Danish University Colleges Adamanzanes as Bifunctional 64Cu2+ Chelators for Radio-Visualisation and Therapy Chelation and Transchelation Holm-Jørgensen, Jacob Rørdam

Danish University Colleges

Adamanzanes as Bifunctional 64Cu2+ Chelators for Radio-Visualisation and Therapy Chelation and Transchelation

Holm-Jørgensen, Jacob Rørdam

Publication date:

2009

Document Version Peer reviewed version

Link to publication

Citation for pulished version (APA):

Holm-Jørgensen, J. R. (2009). Adamanzanes as Bifunctional 64Cu2+ Chelators for Radio-Visualisation and Therapy: Chelation and Transchelation.

General rights

Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of private study or research.

• You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal

Download policy

If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Download date: 19. Sep. 2022

Adamanzanes as Bifunctional

64 Cu ²⁺ Chelators for Radio- Visualization and Therapy

Chelation and Transchelation

Adamanzanes as Bifunctional ⁶⁴Cu²⁺ Chelators for Radio-Visualization and Therapy

Chelation and Transchelation

Submitted: 15-02-2009 Defended: 17-04-2009

Assessment committee:

Professor Erik Larsen, IGM, KU-LIFE (Chair)

Associate Professor Pauli Kofod, University College, Zealand Physician Tomas Ohlsson, Lund University Hospital, Sweden

© Jacob Rørdam Holm-Jørgensen, 2009

Department of Basic Sciences and Environment Faculty of Life Sciences

University of Copenhagen Frederiksberg, Denmark

Printed by Grafisk Produktion Ribe ISBN: 978-87-993033-1-1

Jacob Rørdam Holm-Jørgensen Ph.D. Dissertation

Adamanzanes as Bifunctional

64 Cu ²⁺ Chelators for Radio- Visualization and Therapy

Chelation and Transchelation

Department of Basic Sciences and Environment Faculty of Life Sciences

University of Copenhagen Frederiksberg

Denmark

2009

Acknowledgements

The work presented in this PhD dissertation has been performed under the research school on Metal Ions in Biological Systems.

The research has been performed mainly at the Bioinorganic Chemistry Group, Department of Basic Sciences and Environment, Faculty of Life Sciences, University of Copenhagen. The ⁶⁴Cu experiments and MS have been performed at the Hevesy Laboratory, Radiation Research Division of Risø National

Laboratory for Sustainable Energy at the Technical University of Denmark.

I would like to thank my supervisors Morten J. Bjerrum and Mikael Jensen for

backing my unorthodox approach on coordination chemistry and for guiding me back when that approach let me too far away.

I would also like to thank Palle Rasmussen for inspiring me on how to analyze my ligand challenge results and Louise Holm for a technical assistance with the Cytochrome c adamanzane couplings.

The Danish Chemical Society is thanked for the travelling grant I received for the 1^st EuCheMS conference in Budapest, Hungary.

I have also received help from most of the Hevesy Laboratory and the Bioinorganic Chemistry group for which I am grateful.

Jacob Rørdam Holm-Jørgensen Frederiksberg, February 2009

Preface

Terminology

Stability

Molecular stability is central to this project and stability can means two things, which often are used as if they were interchangeable: Either thermodynamic stability, which determine the equilibrium point, or kinetic stability, which determine the rate of reaction. Unless otherwise specifically stated, in this thesis “stability” is used for thermodynamic stability and “robustness” is used for kinetic stability.

Ligand abbreviation

The rules for ligand abbreviation are described in section IR-4.4.4 of the IUPAC Red book [2]. According to them ligands should be written in lower case letters (e.g. ida and edta). However, very few authors follow these recommendations for ligand abbreviation and write complexane ligands in all-caps style, common for acronyms (e.g. IDA and EDTA). Therefore, to enable also non-chemists to read and understand the thesis, I have chosen to ignore the IUPAC recommendations and follow the crowd.

Sometimes it is good not to be prepared

The first sample measured several times in a diode array spectrophotometer was measured with increasing time intervals because I had not yet learned to use the kinetics program. If I had, I would have programmed it to measure every minute and would have seen very fast chelation, compared to earlier experiments, but a smooth chelation curve and no indication of why the rate had increased. Instead, I manually pressed

“Sample” each time, while looking at a watch and therefore regularly increased the duration between sampling. The resulting curve could have been interpreted as smooth with a human touch/error, if I had not analysed it by differentiation to show the rate. There it was evident that the rate dropped steeply when the duration between measurements increased. The next experiment not only confirmed the light hypothesis, but also showed that the effect of irradiation was surprisingly high, considering that spectrophotometers are designed to minimize induction of photochemistry, though diode array spectrophotometers should not be used if photochemistry is known to be a problem. Therefore, it is possible that the effect of UV light would not have been discovered, if a monochromatic spectrophotometer, which irradiates the samples one

wavelength a time, had been used instead of a diode array spectrophotometer, which splits up the light beam after it has passed through the sample.

Summary

Among the PET and SPECT imaging isotopes, the isotopes of copper have a special position, both because of the favourable binding characteristics of the metal and because of the variety of tracer half-lives and decay radiations available within the same labelling chemistry. For many molecular targeting strategies, the less- than-two-hour half-lives of the standard PET isotopes (¹¹C, ¹³N, ¹⁵O and ¹⁸F) are suboptimal. Both ⁶¹Cu (T½ = 3.4 h) and ⁶⁴Cu (T½ = 12.7 h) can provide tracers better suited to biodistributions peaking after several hours.

Of special interest is ⁶⁴Cu, as it is suitable for diagnosis as well as therapy.

The almost universal strategy of chelation to vector molecules via bifunctional groups applies easily to copper labelling of peptide and protein vectors. However, the chelation must be swift to exploit the short- lived isotopes. This is impeded by the risk of transchelation to other copper binding molecules. The quantitative overweight of competing copper chelating molecules in vivo is so intense that kinetic stability rather than thermodynamic stability is of importance when trying to avoid transchelation. This has been shown to be a problem even with state-of-the art “stable” chelators such as DOTA and TETA [24].

Therefore, the chelation must be very robust, which prolongs chelation procedures.

Bowl-adamanzanes are known to form Cu(II)-complexes with exceptional kinetic stability due to the

cryptate effect and their rigidity. Heat increases the chelation rate but can also denature the vector, depending on the type and complexity of it. Therefore, the harsh reaction conditions currently employed to form the complex under carrier-free production impede the use of preformed bifunctional chelator-vector aggregate (the precursor) before addition of the tracer.

This project has focused on investigation of the stability and robustness of Cu(II) complexes of two adamanzanes: [2⁴.3¹]adz and [3⁵]adz, and on investigation of the Cu(II) chelation properties of [2⁴.3¹]adz, when it is functionalized and vector bound.

The robustness of [Cu([2⁴.3¹]adz)]²⁺ was found to exceed that of [Cu([3⁵]adz)]²⁺ by at least a factor 10. In competition with equimolar concentrations of complexanes such EDTA, HEDTA and NTA, [Cu([3⁵]adz)]²⁺

had an initial rate of transchelation of 3-7.5 % per day.

The stability constant of [Cu([3⁵]adz)]²⁺ was in this study measured to be high (10¹⁷-10²⁰ M^-1), but the extreme robustness of [Cu([2⁴.3¹]adz)]²⁺ prevented measurement of its stability constant.

As an alternative to high pH and temperature, it is here shown that UV light can induce an excited state of the chelator with a proposed substantially increased acidity. The deprotonated chelator can then fully chelate the copper ion and release the photo-energy as heat. The result is significant increase in the chelation rate of Cu²⁺. This induction can be inhibited by blocking the UV light and by presence of redox interfering

molecules.

Though the method can be employed for chelation of Cu²⁺ to vector bound adamanzanes, its value in production of copper-based tracer molecules is limited by the denaturing effect of UV light on proteins.

Sammenfatning

(Summary in Danish)

Kobberisotoperne har en særlig position blandt de radioisotoper som kan bruges til PET og SPECT. Dette skyldes både kobberioners favorable bindingsegenskaber og de mange forskellige halveringstider og hen- faldstyper som er mulige med samme kemiske egenskaber. I mange tilfælde er det vanskeligt at anvende standard PET-isotoperne (¹¹C, ¹³N, ¹⁵O and ¹⁸F) til at målrette specifikke molekyler pga. deres halveringstid på under 2 timer. Modsat er både ⁶¹Cu (T½ = 3.4 h) og ⁶⁴Cu (T½ = 12.7 h) egnet som sporstoffer for biodistri- butioner som topper efter adskillige timer. Af særlig interesse er ⁶⁴Cu eftersom den isotop kan anvendes til såvel diagnostik som terapi.

Den nærmest universelle strategi med at danne et chelat med en bifunktionel gruppe koblet til et vektormole- kyle, kan let anvendes til kobbermærkning af peptider og proteiner. Det er dog afgørende at chelatet formes hurtigt for at kunne udnytte isotoperne med kort halveringstid, og at det er robust, så det kan modstå konkur- rencen fra andre kobberbindende molekyler in vivo. Dette besværliggøres ved risikoen for transchelering til andre kobberbindende molekyler. Den kvantitative overvægt af konkurrerende molekyler in vivo er så mas- siv at kinetisk stabilitet snarere end termodynamisk stabilitet betyder noget for at undgå transchelering. Dette har vist sig at være et problem selv for chelatorer som DOTA og TETA, der ellers er kendt for at danne me- get ”stabile” kobberchelater [24].

Cu²⁺ and skål-adamanzaner er kendt for at danne komplekser med exceptionel kinetisk stabilitet pga. kryptat- effekten og deres rigiditet, men hårde reaktionsbetingelser bliver brugt for at danne chelatet under carrier-frie betingelser. Disse reaktionsbetingelser besværliggør brug af præformet bifunktionel chelator-vektor-aggregat før sporstoffet tilsættes. Varme øger cheleringshastigheden, men kan også denaturere vektoren afhængig af dens type og kompleksitet.

Dette projekt har fokuseret på at undersøge stabiliteten of robustheden af to adamanzaners ([2⁴.3¹]adz og [3⁵]adz) Cu(II)komplekser og på at udforske Cu(II)cheleringsegenskaberne af [2⁴.3¹]adz, når den er funk- tionaliseret og bundet til en vektor.

[Cu([2⁴.3¹]adz)]²⁺ blev målt til at være mindst 10 fold så robust som [Cu([3⁵]adz)]²⁺. I konkurrence med til- svarende koncentrationer af komplexanerne EDTA, HEDTA og NTA havde [Cu([3⁵]adz)]²⁺ en initial trans- cheleringshastighed på 3-7,5 % per dag.

Stabilitetskonstanten for [Cu([3⁵]adz)]²⁺ blev her målt til at være høj (10¹⁷-10²⁰ M^-1), men den ekstreme robusthed af [Cu([2⁴.3¹]adz)]²⁺ forhindrede måling af dens stabilitetskonstant.

Som et alternativ til høj pH og temperatur bliver det her vist at UV-lys kan inducere en exciteret tilstand af chelatoren antageligt med en betydelig højere surhed. Den deprotonerede chelator kan derefter danne et komplet chelat med kobberionen og afgive fotoenergien som varme. Resultatet er en signifikant forøgelse af Cu²⁺-cheleringshastigheden. Denne induktion kan hæmmes ved at blokere UV-lyset og af molekyler som griber ind i redoxprocessen.

Selvom metoden kan anvendes til chelering af Cu²⁺ til vektorbundne adamanzaner, er dens værdi i produk- tionen af kobberbaserede sporstofmolekyler begrænset af den denaturerende effekt af UV-lys på proteiner.

Abbreviations

Abbreviation Full name Defined at page #

[2⁴.3¹]adz 1,4,7,10-Tetraazabicyclo[5.5.3]pentadecane 34

[3⁵]adz 1,5,9,13-Tetraazabicyclo[7.7.3]nonadecane 34

ABM Antigen-binding molecule 16

APD Avalanche photodiodes 2

ATSM diacetyl-bis(N⁴-methylthiosemicarbazone) 12

BFC Bifunctional chelator 5

BLG β-lactoglobulin

81

CAT Computed axial tomography (now CT) 3

CBTE2A 1,4,8,11-Tetraazabicyclo[6.6.2]hexadecane-4,11-diacetic acid 18

Cc Cytochrome c 47

CT Computed tomography 3

D Absorbed dose 14

DMSA Dimercaptosuccinic acid 25

DOPA 3,4-dihydroxyphenylalanine 11

DOTA 1,4,7,10-tetraazacyclododecane-N,N’,N’’,N’’’-tretraacetic acid 11

DTPA Diethylenediaminetetraacetic acid 24

EDC 1-Ethyl-3-(3-dimethylaminopropyl)-carbodiimide hydrochloride 47

EDTA Ethylenediamine tetraacetic acid 14

ESI

Electrospray ionization

44

Fab Fragment antigen-binding 15

FDG ¹⁸F-2-fluoro-2-deoxyglucose 3

FLT 3'-deoxy-3'-¹⁸F-fluorothymidine 11

FPLC Fast protein liquid chromatography 77

GRPR Gastrin-releasing peptide receptor 17

HEDTA

N-(hydroxyethyl)-ethylenediaminetriacetic acid 46

IDA Iminodiacetic acid 86

Kf Formation constant 22

LE Labelling efficiency 8

LMCT Ligand to metal charge transfer (also known as LMET) 20

LMET Ligand to metal electron transfer 20

mAb Monoclonal antibody 15

MAG3 R-[N-mercaptoacetylglycylglycyl)-γ-amino-butyrate 27

MRI Magnetic resonance imaging 3

MRT Magnetic resonance tomography (now MRI) 3

MS Mass spectrometry 44

MTD Maximum tolerated dose 14

NHS N-Hydroxysuccinimide 47

NLS Nucleus seeking sequence 17

NOC NaI³-octretide 17

NTA

Nitrilotriacetic acid 46

PBS Phosphate buffered Saline 46

PET Positron emission tomography 2

PRRT Peptide receptor radionuclide therapy 5

PTSM pyruvaldehyde-bis(N⁴-methy1thiosemicarbazone) 12

RIT Radioimmunotherapy 5

scFv Single Chain Variable Fragment 15

SA Specific activity 8

SOD Superoxide dismutase 9

SPECT Single photon emission computed tomography (also SPET) 3

SSTR Somatostatin receptor 17

TATE [D-Phe¹-Tyr³]-octreotate 17

TETA 1,4,8,11-tetraazacyclotetradecane-N,N',N'',N'''-tetraaceticacid 13

TOC [D-Phe¹-Tyr³]-octreotide 11

TOF Time of flight 2

TRT Targeted radionuclide therapy 5

TSH Thyroid-stimulating hormone (also known as thyrotropin) 5

Table of contents

Acknowledgements... I Preface ... II Summary ... III Sammenfatning (Summary in Danish) ... IV Abbreviations ... V Table of contents ... VII

1 Introduction ... 1

1.1 Visualization techniques ... 1

1.1.1 Positron emission ... 1

1.1.2 Positron emission tomography (PET) ... 2

1.1.3 Other visualization techniques ... 3

1.1.4 Combined visualization techniques ... 4

1.2 Radionuclide therapy ... 4

1.2.1 Non targeted ... 4

1.2.2 Targeted ... 5

1.3 Radionuclides for PET and TRT ... 5

1.3.1 Properties of radionuclides ... 5

1.3.2 Radionuclides for direct labelling ... 9

1.3.3 Radionuclides for indirect labelling ... 11

1.3.4 Rarely used isotopes ... 14

1.4 Dose limitation and radiosensitive organs ... 14

1.5 Immuno-PET and RIT ... 14

1.5.1 Antibodies ... 15

1.5.2 Other antigen binding molecules ... 15

1.5.3 Internalization and degradation... 16

1.5.4 Shedding of antigens ... 16

1.6 Peptides for PET and PRRT... 17

1.7 Pre-treatment enhancement ... 18

1.7.1 Streptavidin/avidin-biotin ... 19

1.7.2 Bi-specific mAbs ... 19

1.7.3 Other pre-treatment enhancements ... 19

1.8 Chemical radiosensitizers ... 19

1.9 Copper chelation ... 20

1.9.1 Coordination chemistry of copper ... 20

1.9.2 Chelation techniques ... 21

1.10 Kinetic stability (robustness) of copper chelates ... 21

1.10.1 Kinetic properties of metal and chelator ... 21

1.10.2 Effects of metal reduction on kinetic stability ... 22

1.10.3 Measurements of kinetic stability ... 24

1.11 Copper chelating molecules ... 25

1.11.1 Ideal BFC’s ... 25

1.11.2 Non BFC’s ... 25

1.11.3 Acyclic BFC’s ... 26

1.11.4 Monocyclic BFC’s ... 27

1.11.5 Bicyclic BFC’s ... 31

1.12 Adamanzanes ... 33

1.12.1 Definition and nomenclature... 33

1.12.2 Coordination chemistry of adamanzanes ... 35

1.13 Copper adamanzane complexes ... 36

1.13.1 Coordination of Cu²⁺ to adamanzanes ... 36

1.13.2 Crystal structures of copper adamanzane complexes ... 38

1.13.3 Stability of copper adamanzane complexes ... 39

2 Aim of Project ... 41

3 Materials and Methods ... 42

3.1 Experimental design ... 42

3.1.1 Copper isotopes ... 42

3.1.2 Irradiation setup ... 42

3.2 Synthesis of adamanzanes ... 44

3.2.1 Caution ... 44

3.2.2 Ligand synthesis ... 44

3.2.3 Functionalization of [2⁴.3¹]adz ... 44

3.3 Chelation of stable copper, by heat and by light ... 44

3.3.1 Synthesis of [Cu(adz)]²⁺ in mg scale ... 44

3.3.2 Real time measurement of Cu²⁺ chelation by adamanzanes ... 45

3.4 Kinetic stability ... 46

3.4.1 Pilot studies ... 46

3.4.2 5^th ligand stability ... 46

3.4.3 Transchelation to competing chelators ... 47

3.5 Binding of 1b to Cytochrome c. and to β-lactoglobulin ... 47

3.5.1 Cytochrome c. ... 47

3.5.2 β-lactoglobulin ... 48

3.6 Radiochemistry ... 48

3.7 Mass spectrometry ... 49

4 Results ... 50

4.1 UV-visible spectra and pH ... 50

4.1.1 pH changes the spectrum of [Cu( [2⁴.3¹]adz)X]ⁿ⁺ in the UV range ... 50

4.1.2 pH changes the spectrum of Cu²⁺ + [2⁴.3¹]adz ... 50

4.2 Stability of [Cu([2

⁴

.3

¹

]adz)]

²⁺

and [Cu([3

⁵

]adz)]

²⁺

... 51

4.2.1 Expected equilibria and equations: ... 51

4.2.2 5^th ligand effect ... 52

4.2.3 Adjusted and additional equilibria and equations: ... 53

4.2.4 [Cu([3⁵]adz)]²⁺ + EDTA ... 54

4.2.5 [Cu([3⁵]adz)]²⁺ + HEDTA / NTA ... 54

4.2.6 [Cu([2⁴.3¹]adz)]²⁺ + HEDTA / NTA ... 58

4.2.7 Transchelation rate comparison ... 61

4.3 Light induces chelation of copper ... 61

4.3.1 First evidence of light induced chelation ... 62

4.3.2 A comparison between coordination of copper to [3⁵]adz and to [2⁴.3¹]adz ... 63

4.3.3 Measurement of light induction ... 64

4.3.4 Blocking UV-light with λ < 220 nm prevents the secondary process... 66

4.3.5 Only UV light induces chelation of Cu²⁺ to [2⁴.3¹]adz ... 66

4.4 Copper binding molecules such as anions alter the chelation rate ... 68

4.4.1 UV light absorbing molecules decrease the effect of UV-irradiation ... 68

4.4.2 Some anions can chemically affect the chelation rate ... 69

4.4.3 Heat enhances light induced chelation and induces light independent chelation ... 71

4.4.4 Dissolved O2 ... 71

4.5 N,N’-(CH

2

COOH)

2

[2

⁴

.3

¹

]adz becomes partly decarboxylated by UV-irradiation ... 73

4.6 N,N’-(CH

₂

CO

₂

CH

₂

CH

₃

)

₂

[2

⁴

.3

¹

]adz is almost resistant to UV-irradiation ... 76

4.7 N,N’-(CH

2

COOH)

2

[2

⁴

.3

¹

]adz can be coupled to Cytochrome c. and purified ... 77

4.8 Cytochrome c is reduced by UV light ... 79

4.9 Radiochemistry ... 80

4.9.1 N-CH2COOH[2⁴.3¹]adz-Cc binds ⁶⁴Cu²⁺ via the adamanzane when irradiated ... 80

4.9.2 Irradiated β-Lactoglobulin binds ⁶⁴Cu²⁺ independent of the adamanzane ... 81

4.9.3 Cytochrome c also binds ⁶⁴Cu²⁺ if the UV photon flux is too high ... 82

5 Discussion ... 85

5.1 UV spectra of copper adamanzane complexes ... 85

5.2 Stability and robustness of copper adamanzane complexes ... 86

5.2.1 5^th ligand effects of added anions ... 86

5.2.2 5^th ligand effects due to pH change ... 86

5.2.3 Evaporation ... 87

5.2.4 Robustness and stability constant ... 87

5.3 Ammonia, TRIS and PBS ... 88

5.4 Light induced chelation of copper ... 89

5.5 Proposed reaction scheme for chelation of Cu

²⁺

by [2

⁴

.3

¹

]adz ... 89

5.6 The effects of bromide (and other dissolved molecules) ... 92

5.6.1 UV filter effect (physical inhibition of light induction) ... 92

5.6.2 Competitive coordination... 93

5.6.3 Redox interference by acido ligands (chemical inhibition of light induction) ... 94

5.6.4 Redox interference by oxygen ... 95

5.6.5 pH stabilization ... 95

5.7 Effect of side chains ... 95

5.7.1 Chelation and redox chemistry ... 95

5.7.2 Linkage to vector ... 97

5.8 Radiochemistry ... 97

5.8.1 In darkness ... 97

5.8.2 Low flux ... 97

5.8.3 High flux ... 98

6 Conclusions ... 100

6.1 Discoveries ... 100

6.1.1 5^th ligand effect on stability and robustness of chelates from other solutes ... 100

6.1.2 Robustness of [Cu([3⁵]adz)]²⁺ and [Cu([2⁴.3¹]adz)]²⁺ ... 100

6.1.3 Estimation of the stability constant of [Cu([3⁵]adz)]²⁺ ... 100

6.1.4 UV light catalysis of chelation ... 100

6.1.5 Redox interference ... 100

6.2 Hypotheses ... 100

6.3 Goals ... 101

6.3.1 Short term ... 101

6.3.2 Long term ... 101

7 Perspectives... 102

8 References ... 105

A Equations ... 118

B Mass spectra ... 120

C Supplementing figures to figure 4.17... 130

1 Introduction

1.1 Visualization techniques

1.1.1 Positron emission

Some radioisotopes emit positrons, which are the positively charged equivalents to electrons, having the same physical properties except for the charge. Emitted positrons are also called β⁺ particles until they loose their kinetic energy. Though the total energy of a β⁺ emission is characteristic for the radioactive isotope, the energy of the β⁺ particle is not discrete as the released energy is divided between the β⁺ particle and a neutrino νe. Therefore, measurements of the kinetic energy of β⁺ particles result in spectra with a characteristic maximum Emax and average Ē. When a positron has lost most of its kinetic energy by interaction with atoms, it will merge with an electron resulting in an annihilation of both. Their masses become conversed of into electromagnetic energy in form of two photons, or rarely three or more [83].

Figure 1.1: Drawing of the principal parts of a PET scanner (from Wikipedia). When the positron and an electron annihilates (at the green dot), and the resulting γ-rays leaves the patient in a perpendicular direction, they can be detected by the ring, which sends enhanced signals to the processing unit. By the use of algorithms an image can be constructed, which often is overlaid with another image showing tissue structure with high spatial resolution.

The combined energy of the photons is always the same as the mass energy of the electron and positron:

E = mc² = 2 · 9.1091·10^-31 kg · (2.9979·10⁸ m/s)² = 163734·10^-13 J = 1022062 eV ≈ 1.022 MeV

If the electron and positron collide with equal and opposite kinetic energy, the two photons will be emitted 180° apart, but as this is not always the case, there is often a slight angular deviation (non-collinearity) resulting in a standard deviation of 0.25° to conserve momentum [39].

1.1.2 Positron emission tomography (PET)

The annihilation radiation can be detected in a PET scanner (figure 1.1). Only near (within nanoseconds) coincidental photons of 511 keV are recorded and assumed to origin from the same annihilation, placing it between the two recording detectors on the so-called line of response. The width of the line depends on the diameter of the detector ring because of the angular deviation, thus making a small ring preferable for high resolution. However, patient size limits how small it can be. The detector ring is made of scintillator material, converting the γ-photons into visible light, and photomultiplier tubes or avalanche photodiodes (APDs) enhancing and recording the light [133]. Unfortunately, a major part of the annihilation radiation is spread because of the Compton Effect and only a small part of the

remainder

actually hits the detector. Dead time and patient dose limit the speed of imaging. The data from the scanner is then converted to a 3D image (tomography) by a computer. The algorithm used for the calculations has great influence on noise reduction and image resolution [113,185]. As the detectors are not infinitely small, the tracer distribution is sampled on a voxel grid. The result is similar to pixilation on a monitor; a voxel has one value. This and the limitation in resolution both contribute to the partial volume effect which have several adverse effects on quantification.

Primarily it will make a small tumour look larger but less intense [202]. Present scanners have a spatial resolution of 4-6 mm entailing a 7 mm detection limit of tumours [148]. In contrast to other in vivo methods for measuring radioactivity, PET measures Bq/voxel and not a relative value [162].

Time of flight (TOF) scanners, where the small difference between the detection times of the two photons is used to calculate the point of annihilation, is being developed [133]. This would potentially reduce the noise and enable improved image quality or reduction of the radioactive dose administered to the patient. TOF is dependent on fast detectors and the most recently developed scintillators are made of Lu2SiO5, which has a high maximal count rate due to a relatively high light output and short decay time [162]. Another possibility is resistive plate chamber technology that can measure the incoming photons in three dimensions [17].

In animal experiments, “patient” size is not necessarily a restriction as mice are the vast majority of the mammals used in biomedical research. This enables use of micro-PET-CT scanners with a CT (Computed Tomography) resolution of 300 µm and a PET resolution of 1 mm [125]. In micro-PET non-collinearity of annihilation photons contributes less to blurring, but the positron range, on the other hand, contributes more, compared to standard PET [18]. This is because the positron range depends on the energy of the positron and thus is the same in humans and mice. As anatomical distances are smaller in mice, isotopes emitting low energetic positrons are preferable.

Since the annihilation radiation is a result of a positron emission, a PET scanning is visualising the location of the positron emitter. The positron emitter can be coupled to or be a part of a molecule, which location or site of metabolism is of importance. No accurate anatomical information is acquired [216].

The by far most widely used tracer is ¹⁸F-2-fluoro-2-deoxyglucose (FDG) [183]. FDG is a glucose analogue and is distributed to and taken up by the cells like glucose. Intracellular FDG is phosphorylated by

hexokinase preventing it from leaving the cell again (except in liver cells) and, in contrast to glucose, FDG cannot be converted to fructose 6-phosphate by phosphoglucose isomerase. The result is an accumulation of FDG in glucose consuming cells; in a resting patient especially the brain and, in case of a pathological process, neoplasm’s and inflamed areas. Furthermore, a number of normal tissues can have an apparently increased FDG uptake [183].

1.1.3 Other visualization techniques

Single photon emission computed tomography (SPECT) is comparable with PET in the sense that they both depend on emission from tracer molecules, but where PET uses the combined detection of two photons to calculate the site of decay, this is not possible in SPECT. Instead, collimators are used to filter off γ-rays not perpendicular to the detector to enable an acceptable resolution of the images. Unfortunately, this method removes the majority of the radiation, thus increasing the dose required for imaging compared to PET.

Alternatively, a pinhole can used enabling high resolution but only of small areas [14].

Computed Tomography (CT), formerly known as computed axial tomography (CAT) is a designation for a 3D image computed from a series of 2D X-ray images. The X-rays images are taken around an axis of rotation and during the scanning; the patient or object is moved along that axis. CT has high spatial

resolution, and contrast agents can be applied for additional information. However, a CT scanning exposes the patients to a moderate to high dose of ionising radiation, and concern has been raised about the increased exposure of the population [109]. In this context, EU has prohibited radiographic screening except for mammography [1]. Today most new CT scanners are combined PET-CT scanners and combining PET and CT gives a huge advantage.

Magnetic resonance imaging (MRI), formerly referred to as magnetic resonance tomography (MRT), is an imaging method that exploits the nuclear spin of hydrogen in an electric field. Contrast agents, often based on gadolinium, can be applied to obtain images visualising metabolic or pathologic processes [7,8], and functional MRI can visualize oxygen consumption since oxyhemoglobin and deoxyhemoglobin has different magnetic properties [9]. However, MR scanners are highly expensive and the extreme magnetic field makes them dangerous for people containing ferro- or paramagnetic components. Especially pacemakers and implantable cadioverter defibrillators increase the number of MRI restricted patients [156].

1.1.4 Combined visualization techniques

PET gives metabolic information but at a low spatial resolution. CT gives high spatial resolution but no metabolic information. Combining the images gives detailed information of where a metabolic process takes place. A few years ago PET and CT scans were performed separately and the images aligned afterwards, but accurate alignment is only achievable for brain images due to differences in patient positioning [216]. In contemporary scanners the patient is scanned in the same position making image alignment considerably easier even for head and neck tumours [68], although image processing by algorithms is still necessary [113].

Moreover, the CT scan can be used for attenuation correction improving the quality of the PET images and reducing scan time by up to 40 % [216]. Nonetheless, MRI is still superior to PET-CT in detecting bone metastases [188], and the effective dose for an ¹⁸F-FDGPET-CT scanning is up to 25 mSv, of which ¹⁸F- FDG is only responsible for 4 mSv [187,193].

Another promising combination is PET-MRI as it potentially can produce images with even higher spatial resolution than PET-CT, and with the exposure coming from the PET tracer only [193]. Earlier scanning times of more than an hour for both MRI and PET prevented a dual scan. A contemporaneously PET and MRI scanning was impossible because the photomultiplier tubes used in PET scanners are ineffective in a strong magnetic field. However, new flat panel positron-sensitive photomultiplier tubes are less sensitive to magnetic fields, and in combination with long optic fibre light guides, simultaneous imaging of rats is possible [175].

1.2 Radionuclide therapy

1.2.1 Non targeted

Radionuclide therapy is treatment of pathological conditions such as cancer or arthritis by unsealed sources of radionuclides [235]. In the most common radiotherapy, external beam radiation, a beam from a γ-source is directed against the target tissue. This, of cause, will also affect the healthy tissue that the beam passes on its way through the body, and will only hit tumour tissue when the radiologists know the precise location.

Another non-targeted radiotherapy is brachytherapy, where the radioactive material is placed against the tumour. This method is also both common and requires knowledge of the precise location of the tumour [219,235]. Radioembolization is normally not considered a brachytherapy, though it does live up to the Wikipedia definition by being “a form of radiotherapy where a radioactive source is placed inside or next to the area requiring treatment”. Radioembolization is effective in treatment of unresectable liver cancers, because they are mainly perfused by the artery, in contrast to healthy liver tissue, which receives most of its blood from the portal vein [219]. Radiosynoviorthesis is another brachytherapy-like treatment where colloids containing β-emitting radionuclides are injected into the articular cavity. There, they are phagocytized by the synovial lining cells and the surrounding synovial tissue is irradiated [110].

1.2.2 Targeted

Targeted radionuclide therapy (TRT) is also known as endoradiotherapy. It differs from the other types of radiotherapy in that the radionuclide is administered systemically and either by itself or carried by a vector molecule concentrates in the target tissue. The vector is typically an antibody or a derivative, which makes the treatment radioimmunotherapy (RIT) [220], or it is a peptide, often a hormone analogue, and then the treatment is peptide receptor radionuclide therapy (PRRT) [65]. Other forms of vector molecules can be inorganic complexes, steroids, nucleoside or magnetic nanoparticles [40,149,235]. When ¹³¹I is used for ablation of thyroid tissue, no vector is required as thyroid tissue has active uptake of iodine, but recombinant thyroid-stimulating hormone (TSH) can be used for optimising treatment [10].

In TRT it is important to decide the number of radionuclides per vector molecule. If the vector molecules are very small and addition of another chelate would have adverse effects on the distribution in the body, then only one radionuclide per vector molecule is possible. On the other hand, if there are few receptors on the surface of the target cells and they thus are easily saturated, then it might be necessary to bind more than one radionuclide per vector molecule. Large vector molecules such as antibodies can easier tolerate more

chelators per vector molecule but if the number becomes too high, the immunoreactivity of the labelled antibody decreases [30]. Instead of conjugating chelators at multiple sites on the vector molecule, it is possible to increase the number of radionuclides per vector by attaching a polymer of chelating molecules [70].

1.3 Radionuclides for PET and TRT

1.3.1 Properties of radionuclides

Radionuclides can be divided into two groups depending on the labelling technique. Direct labelling by substituting a stable atom with a positron emitting one, and indirect labelling, which require a bifunctional chelator (BFC) to bind the radionuclide to the molecule of interest (figure 1.2).

Figure 1.2: Direct and indirect labelling of vector molecules with tracers. DOTA-TOC is shown in figure 1.3.

For all PET nuclides, it is crucial that the daughter nuclide is stable. Otherwise, the result of the decay would be unbounded radionuclides in the body as the decay usually breaks the chemical bond holding the parent

Tracer

Vector molecule

Tracer Chelator

Linker

Direct labelling

E.g. ¹⁸FDG

Indirect labelling via bifunctional chelator

E.g. DOTA-TOC Vector molecule

nuclide in place. This would not only disturb imaging but also increase the dose to the organs where the daughter nuclide accumulates, an increase not dependent on the parent compound. For TRT, unstable

daughter nuclides can be acceptable, if the half-life is short enough and the emission type suitable for therapy [153].

It is important for both PET and TRT to choose the right radionuclide. There are 5 characteristics of the radionuclides, which have to be weighed against each other:

• Emission type

• Half life

• Ease of production

• Chemistry (how to bind, and does it stay?)

• Biological behaviour (where do free radionuclides go?)

Emission type

Radionuclides for PET has to be positron emitting and other emission types are potentially harmful. For TRT the choice of emission type is somewhat more complicated. The purpose of TRT is to deliver a lethal dose of radiation to the pathogenic cells while minimising the dose delivered to healthy tissue. With this in mind, one could think that emissions with a short range and a high linear energy transfer (LET) would be optimal in all cases. However, neoplastic tissue is by nature heterogeneous and huge variations in receptor density can be expected both within a tumour and among tumours in the same patient [235]. If the radiation has a longer range than a cell radius, then most of the dose a cell receives will come from radionuclide on/in surrounding cells. That is the crossfire effect. The more heterogeneous the tumour is, the more important is the crossfire effect to ensure irradiation of all parts of the tumour.

The crossfire effect results in a lower dose at the edge of the tumour than in the centre. To keep the radiation within the tumour, the range of the radiation must fit the tumour size. Long range β-emitters such as ⁹⁰Y, are ideal for large tumours such as 500 g hepatomas (liver tumour), but in a 1 cm diameter tumour, less than 10

% of the β-particle energy from ⁹⁰Y is kept within the tumour. If a short-range β-emitter such as ¹³¹I is used, more than 50 % of the β-particle energy is absorbed within a 1 cm tumour. In smaller tumours, the

differences become even larger [100]. On the other hand, the larger the “cold” areas with low receptor density are, the lower is the effect of short-range emissions.

Long-range β-particles have low LET, which results in few ionizations per distance the β-particle travels.

This can be compensated by increasing the number of hits from the crossfire effect but a higher total dose is required to kill a cell since the relative biological effectiveness increases with LET (until a maximum at more than 100 keV/µm). Moreover, the effect of low LET radiation is dependent on the oxygen tension of the tissue [218]. This oxygen dependence is a result of spatial rather than temporal effects, so in a hypoxic tumour an increase in rate is less effective [89]. This does not mean that the rate is unimportant, as low doses

of low LET radiation can induce radioresistance in bystanding cells [102]. On the other hand, high LET radiation generates toxins potent enough to kill bystanding cells [26].

Some heterogeneity in the distribution of the radionuclides within a tumour during treatment is dependent on the vector molecule. If the vector binds rapidly and robustly to binding sites on/in the tumour, it will

concentrate in the periphery of the tumour. As low affinity for tumour binding sites would result in poor tumour to healthy tissue ratio, the optimal vector has a high affinity but slow association rate to the binding sites [69].

Alpha-particles have a very high LET. In a comparison of mean LET, the α-emitter ²¹¹At has about 97 keV/µm, where ⁹⁰Y only has 0.22 keV/µm [235]. The high LET and short range (a few cell diameters ~ less than 50 nm) make α-particles suitable for treatment of micro tumours, single cell layer tumours or loose cancer cells (e.g. leukaemia). The most prevalent α-emitting isotopes in TRT are ²¹¹At, ²¹²Bi and ²¹³Bi. They all have short half-life demanding fast synthesis, delivery and biodistribution of the compound [46].

In between the α-particle emitters and long range β-particle emitters, there are the medium and short range β- particle emitters. A rising star is ¹⁷⁷Lu, which has shown encouraging results with for instance somatostatin analogues [121].

Auger and Coster-Kronig electrons have the shortest range of all particles used for TRT. They are created as a result of either electron capture or internal conversion decay. Both these decay types lead to an electron vacancy in the inner shell and the reorganization that follows can transfer the excess in binding energy to emission of electrons from the atom. If an electron is emitted from an inner shell, it is an Auger electron and a Coster-Kronig electron if it comes from an outer shell but the common name for all of them is Auger electrons. Typically, a cascade of 5 – 30 Auger electrons is emitted per decay [31].

These low energy electrons have subcellular range and thus only affect the target cell and its immediate neighbours. Moreover, they exert a high LET effect but only if located in the cell nucleus [31]. This makes them less cytotoxic to non-target cells as free radionuclides or digested but still chelated nuclides are unlikely to enter to nucleus of other cells after eventually having diffused out of the tumour. On the other hand, it increases the requirements of the vector molecules as it has to both get the radionuclides inside the target nuclei in time and keep them there until they have decayed. Examples of Auger electron emitting isotopes can be found in table 1.2 and table 1.3.

Half-life

If the half-life is too short, the radionuclide might not have time to concentrate in the tumour and if it is too long, the radionuclide might leave the tumour again before decay. For PET, only the tumour concentration is important but for TRT the tumour cell concentration is also important. In case of Auger electron emitters the therapeutic time window is even smaller as they not only have to be located inside the tumour but inside the cell nuclei of the tumour. The therapeutic window of a cancer cell depends on its division rate. Each division halves the radionuclide concentration in the cells and dividing cells are more vulnerable than non-dividing cells. It takes a lower dose to make a cell stop dividing than to kill it, so a low dose will actually make the

cells harder to kill later with an increased dose. There are two strategies to make the radionuclide half-life and the therapeutic time window fit: Either to choose an isotope with a suitable half-life, or to change the therapeutic time window. The later can be done by chemical modifications of the vector molecule (e.g., making it more or less lipophilic, or changing its size) or by changing the behaviour of the tumour (e.g.: TSH stimulation of thyroid tumours). Large malign tumours have limited vascularisation and increased interstitial pressure since they do not have a functional lymphatic drainage system [161,218]. This results in a lower diffusion rate of especially large vector molecules such as antibodies. Therefore, successful treatment of a tumour by itself changes the pharmacokinetics as the tumour becomes smaller.

Ease of production and cost

It is an advantage for research if the radionuclide can be produced locally and if it is affordable. ⁶⁴Cu and

67Cu have shown near identical therapeutic properties both in vitro and in vivo, even under conditions where the effect of the Auger electrons of ⁶⁴Cu was limited. However, where ⁶⁴Cu is relatively inexpensive and can easily be produced on a small medical cyclotron, the high energies required for ⁶⁷Cu production increase its cost and restricts its availability [44].Only a few accelerators able to produce ⁶⁷Cu exist. Therefore, ⁶⁴Cu will often be preferred unless special circumstances points in the direction of ⁶⁷Cu. Besides the energy level required for production, the prize of target material and purification process affects the cost of the radionuclide. The use of some radionuclides is limited by the specific activity (GBq/mmol) they can be produced with, while others can be produced carrier-free [235]. If however, a radionuclide has shown excellent properties in therapy, the prize and availability are not limiting for the use of the nuclide.

Chemistry

If a vector molecule is required, the chemical properties of the radionuclide determine how it can be bound.

Halogens can be bound directly to organic molecules by substitution, while metals require a BFC. For research a range of BFC’s can be tested for optimal binding of the radiometal but for clinical purposes a radiometal must be chosen that can be used with the available pre-tested BFC’s such as DOTA (see page 11).

Some metal ions including Cu²⁺ favour nitrogen donors, while others such as Y³⁺, favour oxygen donors [151]. Nonetheless, DOTA, which binds through nitrogen donors, is the most employed ⁹⁰Y³⁺ chelator.

The labelling efficiency (LE) must be high. Otherwise, a high surplus of ligand is required to ensure that the concentration of free radiometal is insignificant, and that results in a low specific activity (SA = Bq/mole ligand). A low specific activity is a problem if the concentration of target receptors is limiting for the

therapeutic effect, and if binding to the receptors causes a biological response great enough to be the limiting factor on dose size. Moreover, the radiometal must be able to bind robustly to the BFC to avoid dissociation before decay [198]. A high LE does not guarantee in vivo stability, as several comparisons have shown no correlation between LE and in vivo stability [37,151,236]. In fact, a high LE is characteristic of a labile binding as it makes it easier to form a complex under mild conditions [236].

Biological behaviour

No chelating molecule has an infinite formation constant (Kf), and even if that existed, the

radiopharmaceutical could be degraded in vivo, thereby releasing the radionuclide either in free form or by separating the chelator from the vector. Consequently, it is important to consider where free and chelated radionuclides go if released from the vector. Some radionuclides such as ⁹⁰Y³⁺, ¹⁵³Sm³⁺ and ¹⁷⁷Lu³⁺ are known to be ‘bone seekers’ [137]. Radiopharmaceuticals with these nuclides have increased risk of having the bone marrow as the dose-limiting organ. I^- will normally accumulate in the thyroid but this can be prevented by co-administration of an excess of cold NaI.

Natural copper is a part of a range of enzymes and is the third most abundant trace metal in the body [20]. As free copper ions are capable of catalysing the formation of highly reactive hydroxyl radicals in a redox cycle between Cu(I) and Cu(II) [21,82] the free copper ion concentration is maintained at low levels in vivo [144].

On the other hand, as copper is essential for survival for all living cells, the total concentration is not insignificantly low, but kept tightly regulated by a large number of chelating molecules: transporters, chaperones and enzymes [169]. Free copper ions will be distributed on all these copper binding molecules trapping them in the cell, but often transchelated copper (copper which has gone from been part of one chelate to another) accumulates in the liver in superoxide dismutase (SOD) [13,51]. This indicates that transchelation takes place outside the cells or in the liver, as only cells in the liver and the intestine tissue transport copper out of the cells [131]. If the copper is located in the nucleus at the time of transchelation, there is a chance that it will bind to the DNA. Malignant cells have a higher tendency to bind divalent metal ions such as Cu²⁺ to their DNA, and furthermore, if ⁶⁴Cu is bound, it is more lethal to malignant cells than to normal cells [6].

The rest of the radionuclides used for TRT are not known to be characterized by a tendency to concentrate in a specific organ.

1.3.2 Radionuclides for direct labelling

Direct labelling is an advantage when dealing with small molecules having a high metabolic turnover (figure 1.2). A bond to a bifunctional chelator containing a positron emitter would change the chemical properties of the labelled compound and the fast distribution of most small molecules permits the use of an isotope with a short half-life [162]. As examples: ¹³N can as ¹³N2 be used to measure regional perfusion and aeration of the lungs as N2 will diffuse into aerated alveolar units on first pass and pulmonary blood flow, and lung water can be accurately measured with H2

15O [155,177]. Many organic compounds are available labelled with ¹¹C.

For example: ¹¹C labelled amine precursors and acetate for tumour imaging [64,150]. A disadvantage is that the half-lives are so short that the radionuclides have to be produced in house on a cyclotron, and even that will not allow time for a more complex chemical synthesis afterwards [162]. In the other end, ¹²⁴I has a half- life of 4.18 days allowing plenty of time for synthesis but it also leads to an increased dose to the patient.

Furthermore, if the compound is metabolized, the ¹²⁴I could end up in the thyroid, unless cold NaI is

administered [166]. If the loss is insignificant, a ¹²⁴I labelled compound can be used to show the distribution

and effect of the same compound used for ¹³¹I-radiotherapy [220]. If the target organ is the thyroid or metastases from a thyroid carcinoma, free iodine in form of ¹²⁴I-NaI can be used and loss of iodine from labelled compounds is not a concern [67,195].

Table 1.1: PET radionuclide decay properties for isotopes with half-lives > 2 minutes [141,164,174,228]

The average positron range is calculated according to the continuous slowing-down approximation (CSDA), i.e. the range is considered to be the distance along the path travelled by an ‘average electron’ until the electron stops.

Z Radio- nuclide

Half life

Percentage β⁺ decay (%)

Other emissions

Maximum β⁺ energy (MeV)

Average posi- tron range in H2O (mm)

6 ¹¹C 20.334 min 99.8 X; eau 0.9602 1.22

7 ¹³N 9.965 min 99.8 X; eau 1.1985 1.73

8 ¹⁵O 122.24 s 99.9 X; eau 1.732 2.96

9 ¹⁸F 1.8291 h 96.7 X; eau 0.6335 0.64

19 ³⁸K 7.636 min 99.8 γ; X; eau 4.8911 5.52

22 ⁴⁵Ti 184.8 min 84.83 γ; X; eau 1.0280 1.47

26 ⁵²Fe 8.275 h 55.49 γ; X; eau 0.804 1.01

27 ⁵⁵Co 17.53 h 76 γ; X; eau 1.498 2.11

29 ⁶⁰Cu 23.7 min 93 γ; X; eau 3.772 4.20

61Cu 3.333 h 61 γ; X; eau 1.2152 1.77

62Cu 9.67 min 97.43 γ; X; eau 2.926 6.06

64Cu 12.7 h 17.40 γ; X; eau; β^- 0.6531 0.75

31 ⁶⁶Ga 9.49 h 56.0 γ; X; eau 4.153 8.37

68Ga 67.71 min 89.14 γ; X; eau 1.8991 3.46

33 ⁷²As 26.0 h 87.8 γ; X; eau 3.334 5.28

74As 17.77 d 29 γ; X; eau; β^- 1.5405 1.47

35 ⁷⁶Br 16.2 h 55 γ; X; eau 3.941 5.33

77Br 57.036 h 0.74 γ; X; eau 0.343 0.29

37 ⁸¹Rb 4.576 h 0.271 γ; X; eau 1.215 1.44

82Rb 1.273 m 95.43 γ; X; eau 3.378 6.96

38 ⁸⁶Y 14.74 h 31.9 γ; X; eau 3.141 2.57

39 ⁸⁹Zr 78.41 22.74 γ; X; eau 0.902 1.27

40 ^94mTc 52.0 min 70.2 γ; X; eau 2.439 4.75

43 ^110mIn 4.9 h 0.033 γ; X; eau 0.569 0.60

53 ¹²⁴I 4.176 d 22.8 γ; X; eau 2.1376 3.41

For most metabolic measurements, the half-life of ¹⁸F is ideal, and if synthesis is rapid and efficient, compounds can be produced at other locations. FDG is used to measure energy metabolism but in case of brain tumours, protein metabolism can be more interesting due to the high background of FDG-PET in brain tissue [104]. Another advantage of labelled amino acids is a lower uptake in inflammatory sites.

[¹¹C]methionine was the first labelled amino acid to be established but tyrosine is a better indicator of protein synthesis and thus believed to better reflect increase in proliferation rate and malignancy of cancer cells [104]. Tyrosine is available both as [¹¹C]tyrosine and [¹⁸F]fluorethyl-tyrosine for PET [103]. A third popular use of ¹⁸F is 3'-deoxy-3'-¹⁸F-fluorothymidine (FLT). Like FDG, FLT is actively transported over the cell membrane and phosphorylated, trapping the tracer intracellularly. The rate of uptake appears to be dependent on the activity of thymidine kinase 1, which, in general, is 10-fold increased in cancer cells [103]. However, it must always be considered whether FLT is superior to FDG [101].

18F can also replace ¹¹C as tracer for labelling choline, which is required for building cell membranes and be used for imaging neuroendocrine tumours with [¹⁸F]3,4-dihydroxyphenylalanine([¹⁸F]DOPA) [90,157].

1.3.3 Radionuclides for indirect labelling

Indirect labelling enables use of tracer isotopes with a wide range of half-lives making it possible to choose an isotope with decay characteristics that fits the distribution kinetics of the vector molecule. Moreover, it is often possible to do an isotope switch and replace the positron-emitting isotope with one suitable for

radionuclide therapy. Unfortunately, attachment of a metal ion chelate affects the distribution of the vector molecule more than direct labelling and chelate stability becomes more important the longer the half life of the tracer. This means that not only the radioisotope but also the chelator and the linker have to be

considered (figure 1.2) [58,128,139].

Gallium has the advantage of being foreign to the body and gallium chelates are thus less exposed to in vivo competition from other chelating molecules. Previously, where an in house cyclotron was a rarity, ⁶⁸Ga (T½ = 68 min) was especially interesting since ⁶⁸Ga can be obtained from a ⁶⁸Ge/⁶⁸Ga generator but in a few years most facilities having a PET scanner can be expected to also have access to a cyclotron [138]. For planning and evaluation of radiotherapy with the somatostatin analogue [1,4,7,10-tetraazacyclododecane-

N,N’,N’’,N’’’-tretraacetic-acid-D-Phe¹-Tyr³]-octreotide (DOTA-TOC, figure 1.3) for instance with ⁹⁰Y (T½ = 64.0 h) ⁶⁸Ga-DOTA-TOC has shown good results [92,118].

If available from an in house cyclotron, ⁶⁶Ga (T½ = 9.5 h) is attractive for tumour imaging with tracer molecules having a slow (peaking hours after administration) tumour uptake and clearance, due to its longer half live [140]. However, isotope switch to gallium β-emitters (e.g. ⁷²Ga; T½ = 14.1 hour; 100 % β^-) for TRT use, is prevented by the high-energy γ-emissions of these isotopes. Yttrium in form of ⁹⁰Y is widely used for targeted radionuclide therapy for both cancer and autoimmune diseases [110,145]. The lack of penetrating γ- emission from ⁹⁰Y prevents exposure to remote organs and the surroundings but it also prevents imaging of the distribution. ¹¹¹In (T½ = 67.3 hours; 100 % γ) has been used as an yttrium analogue to follow the

distribution by SPECT, but the higher chelation stability of indium leads to an underestimation of the yttrium accumulation in bone marrow, a dose limiting organ [130]. Instead ⁸⁶Y (T½ = 14.7 h) can be employed, not only having identical chemistry but also enabling quantitative measuring of the

biodistribution, which compensates for the difference in half-life [130]. With the development of better chelating molecules, differences in biodistribution will become insignificant and the emission type vs. half- life will be decisive for the isotope choice. When using ⁸⁶Y for quantification, it is important to take the multiple high energy γ-rays emitted coincidentally with the annihilation radiation into account, as attenuation correction on these false coincidences can result in apparent activity. This can be avoided by a simple extra correction [170].

Table 1.2: Metals that can form stable coordination complexes with probe properties of biomedical interest noted [33].

Among the PET imaging isotopes, the isotopes of copper have a special position, because of the favourable binding characteristics of the metal and the variety of tracer half-lives and decay radiations available within the same labelling chemistry. The binding characteristics will be discussed under “Coordination chemistry of copper” so this chapter will focus on the decay characteristics and use of copper PET isotopes.

62Cu (T½ = 9.73 min) is the copper PET isotope with the shortest half-life. The T½ limits its use to perfusion studies such as with pyruvaldehyde bis(N⁴-methy1thiosemicarbazone)Cu(II) (PTSM)Cu(II), since the diagnostic time window is 2-3 times the half-life of the tracer [92]. It can be produced by a ⁶²Zn/⁶²Cu generator [78]. Similarly ⁶⁰Cu (T½ = 23.7 min) is used for hypoxia imaging with diacetyl-bis(N⁴- methylthiosemicarbazone) (ATSM), which is trapped selectively in hypoxic areas [122,141]. The high positron range is insignificant in large animals and humans, but in small animals, correction with a maximum

a posteriori reconstruction algorithm is an advantage [185]. Both ⁶¹Cu (T½ = 3.33 h) and ⁶⁴Cu (T½ = 12.7 h) can provide tracers better suited to biodistributions peaking after several hours. If optimal biodistribution is reached within a few hours as with Cu-1,4,8,11-tetraazacyclotetradecane-N,N',N'',N'''-tetraaceticacid (Cu- TETA), ⁶¹Cu labelling provide the best images due to the higher positron yield and shorter half life [141]. If prolonged chemical synthesis is required to produce the final compound, ⁶⁴Cu might be the only copper isotope usable, and compounds with very slow biodistribution such as labelled antibodies also require the long half-life of ⁶⁴Cu [44]. Furthermore, most radiochemical experiments benefits from a long half-life of the tracer, enabling several experiments before it disappears. ⁶⁴Cu also has the lowest positron range of the four copper PET isotopes, which should result in the highest spatial resolution. The price is longer detection times and higher radiation doses to the patient because of the low sensitivity of 0.98 cps/Bq/ml compared to the 3.43 cps/Bq/ml of ⁶¹Cu. Another factor is that more copper ions have time to transchelate before decay [230].

In deciding which copper isotope to use for imaging, all the mentioned factors must be weighed against each other and if chelation is swift and robust, one has the option of isotope switch.

In the end, ⁶⁴Cu does stand out, as it can be used for therapy as well. Both the β^- (40 %) and Auger electrons resulting from electron capture (41 %) have therapeutic potential [20].

Table 1.3: Selection of radionuclides used for TRT [42,96,158,165,235]

Z Radio- nuclide

Half life Therapeutic emissions

Maximal range Range of mean energy particle

LET Other emissions

29 ⁶⁴Cu 12.7 h eau; 1.31 µm 261 nm high γ; X

β^- 2.16 mm 0.42 mm Low

β⁺ 2.54 mm 0.75 mm Low

67Cu 61.83 h β^- (eau) 2.07 mm 0.26 mm low γ; X

39 ⁹⁰Y 64.00 h β^- (eau) 11.41 mm 4.01 mm low Γ

43 ¹¹¹In 2.805 d eau (in av. 15) 623 µm 4.13 µm high γ; X

51 ¹¹⁹Sb 38.19 h eau high X

53 ¹²⁵I 59.40 d eau (in av. 25) 23,1 µm high X

131I 8.025 d β^- (eau) 3.34 mm 0.39 mm low γ; X

62 ¹⁵³Sm 46.50 h β^- (eau) 3.34 mm 0.54 mm low γ; X

68 ¹⁶⁹Er 9.40 d β^- (eau) 1.06 mm 0.14 mm low γ; X

71 ¹⁷⁷Lu 6.647 d β^- (eau) 1.76 mm 0.24 mm low γ; X

75 ¹⁸⁶Re 3.718 d β^- (eau) 4.74 mm 1.04 mm low γ; X

83 ²¹²Bi 60.55 m α (β^-; eau) 45,1 µm high γ; X

213Bi 45.59 m α (β^-; eau) 42.6 µm high γ; X

85 ²¹¹At 7.214 h α (eau) 42.6 µm high γ; X

1.3.4 Rarely used isotopes

Several other positron-emitting isotopes are available but not normally used in clinic, either because they are difficult to produce in sufficient purity or because other isotopes have better decay characteristics. The three alkali metal radionuclides, ³⁸K (T½ = 7.6 min), ⁸¹Rb (T½ = 4.6 h) and ^82mRb (T½ = 6.3 h), have been used as blood tracers. The remaining isotopes are mostly longer living, suitable for halogenation of organic

compounds (e.g. ^34mCl, ^{75, 76, 77}Br) or for labelling with a bifunctional chelator (e.g. ^{51, 52}Mn, ⁵²Fe, ⁵⁵Co, ^94mTc and ¹¹⁸Sb) [16,164,174].

1.4 Dose limitation and radiosensitive organs

Unfortunately, some of the energy from the radioactive decays is delivered to healthy body tissue with damaging effect. As some organs are more sensitive to radiation than others are, it is necessary to estimate the absorbed dose (D) of potential limiting organs. This is done by measuring the activity at appropriate time intervals and integrating the results. This is converted to absorbed energy and divided by the mass of the organ to yield D with the unit Gy (J/kg). The effective dose can then be calculated by multiplying with the tissue weighting factor [111]. D should not be confused with the administered dose, which is measured in GBq. The estimation of D in TRT relies on co-administration of a radionuclide useful for imaging or

modelling of tissue distribution if the radionuclide used for treatment cannot be used for imaging [27]. Based on the distribution patterns in different TRT a maximum tolerated dose (MTD) can be estimated [55].

The exposure to normal tissues are rarely a limitation to PET as the doses required for imaging are relatively small and the half-lives of the preferred isotopes are short. TRT, on the other hand, requires doses high enough to obtain therapeutic tumour doses but withoutdose-limiting toxicity to the rest of the body. One might think that the higher the dose, the higher the therapeutic effect, and that treatment should be with the maximum tolerated dose, but this is not always the case with TRT [55,154], and there is still no international consensus on guidelines for administration of radioiodine in treatment of thyroid cancer patients [27].

In order to minimize non-target organ D it has been tried to co-administrate ethylenediamine tetraacetic acid (EDTA) as a chelator to remove ⁹⁰Y lost from the radiopharmaceutical by facilitating renal clearance [184].

Another strategy to minimize non-target organ D is removal of unbound labelled compounds after optimal targeting of tumour tissue [126]. It should also be possible to increase the maximum tolerated dose by using a mixture of radionuclides, or repeating treatment but with a radionuclide having a different dose-limiting organ. For example: ¹⁷⁷Lu (bone marrow) and ⁶⁴Cu (liver). Lysine, poly-glutamic acid and Gelofusine (a gelatin-based plasma expander) have shown potential as inhibitors of renal reabsorption of radiolabeled peptides, thus reducing the renal dose [74].

1.5 Immuno-PET and RIT

Both immuno-PET and RIT relies on antibodies or fractions of them as vector molecules. A major difference between the two techniques besides the purpose is that in immuno-PET a high resolution with a clear

distinction between highly labelled tumour tissue and the surroundings is only required for the time it takes