases. Original data and literature review. Br J Cancer 2004, 90(12):2344-2348. 42. Kraus MH, Issing W, Miki T, Popescu NC, Aaronson SA: Isolation and characterization of ERBB3, a third member of the ERBB/epidermal growth factor receptor family: evidence for overexpression in a subset of human mammary tumors. Proc Natl Acad Sci U S A 1989, 86(23):9193-9197. 43. Plowman GD, Culouscou JM, Whitney GS, Green JM, Carlton GW, Foy L, Neubauer MG, Shoyab M: Ligand-specific activation of HER4/p180erbB4, a fourth member of the epidermal growth factor receptor family. Proc Natl Acad Sci U S A 1993, 90(5):1746-1750. 44. Guy PM, Platko JV, Cantley LC, Cerione RA, Carraway KL, 3rd: Insect cell-expressed p180erbB3 possesses an impaired tyrosine kinase activity. Proc Natl Acad Sci U S A 1994, 91(17):8132-8136. 45. Citri A, Skaria KB, Yarden Y: The deaf and the dumb: the biology of ErbB-2 and ErbB-3. Exp Cell Res 2003, 284(1):54-65.

56 46. Wiseman SM, Makretsov N, Nielsen TO, Gilks B, Yorida E, Cheang M, Turbin D, Gelmon K, Huntsman DG: Coexpression of the type 1 growth factor receptor family members HER-1, HER-2, and HER-3 has a syn- ergistic negative prognostic effect on breast carcinoma survival. Cancer 2005, 103(9):1770-1777. 47. Ekberg T, Nestor M, Engstrom M, Nordgren H, Wester K, Carlsson J, Anniko M: Expression of EGFR, HER2, HER3, and HER4 in metastatic squamous cell carcinomas of the oral cavity and base of tongue. Int J Oncol 2005, 26(5):1177-1185. 48. Yarden Y, Sliwkowski MX: Untangling the ErbB signalling network. Nat Rev Mol Cell Biol 2001, 2(2):127-137. 49. Zhang W, Liu HT: MAPK signal pathways in the regulation of cell proliferation in mammalian cells. Cell Res 2002, 12(1):9-18. 50. Thomas SM, Coppelli FM, Wells A, Gooding WE, Song J, Kassis J, Dren- ning SD, Grandis JR: Epidermal growth factor receptor-stimulated activation of phospholipase Cgamma-1 promotes invasion of head and neck squamous cell carcinoma. Cancer Res 2003, 63(17):5629-5635. 51. Wells A, Grandis JR: Phospholipase C-gamma1 in tumor progression. Clin Exp Metastasis 2003, 20(4):285-290. 52. Datta SR, Brunet A, Greenberg ME: Cellular survival: a play in three Akts. Genes Dev 1999, 13(22):2905-2927. 53. Guyton A, Hall J: Human Physiology and Mechanisms of Disease, 6 edn. Philadelphia: W.B Saunders Company; 1996. 54. Yokota T, Milenic DE, Whitlow M, Schlom J: Rapid tumor penetration of a single-chain Fv and comparison with other immunoglobulin forms. Cancer Res 1992, 52(12):3402-3408. 55. Shockley TR, Lin K, Nagy JA, Tompkins RG, Dvorak HF, Yarmush ML: Penetration of tumor tissue by antibodies and other immunoproteins. Ann N Y Acad Sci 1991, 618:367-382. 56. Dvorak HF, Nagy JA, Dvorak AM: Structure of solid tumors and their vasculature: implications for therapy with monoclonal antibodies. Cancer Cells 1991, 3(3):77-85. 57. Kohler G, Milstein C: Continuous cultures of fused cells secreting antibody of predefined specificity. Nature 1975, 256(5517):495-497. 58. Walsh G: Biopharmaceutical benchmarks--2003. Nat Biotechnol 2003, 21(8):865-870. 59. Albanell J, Codony J, Rovira A, Mellado B, Gascon P: Mechanism of action of anti-HER2 monoclonal antibodies: scientific update on trastuzumab and 2C4. Adv Exp Med Biol 2003, 532:253-268. 60. Sliwkowski MX, Lofgren JA, Lewis GD, Hotaling TE, Fendly BM, Fox JA: Nonclinical studies addressing the mechanism of action of trastuzumab (Herceptin). Semin Oncol 1999, 26(4 Suppl 12):60-70. 61. Codony-Servat J, Albanell J, Lopez-Talavera JC, Arribas J, Baselga J: Cleavage of the HER2 ectodomain is a pervanadate-activable process that is inhibited by the tissue inhibitor of metalloproteases-1 in breast cancer cells. Cancer Res 1999, 59(6):1196-1201.

57 62. Gennari R, Menard S, Fagnoni F, Ponchio L, Scelsi M, Tagliabue E, Cas- tiglioni F, Villani L, Magalotti C, Gibelli N et al: Pilot study of the mechanism of action of preoperative trastuzumab in patients with primary operable breast tumors overexpressing HER2. Clin Cancer Res 2004, 10(17):5650-5655. 63. Mimura K, Kono K, Hanawa M, Kanzaki M, Nakao A, Ooi A, Fujii H: Trastuzumab-mediated antibody-dependent cellular cytotoxicity against esophageal squamous cell carcinoma. Clin Cancer Res 2005, 11(13):4898-4904. 64. Yakes FM, Chinratanalab W, Ritter CA, King W, Seelig S, Arteaga CL: Herceptin-induced inhibition of phosphatidylinositol-3 kinase and Akt Is required for antibody-mediated effects on p27, cyclin D1, and anti- tumor action. Cancer Res 2002, 62(14):4132-4141. 65. Petit AM, Rak J, Hung MC, Rockwell P, Goldstein N, Fendly B, Kerbel RS: Neutralizing antibodies against epidermal growth factor and ErbB-2/neu receptor tyrosine kinases down-regulate vascular endothe- lial growth factor production by tumor cells in vitro and in vivo: angiogenic implications for signal transduction therapy of solid tumors. Am J Pathol 1997, 151(6):1523-1530. 66. Izumi Y, Xu L, di Tomaso E, Fukumura D, Jain RK: Tumour biology: herceptin acts as an anti-angiogenic cocktail. Nature 2002, 416(6878):279-280. 67. Baselga J, Norton L, Albanell J, Kim YM, Mendelsohn J: Recombinant humanized anti-HER2 antibody (Herceptin) enhances the antitumor activity of paclitaxel and doxorubicin against HER2/neu overexpressing human breast cancer xenografts. Cancer Res 1998, 58(13):2825- 2831. 68. Pegram M, Hsu S, Lewis G, Pietras R, Beryt M, Sliwkowski M, Coombs D, Baly D, Kabbinavar F, Slamon D: Inhibitory effects of combinations of HER-2/neu antibody and chemotherapeutic agents used for treatment of human breast cancers. Oncogene 1999, 18(13):2241-2251. 69. Sato S, Kajiyama Y, Sugano M, Iwanuma Y, Sonoue H, Matsumoto T, Sasai K, Tsurumaru M: Monoclonal antibody to HER-2/neu receptor enhances radiosensitivity of esophageal cancer cell lines expressing HER-2/neu oncoprotein. Int J Radiat Oncol Biol Phys 2005, 61(1):203- 211. 70. Liang K, Lu Y, Jin W, Ang KK, Milas L, Fan Z: Sensitization of breast cancer cells to radiation by trastuzumab. Mol Cancer Ther 2003, 2(11):1113-1120. 71. Agus DB, Akita RW, Fox WD, Lewis GD, Higgins B, Pisacane PI, Lofgren JA, Tindell C, Evans DP, Maiese K et al: Targeting ligand-activated ErbB2 signaling inhibits breast and prostate tumor growth. Cancer Cell 2002, 2(2):127-137. 72. Holliger P, Hudson PJ: Engineered antibody fragments and the rise of single domains. Nat Biotechnol 2005, 23(9):1126-1136. 73. Froidevaux S, Eberle AN: Somatostatin analogs and radiopeptides in cancer therapy. Biopolymers 2002, 66(3):161-183.

58 74. van der Lely AJ, de Herder WW, Krenning EP, Kwekkeboom DJ: Octreo- scan radioreceptor imaging. Endocrine 2003, 20(3):307-311. 75. Kwekkeboom DJ, Mueller-Brand J, Paganelli G, Anthony LB, Pauwels S, Kvols LK, O'Dorisio T M, Valkema R, Bodei L, Chinol M et al: Overview of results of peptide receptor radionuclide therapy with 3 radiolabeled somatostatin analogs. J Nucl Med 2005, 46 Suppl 1:62S-66S. 76. Ginj M, Hinni K, Tschumi S, Schulz S, Maecke HR: Trifunctional somatostatin-based derivatives designed for targeted radiotherapy using auger electron emitters. J Nucl Med 2005, 46(12):2097-2103. 77. Park BW, Zhang HT, Wu C, Berezov A, Zhang X, Dua R, Wang Q, Kao G, O'Rourke DM, Greene MI et al: Rationally designed anti-HER2/neu peptide mimetic disables P185HER2/neu tyrosine kinases in vitro and in vivo. Nat Biotechnol 2000, 18(2):194-198. 78. Nygren PA, Skerra A: Binding proteins from alternative scaffolds. J Immunol Methods 2004, 290(1-2):3-28. 79. Hosse RJ, Rothe A, Power BE: A new generation of protein display scaffolds for molecular recognition. Protein Sci 2006, 15(1):14-27. 80. Griffiths AD, Williams SC, Hartley O, Tomlinson IM, Waterhouse P, Crosby WL, Kontermann RE, Jones PT, Low NM, Allison TJ et al: Isola- tion of high affinity human antibodies directly from large synthetic repertoires. Embo J 1994, 13(14):3245-3260. 81. Moks T, Abrahmsen L, Nilsson B, Hellman U, Sjoquist J, Uhlen M: Staphylococcal protein A consists of five IgG-binding domains. Eur J Biochem 1986, 156(3):637-643. 82. Jansson B, Uhlen M, Nygren PA: All individual domains of staphylococcal protein A show Fab binding. FEMS Immunol Med Microbiol 1998, 20(1):69-78. 83. Nilsson B, Moks T, Jansson B, Abrahmsen L, Elmblad A, Holmgren E, Henrichson C, Jones TA, Uhlen M: A synthetic IgG-binding domain based on staphylococcal protein A. Protein Eng 1987, 1(2):107-113. 84. Moks T, Abrahmsen L, Holmgren E, Bilich M, Olsson A, Uhlen M, Pohl G, Sterky C, Hultberg H, Josephson S et al: Expression of human insulin- like growth factor I in bacteria: use of optimized gene fusion vectors to facilitate protein purification. Biochemistry 1987, 26(17):5239-5244. 85. Andersson C, Sandberg L, Murby M, Sjolander A, Lovgren-Bengtsson K, Stahl S: General expression vectors for production of hydrophobically tagged immunogens for direct iscom incorporation. J Immunol Methods 1999, 222(1-2):171-182. 86. Jansson M, Li YC, Jendeberg L, Anderson S, Montelione BT, Nilsson B: High-level production of uniformly 15N- and 13C-enriched fusion proteins in Escherichia coli. J Biomol NMR 1996, 7(2):131-141. 87. Nord K, Nilsson J, Nilsson B, Uhlen M, Nygren PA: A combinatorial library of an alpha-helical bacterial receptor domain. Protein Eng 1995, 8(6):601-608. 88. Nord K, Nord O, Uhlen M, Kelley B, Ljungqvist C, Nygren PA: Recombi- nant human factor VIII-specific affinity ligands selected from phage-

59 displayed combinatorial libraries of protein A. Eur J Biochem 2001, 268(15):4269-4277. 89. Nord K, Gunneriusson E, Ringdahl J, Stahl S, Uhlen M, Nygren PA: Bind- ing proteins selected from combinatorial libraries of an alpha-helical bacterial receptor domain. Nat Biotechnol 1997, 15(8):772-777. 90. Hansson M, Ringdahl J, Robert A, Power U, Goetsch L, Nguyen TN, Uhlen M, Stahl S, Nygren PA: An in vitro selected binding protein (affibody) shows conformation-dependent recognition of the respiratory syncytial virus (RSV) G protein. Immunotechnology 1999, 4(3-4):237- 252. 91. Ljungquist C, Jansson B, Moks T, Uhlen M: Thiol-directed immobilization of recombinant IgG-binding receptors. Eur J Biochem 1989, 186(3):557-561. 92. Engfeldt T, Renberg B, Brumer H, Nygren PA, Karlstrom AE: Chemical synthesis of triple-labelled three-helix bundle binding proteins for specific fluorescent detection of unlabelled protein. Chembiochem 2005, 6(6):1043-1050. 93. Nord K, Gunneriusson E, Uhlén M, Nygren PA: Ligands selected from combinatorial libraries of protein A for use in affinity capture of apoli- poprotein A-1M and taq DNA polymerase. J Biotechnol 2000, 80(1):45- 54. 94. Kehoe JW, Kay BK: Filamentous phage display in the new millennium. Chem Rev 2005, 105(11):4056-4072. 95. Wittrup KD: Protein engineering by cell-surface display. Curr Opin Biotechnol 2001, 12(4):395-399. 96. Eckart MR, Bussineau CM: Quality and authenticity of heterologous proteins synthesized in yeast. Curr Opin Biotechnol 1996, 7(5):525-530. 97. Lipovsek D, Pluckthun A: In-vitro protein evolution by ribosome display and mRNA display. J Immunol Methods 2004, 290(1-2):51-67. 98. Schaffitzel C, Hanes J, Jermutus L, Pluckthun A: Ribosome display: an in vitro method for selection and evolution of antibodies from libraries. J Immunol Methods 1999, 231(1-2):119-135. 99. Takahashi TT, Austin RJ, Roberts RW: mRNA display: ligand discovery, interaction analysis and beyond. Trends Biochem Sci 2003, 28(3):159- 165. 100. Cerchia L, Hamm J, Libri D, Tavitian B, de Franciscis V: Nucleic acid aptamers in cancer medicine. FEBS Lett 2002, 528(1-3):12-16. 101. Kassis AI, Adelstein SJ: Radiobiologic principles in radionuclide therapy. J Nucl Med 2005, 46 Suppl 1:4S-12S. 102. Adelstein SJ, Kassis AI, Bodei L, Mariani G: Radiotoxicity of iodine-125 and other auger-electron-emitting radionuclides: background to therapy. Cancer Biother Radiopharm 2003, 18(3):301-316. 103. Lobachevsky PN, Martin RF: DNA breakage by decay of Auger electron emitters: experiments with 123I-iodoHoechst 33258 and plasmid DNA. Radiat Res 2005, 164(6):766-773. 104. Tolmachev V, Orlova A, Lundqvist H: Approaches to improve cellular retention of radiohalogen labels delivered by internalising tumour-

60 targeting proteins and peptides. Curr Med Chem 2003, 10(22):2447- 2460. 105. Sundberg AL, Blomquist E, Carlsson J, Steffen AC, Gedda L: Cellular retention of radioactivity and increased radiation dose. Model experiments with EGF-dextran. Nucl Med Biol 2003, 30(3):303-315. 106. Sundberg AL, Orlova A, Bruskin A, Gedda L, Carlsson J, Blomquist E, Lundqvist H, Tolmachev V: [(111)In]Bz-DTPA-hEGF: Preparation and in vitro characterization of a potential anti-glioblastoma targeting agent. Cancer Biother Radiopharm 2003, 18(4):643-654. 107. Sundberg AL, Gedda L, Orlova A, Bruskin A, Blomquist E, Carlsson J, Tolmachev V: [177Lu]Bz-DTPA-EGF: Preclinical characterization of a potential radionuclide targeting agent against glioma. Cancer Biother Radiopharm 2004, 19(2):195-204. 108. Reist CJ, Garg PK, Alston KL, Bigner DD, Zalutsky MR: Radioiodination of internalizing monoclonal antibodies using N-succinimidyl 5-iodo-3- pyridinecarboxylate. Cancer Res 1996, 56(21):4970-4977. 109. FitzGerald DJ, Kreitman R, Wilson W, Squires D, Pastan I: Recombinant immunotoxins for treating cancer. Int J Med Microbiol 2004, 293(7- 8):577-582. 110. Kreitman RJ: Toxin-labeled monoclonal antibodies. Curr Pharm Bio- technol 2001, 2(4):313-325. 111. Trail PA, Bianchi AB: Monoclonal antibody drug conjugates in the treatment of cancer. Curr Opin Immunol 1999, 11(5):584-588. 112. Park JW, Benz CC, Martin FJ: Future directions of liposome- and im- munoliposome-based cancer therapeutics. Semin Oncol 2004, 31(6 Suppl 13):196-205. 113. Imam SK: Molecular nuclear imaging: the radiopharmaceuticals (review). Cancer Biother Radiopharm 2005, 20(2):163-172. 114. Lundqvist H, Tolmachev V: Targeting peptides and positron emission tomography. Biopolymers 2002, 66(6):381-392. 115. Kairemo KJ: Positron emission tomography of monoclonal antibodies. Acta Oncol 1993, 32(7-8):825-830. 116. Gunneriusson E, Nord K, Uhlen M, Nygren P: Affinity maturation of a Taq DNA polymerase specific affibody by helix shuffling. Protein Eng 1999, 12(10):873-878. 117. Haigler HT, Maxfield FR, Willingham MC, Pastan I: Dansylcadaverine inhibits internalization of 125I-epidermal growth factor in BALB 3T3 cells. J Biol Chem 1980, 255(4):1239-1241. 118. Adams GP, Schier R, McCall AM, Crawford RS, Wolf EJ, Weiner LM, Marks JD: Prolonged in vivo tumour retention of a human diabody targeting the extracellular domain of human HER2/neu. Br J Cancer 1998, 77(9):1405-1412. 119. Adams GP, Schier R, McCall AM, Simmons HH, Horak EM, Alpaugh RK, Marks JD, Weiner LM: High affinity restricts the localization and tumor penetration of single-chain fv antibody molecules. Cancer Res 2001, 61(12):4750-4755.

61 120. Mogensen CE, Solling: Studies on renal tubular protein reabsorption: partial and near complete inhibition by certain amino acids. Scand J Clin Lab Invest 1977, 37(6):477-486. 121. Bernard A, Ouled Amor A, Lauwerys R: Charge-dependent renal uptake of beta 2-microglobulin in conscious rats. Scand J Clin Lab Invest 1992, 52(5):415-423. 122. Adams GP: Improving the tumor specificity and retention of antibody- based molecules. In Vivo 1998, 12(1):11-21. 123. Sundberg-Smith LJ, Doherty JT, Mack CP, Taylor JM: Adhesion stimu- lates direct PAK1/ERK2 association and leads to ERK-dependent PAK1 Thr212 phosphorylation. J Biol Chem 2005, 280(3):2055-2064. 124. Le XF, Lammayot A, Gold D, Lu Y, Mao W, Chang T, Patel A, Mills GB, Bast RC, Jr.: Genes affecting the cell cycle, growth, maintenance, and drug sensitivity are preferentially regulated by anti-HER2 antibody through phosphatidylinositol 3-kinase-AKT signaling. J Biol Chem 2005, 280(3):2092-2104. 125. Orlova A, Magnusson M, Eriksson T, Nilsson M, B. L, Höiden-Guthenberg I, Widström C, Carlsson J, Tolmachev V, Ståhl S et al: Tumor imaging using a picomolar affinity HER2 binding Affibody molecule. Cancer Research 2006, In press. 126. van Engeland M, Nieland LJ, Ramaekers FC, Schutte B, Reutelingsperger CP: Annexin V-affinity assay: a review on an apoptosis detection system based on phosphatidylserine exposure. Cytometry 1998, 31(1):1-9. 127. Kaufmann SH, Earnshaw WC: Induction of apoptosis by cancer chemotherapy. Exp Cell Res 2000, 256(1):42-49. 128. Yaginuma Y, Westphal H: Abnormal structure and expression of the p53 gene in human ovarian carcinoma cell lines. Cancer Res 1992, 52(15):4196-4199. 129. Johansson L, Nilsson K, Carlsson J, Larsson B, Jakobsson P: Radiation effects on cultured human lymphoid cells. Analysis using the growth extrapolation method. Acta Radiol Oncol 1981, 20(1):51-59. 130. Hall E: Radiobiology for the radiologist. Philadelphia: Lippincott Wil- liams & Wilkins; 2000. 131. Zalutsky MR, Pozzi OR: Radioimmunotherapy with alpha-particle emitting radionuclides. Q J Nucl Med Mol Imaging 2004, 48(4):289-296. 132. Palm S, Humm JL, Rundqvist R, Jacobsson L: Microdosimetry of astatine-211 single-cell irradiation: role of daughter polonium-211 diffu- sion. Med Phys 2004, 31(2):218-225. 133. Palm S, Back T, Claesson I, Delle U, Hultborn R, Jacobsson L, Kopf I, Lindegren S: Effects of the alpha-particle emitter At-211 and low-dose- rate gamma-radiation on the human cell line Colo-205 as studied with a growth assay. Anticancer Res 1998, 18(3A):1671-1676. 134. Palm S, Andersson H, Back T, Claesson I, Delle U, Hultborn R, Jacobsson L, Kopf I, Lindegren S: In vitro effects of free 211At,211At-albumin and 211At-monoclonal antibody compared to external photon irradiation on two human cancer cell lines. Anticancer Res 2000, 20(2A):1005-1012.

62 135. Persson MI, Gedda L, Jensen HJ, Lundqvist H, Malmstrom PU, Tolmachev V: Astatinated trastuzumab, a putative agent for radionuclide immunotherapy of ErbB2-expressing tumours. Oncol Rep 2006, 15(3):673- 680. 136. Wilbur DS, Vessella RL, Stray JE, Goffe DK, Blouke KA, Atcher RW: Preparation and evaluation of para-[211At]astatobenzoyl labeled anti- renal cell carcinoma antibody A6H F(ab')2. In vivo distribution comparison with para-[125I]iodobenzoyl labeled A6H F(ab')2. Nucl Med Biol 1993, 20(8):917-927. 137. Orlova A, Hoglund J, Lubberink M, Lebeda O, Gedda L, Lundqvist H, Tolmachev V, Sundin A: Comparative biodistribution of the radiohalo- genated (Br, I and At) antibody A33. Implications for in vivo dosimetry. Cancer Biother Radiopharm 2002, 17(4):385-396. 138. Garg PK, Harrison CL, Zalutsky MR: Comparative tissue distribution in mice of the alpha-emitter 211At and 131I as labels of a monoclonal antibody and F(ab')2 fragment. Cancer Res 1990, 50(12):3514-3520. 139. Hadley SW, Wilbur DS, Gray MA, Atcher RW: Astatine-211 labeling of an antimelanoma antibody and its Fab fragment using N-succinimidyl p-astatobenzoate: comparisons in vivo with the p-[125I]iodobenzoyl conjugate. Bioconjug Chem 1991, 2(3):171-179. 140. Wilbur DS, Chyan MK, Hamlin DK, Kegley BB, Risler R, Pathare PM, Quinn J, Vessella RL, Foulon C, Zalutsky M et al: Reagents for astatina- tion of biomolecules: comparison of the in vivo distribution and stability of some radioiodinated/astatinated benzamidyl and nido-carboranyl compounds. Bioconjug Chem 2004, 15(1):203-223. 141. Larsen RH, Slade S, Zalutsky MR: Blocking [211At]astatide accumula- tion in normal tissues: preliminary evaluation of seven potential compounds. Nucl Med Biol 1998, 25(4):351-357. 142. Orlova A, Nilsson FY, Wikman M, Widstrom C, Stahl S, Carlsson J, Tol- machev V: Comparative In Vivo Evaluation of Technetium and Iodine Labels on an Anti-HER2 Affibody for Single-Photon Imaging of HER2 Expression in Tumors. J Nucl Med 2006, 47(3):512-519.

63 Acta Universitatis Upsaliensis Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Medicine 127

Editor: The Dean of the Faculty of Medicine

A doctoral dissertation from the Faculty of Medicine, Uppsala University, is usually a summary of a number of papers. A few copies of the complete dissertation are kept at major Swedish research libraries, while the summary alone is distributed internationally through the series Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Medicine. (Prior to January, 2005, the series was published under the title “Comprehensive Summaries of Uppsala Dissertations from the Faculty of Medicine”.)

ACTA UNIVERSITATIS UPSALIENSIS Distribution: publications.uu.se UPPSALA urn:nbn:se:uu:diva-6678 2006