Final Report Summary - TARCC (Targeting alpha-particle emitting radionuclides to combat cancer)
Project context and objectives:
Work package one (WP1): Radionuclide chemistry
One important point for the development of the field of targeted alpha therapy is the reliable support with alpha emitting radionuclides and the establishment of robust and fast labelling protocols for labelling of potent antibodies and peptides with these radionuclides. Therefore one important objective was the optimisation of the production of alpha-emitting radionuclides (e.g. actinium-225, astatine-211 and bismuth-212) and the development of new methods for radiolabelling biomolecules (e.g. aminoacids, peptides or antibodies) with alpha-emitting radionuclides. The second objective was the evaluation of the stability of radioimmunoconjugates. Therefore the sensitivity of alpha-emitting radiopharmaceuticals to self-radiolysis was determined and ways of reduction of that was investigated. The third objective was the development and evaluation of in situ generators in the respect of applicability in targeted alpha therapy.
Therefore, the main goals were to:
1. improve the availability of a-emitting radionuclides (actinium-225/bismuth-213 generators, thorium-228/lead-212/bismuth-212 generators, astatine-211)
2. improve yields, specific activity and description of optimised radiolabelling protocols
3. improve the yield of astatine-211 labelling and stability of astatine-211-labelled products
4. improve the quality of a-radionuclide radiopharmaceuticals (radiochemical purity, protection from radiolysis)
5. study longer half-life radionuclides: in situ generators (e.g. lead-212), including encapsulation in nano-colloidal vectors.
WP2: Vectors
The major objective of this WP was the development and evaluation of new vectors for well defined targets. These targets should be relevant, easily accessible, specific (or at least highly expressed) and highly expressed. In addition, the vectors should have different pharmacokinetics matching the physical half-lives of the different alpha-emitting radionuclides (46 min to 10 days).
Several families of vectors were addressed and developed, namely peptides, hybrid peptides and amino acids. Labelling protocols (along with WP1) were developed with the aim to be translated into the clinic. In vitro pharmacologic studies such as binding affinities, cell uptake (binding and internalisation) in culture and cell retention (efflux) were performed. The best performing vectors were studied in available animal models (within and along with WP3).
WP3: Animal models for preclinical alpha radionuclide therapy
Models for human hematologic and solid tumours relevant to the clinical situation have been established for optimisation of alpha-radionuclide therapy depending on the characteristics of the vector and the mode of application. These strategies aimed at a high extent of tumour targeting combined with high therapeutic efficacy and low overall toxicity.
Growth of tumours was evaluated by bioluminescence imaging following transfection of cells with firefly luciferase. Tumour targeting was evaluated by biodistribution studies, scintigraphic imaging and autoradiography. The collected data served for dosimetry calculations.
Therapeutic efficacy was investigated by monitoring tumour growth with ultrasound, bioluminescence imaging or magnetic resonance imaging (MRI) and particularly by survival of animals.
Target organ toxicity according to the targeting modality and vectors were defined and long-term toxicity studies were performed (bone marrow, liver, kidneys). Bone marrow toxicity was evaluated by blood cell counts and analysis of chromosomal aberrations in bone marrow cells. Liver toxicity was examined via liver specific enzyme activities. Urea and plasma creatinine levels were used for determination of kidney toxicity. Moreover dose-limiting organs like kidney and liver were examined histopathologicaly. Therefore, the main goals were to:
1. evaluate targeting, therapeutic efficacy and toxicity following systemic application of peptide or antibody alpha-emitter conjugates in hematologic diseases like multiple myeloma
2. evaluate targeting, therapeutic efficacy and toxicity following systemic application of phenylalanine peptide or antibody alpha-emitter conjugates in gliomas, prostate cancer and colon cancer
3. evaluate pretargeting strategies in colorectal tumours
4. evaluate targeting, therapeutic efficacy and toxicity following intracavitary application of phenylalanine, peptide or antibody alpha-emitter conjugates (intratumoural in gliomas, intraperitoneal in gastric cancer and breast cancer, intravesical in bladder cancer).
WP4: Dosimetry and radiobiology
Important objectives of the project were dosimetry and radiobiology. The absorbed dose (radiation dose) to the tumour and normal tissues are fundamental to be able to understand and to interpret results such as therapeutic effectiveness and toxicity obtained from experimental tumour treatment studies. Much work on dosimetric methods has previously been performed for radionuclides emitting electrons and photons. For alpha particle emitting radionuclides much less effort has been made in establishing methods and models for the relatively short-ranged alpha particles. More knowledge is desired concerning the need of microdosimetric approaches in some tissue types related to the biodistribution of the radionuclide within the tissues. Concerning radiobiology, a higher biological effectiveness of alpha particles compared to electrons and photons is expected, the biological effects and toxicity towards critical organs are studied and the maximum tolerated activities and doses should be defined. The goals of these studies were to:
1. study energy delivery for cell experiments through a microdosimetric approach
2. establish animal dosimetric models adapted to mice and rats
3. establish a detailed tissue dosimetry model
4. study the dose-effect relationship for major risk organs
5. assess the biological effects and relative biological effectiveness for cell and animal experiments
6. investigate the effects of dose-rate and hypoxia in tumour cell experiments
7. determine the maximum tolerated activities and doses by toxicity studies.
Ultimately, optimal treatment protocols for future preclinical and clinical studies of targeted alpha-radionuclide therapy will be proposed from the results of the dosimetry and radiobiology studies.
Project results:
Radionuclide chemistry
One of the main goals of WP1 was the establishment of the astatine-211 production. Until now the astatine-211 is routinely produced in small amounts and gives a clear perspective for the large scale production. For the separation of astatine-211 the dry distillation (which seems to be more suitable for routine) and wet extraction have been performed. In the frame of the development of novel astatine-211 labelling techniques different reaction ways were evaluated. The radiolabelling of peptides with astatine-211 has been performed with electrophilic and nucleophilic substitution reactions, latter reaction leads to higher yields and fewer by products. The collection of available published data and new experimental developments permitted Centre National de la Recherche Scientifique (CNRS) to determine the Pourbaix Diagram of astatine in non-complexing medium which is the basis of understanding of astatine-211 labelling chemistry.
Astatine-211 is produced using the Bi-209(alpha,2n) reaction. At Arronax, the alpha beam is delivered at a fixed energy. It is then necessary to install a beam energy degrader to decrease the energy from 68 MeV down to 28.3 MeV. This energy has been calculated taking into account the beam energy spreading as well as the production of the astatine-210 which decays mainly to polonium-210, an alpha emitter with long period (138 days), which is a bone seeker. At this energy, an astatine-210/ astatine-211 ratio below 0.01% must be obtained, which is the usual limit for mother-nuclides in eluates from generator-based radionuclides for medical applications. The target is obtained by evaporation under vacuum of high purity bismuth. The support is made of aluminium nitride (AlN), a ceramic which as good thermal properties. A typical target has an area of 14 cm2. Its thickness of 40 µm can be obtained within a few hours. The homogeneity of the deposit is checked using profilometry. For the irradiation, the target is tilted at 15° in order to lower the power density on target. At the end of bombardment, the target is sent to hot cells for treatment through a rabbit system. The first irradiation has not yet been performed but it will start as soon as the energy degrader is installed (September 2011).
The use of astatine in hypervalent state was proposed as an alternative radiolabelling method to succinimidyl-astatobenzoate (SAB). The feasibility of the radiolabelling was first demonstrated by the use of a simple aryltin precursor. The method was then transferred to a bifunctionalised precursor allowing the introduction of hypervalent astatine and the coupling to proteins. These precursors were composed of a tin group on a hexafluorocumyl alcohol allowing the introduction of astatine and the formation of the hypervalent bonds. A maleimide was introduced for the protein coupling with the sulphydryl groups. A two-step procedure was set up to obtain the hypervalent astatinated compounds. In the first step, the tin precursor in methanol/acetic acid was heated at 100°C over 30 min in the presence of astatine-211 and N-chlorosuccinimide to give the monovalent astatinated species in good yield (90-95%). N-bromosuccinimide in chloroform was then added and the hypervalent species were obtained after 30 min at 60°C (95% yield).
The bifunctionalised compound was coupled to bovine serum albumin for in vitro stability evaluation. The radiolabelled BSA was incubated over 12h in human serum and no deastatination was observed (stability more than 99%). Manual and automated syntheses of the SAB prosthetic group from a tin precursor have also been improved.
Vectors
Amino acids
Astatine-211-labelling of phenylalanine and phenylalanine-analogues
Astatine-labelling by the Cu+-assisted nucleophilic halogen exchange method is highly reproducible and provides adequate yields for experimental and, eventually, clinical applications. 2- and 4-[Astatine]-Phe are efficiently and specifically transported by the pertinent amino acid transport systems. 4-[Astatine-211]-Phe is the most promising derivative to be further evaluated for intracavitary experimental glioma treatment.
Peptides
Somatostatin analogues
Somatostatin receptors are overexpressed on a variety of human tumours, in particular neuroendocrine tumours. Somatostatin-based agonists are successful in the clinic as imaging and targeted radionuclide agents. Agonists which internalise when bound to their receptors were considered as ideal radiovectors because of the long retention in the tumour. The rationale to study radioantagonists was the fact that most G-protein coupled receptor antagonists recognise many more receptor binding sites compared to agonists. The idea was to have very high early uptake which matches with the half-lives of short-lived positron- and alpha emitters. It finally and unexpectedly turned out that most of our antagonist studied showed not only higher uptake in the tumour but also longer retention which now allows the use of broad spectrum of half-lives. A variety of peptides with antagonistic properties were developed along with our colleagues from the Salk Institute, La Jolla (Dr. Jean Rivier) and the University of Berne (Prof. Jean Claude Reubi). The synthetic strategy was solid phase peptide synthesis. The unnatural amino acids (D-Aph(Cbm) and Aph(Hor)) were synthesised according to published procedures or peptides containing these amino acids were received from the Salk Institute. As shown below the octapeptides had a change in chirality of the aa-1 and aa-3 compared to the well known octreotide based agonists. The peptides were characterised by ESI-MS and their purity confirmed by HPLC. About 20 new peptides were synthesised and 12 studied thoroughly in vitro and in three mice models.
Somatostatin-based antagonists with high hydrophilicity and high affinity
JR11 and ANT1 were synthesised by partner two (UBH), Radiological Chemistry Division, according to an internal good manufacturing practices (GMP) protocol, careful analysis followed (residual solvents, TFA-content) and acute toxicity studies were performed. This procedure was accepted by the local ethics committee (Basel and Freiburg) and allowed first clinical studies using indium-111-DOTA-ANT1. The data although very promising indicated that improved affinity and higher hydrophilicity may help to improve pharmacokinetics even further. We finally came up with the three new compounds shown below. They are hydrophilic (compared to [Indium-111]-DOTA-ANT) and show improved somatostatin receptor two affinity.
The preclinical data, in particular with Indium-111-DOTA-JR11 (manuscript in preparation) encouraged us to start early human use studies at the university hospital Freiburg (partner 11), department of Nuclear medicine (Prof. W. Weber; Dr D. Wild). The improvement of DOTA-JR11 was also labelled with lutetium-177 and studied in comparison with Lutetium-177-DOTA-TATE in three patients. Tumour uptake was 2-4-fold higher with the antagonist; kidney uptake was also somewhat higher but the therapeutic window is still distinctly wider for the antagonist Indium-111-DOTA-JR11.
The pharmacokinetics of the antagonists is fast; they can be labelled with bismuth-213 and actinium-225 using very similar protocols than those for indium-111, yttrium-90 and lutetium-177. First clinical studies with bismuth-213 were planned along with the department of Nuclear Medicine, University Heidelberg and partner eight (JRC-ITU). The studies have been delayed due to the lack of GMP-JR11.
These conjugates were labelled with the radiometals mentioned above and biodistribution was studied with the positron emitters copper-64 and gallium-68. (see Fani M et al , 'PET of Somatostatin Receptor-Positive Tumours Using Cu-64 and Ga-68-Somatostatin Antagonists: The Chelate Makes the Difference' J. Nucl. Med. 2011;52:1110-1118)
Precursor synthesis for chemical ligation and astatine-211-labelling
Synthesis of precursors for selective astatine-211-labelling of peptides (Substance P)
Substance P is a ligand binding to neurokinin type one (NK-1) receptors, which are highly overexpressed in malignant gliomas. A few potential peptide vectors were recently developed. Two of these peptides, DOTAGA-substance P and DOTA-[Thi-8, Met(O2)-11]-substance P were shown to be suitable for diagnostic labelling with indium-111 and therapeutic labelling with yttrium-90, lutetium-177 and bismuth-213. The modification of substance P (introduction of Thi-8 for Phe-8 and Met-11 replaced by the corresponding sulfone) led to improved radiolytic and improved metabolic stability.
The DOTA-conjugated substance-P derivative was successfully studied in the clinic in different brain tumour patients, labelled with bismuth-213 (Eur J Nucl Med Mol Imaging, 2010;37(7):1335-44).
Astatination of substance P derivative by selective chemical ligation
The idea of the project is to find a route for routine application of astatine-211-labelled substance P derivatives. The strategy is to conjugate astatine-211-labelled benzaldehyde onto an aminooxy functionalised [Thi-8,Met(O2)-11]-substance P via oxime formation.
The advantage of an aminooxy functionalised peptide is the high reactivity with carbonyl functionalities and the resulting stable oxime formation. In addition, the major advantage is that a carbonyl-moiety can be selectively linked in the presence of an aminooxy group due to pKa differences of the aminooxy functionalised compound (pKa µ 5) and the amino functionalised compound (pKa µ 11). This means that at a low pH of 3-4, the amino group is fully protected by the proton whereas the aminooxy group is still reactive. The conjugation kinetics was optimised using 4-Fluorine-19-benzaldehyde.
On the other hand the problem with the aminooxy precursor is its high reactivity with all kind of carbonyl compounds such as acetone in solvents and formaldehyde from the air. This all had to be understood and optimised for improved synthesis. Attempts to introduce 'weak' protecting groups which are being cleaved during the conjugation step were not successful.
Development of metabolically stabilised GASTRIN derivatives (partners 2/11, 6)
Non-ionic and mixed ionic/non-ionic spacer DOTA-gastrin-conjugates as possible precursors for SST/CCKB receptors binding hybrid
To date there is no minigastrin analogue suitable for radiopeptide therapy of medullary thyroid carcinoma and other gastrin receptor expressing tumours such as neuroendocrine tumours and SCLC. The main limiting factor is either their low tumour uptake or/and too high kidney retention which has been related to the N-terminal Glu-s ; deletion of these Glu-s lowered the kidney uptake but the metabolic stability deteriorated. The goal of this work was to develop radiogastrin derivatives with high metabolic stability, high affinity and low kidney uptake. The strategy was to replace the hydrophilic, negatively charged spacers (Glu-s) with hydrophilic but uncharged spacers.
In addition, DOTA-conjugated minigastrin analogues with so called mixed ionic/non-ionic spacers were synthesised manually on solid phase. Gastrin receptor affinities were determined by receptor autoradiography in cooperation with Prof. Reubi, university of Berne. 111In-labelled peptides were evaluated in vitro in AR4-2J cell line and in human serum. The best peptides were studied in rats bearing the AR42J tumour.
We developed DOTA-coupled minigastrin analogues with ionic, non-ionic and mixed ionic/non-ionic spacers with high stability in human serum. All conjugates showed high binding affinity to the gastrin receptor (IC-50-values between 0.5 and 4.8 nM). All radiopeptides showed receptor-specific internalisation. The metabolic stability remains high within the series. Feasibility of labelling with bismuth-213 was confirmed. Further biodistribution studies in animal models are needed in order to determine which type of spacer is superior in relation to best tumour-to-kidney ratio.
Synthesis of hybrid peptides (partner 2/11, 6) and chelated F3 (partner 3)
Synthesis of hybrid peptides
The first step of the development of hybrid peptide was the selection of somatostatin and gastrin synthons. TOC as the most widely used somatostatin analogue and PP-F10 as one of the most promising minigastrin analogues were selected. Several attempts to couple somatostatin and CCK2/gastrin synthons were performed.
Synthesis of chelated F3 (Partner TUM): 213Bi-and 225Ac-labeled substance peptide F3
The tumour homing peptide F3 is effectively internalised into the nucleus of tumour cells upon binding to nucleolin and therefore can be used as an appropriate carrier for alpha-particle emitting isotopes in targeted tumour therapy. We have successfully labelled DTPA-chelated F3 peptide with the alpha-emitter bismuth-213 and demonstrated therapeutic efficacy of bismuth-213-DTPA-F3 conjugates in a nude mouse model of peritoneal carcinomatosis. In our effort to expand the study we have now labelled DOTA-chelated F3 with the alpha-emitter actinium-225 and compared cytotoxicity and therapeutic efficacy of both conjugates. ID50 values of bismuth-213-DTPA-F3 (53 kBq/ml) and actinium-225-DOTA-F3 (67 Bq/ml) - as determined via clonogenic assays - differed by a factor of 1,000. Repeated therapy with six x 1.85 kBq actinium-225-DOTA-F3 or six x 1.85 MBq bismuth-213-DTPA-F3 prolonged median survival to 95 days and 97 days, respectively, with negligible kidney damage (see WP3) (manuscript in preparation).
Antibody production and characterisation
Monoclonal antibodies directed against murine CD138 have been generated by immunisation against peptides extracted from the murine CD138 sequence. The peptides were chosen in an area similar to that recognised by the anti-human CD138. Nine hybridomas recognised murine multiple myeloma by flow cytometry. Subcloning was performed and affinity tests are in nM range.
Bivalent hapten synthesis and labelling protocols
Bivalent haptens
By definition a hapten is a low molecular weight (generally less than 1 000 Da) compound that is recognised by specific antibodies. Usually, haptens are first coupled to carrier proteins, such as serum albumin or keyhole limpet hemocyanin, to form an immunogenic conjugate capable of triggering an antibody response in animals. Then monoclonal antibodies may be obtained by classical hybridoma technology. The DTPA-indium complex has been used in earlier preclinical and clinical work by partner one (INSERM) and collaborating teams. The histamine-succinyl-glycine (HSG) pseudo-peptide has also been proposed as a useful hapten in the context of pretargeting and interest has recently shifted to this hapten because it may be used to derive pretargeting vectors that may be labelled with a variety of radionuclides. Bivalent haptens have been shown to be superior to monovalent ones for in vivo pretargeting of tumours. Small molecular weight molecules comprising two HSG haptens and one or more sites for radioactive labelling have been designed and synthesised.
Radiolabelling of bivalent haptens
Another bivalent HSG hapten, provided by the Immunomedics Company (Morris Plains, NJ, USA) was used in the project. It was radiolabelled with radioactive iodine (iodine-125) for biodistribution studies and with lutetium-177 and bismuth-213 for preclinical therapy studies (Griffiths GL, Chang CH, McBride WJ, Rossi EA, Sheerin A, Tejada GR, Karacay H, Sharkey RM, Horak ID, Hansen HJ, Goldenberg DM. Reagents and methods for PET using bispecific antibody pretargeting and 68Ga-radiolabeled bivalent hapten-peptide-chelate conjugates. J Nucl Med. 2004, 45: 30-39).
Animal models for preclinical alpha radionuclide therapy
Multiple myeloma
Systemic application of alpha-emitter conjugates in hematologic disease a syngeneic murine model of multiple myeloma was established by intravenous transfer of 5T33 myeloma cells that disseminated into the bone marrow. Syndecan-1 (CD138), a heparan sulfate proteoglycan, constantly expressed on tumour cells in multiple myeloma served as surface antigen in targeted therapy. Mice were treated with bismuth-213-labeled anti-mCD138 at four different activities (1.85 3.7 7.4 and 11.1 MBq). The group treated with 3.7 MBq exhibited a median survival of more than 300 days compared with 45.5 days for the control group. The highest activity (11.1 MBq) turned out toxic while the lowest activity of 1.85 MBq showed no therapeutic effect. 7.4 MBq induced a median survival of 227 days. With activities of 3.7 and 7.4 MBq, surviving mice exhibited a transient haematological toxicity and only the injected activity of 7.4 MBq caused a transient rise of Flt-3L indicative of low myelotoxicity.
Prostate cancer
In a human androgen-independent prostate carcinoma model (PC-3-tumour model) therapeutic studies were undertaken using two novel bismuth-213 or lutetium-177 labelled bombesine-based peptides (DOTA-PEG4-bombesin [DOTA-PESIN] and DO3A-CH2CO-8-aminooctanoyl-Q-W-A-V-G-H-L-M-NH2 [AMBA]). For therapy and toxicity studies, mice were injected with lutetium-177-DOTA-PESIN, bismuth-213-DOTA-PESIN or bismuth-213-AMBA, whereas control groups were left untreated or were given nonradioactive lutetium-175-DOTA-PESIN. All radiopeptides showed a rapid blood clearance and tumour uptake. The maximum tolerated dose (MTD) of bismuth-213-DOTA-PESIN and bismuth-213-AMBA was 25 MBq and the MTD of lutetium-177-DOTA-PESIN was 112 MBq, corresponding to kidney doses of 6-11 Gy. At these dose levels, lutetium-177-DOTA-PESIN showed no, bismuth-213-DOTA-PESIN slight and bismuth-213-AMBA marked kidney damage. At MTD, bismuth-213-DOTA-PESIN and bismuth-213-AMBA were significantly more effective than lutetium-177-DOTA-PESIN. These preclinical data show that alpha-therapy with bismuth-213-DOTA-PESIN or bismuth-213-AMBA is more effective than therapy with beta-particle emitting lutetium-177-DOTA-PESIN. In addition, therapy with bismuth-213-DOTA-PESIN has a better safety profile than treatment with bismuth-213-AMBA and represents an important new approach for treating recurrent prostate cancer.
Colorectal cancer
Subcutaneous colorectal µtumours or hepatic µmetastases were established for therapeutic studies using the affinity enhancement system (AES) technology comprising non radioactive bispecific antibodies and a radiolabelled (bismuth-213 or astatine-211) bivalent hapten.
LS174T human colonic carcinoma cells (1 million) were inoculated subcutaneously or via the portal vein (hepatic metastasis model) in NMRI-nu mice. The tumour implantation rate was 100% using aggregated cells and only 26.6% with isolated cells. Photons emitted by one million cells were detected by bioluminescence imaging immediately following injection and allowed visual confirmation of hepatic distribution. Therefore, hepatic tumour graft via the portal route and bioluminescence imaging provides a reliable animal model and permits sensitive in vivo detection and follow-up of hepatic metastases. The hepatic model appears to reproduce more closely colon cancer spread compared to the subcutaneous one. The hepatic model is of particular interest to study radioimmunotherapy.
Pharmacokinetic distribution of the TF2 bispecific anti-Carcino-Embryonic-Antigen (CEA) x anti histamine-succinyl-glycine (HSG) antibody was evaluated. For various antibody / hapten molar ratios (10:1 and 2:1) and pretargeting time intervals (5, 10, 15, 24 h), biodistribution studies were performed in the subcutaneous model with TF2 and Iodine-125-labeled bivalent HSG peptide IMP-288 to optimise tumour uptake and tumour / organ uptake ratios.
Activity blood clearance was very fast. Using an antibody / peptide ratio of 10:1 and a pretargeting time interval between 15 and 24 h, the peak of tumour uptake reached 9.13 ± 4.4 % injected dose per gram (%ID/g) in subcutaneous tumours two h after hapten injection. Reducing the time interval between antibody and peptide injections increased tumour uptake, but activity uptake also increased in most tissues, particularly in liver and spleen. Tumour uptake 1h after hapten injection was higher in hepatic metastases (11-26% ID/g) than in subcutaneous tumours (7-12% ID/g). The results of the biodistribution studies and dosimetric calculations revealed that pretargeting must be finely tuned for therapeutic application.
For pretargeted radioimmunotherapy IMP-288 targeting TF2 was labelled with bismuth-213 to a specific activity of 74 MBq/nmol. Seven days after tumour inoculation, groups of four animals were selected with matched bioluminescence signals. Four control groups were set: no treatment, TF2 alone, 0.2 or 0.1 nmol of labelled peptide only. Two treatment groups were set: pretargeting with two nmol of TF2 followed by 0.2 or 0.1 nmol of bismuth-213-labelled peptide IMP-288 with a pretargeting time interval of 18 h. Haematological toxicity was monitored by weekly blood sampling. Tumour growth was monitored by in vivo bioluminescence imaging (every four to six days). Animals were sacrificed when a 20% weight loss or cachexia was observed or when the bioluminescence signal reached 20x107 counts. Increase of the bioluminescence signal was delayed in treated animals (TF2 + bismuth-213-labeled IMP-288) with tumour growth starting 10 days after treatment whereas tumours grew continuously in control groups. At day 23 after treatment, the lowest bioluminescence signal was observed in the group treated with two nmol TF2 and 0.1 nmol of labelled IMP-288. Animals were euthanised when the bioluminescence signal reached 20x107 photons. The Kaplan-Meier curve showed 0% survival in the untreated group versus 75% in the group treated with two nmol TF2 and 0.1 nmol IMP-288. Toxicity was acceptable with a transient decrease of blood cells six days after treatment and complete recovery at day 13.
Brain tumours
Increased amino acid transport in brain tumours is used for diagnostic purposes. It has been shown that astatine-211-labelled phenylalanine (At-Phe) is taken up by glioblastoma cells in vitro. It was tested, whether systemic intravenous treatment of rats with intracranially implanted glioblastomas with At-Phe would have a beneficial effect. At-Phe was prepared via nucleophilic halogen exchange on L-iodophenylalanine. The rat glioblastoma cell line BT4Ca was implanted into the prefrontal cortex of female BDIX rats by stereotaxic µinjection (10 000 cells/3 µlitre; n=63). At-Phe (2 MBq) or phosphate buffered saline (PBS) were injected intravenously three days after implantation. A third group was treated twice, on days three and 10 after implantation. Rat?s health condition was assessed each day by using a score system (e.g. weight, pupillary and motor reflexes) and rats were sacrificed when showing a premortal health status. Additionally, rats were sacrificed on days 6, 10, 13 and 17 after transplantation to measure the tumour volume and area of necrosis. Furthermore, the proliferation index was assessed after immunohistological staining for Ki-67.
The mean survival time of the rats who had received two applications of At-Phe was significantly higher compared to control rats (less than 0.05). The score-analysis showed that At-Phe treated rats had a significant better health condition compared to PBS treated rats (less than 0.05). Analysis of the tumour volume on days 6, 10, 13 and 17 after transplantation showed no differences between groups, but rats with a premortal health status had a significantly larger tumour than rats in good health condition (less than 0.05). Overall the proliferation index was not different between groups, but necrosis was marginally larger after At-Phe treatment (less than 0.05 on day 10).
Gastric cancer
In a gastric cancer peritoneal carcinomatosis model established by intraperitoneal injection of 1x107 HSC45-M2 gastric cancer cells, expressing mutant d9-E-cadherin therapeutic efficacy and toxicity of bismuth-213-d9MAb (targeting mutant d9-E-cadherin) were evaluated in comparison to beta-particle emitting lutetium-177-d9MAb conjugates. Biodistribution data showed an uptake of bismuth-213-d9MAb in the tumour nodules up to 40% ID/g at 45 min after injection. In all other tissues the uptake of bismuth-213-d9MAb was below 3.2% ID/g.
For optimisation of radioimmunotherapy mice were injected intraperitoneally with 0.37 1.85 7.4 or 22.2 MBq of bismuth-213-d9MAb at days one or eight after tumour cell inoculation.
Application of 1.85 MBq of bismuth-213-d9MAb was most efficient in terms of survival. At the end of the observation period almost 90% of the mice were still alive and showed no signs of disease. Bismuth-213-d9MAb activities both higher and lower than 1.85 MBq proved to be less efficient concerning prolongation of survival. Therapy with bismuth-213-d9MAb at day eight was less efficient than treatment on day 1. Tumour growth as well as regression after therapy could be visualised by bioluminescence imaging.
Toxicity of bismuth-213-d9MAb therapy was negligible at therapeutically effective activities. White blood cell counts were reduced within two days after therapy but recovered to pre-treatment values at day 21. Reduction was maximal after application of 22.2 MBq and minimal after 1.85 MBq. Chromosomal aberrations in bone marrow cells could only be observed at day one after therapy. Following autopsy none of these mice showed any signs of organ toxicity. Therefore, activities of up to 7.4 MBq of bismuth-213-d9MAb are thought to cause no toxic side effects.
Treatment of mice with 1.85 7.4 and 14.8 MBq of lutetium-177-d9MAb also significantly prolonged median survival. Therapy with 7.4 MBq of lutetium-177-d9MAb was most effective: more than 50% of the animals survived longer than 250 d. Therapy with 1.85 7.4 and 14.8 MBq of lutetium-177-d9MAb at day eight after tumour cell inoculation also resulted in statistically significant improvement of median survival. Intraperitoneal injection of lutetium-177-d9MAb induced a transient reduction in leukocyte counts at activities more than 1.85 MBq. Injection of 7.4 MBq and 14.8 MBq of lutetium-177-d9MAb rapidly decreased leukocyte counts to 40% at day 10 and 6% at day 17 of pre-treatment values. Leukocyte numbers did not recover to pre-treatment values after application of 14.8 MBq. Following i.p. injection of mice with 7.4 and 14.8 MBq lutetium-177-d9MAb 3% and 13% of the mice, respectively, developed a massive cervical, axillary and femoral-inguinal lymphadenopathy confirmed as lymphoblastic lymphoma by histological examination. One animal with enlarged lymph nodes had developed a plasmocytoma. Furthermore the spleen of this animal showed a follicular hyperplasia, the kidney a severe proliferative glomerulonephritis and the liver a hepatocarcinoma. Comparison of the results obtained with lutetium-177-d9MAb and bismuth-213-d9MAb demonstrate therapeutic efficacy for both radionuclides. However, lutetium-177-d9MAb turned out to be much more toxic at therapeutically effective doses because of long-lasting circulation of the activity in the blood after resorption from the intraperitoneal cavity.
In order to optimise intraperitoneal radioimmunotherapy, intraperitoneal retention and biodistribution of Fab-fragments, intact IgG and IgM coupled to indium-111 or bismuth-213 were analysed in tumour free mice after i.p. injection. The results indicate a rapid resorption of Fab-fragments from the peritoneal cavity resulting in a high kidney accumulation. IgG remained in the circulation for more than 72 h after resorption from the peritoneal cavity. IgM, in contrast showed a long-lasting retention in the peritoneal cavity. After resorption IgM is rapidly trapped in the liver resulting in a very low activity in the circulation at all time points. These results could be confirmed by scintigraphic imaging. Therefore high molecular weight IgM antibodies seem to be the preferable radionuclide carriers for intraperitoneal injection.
Breast cancer
A dimer of the vascular tumour homing peptide F3 was chemically coupled to the alpha-emitter bismuth-213 (bismuth-213-DTPA-[F3]2). It was found that bismuth-213-DTPA-[F3]2 accumulate in the nucleus of tumour cells in vitro and in intraperitoneally growing tumours in vivo. To study the anti-tumour effect of bismuth-213-DTPA-[F3]2 mice bearing intraperitoneally growing xenograft tumours were treated with bismuth-213-DTPA-[F3]2. SCID-mice were injected i.p. with 10 million MDA-MB-435 cells expressing firefly luciferase. In the tumour prevention study groups of eight mice were injected every second day with 1.85 MBq bismuth-213-DTPA-[F3]2, 1.85 MBq bismuth-213-DTPA or 100 µlitre PBS between day four and 14 after inoculation of the tumour cells. In the tumour reduction study groups of eight mice were injected i.p. with 1.85 MBq bismuth-213-DTPA-[F3]2 1.85 MBq bismuth-213-DTPA or 100 µlitre PBS between the days 16 and 26 after inoculation of the tumour cells.
The survival time of the animals was increased from 51 to 93.5 days in the prevention study and from 57 days to 78 days in the tumour reduction study after treatment with bismuth-213-DTPA-[F3]2. No toxicity of the treatment was observed. The serum concentration of creatinine was determined in all tumour bearing mice when they were sacrificed after therapy. No significant elevation of the serum creatinine level after treatment with bismuth-213-DTPA-[F3]2 at time of sacrifice was observed.
In a second study stable conjugates of the vascular tumour-homing peptide F3 with actinium-225 were generated to accomplish targeted delivery into tumour cells. The aim of this study was to determine the therapeutic window of actinium-225-DOTA-F3 and to compare it with bismuth-213-DTPA-F3.
Anti-tumour activity of bismuth-213-DTPA-F3 and actinium-225-DOTA-F3 was determined in-vitro using clonogenic assays. Activities with equivalent anti-tumour effects in vitro were then used for treatment of mice with intra-peritoneal MDA-MB-435 xenograft tumours in vivo. Survival of animals was analysed and therapy monitoring was performed by optical imaging and histological analysis. Nephrotoxic effects were analysed by histology.
In vitro the ID50 of bismuth-213-DTPA-F3 and actinium-225-DOTA-F3 was determined as 53 kBq/ml and 67 Bq/ml, respectively. In vivo, therapy with six x 1.85 kBq actinium-225-DOTA-F3 or six x 1.85 MBq bismuth-213i-DTPA-F3 or PBS were compared. Median survival in the PBS group was 60 days. Median survival in the group treated with actinium-225-DOTA-F3 was 95 days and 97 days in the group treated with bismuth-213-DTPA-F3. two bismuth-213-DTPA-F3 and actinium-225-DOTA-F3 had different effects on tumour growth as determined by optical imaging. bismuth-213-DTPA-F3 reduced the tumour mass at early time points until 30 days after treatment whereas actinium-225-DOTA-F3 has a prolonged anti-tumour effect leading to a reduction of tumour mass until the death of animals. Bismuth-213-DTPA-F3 was more effective in reducing the number of metastases.
Bladder cancer
An orthotopic human bladder carcinoma mouse model using the EGFR-overexpressing luciferase transfected bladder carcinoma cell line EJ28 was successfully established by intravesical instillation of two million EJ28 luc cells / 100 µlitre after gentle electrocautery of the murine bladder. Therapeutic efficacy of intravesically instilled bismuth-213-anti-EGFR-MAb was compared with standard mitomycin C chemotherapy. Mice without therapy and those treated with unlabeled anti-EGFR-MAb reached a median survival of 41 d and 89 d, respectively. Mice that underwent therapy with 0.925 MBq of bismuth-213-anti-EGFR-MAb one h, seven d or 14 d after cell instillation survived more than 300 d in 90%, 80% and 40% of cases, respectively. The results were visualised by bioluminescence imaging. Therapy with 0.37 MBq one h or seven d after tumour cell inoculation resulted in survival more than 300 d in 90% and 50% of mice, respectively. Mitomycin C treatment after one h and seven d prolonged survival more than 300 d in 40% and 50%, respectively, however turned out to be nephrotoxic. In contrast, no signs of nephrotoxicity could be observed following bismuth-213-anti-EGFR-MAb treatment. The study suggests that radioimmunotherapy using intravesically instilled bismuth-213-anti-EGFR-MAb is a promising option for treatment of bladder cancer in patients.
In a second study therapeutic efficacy of a fractionated intravesical treatment with bismuth-213-anti-EGFR-MAb in an advanced tumour stage was evaluated. Radioimmunotherapy was initiated between days 14 and 36 after cell instillation when tumours could be clearly detected via bioluminescence imaging. The first group received three intravesical applications of 0.46 MBq of bismuth-213-anti-EGFR-Mab each in a time interval of four days. The second group was instilled with the unlabelled anti-EGFR-MAb and the third group remained untreated. Tumour development was monitored by bioluminescence imaging and survival was observed up to 300 days.
60% of the animals of the group treated three times with 0.46 MBq of bismuth-213-anti-EGFR-Mab in a time interval of four days showed a reduction of tumour size while in the untreated group the tumour sizes increased in all animals. The group receiving no treatment or unlabelled anti-EGFR-MAb showed a mean survival of 44 days after tumour cell inoculation while median survival in the bismuth-213 treated group was 88 days. The urothelia of the tumour-free animals treated three times with 0.46 MBq bismuth-213-anti-EGFR-MAb did not show any pathological findings.
Fractionated intravesical therapy with the alpha-emitter bismuth-213 coupled to anti-EGFR-MAb effectively reduces tumour size without radiation damage to the normal urothelium. Thus, fractionated therapy is a very promising approach for successful treatment of bladder carcinoma in an advanced stage.
Dosimetry and radiobiology
Microdosimetry of alpha particles for cell experiments
Two Monte Carlo (MC) codes, MCNPX and GEANT4, were setup and after validation used by two of the partners, together with conventional dosimetric methods, namely microdosimetric calculations for astatine-211 in a spherical cell and in a non-symmetrical cell; microdosimetric calculations were performed for astatine-211 using GEANT4. Two cell geometry models were tested: a regular mathematical cell model of a spherical cell with a concentric spherical cell nucleus and a voxel-based virtual cell model of a non-symmetrical cell.
Astatine-211 was located homogeneously in the nucleus, the cytoplasm, on the cell surface or outside the cell. There were differences between the low energy electromagnetic model (G4-LEM) and the standard electromagnetic model (G4-SEM) in mean specific energy and S values. Compared to published data the mean specific energy demonstrated less than 2% difference for G4-LEM, less than 6% difference for G4-SEM for sources in the nucleus, cytoplasm and on the cell surface. For astatine-211 located outside the cell differences of less than 4% and more than 10%, respectively, were found. Compared to published S values less than 6% and less than 4% differences were found for astatine-211 and less than 2% and less than 3% differences for its daughter polonium-211. Results from the non-symmetrical cell reflect the influence of irregular shape and heterogeneous medium.
Tumour control probability (TCP) for uniform and non-uniform activity distributions
In a study on cell clusters, we demonstrated how TCP depends on the radionuclide distribution in the tumour, both macroscopically and at the sub-cellular level for astatine-211, yttrium-90, lutetium-177, rhodium-103m. The radionuclides were uniformly distributed within the sub-cellular compartment and they were uniformly, normally or log-normally distributed among the cells in the tumour. In summary, the results demonstrate that the radionuclide distribution both within the cells but also within the tumour clearly affect the absorbed dose and the TCP for astatine-211 and also for the low-energy electron emitter rhodium-103m, while this was almost negligible for the other radionuclides studied (lutetium-177, yttrium-90). TCP for astatine-211 and rhodium-103m was affected by the radionuclide distribution to a great extent when the radionuclides were in the cell nucleus and to lower extent when the radionuclides were distributed on the cell membrane or in the cytoplasm. When the radionuclide was distributed in the nucleus, the cumulated activities required for TCP=0.99 increased when the activity distribution became more heterogeneous for astatine-211 and rhodium-103m, with larger increase for normally distributed radionuclides compared to log-normally distributed. When the activity was distributed on the cell membrane, the cumulated activities required for TCP=0.99 were not affected for astatine-211 and rhodium-103m when the activity distribution became more heterogeneous.
Dosimetric models for mouse and rat
Mouse dosimetry using MCNPX and GEANT4
A mouse geometric model (Bitar model) was defined based on 3D reconstruction of sequential slices of a frozen 30 g NMRI nude mouse. Calculations were compared with those of the 28 g Digimouse model. Mouse dosimetric models were developed for bismuth-213 and astatine-211. S-values were calculated from absorbed energy distributions obtained with MCNPX and GEANT4, assuming uniform activity distribution in each organ and contribution of all daughter radionuclides except bismuth-207. Absorbed fractions were high (0.92-0.99) showing that alpha-particles deliver most of their energy in the source organ. For some larger organs, self-dose S-values were similar for both mouse models (difference less than 6%). For the other organs there were differences up to 2000 %. Cross-dose S-values varied within 3-5000%. These discrepancies can be explained by the variations in organ size, shape and location between the two mouse models. Comparing GEANT4 and MCNPX, differences in S-values for bismuth-213 were less than 5% for all organs. When both self-dose and cross-dose were considered, differences of up to 40% were found, but the largest differences were found for very low S-values with low impact on total absorbed dose. For high S-values, relative differences were less than 15%.
Dosimetry of bismuth-213-BSA in mice
Application on biodistribution data: Based on pharmacokinetic data of bismuth-213-BSA in non-tumour-bearing nude mice time activity curves and cumulated activity were derived for each organ. Absorbed dose based on MC S-values were in general similar to those derived from calculations assuming that all emitted energy from particles was absorbed within the organ. The highest differences were found for lungs (8%). Comparisons between the two MC-techniques showed that the highest difference was observed for the kidneys (10% with Bitar model). By comparing the two mouse models the absorbed dose was similar in organs containing high activity (less than 4% difference), with larger discrepancies in organs containing low activity (up to 110 % difference).
Rat dosimetry
No new rat geometrical model was developed because no partner had any need for such a model. Furthermore, since rats are much larger than mice mean absorbed doses can accurately be estimated assuming that all alpha particles deliver their energy within each organ. Detailed tissue dosimetry, however, might require better knowledge in morphology and most probably specific models must be defined for each rat type and size/age, as for the mouse models.
Biodistribution and dosimetry of free astatide-211, iodide-14 and iodide-131 in rats
Low amounts of astatide-211 are released from the radiopharmaceutical in vivo. It is then important to know the biodistribution and the absorbed dose to normal tissues for astatide-211 in order to determine potential risk organs when using astatine-211-labelled radiopharmaceuticals for treatment. Very limited knowledge on biodistribution of astatide-211 is available for rats. Rats were injected simultaneously with iodide-125 and iodide-131, or free astatide-211 and the activities and activity concentrations in organs and tissues were determined and mean absorbed doses were calculated. The biodistribution of free astatide-211 was different compared to iodide-125- and iodide-131, while it was similar for 125I and iodide-131. The uptake of all three radionuclides was highest in the thyroid (400 %IA/g for iodide-125, 430 %IA/g for iodide-131 and 75 %IA/g for astatide-211 after 18 h). In all other organs and tissues the activity concentration and retention of astatine-211 was in general higher than that of iodide-125 and iodide-131. The thyroid received the highest absorbed dose per injected activity (400 mGy/MBq for iodide-125 and 19000 mGy/MBq for iodide-131 after seven days and 17000 mGy/MBq for astatide-211 after 24 h). Otherwise, the highest mean absorbed doses were found in stomach, heart and small intestine for iodide-125 and iodide-131 and in stomach, lungs and spleen for astatine-211. Astatine-211 delivered a much higher absorbed dose per injected activity to all extrathyroidal organs compared to iodide-125 and iodide-131.
Translation of data from astatine-211 in rats and mice to dosimetry of astatine-211 in humans
The possibility to translate astatine-211 biodistribution data between different species, especially from rat and mouse to human was studied, since biodistribution and dosimetric data for humans are difficult to obtain. The absorbed dose of astatine-211 and iodine-131 to different tissues in man was estimated based on extrapolation of biodistribution data obtained from rat and mouse. The highest estimated absorbed dose per unit activity administered from both species to man was obtained for the thyroid and stomach together with the pituitary (rat) and the salivary glands (mouse), all expressing NIS. The estimated effective dose to man differed by a factor of two depending on the species. The biodistribution of astatide-211 is different from that of iodide in animals and the iodide kinetic models (ICRP53) cannot be applied to astatine-211 in man. The iodine-125 values from mice are closer to data in ICRP53 than from rat. Thus, biodistribution and dosimetry data differ between species, with a clear difference between mouse and rat.
Detailed tissue models
The initial plans were to develop geometrical models for tissue scale dosimetry for several tissue types, especially for the critical organs. The thyroid model is finalised and the work with defining a model of the kidneys is ongoing but will take a longer time than expected due to the high complexity with the need to define the many various parts of the nephron involved in late side effects in radionuclide therapy.
Detailed dosimetric model of the thyroid
One of the main risk organs when using astatine-211-labelled radiopharmaceuticals is the thyroid and therefore much effort was used for this model. The high uptake of free astatide-211 in thyroid (see above) indicates that thyroid-blocking should be used when astatine-211-labelled pharmaceuticals are applied for therapy. A general thyroid follicle was constructed using spherical geometry, with a follicle lumen with a diameter of 10-500 µm, surrounded by a single layer of follicle cells with 4-8 µm (diameter) spherical nuclei. Follicle models were defined for mouse, rat and humans. Calculations were performed for alpha-particles using MCNPX. The target was the follicle cell nucleus, with astatine-211 in the follicle lumen, follicle cells or nuclei. The mean specific energy with astatine-211 within the follicle lumen was 0.17 1.1 and 2.0 mGy, for human, rat and mouse, respectively, with corresponding single-hit mean specific energy 370, 610 and 1200 mGy. Approximately 45% of the alpha particles in a astatine-211 decay deposit all the kinetic energy within the follicle lumen for the human model, compared to 7% and less than 0.5% for the rat and mouse model, respectively. The results demonstrate the importance of taking the range of the alpha particles and the spatial distribution of astatine-211 into account for dosimetric evaluation of thyroid exposure. The dosimetry for 211At in the thyroid cannot readily be translated between the species.
Radiobiological effects in major risk organs
Thyroid toxicity of astatine-211 and iodine-131 in mice studied by gene expression analysis
Female BALB/c mice were i.v. injected with astatine-211 or iodine-131 and the absorbed dose to the thyroid was 0.05-32 Gy and 0.85-17 Gy, respectively. Gene expression profiling (Illumina platform and Nexus) was performed using total RNA from thyroids removed 24 h after injection.
Analysis of astatine-211 irradiated thyroids revealed distinct gene expression profiles compared to non-irradiated controls, with more differentially expressed transcripts at lower absorbed doses (1225 and 1636 transcripts for 0.05 and 0.5 Gy, respectively). Down-regulation of gene expression was more frequent at lower and up-regulation at higher absorbed doses, with an intermediate phase around 1.4 Gy. A high amount of differentially expressed transcripts was unique for 0.5 and 1.4 Gy. Less than 7.5% of the transcripts detected at 0.05 11 and 32 Gy were unique. Of the 130 transcripts differentially expressed in all irradiated groups, 43 transcripts were up-regulated for 1.4 Gy, but down-regulated in all other groups. Biological processes triggered for all absorbed doses were related to transport, metabolism and muscle contraction. Briefly, high absorbed doses affected genes involved in cytoskeleton reorganisation, while low doses produced radiation-induced changes involved in cellular maintenance and metabolism. Interestingly, the type of cellular response to astatine-211 exposure marked a transition phase at 1.4 Gy.
Corresponding results after iodine-131 exposure: Compared with non-irradiated controls, 497 transcripts were altered for 0.85 Gy (more down), 546 transcripts for 8.5 Gy and 90 transcripts for 17 Gy (more up). Of these, 17 transcripts were affected in all irradiated groups. Totally, 36, 31 and 41 biological processes were affected for 0.85 8.5 and 17 Gy, respectively; only one was found in all groups. The cellular response to radiation strongly differs with absorbed dose. For 0.85 Gy, anti-apoptosis, signal transduction and transcription were affected, while 17 Gy affected metabolic functions, immune response and chemotaxis. Inflammatory response was seen for both 8.5 and 17 Gy. Effects on cell cycle arrest, muscle contraction and biogenesis were seen for 8.5 Gy alone. Influence on cell growth, cell adhesion and negative regulation of apoptosis were seen for both 0.85 and 8.5 Gy. Briefly, low absorbed dose activated cellular repair mechanisms, whereas higher doses activated cellular protection mechanisms such as repair and apoptosis.
The response was very different for astatine-211 and iodine-131. In general, irradiation by astatine-211 affected the expression of more genes than by iodine-131. Iodine-131 affected genes involved in cell cycle functions and inflammatory processes, while astatine-211 affected genes involved in transport processes. The dose response relationships were also clearly different between the radionuclides. In conclusion, the results indicate that astatine-211 and iodine-131 irradiation of the thyroid affect the cell signalling pathways differently, with more dramatic effects obtained from astatine-211 than from iodine-131.
Radiobiological effects of astatine-211 and iodine-131 on kidney, liver, spleen and lung in mice studied by gene expression analysis
In a similar way as described above BALB/c mice were i.v. injected with astatine-211 or iodine-131. Kidneys, liver, lungs and spleen were removed after 24 h and total RNA was extracted and analysed. Results for iodine-131: A surprisingly strong response was observed despite the low absorbed doses delivered (0.1-9.7 mGy). The number of transcripts affected was lowest for kidney cortex (260) and highest for lung (857). Most affected transcripts were specific for the different absorbed doses and few were found in more than one tissue. For the transcripts affected at all dose levels the response was in general independent of dose and only a few transcripts showed increasing or decreasing regulation with increasing absorbed dose. Affected biological processes were primarily associated with the normal functions of the each tissue. Immune response only was affected in all tissues and processes related to metabolism were affected in several tissue types. Evaluation of the results for astatine-211 is on-going. Briefly, the response was very different for the two radionuclides. In general, astatine-211 affected the expression of more genes than iodine-131 compared to non-exposed mice. Furthermore, the two radionuclides affected expression of different genes involved in different biological processes.
The relative biological effects and effectiveness of alpha emitters
Studies on relative biological effectiveness (RBE) should compare two radiation types while keeping parameters like time of exposure and dose rate constant. Such studies are very difficult to perform with radiopharmaceuticals due to the limited number of radionuclides emitting high and low-LET radiation, radiolabelling techniques and compounds available. Preferably, the same compound should be used and labelled with two radionuclides with similar half-life and labelling technique, while the tumour binding properties are retained. Furthermore, different end-points will give different results.
The focus has been on studies with alpha-emitting radionuclides and to get data on therapeutic effects and toxicity. Few corresponding studies on reference radionuclides were done, partly due to lack of suitable compounds. In the future, other less strict RBE-like parameters need to be defined for appreciation of the higher biological effects of alpha-emitting radionuclides.
Radiobiological effects of bismuth-213-MAb on cancer cells studied by quantification of gamma-H2AX
Cytotoxicity of alpha-particles is particularly based on induction of DNA double-strand breaks (DSBs). Methods detecting DSBs might be useful to estimate absorbed doses in tumour and normal tissue. At sites of DSBs, histones H2AX are phosphorylated to gamma-H2AX that can be detected by immunofluorescence, flow cytometry and Western blotting. Cultured EJ28 bladder cancer cells were exposed to different activity concentrations of 213Bi-anti-EGFR-MAb for three h and gamma-H2AX was quantified at different times after exposure. The number of gamma-H2AX foci per cell and the number of gamma-H2AX foci positive cells increased with increasing 213Bi activity concentration. Incubation at either 4°C or 37°C unexpectedly gave similar results. The number of gamma-H2AX foci only gradually decreased with time after incubation. Because gamma-H2AX directs repair factors to the sites of DNA-DSBs and rapidly becomes dephosphorylated after repair, the results suggest that DSB repair is impaired in EJ28 cells. Quantification of DSBs via gamma-H2AX foci might allow for estimation of radiation dose and cytotoxicity.
Radiobiological effects of bismuth-213-MAb on cancer cells in vitro determined by gene expression analysis
Whole genome gene expression profiling was performed at 6, 24 and 48 h after incubation of HSC45-M2 gastric cancer cells with bismuth-213-d9MAb. Irradiation induced both up- and down-regulation of approximately 1,000 genes at each time point analysed. Eight genes appeared up-regulated and 12 genes down-regulated throughout. Among the genes showing continuous up-regulation, COL4A2, NEDD9 and C3 have not been associated with cellular response to high LET radiation so far. The same holds true for WWP2, RFX3, HIST4H4 and JADE1 that showed continuous down-regulation. The consistently up-regulated (ITM2C, FLJ11000, MSMB) and down-regulated (HCG9, GAS2L3, FLJ21439) genes have not been associated with any biological process or molecular function so far. Thus, these findings revealed interesting new targets for selective elimination of tumour cells and new insights regarding response of tumour cells to alpha-emitter exposure.
Effects of dose-rate and hypoxia
Hypoxic cells within tumours are radioresistant, but high LET alpha-particles are assumed to damage cells independently of their oxygenation status. Therefore, the cytotoxicity of bismuth-213-anti-EGFR-MAb was compared with that of photon irradiation at both normoxic and hypoxic conditions in squamous epithelium carcinoma CAL33 cells. The hypoxia marker HIF-1alpha was assayed via Western blotting at different time points after incubation in the hypoxia chamber. CAL33 cells were incubated with bismuth-213-anti-EGFR-MAb (0-1.48 MBq/ml) or irradiated with photons (0-12 Gy) under normoxic or hypoxic conditions. Cell survival, analysed by clonogenic assay and WST colorimetric assay, decreased with increasing activity concentration (bismuth-213) or dose (photons). After photon irradiation survival of hypoxic cells was significantly higher than of normoxic cells. In contrast, similar survival was found in normoxic and hypoxic cells after incubation with 213Bi-anti-EGFR-MAb. Thus, alpha-emitters such as bismuth-213 are effective tools for eradication of hypoxic tumour cells.
Maximum tolerated activities for alpha-emitting vectors
Toxic effects were obtained in the liver using bismuth-213-labelled antibodies given i.v. while other bismuth-213-labelled antibodies gave bone marrow toxicity after i.p. administration. For smaller peptides toxicity was found for the kidneys and bone marrow, with MTAs of 25 MBq of bismuth-213 and 20-30 kBq of actinium-225. With a time period allowing recovery of side effects more fractions could be given without acute toxicity.
Optimal therapy strategies and protocols
Based on the results of the TARCC project some strategies for finding optimal treatment protocols can be drawn. The biodistribution of astatine-211-, bismuth-213- and actinium-225-labelled radiopharmaceuticals studied were very different and the critical organs should be defined for each radiopharmaceutical. The dosimetry for the critical organs must be further studied, if possible on the cellular level for some organs and the maximum tolerated activity or tolerance dose should be found. Based on data from the project we can conclude that at least the following critical organs are found for i.v. or i.p. administration: thyroid and bone marrow for astatine-211-labelled compounds, liver and bone marrow for bismuth-213-labelled proteins and peptides, kidneys for actinium-225-labelled proteins and kidneys for bismuth-213- and actinium-225-labelled peptides. For intracavitary administration other critical organs should also be considered. Thus, techniques blocking thyroid and kidney uptake should be used when appropriate. Furthermore, the promising results showing that alpha particles can eradicate also hypoxic tumour cells indicate that therapy with alpha emitting radiopharmaceuticals should be considered also for patients with reduced tumour perfusion and hypoxia.
Conclusions:
From the results obtained in the TARCC project, it could be derived that alpha-radionuclide conjugates are superior to beta-radionuclide conjugates in treatment of minimal disease and disseminated tumour cells. Many different models have been explored, in both preclinical experiments to demonstrate anti-tumour efficacy and toxicity and in in vitro models that explored more fundamental aspects of alpha radionuclide therapy. Efficacy was clearly established in models of minimal disease, as expected, with minimal toxicity. It was also confirmed that the cytotoxic effect of alpha particles is independent of hypoxia and that response of the cells to this particular kind of irradiation differs from that observed with low linear energy transfer radiations (photons, electrons). These results clearly demonstrate that alpha radionuclide therapy has the potential of killing even the last tumour cell, including cancer stem cells, which are known to be resistant to chemotherapy and classical radiotherapy. This aspect will certainly be the objective of future research.
Future clinical trials could be developed with a maximum safety, despite of the high intrinsic toxicity of alpha particles.
Ultimately, targeted alpha-radionuclide therapy appears as a promising approach that could reduce side-effects resulting from current targeted radionuclide therapy (TRT), a treatment option that is generally well tolerated giving patients a fast and clear improvement in quality of life.
Project website: http://www.tarcc.org
Email: contact@tarcc.org
Work package one (WP1): Radionuclide chemistry
One important point for the development of the field of targeted alpha therapy is the reliable support with alpha emitting radionuclides and the establishment of robust and fast labelling protocols for labelling of potent antibodies and peptides with these radionuclides. Therefore one important objective was the optimisation of the production of alpha-emitting radionuclides (e.g. actinium-225, astatine-211 and bismuth-212) and the development of new methods for radiolabelling biomolecules (e.g. aminoacids, peptides or antibodies) with alpha-emitting radionuclides. The second objective was the evaluation of the stability of radioimmunoconjugates. Therefore the sensitivity of alpha-emitting radiopharmaceuticals to self-radiolysis was determined and ways of reduction of that was investigated. The third objective was the development and evaluation of in situ generators in the respect of applicability in targeted alpha therapy.
Therefore, the main goals were to:
1. improve the availability of a-emitting radionuclides (actinium-225/bismuth-213 generators, thorium-228/lead-212/bismuth-212 generators, astatine-211)
2. improve yields, specific activity and description of optimised radiolabelling protocols
3. improve the yield of astatine-211 labelling and stability of astatine-211-labelled products
4. improve the quality of a-radionuclide radiopharmaceuticals (radiochemical purity, protection from radiolysis)
5. study longer half-life radionuclides: in situ generators (e.g. lead-212), including encapsulation in nano-colloidal vectors.
WP2: Vectors
The major objective of this WP was the development and evaluation of new vectors for well defined targets. These targets should be relevant, easily accessible, specific (or at least highly expressed) and highly expressed. In addition, the vectors should have different pharmacokinetics matching the physical half-lives of the different alpha-emitting radionuclides (46 min to 10 days).
Several families of vectors were addressed and developed, namely peptides, hybrid peptides and amino acids. Labelling protocols (along with WP1) were developed with the aim to be translated into the clinic. In vitro pharmacologic studies such as binding affinities, cell uptake (binding and internalisation) in culture and cell retention (efflux) were performed. The best performing vectors were studied in available animal models (within and along with WP3).
WP3: Animal models for preclinical alpha radionuclide therapy
Models for human hematologic and solid tumours relevant to the clinical situation have been established for optimisation of alpha-radionuclide therapy depending on the characteristics of the vector and the mode of application. These strategies aimed at a high extent of tumour targeting combined with high therapeutic efficacy and low overall toxicity.
Growth of tumours was evaluated by bioluminescence imaging following transfection of cells with firefly luciferase. Tumour targeting was evaluated by biodistribution studies, scintigraphic imaging and autoradiography. The collected data served for dosimetry calculations.
Therapeutic efficacy was investigated by monitoring tumour growth with ultrasound, bioluminescence imaging or magnetic resonance imaging (MRI) and particularly by survival of animals.
Target organ toxicity according to the targeting modality and vectors were defined and long-term toxicity studies were performed (bone marrow, liver, kidneys). Bone marrow toxicity was evaluated by blood cell counts and analysis of chromosomal aberrations in bone marrow cells. Liver toxicity was examined via liver specific enzyme activities. Urea and plasma creatinine levels were used for determination of kidney toxicity. Moreover dose-limiting organs like kidney and liver were examined histopathologicaly. Therefore, the main goals were to:
1. evaluate targeting, therapeutic efficacy and toxicity following systemic application of peptide or antibody alpha-emitter conjugates in hematologic diseases like multiple myeloma
2. evaluate targeting, therapeutic efficacy and toxicity following systemic application of phenylalanine peptide or antibody alpha-emitter conjugates in gliomas, prostate cancer and colon cancer
3. evaluate pretargeting strategies in colorectal tumours
4. evaluate targeting, therapeutic efficacy and toxicity following intracavitary application of phenylalanine, peptide or antibody alpha-emitter conjugates (intratumoural in gliomas, intraperitoneal in gastric cancer and breast cancer, intravesical in bladder cancer).
WP4: Dosimetry and radiobiology
Important objectives of the project were dosimetry and radiobiology. The absorbed dose (radiation dose) to the tumour and normal tissues are fundamental to be able to understand and to interpret results such as therapeutic effectiveness and toxicity obtained from experimental tumour treatment studies. Much work on dosimetric methods has previously been performed for radionuclides emitting electrons and photons. For alpha particle emitting radionuclides much less effort has been made in establishing methods and models for the relatively short-ranged alpha particles. More knowledge is desired concerning the need of microdosimetric approaches in some tissue types related to the biodistribution of the radionuclide within the tissues. Concerning radiobiology, a higher biological effectiveness of alpha particles compared to electrons and photons is expected, the biological effects and toxicity towards critical organs are studied and the maximum tolerated activities and doses should be defined. The goals of these studies were to:
1. study energy delivery for cell experiments through a microdosimetric approach
2. establish animal dosimetric models adapted to mice and rats
3. establish a detailed tissue dosimetry model
4. study the dose-effect relationship for major risk organs
5. assess the biological effects and relative biological effectiveness for cell and animal experiments
6. investigate the effects of dose-rate and hypoxia in tumour cell experiments
7. determine the maximum tolerated activities and doses by toxicity studies.
Ultimately, optimal treatment protocols for future preclinical and clinical studies of targeted alpha-radionuclide therapy will be proposed from the results of the dosimetry and radiobiology studies.
Project results:
Radionuclide chemistry
One of the main goals of WP1 was the establishment of the astatine-211 production. Until now the astatine-211 is routinely produced in small amounts and gives a clear perspective for the large scale production. For the separation of astatine-211 the dry distillation (which seems to be more suitable for routine) and wet extraction have been performed. In the frame of the development of novel astatine-211 labelling techniques different reaction ways were evaluated. The radiolabelling of peptides with astatine-211 has been performed with electrophilic and nucleophilic substitution reactions, latter reaction leads to higher yields and fewer by products. The collection of available published data and new experimental developments permitted Centre National de la Recherche Scientifique (CNRS) to determine the Pourbaix Diagram of astatine in non-complexing medium which is the basis of understanding of astatine-211 labelling chemistry.
Astatine-211 is produced using the Bi-209(alpha,2n) reaction. At Arronax, the alpha beam is delivered at a fixed energy. It is then necessary to install a beam energy degrader to decrease the energy from 68 MeV down to 28.3 MeV. This energy has been calculated taking into account the beam energy spreading as well as the production of the astatine-210 which decays mainly to polonium-210, an alpha emitter with long period (138 days), which is a bone seeker. At this energy, an astatine-210/ astatine-211 ratio below 0.01% must be obtained, which is the usual limit for mother-nuclides in eluates from generator-based radionuclides for medical applications. The target is obtained by evaporation under vacuum of high purity bismuth. The support is made of aluminium nitride (AlN), a ceramic which as good thermal properties. A typical target has an area of 14 cm2. Its thickness of 40 µm can be obtained within a few hours. The homogeneity of the deposit is checked using profilometry. For the irradiation, the target is tilted at 15° in order to lower the power density on target. At the end of bombardment, the target is sent to hot cells for treatment through a rabbit system. The first irradiation has not yet been performed but it will start as soon as the energy degrader is installed (September 2011).
The use of astatine in hypervalent state was proposed as an alternative radiolabelling method to succinimidyl-astatobenzoate (SAB). The feasibility of the radiolabelling was first demonstrated by the use of a simple aryltin precursor. The method was then transferred to a bifunctionalised precursor allowing the introduction of hypervalent astatine and the coupling to proteins. These precursors were composed of a tin group on a hexafluorocumyl alcohol allowing the introduction of astatine and the formation of the hypervalent bonds. A maleimide was introduced for the protein coupling with the sulphydryl groups. A two-step procedure was set up to obtain the hypervalent astatinated compounds. In the first step, the tin precursor in methanol/acetic acid was heated at 100°C over 30 min in the presence of astatine-211 and N-chlorosuccinimide to give the monovalent astatinated species in good yield (90-95%). N-bromosuccinimide in chloroform was then added and the hypervalent species were obtained after 30 min at 60°C (95% yield).
The bifunctionalised compound was coupled to bovine serum albumin for in vitro stability evaluation. The radiolabelled BSA was incubated over 12h in human serum and no deastatination was observed (stability more than 99%). Manual and automated syntheses of the SAB prosthetic group from a tin precursor have also been improved.
Vectors
Amino acids
Astatine-211-labelling of phenylalanine and phenylalanine-analogues
Astatine-labelling by the Cu+-assisted nucleophilic halogen exchange method is highly reproducible and provides adequate yields for experimental and, eventually, clinical applications. 2- and 4-[Astatine]-Phe are efficiently and specifically transported by the pertinent amino acid transport systems. 4-[Astatine-211]-Phe is the most promising derivative to be further evaluated for intracavitary experimental glioma treatment.
Peptides
Somatostatin analogues
Somatostatin receptors are overexpressed on a variety of human tumours, in particular neuroendocrine tumours. Somatostatin-based agonists are successful in the clinic as imaging and targeted radionuclide agents. Agonists which internalise when bound to their receptors were considered as ideal radiovectors because of the long retention in the tumour. The rationale to study radioantagonists was the fact that most G-protein coupled receptor antagonists recognise many more receptor binding sites compared to agonists. The idea was to have very high early uptake which matches with the half-lives of short-lived positron- and alpha emitters. It finally and unexpectedly turned out that most of our antagonist studied showed not only higher uptake in the tumour but also longer retention which now allows the use of broad spectrum of half-lives. A variety of peptides with antagonistic properties were developed along with our colleagues from the Salk Institute, La Jolla (Dr. Jean Rivier) and the University of Berne (Prof. Jean Claude Reubi). The synthetic strategy was solid phase peptide synthesis. The unnatural amino acids (D-Aph(Cbm) and Aph(Hor)) were synthesised according to published procedures or peptides containing these amino acids were received from the Salk Institute. As shown below the octapeptides had a change in chirality of the aa-1 and aa-3 compared to the well known octreotide based agonists. The peptides were characterised by ESI-MS and their purity confirmed by HPLC. About 20 new peptides were synthesised and 12 studied thoroughly in vitro and in three mice models.
Somatostatin-based antagonists with high hydrophilicity and high affinity
JR11 and ANT1 were synthesised by partner two (UBH), Radiological Chemistry Division, according to an internal good manufacturing practices (GMP) protocol, careful analysis followed (residual solvents, TFA-content) and acute toxicity studies were performed. This procedure was accepted by the local ethics committee (Basel and Freiburg) and allowed first clinical studies using indium-111-DOTA-ANT1. The data although very promising indicated that improved affinity and higher hydrophilicity may help to improve pharmacokinetics even further. We finally came up with the three new compounds shown below. They are hydrophilic (compared to [Indium-111]-DOTA-ANT) and show improved somatostatin receptor two affinity.
The preclinical data, in particular with Indium-111-DOTA-JR11 (manuscript in preparation) encouraged us to start early human use studies at the university hospital Freiburg (partner 11), department of Nuclear medicine (Prof. W. Weber; Dr D. Wild). The improvement of DOTA-JR11 was also labelled with lutetium-177 and studied in comparison with Lutetium-177-DOTA-TATE in three patients. Tumour uptake was 2-4-fold higher with the antagonist; kidney uptake was also somewhat higher but the therapeutic window is still distinctly wider for the antagonist Indium-111-DOTA-JR11.
The pharmacokinetics of the antagonists is fast; they can be labelled with bismuth-213 and actinium-225 using very similar protocols than those for indium-111, yttrium-90 and lutetium-177. First clinical studies with bismuth-213 were planned along with the department of Nuclear Medicine, University Heidelberg and partner eight (JRC-ITU). The studies have been delayed due to the lack of GMP-JR11.
These conjugates were labelled with the radiometals mentioned above and biodistribution was studied with the positron emitters copper-64 and gallium-68. (see Fani M et al , 'PET of Somatostatin Receptor-Positive Tumours Using Cu-64 and Ga-68-Somatostatin Antagonists: The Chelate Makes the Difference' J. Nucl. Med. 2011;52:1110-1118)
Precursor synthesis for chemical ligation and astatine-211-labelling
Synthesis of precursors for selective astatine-211-labelling of peptides (Substance P)
Substance P is a ligand binding to neurokinin type one (NK-1) receptors, which are highly overexpressed in malignant gliomas. A few potential peptide vectors were recently developed. Two of these peptides, DOTAGA-substance P and DOTA-[Thi-8, Met(O2)-11]-substance P were shown to be suitable for diagnostic labelling with indium-111 and therapeutic labelling with yttrium-90, lutetium-177 and bismuth-213. The modification of substance P (introduction of Thi-8 for Phe-8 and Met-11 replaced by the corresponding sulfone) led to improved radiolytic and improved metabolic stability.
The DOTA-conjugated substance-P derivative was successfully studied in the clinic in different brain tumour patients, labelled with bismuth-213 (Eur J Nucl Med Mol Imaging, 2010;37(7):1335-44).
Astatination of substance P derivative by selective chemical ligation
The idea of the project is to find a route for routine application of astatine-211-labelled substance P derivatives. The strategy is to conjugate astatine-211-labelled benzaldehyde onto an aminooxy functionalised [Thi-8,Met(O2)-11]-substance P via oxime formation.
The advantage of an aminooxy functionalised peptide is the high reactivity with carbonyl functionalities and the resulting stable oxime formation. In addition, the major advantage is that a carbonyl-moiety can be selectively linked in the presence of an aminooxy group due to pKa differences of the aminooxy functionalised compound (pKa µ 5) and the amino functionalised compound (pKa µ 11). This means that at a low pH of 3-4, the amino group is fully protected by the proton whereas the aminooxy group is still reactive. The conjugation kinetics was optimised using 4-Fluorine-19-benzaldehyde.
On the other hand the problem with the aminooxy precursor is its high reactivity with all kind of carbonyl compounds such as acetone in solvents and formaldehyde from the air. This all had to be understood and optimised for improved synthesis. Attempts to introduce 'weak' protecting groups which are being cleaved during the conjugation step were not successful.
Development of metabolically stabilised GASTRIN derivatives (partners 2/11, 6)
Non-ionic and mixed ionic/non-ionic spacer DOTA-gastrin-conjugates as possible precursors for SST/CCKB receptors binding hybrid
To date there is no minigastrin analogue suitable for radiopeptide therapy of medullary thyroid carcinoma and other gastrin receptor expressing tumours such as neuroendocrine tumours and SCLC. The main limiting factor is either their low tumour uptake or/and too high kidney retention which has been related to the N-terminal Glu-s ; deletion of these Glu-s lowered the kidney uptake but the metabolic stability deteriorated. The goal of this work was to develop radiogastrin derivatives with high metabolic stability, high affinity and low kidney uptake. The strategy was to replace the hydrophilic, negatively charged spacers (Glu-s) with hydrophilic but uncharged spacers.
In addition, DOTA-conjugated minigastrin analogues with so called mixed ionic/non-ionic spacers were synthesised manually on solid phase. Gastrin receptor affinities were determined by receptor autoradiography in cooperation with Prof. Reubi, university of Berne. 111In-labelled peptides were evaluated in vitro in AR4-2J cell line and in human serum. The best peptides were studied in rats bearing the AR42J tumour.
We developed DOTA-coupled minigastrin analogues with ionic, non-ionic and mixed ionic/non-ionic spacers with high stability in human serum. All conjugates showed high binding affinity to the gastrin receptor (IC-50-values between 0.5 and 4.8 nM). All radiopeptides showed receptor-specific internalisation. The metabolic stability remains high within the series. Feasibility of labelling with bismuth-213 was confirmed. Further biodistribution studies in animal models are needed in order to determine which type of spacer is superior in relation to best tumour-to-kidney ratio.
Synthesis of hybrid peptides (partner 2/11, 6) and chelated F3 (partner 3)
Synthesis of hybrid peptides
The first step of the development of hybrid peptide was the selection of somatostatin and gastrin synthons. TOC as the most widely used somatostatin analogue and PP-F10 as one of the most promising minigastrin analogues were selected. Several attempts to couple somatostatin and CCK2/gastrin synthons were performed.
Synthesis of chelated F3 (Partner TUM): 213Bi-and 225Ac-labeled substance peptide F3
The tumour homing peptide F3 is effectively internalised into the nucleus of tumour cells upon binding to nucleolin and therefore can be used as an appropriate carrier for alpha-particle emitting isotopes in targeted tumour therapy. We have successfully labelled DTPA-chelated F3 peptide with the alpha-emitter bismuth-213 and demonstrated therapeutic efficacy of bismuth-213-DTPA-F3 conjugates in a nude mouse model of peritoneal carcinomatosis. In our effort to expand the study we have now labelled DOTA-chelated F3 with the alpha-emitter actinium-225 and compared cytotoxicity and therapeutic efficacy of both conjugates. ID50 values of bismuth-213-DTPA-F3 (53 kBq/ml) and actinium-225-DOTA-F3 (67 Bq/ml) - as determined via clonogenic assays - differed by a factor of 1,000. Repeated therapy with six x 1.85 kBq actinium-225-DOTA-F3 or six x 1.85 MBq bismuth-213-DTPA-F3 prolonged median survival to 95 days and 97 days, respectively, with negligible kidney damage (see WP3) (manuscript in preparation).
Antibody production and characterisation
Monoclonal antibodies directed against murine CD138 have been generated by immunisation against peptides extracted from the murine CD138 sequence. The peptides were chosen in an area similar to that recognised by the anti-human CD138. Nine hybridomas recognised murine multiple myeloma by flow cytometry. Subcloning was performed and affinity tests are in nM range.
Bivalent hapten synthesis and labelling protocols
Bivalent haptens
By definition a hapten is a low molecular weight (generally less than 1 000 Da) compound that is recognised by specific antibodies. Usually, haptens are first coupled to carrier proteins, such as serum albumin or keyhole limpet hemocyanin, to form an immunogenic conjugate capable of triggering an antibody response in animals. Then monoclonal antibodies may be obtained by classical hybridoma technology. The DTPA-indium complex has been used in earlier preclinical and clinical work by partner one (INSERM) and collaborating teams. The histamine-succinyl-glycine (HSG) pseudo-peptide has also been proposed as a useful hapten in the context of pretargeting and interest has recently shifted to this hapten because it may be used to derive pretargeting vectors that may be labelled with a variety of radionuclides. Bivalent haptens have been shown to be superior to monovalent ones for in vivo pretargeting of tumours. Small molecular weight molecules comprising two HSG haptens and one or more sites for radioactive labelling have been designed and synthesised.
Radiolabelling of bivalent haptens
Another bivalent HSG hapten, provided by the Immunomedics Company (Morris Plains, NJ, USA) was used in the project. It was radiolabelled with radioactive iodine (iodine-125) for biodistribution studies and with lutetium-177 and bismuth-213 for preclinical therapy studies (Griffiths GL, Chang CH, McBride WJ, Rossi EA, Sheerin A, Tejada GR, Karacay H, Sharkey RM, Horak ID, Hansen HJ, Goldenberg DM. Reagents and methods for PET using bispecific antibody pretargeting and 68Ga-radiolabeled bivalent hapten-peptide-chelate conjugates. J Nucl Med. 2004, 45: 30-39).
Animal models for preclinical alpha radionuclide therapy
Multiple myeloma
Systemic application of alpha-emitter conjugates in hematologic disease a syngeneic murine model of multiple myeloma was established by intravenous transfer of 5T33 myeloma cells that disseminated into the bone marrow. Syndecan-1 (CD138), a heparan sulfate proteoglycan, constantly expressed on tumour cells in multiple myeloma served as surface antigen in targeted therapy. Mice were treated with bismuth-213-labeled anti-mCD138 at four different activities (1.85 3.7 7.4 and 11.1 MBq). The group treated with 3.7 MBq exhibited a median survival of more than 300 days compared with 45.5 days for the control group. The highest activity (11.1 MBq) turned out toxic while the lowest activity of 1.85 MBq showed no therapeutic effect. 7.4 MBq induced a median survival of 227 days. With activities of 3.7 and 7.4 MBq, surviving mice exhibited a transient haematological toxicity and only the injected activity of 7.4 MBq caused a transient rise of Flt-3L indicative of low myelotoxicity.
Prostate cancer
In a human androgen-independent prostate carcinoma model (PC-3-tumour model) therapeutic studies were undertaken using two novel bismuth-213 or lutetium-177 labelled bombesine-based peptides (DOTA-PEG4-bombesin [DOTA-PESIN] and DO3A-CH2CO-8-aminooctanoyl-Q-W-A-V-G-H-L-M-NH2 [AMBA]). For therapy and toxicity studies, mice were injected with lutetium-177-DOTA-PESIN, bismuth-213-DOTA-PESIN or bismuth-213-AMBA, whereas control groups were left untreated or were given nonradioactive lutetium-175-DOTA-PESIN. All radiopeptides showed a rapid blood clearance and tumour uptake. The maximum tolerated dose (MTD) of bismuth-213-DOTA-PESIN and bismuth-213-AMBA was 25 MBq and the MTD of lutetium-177-DOTA-PESIN was 112 MBq, corresponding to kidney doses of 6-11 Gy. At these dose levels, lutetium-177-DOTA-PESIN showed no, bismuth-213-DOTA-PESIN slight and bismuth-213-AMBA marked kidney damage. At MTD, bismuth-213-DOTA-PESIN and bismuth-213-AMBA were significantly more effective than lutetium-177-DOTA-PESIN. These preclinical data show that alpha-therapy with bismuth-213-DOTA-PESIN or bismuth-213-AMBA is more effective than therapy with beta-particle emitting lutetium-177-DOTA-PESIN. In addition, therapy with bismuth-213-DOTA-PESIN has a better safety profile than treatment with bismuth-213-AMBA and represents an important new approach for treating recurrent prostate cancer.
Colorectal cancer
Subcutaneous colorectal µtumours or hepatic µmetastases were established for therapeutic studies using the affinity enhancement system (AES) technology comprising non radioactive bispecific antibodies and a radiolabelled (bismuth-213 or astatine-211) bivalent hapten.
LS174T human colonic carcinoma cells (1 million) were inoculated subcutaneously or via the portal vein (hepatic metastasis model) in NMRI-nu mice. The tumour implantation rate was 100% using aggregated cells and only 26.6% with isolated cells. Photons emitted by one million cells were detected by bioluminescence imaging immediately following injection and allowed visual confirmation of hepatic distribution. Therefore, hepatic tumour graft via the portal route and bioluminescence imaging provides a reliable animal model and permits sensitive in vivo detection and follow-up of hepatic metastases. The hepatic model appears to reproduce more closely colon cancer spread compared to the subcutaneous one. The hepatic model is of particular interest to study radioimmunotherapy.
Pharmacokinetic distribution of the TF2 bispecific anti-Carcino-Embryonic-Antigen (CEA) x anti histamine-succinyl-glycine (HSG) antibody was evaluated. For various antibody / hapten molar ratios (10:1 and 2:1) and pretargeting time intervals (5, 10, 15, 24 h), biodistribution studies were performed in the subcutaneous model with TF2 and Iodine-125-labeled bivalent HSG peptide IMP-288 to optimise tumour uptake and tumour / organ uptake ratios.
Activity blood clearance was very fast. Using an antibody / peptide ratio of 10:1 and a pretargeting time interval between 15 and 24 h, the peak of tumour uptake reached 9.13 ± 4.4 % injected dose per gram (%ID/g) in subcutaneous tumours two h after hapten injection. Reducing the time interval between antibody and peptide injections increased tumour uptake, but activity uptake also increased in most tissues, particularly in liver and spleen. Tumour uptake 1h after hapten injection was higher in hepatic metastases (11-26% ID/g) than in subcutaneous tumours (7-12% ID/g). The results of the biodistribution studies and dosimetric calculations revealed that pretargeting must be finely tuned for therapeutic application.
For pretargeted radioimmunotherapy IMP-288 targeting TF2 was labelled with bismuth-213 to a specific activity of 74 MBq/nmol. Seven days after tumour inoculation, groups of four animals were selected with matched bioluminescence signals. Four control groups were set: no treatment, TF2 alone, 0.2 or 0.1 nmol of labelled peptide only. Two treatment groups were set: pretargeting with two nmol of TF2 followed by 0.2 or 0.1 nmol of bismuth-213-labelled peptide IMP-288 with a pretargeting time interval of 18 h. Haematological toxicity was monitored by weekly blood sampling. Tumour growth was monitored by in vivo bioluminescence imaging (every four to six days). Animals were sacrificed when a 20% weight loss or cachexia was observed or when the bioluminescence signal reached 20x107 counts. Increase of the bioluminescence signal was delayed in treated animals (TF2 + bismuth-213-labeled IMP-288) with tumour growth starting 10 days after treatment whereas tumours grew continuously in control groups. At day 23 after treatment, the lowest bioluminescence signal was observed in the group treated with two nmol TF2 and 0.1 nmol of labelled IMP-288. Animals were euthanised when the bioluminescence signal reached 20x107 photons. The Kaplan-Meier curve showed 0% survival in the untreated group versus 75% in the group treated with two nmol TF2 and 0.1 nmol IMP-288. Toxicity was acceptable with a transient decrease of blood cells six days after treatment and complete recovery at day 13.
Brain tumours
Increased amino acid transport in brain tumours is used for diagnostic purposes. It has been shown that astatine-211-labelled phenylalanine (At-Phe) is taken up by glioblastoma cells in vitro. It was tested, whether systemic intravenous treatment of rats with intracranially implanted glioblastomas with At-Phe would have a beneficial effect. At-Phe was prepared via nucleophilic halogen exchange on L-iodophenylalanine. The rat glioblastoma cell line BT4Ca was implanted into the prefrontal cortex of female BDIX rats by stereotaxic µinjection (10 000 cells/3 µlitre; n=63). At-Phe (2 MBq) or phosphate buffered saline (PBS) were injected intravenously three days after implantation. A third group was treated twice, on days three and 10 after implantation. Rat?s health condition was assessed each day by using a score system (e.g. weight, pupillary and motor reflexes) and rats were sacrificed when showing a premortal health status. Additionally, rats were sacrificed on days 6, 10, 13 and 17 after transplantation to measure the tumour volume and area of necrosis. Furthermore, the proliferation index was assessed after immunohistological staining for Ki-67.
The mean survival time of the rats who had received two applications of At-Phe was significantly higher compared to control rats (less than 0.05). The score-analysis showed that At-Phe treated rats had a significant better health condition compared to PBS treated rats (less than 0.05). Analysis of the tumour volume on days 6, 10, 13 and 17 after transplantation showed no differences between groups, but rats with a premortal health status had a significantly larger tumour than rats in good health condition (less than 0.05). Overall the proliferation index was not different between groups, but necrosis was marginally larger after At-Phe treatment (less than 0.05 on day 10).
Gastric cancer
In a gastric cancer peritoneal carcinomatosis model established by intraperitoneal injection of 1x107 HSC45-M2 gastric cancer cells, expressing mutant d9-E-cadherin therapeutic efficacy and toxicity of bismuth-213-d9MAb (targeting mutant d9-E-cadherin) were evaluated in comparison to beta-particle emitting lutetium-177-d9MAb conjugates. Biodistribution data showed an uptake of bismuth-213-d9MAb in the tumour nodules up to 40% ID/g at 45 min after injection. In all other tissues the uptake of bismuth-213-d9MAb was below 3.2% ID/g.
For optimisation of radioimmunotherapy mice were injected intraperitoneally with 0.37 1.85 7.4 or 22.2 MBq of bismuth-213-d9MAb at days one or eight after tumour cell inoculation.
Application of 1.85 MBq of bismuth-213-d9MAb was most efficient in terms of survival. At the end of the observation period almost 90% of the mice were still alive and showed no signs of disease. Bismuth-213-d9MAb activities both higher and lower than 1.85 MBq proved to be less efficient concerning prolongation of survival. Therapy with bismuth-213-d9MAb at day eight was less efficient than treatment on day 1. Tumour growth as well as regression after therapy could be visualised by bioluminescence imaging.
Toxicity of bismuth-213-d9MAb therapy was negligible at therapeutically effective activities. White blood cell counts were reduced within two days after therapy but recovered to pre-treatment values at day 21. Reduction was maximal after application of 22.2 MBq and minimal after 1.85 MBq. Chromosomal aberrations in bone marrow cells could only be observed at day one after therapy. Following autopsy none of these mice showed any signs of organ toxicity. Therefore, activities of up to 7.4 MBq of bismuth-213-d9MAb are thought to cause no toxic side effects.
Treatment of mice with 1.85 7.4 and 14.8 MBq of lutetium-177-d9MAb also significantly prolonged median survival. Therapy with 7.4 MBq of lutetium-177-d9MAb was most effective: more than 50% of the animals survived longer than 250 d. Therapy with 1.85 7.4 and 14.8 MBq of lutetium-177-d9MAb at day eight after tumour cell inoculation also resulted in statistically significant improvement of median survival. Intraperitoneal injection of lutetium-177-d9MAb induced a transient reduction in leukocyte counts at activities more than 1.85 MBq. Injection of 7.4 MBq and 14.8 MBq of lutetium-177-d9MAb rapidly decreased leukocyte counts to 40% at day 10 and 6% at day 17 of pre-treatment values. Leukocyte numbers did not recover to pre-treatment values after application of 14.8 MBq. Following i.p. injection of mice with 7.4 and 14.8 MBq lutetium-177-d9MAb 3% and 13% of the mice, respectively, developed a massive cervical, axillary and femoral-inguinal lymphadenopathy confirmed as lymphoblastic lymphoma by histological examination. One animal with enlarged lymph nodes had developed a plasmocytoma. Furthermore the spleen of this animal showed a follicular hyperplasia, the kidney a severe proliferative glomerulonephritis and the liver a hepatocarcinoma. Comparison of the results obtained with lutetium-177-d9MAb and bismuth-213-d9MAb demonstrate therapeutic efficacy for both radionuclides. However, lutetium-177-d9MAb turned out to be much more toxic at therapeutically effective doses because of long-lasting circulation of the activity in the blood after resorption from the intraperitoneal cavity.
In order to optimise intraperitoneal radioimmunotherapy, intraperitoneal retention and biodistribution of Fab-fragments, intact IgG and IgM coupled to indium-111 or bismuth-213 were analysed in tumour free mice after i.p. injection. The results indicate a rapid resorption of Fab-fragments from the peritoneal cavity resulting in a high kidney accumulation. IgG remained in the circulation for more than 72 h after resorption from the peritoneal cavity. IgM, in contrast showed a long-lasting retention in the peritoneal cavity. After resorption IgM is rapidly trapped in the liver resulting in a very low activity in the circulation at all time points. These results could be confirmed by scintigraphic imaging. Therefore high molecular weight IgM antibodies seem to be the preferable radionuclide carriers for intraperitoneal injection.
Breast cancer
A dimer of the vascular tumour homing peptide F3 was chemically coupled to the alpha-emitter bismuth-213 (bismuth-213-DTPA-[F3]2). It was found that bismuth-213-DTPA-[F3]2 accumulate in the nucleus of tumour cells in vitro and in intraperitoneally growing tumours in vivo. To study the anti-tumour effect of bismuth-213-DTPA-[F3]2 mice bearing intraperitoneally growing xenograft tumours were treated with bismuth-213-DTPA-[F3]2. SCID-mice were injected i.p. with 10 million MDA-MB-435 cells expressing firefly luciferase. In the tumour prevention study groups of eight mice were injected every second day with 1.85 MBq bismuth-213-DTPA-[F3]2, 1.85 MBq bismuth-213-DTPA or 100 µlitre PBS between day four and 14 after inoculation of the tumour cells. In the tumour reduction study groups of eight mice were injected i.p. with 1.85 MBq bismuth-213-DTPA-[F3]2 1.85 MBq bismuth-213-DTPA or 100 µlitre PBS between the days 16 and 26 after inoculation of the tumour cells.
The survival time of the animals was increased from 51 to 93.5 days in the prevention study and from 57 days to 78 days in the tumour reduction study after treatment with bismuth-213-DTPA-[F3]2. No toxicity of the treatment was observed. The serum concentration of creatinine was determined in all tumour bearing mice when they were sacrificed after therapy. No significant elevation of the serum creatinine level after treatment with bismuth-213-DTPA-[F3]2 at time of sacrifice was observed.
In a second study stable conjugates of the vascular tumour-homing peptide F3 with actinium-225 were generated to accomplish targeted delivery into tumour cells. The aim of this study was to determine the therapeutic window of actinium-225-DOTA-F3 and to compare it with bismuth-213-DTPA-F3.
Anti-tumour activity of bismuth-213-DTPA-F3 and actinium-225-DOTA-F3 was determined in-vitro using clonogenic assays. Activities with equivalent anti-tumour effects in vitro were then used for treatment of mice with intra-peritoneal MDA-MB-435 xenograft tumours in vivo. Survival of animals was analysed and therapy monitoring was performed by optical imaging and histological analysis. Nephrotoxic effects were analysed by histology.
In vitro the ID50 of bismuth-213-DTPA-F3 and actinium-225-DOTA-F3 was determined as 53 kBq/ml and 67 Bq/ml, respectively. In vivo, therapy with six x 1.85 kBq actinium-225-DOTA-F3 or six x 1.85 MBq bismuth-213i-DTPA-F3 or PBS were compared. Median survival in the PBS group was 60 days. Median survival in the group treated with actinium-225-DOTA-F3 was 95 days and 97 days in the group treated with bismuth-213-DTPA-F3. two bismuth-213-DTPA-F3 and actinium-225-DOTA-F3 had different effects on tumour growth as determined by optical imaging. bismuth-213-DTPA-F3 reduced the tumour mass at early time points until 30 days after treatment whereas actinium-225-DOTA-F3 has a prolonged anti-tumour effect leading to a reduction of tumour mass until the death of animals. Bismuth-213-DTPA-F3 was more effective in reducing the number of metastases.
Bladder cancer
An orthotopic human bladder carcinoma mouse model using the EGFR-overexpressing luciferase transfected bladder carcinoma cell line EJ28 was successfully established by intravesical instillation of two million EJ28 luc cells / 100 µlitre after gentle electrocautery of the murine bladder. Therapeutic efficacy of intravesically instilled bismuth-213-anti-EGFR-MAb was compared with standard mitomycin C chemotherapy. Mice without therapy and those treated with unlabeled anti-EGFR-MAb reached a median survival of 41 d and 89 d, respectively. Mice that underwent therapy with 0.925 MBq of bismuth-213-anti-EGFR-MAb one h, seven d or 14 d after cell instillation survived more than 300 d in 90%, 80% and 40% of cases, respectively. The results were visualised by bioluminescence imaging. Therapy with 0.37 MBq one h or seven d after tumour cell inoculation resulted in survival more than 300 d in 90% and 50% of mice, respectively. Mitomycin C treatment after one h and seven d prolonged survival more than 300 d in 40% and 50%, respectively, however turned out to be nephrotoxic. In contrast, no signs of nephrotoxicity could be observed following bismuth-213-anti-EGFR-MAb treatment. The study suggests that radioimmunotherapy using intravesically instilled bismuth-213-anti-EGFR-MAb is a promising option for treatment of bladder cancer in patients.
In a second study therapeutic efficacy of a fractionated intravesical treatment with bismuth-213-anti-EGFR-MAb in an advanced tumour stage was evaluated. Radioimmunotherapy was initiated between days 14 and 36 after cell instillation when tumours could be clearly detected via bioluminescence imaging. The first group received three intravesical applications of 0.46 MBq of bismuth-213-anti-EGFR-Mab each in a time interval of four days. The second group was instilled with the unlabelled anti-EGFR-MAb and the third group remained untreated. Tumour development was monitored by bioluminescence imaging and survival was observed up to 300 days.
60% of the animals of the group treated three times with 0.46 MBq of bismuth-213-anti-EGFR-Mab in a time interval of four days showed a reduction of tumour size while in the untreated group the tumour sizes increased in all animals. The group receiving no treatment or unlabelled anti-EGFR-MAb showed a mean survival of 44 days after tumour cell inoculation while median survival in the bismuth-213 treated group was 88 days. The urothelia of the tumour-free animals treated three times with 0.46 MBq bismuth-213-anti-EGFR-MAb did not show any pathological findings.
Fractionated intravesical therapy with the alpha-emitter bismuth-213 coupled to anti-EGFR-MAb effectively reduces tumour size without radiation damage to the normal urothelium. Thus, fractionated therapy is a very promising approach for successful treatment of bladder carcinoma in an advanced stage.
Dosimetry and radiobiology
Microdosimetry of alpha particles for cell experiments
Two Monte Carlo (MC) codes, MCNPX and GEANT4, were setup and after validation used by two of the partners, together with conventional dosimetric methods, namely microdosimetric calculations for astatine-211 in a spherical cell and in a non-symmetrical cell; microdosimetric calculations were performed for astatine-211 using GEANT4. Two cell geometry models were tested: a regular mathematical cell model of a spherical cell with a concentric spherical cell nucleus and a voxel-based virtual cell model of a non-symmetrical cell.
Astatine-211 was located homogeneously in the nucleus, the cytoplasm, on the cell surface or outside the cell. There were differences between the low energy electromagnetic model (G4-LEM) and the standard electromagnetic model (G4-SEM) in mean specific energy and S values. Compared to published data the mean specific energy demonstrated less than 2% difference for G4-LEM, less than 6% difference for G4-SEM for sources in the nucleus, cytoplasm and on the cell surface. For astatine-211 located outside the cell differences of less than 4% and more than 10%, respectively, were found. Compared to published S values less than 6% and less than 4% differences were found for astatine-211 and less than 2% and less than 3% differences for its daughter polonium-211. Results from the non-symmetrical cell reflect the influence of irregular shape and heterogeneous medium.
Tumour control probability (TCP) for uniform and non-uniform activity distributions
In a study on cell clusters, we demonstrated how TCP depends on the radionuclide distribution in the tumour, both macroscopically and at the sub-cellular level for astatine-211, yttrium-90, lutetium-177, rhodium-103m. The radionuclides were uniformly distributed within the sub-cellular compartment and they were uniformly, normally or log-normally distributed among the cells in the tumour. In summary, the results demonstrate that the radionuclide distribution both within the cells but also within the tumour clearly affect the absorbed dose and the TCP for astatine-211 and also for the low-energy electron emitter rhodium-103m, while this was almost negligible for the other radionuclides studied (lutetium-177, yttrium-90). TCP for astatine-211 and rhodium-103m was affected by the radionuclide distribution to a great extent when the radionuclides were in the cell nucleus and to lower extent when the radionuclides were distributed on the cell membrane or in the cytoplasm. When the radionuclide was distributed in the nucleus, the cumulated activities required for TCP=0.99 increased when the activity distribution became more heterogeneous for astatine-211 and rhodium-103m, with larger increase for normally distributed radionuclides compared to log-normally distributed. When the activity was distributed on the cell membrane, the cumulated activities required for TCP=0.99 were not affected for astatine-211 and rhodium-103m when the activity distribution became more heterogeneous.
Dosimetric models for mouse and rat
Mouse dosimetry using MCNPX and GEANT4
A mouse geometric model (Bitar model) was defined based on 3D reconstruction of sequential slices of a frozen 30 g NMRI nude mouse. Calculations were compared with those of the 28 g Digimouse model. Mouse dosimetric models were developed for bismuth-213 and astatine-211. S-values were calculated from absorbed energy distributions obtained with MCNPX and GEANT4, assuming uniform activity distribution in each organ and contribution of all daughter radionuclides except bismuth-207. Absorbed fractions were high (0.92-0.99) showing that alpha-particles deliver most of their energy in the source organ. For some larger organs, self-dose S-values were similar for both mouse models (difference less than 6%). For the other organs there were differences up to 2000 %. Cross-dose S-values varied within 3-5000%. These discrepancies can be explained by the variations in organ size, shape and location between the two mouse models. Comparing GEANT4 and MCNPX, differences in S-values for bismuth-213 were less than 5% for all organs. When both self-dose and cross-dose were considered, differences of up to 40% were found, but the largest differences were found for very low S-values with low impact on total absorbed dose. For high S-values, relative differences were less than 15%.
Dosimetry of bismuth-213-BSA in mice
Application on biodistribution data: Based on pharmacokinetic data of bismuth-213-BSA in non-tumour-bearing nude mice time activity curves and cumulated activity were derived for each organ. Absorbed dose based on MC S-values were in general similar to those derived from calculations assuming that all emitted energy from particles was absorbed within the organ. The highest differences were found for lungs (8%). Comparisons between the two MC-techniques showed that the highest difference was observed for the kidneys (10% with Bitar model). By comparing the two mouse models the absorbed dose was similar in organs containing high activity (less than 4% difference), with larger discrepancies in organs containing low activity (up to 110 % difference).
Rat dosimetry
No new rat geometrical model was developed because no partner had any need for such a model. Furthermore, since rats are much larger than mice mean absorbed doses can accurately be estimated assuming that all alpha particles deliver their energy within each organ. Detailed tissue dosimetry, however, might require better knowledge in morphology and most probably specific models must be defined for each rat type and size/age, as for the mouse models.
Biodistribution and dosimetry of free astatide-211, iodide-14 and iodide-131 in rats
Low amounts of astatide-211 are released from the radiopharmaceutical in vivo. It is then important to know the biodistribution and the absorbed dose to normal tissues for astatide-211 in order to determine potential risk organs when using astatine-211-labelled radiopharmaceuticals for treatment. Very limited knowledge on biodistribution of astatide-211 is available for rats. Rats were injected simultaneously with iodide-125 and iodide-131, or free astatide-211 and the activities and activity concentrations in organs and tissues were determined and mean absorbed doses were calculated. The biodistribution of free astatide-211 was different compared to iodide-125- and iodide-131, while it was similar for 125I and iodide-131. The uptake of all three radionuclides was highest in the thyroid (400 %IA/g for iodide-125, 430 %IA/g for iodide-131 and 75 %IA/g for astatide-211 after 18 h). In all other organs and tissues the activity concentration and retention of astatine-211 was in general higher than that of iodide-125 and iodide-131. The thyroid received the highest absorbed dose per injected activity (400 mGy/MBq for iodide-125 and 19000 mGy/MBq for iodide-131 after seven days and 17000 mGy/MBq for astatide-211 after 24 h). Otherwise, the highest mean absorbed doses were found in stomach, heart and small intestine for iodide-125 and iodide-131 and in stomach, lungs and spleen for astatine-211. Astatine-211 delivered a much higher absorbed dose per injected activity to all extrathyroidal organs compared to iodide-125 and iodide-131.
Translation of data from astatine-211 in rats and mice to dosimetry of astatine-211 in humans
The possibility to translate astatine-211 biodistribution data between different species, especially from rat and mouse to human was studied, since biodistribution and dosimetric data for humans are difficult to obtain. The absorbed dose of astatine-211 and iodine-131 to different tissues in man was estimated based on extrapolation of biodistribution data obtained from rat and mouse. The highest estimated absorbed dose per unit activity administered from both species to man was obtained for the thyroid and stomach together with the pituitary (rat) and the salivary glands (mouse), all expressing NIS. The estimated effective dose to man differed by a factor of two depending on the species. The biodistribution of astatide-211 is different from that of iodide in animals and the iodide kinetic models (ICRP53) cannot be applied to astatine-211 in man. The iodine-125 values from mice are closer to data in ICRP53 than from rat. Thus, biodistribution and dosimetry data differ between species, with a clear difference between mouse and rat.
Detailed tissue models
The initial plans were to develop geometrical models for tissue scale dosimetry for several tissue types, especially for the critical organs. The thyroid model is finalised and the work with defining a model of the kidneys is ongoing but will take a longer time than expected due to the high complexity with the need to define the many various parts of the nephron involved in late side effects in radionuclide therapy.
Detailed dosimetric model of the thyroid
One of the main risk organs when using astatine-211-labelled radiopharmaceuticals is the thyroid and therefore much effort was used for this model. The high uptake of free astatide-211 in thyroid (see above) indicates that thyroid-blocking should be used when astatine-211-labelled pharmaceuticals are applied for therapy. A general thyroid follicle was constructed using spherical geometry, with a follicle lumen with a diameter of 10-500 µm, surrounded by a single layer of follicle cells with 4-8 µm (diameter) spherical nuclei. Follicle models were defined for mouse, rat and humans. Calculations were performed for alpha-particles using MCNPX. The target was the follicle cell nucleus, with astatine-211 in the follicle lumen, follicle cells or nuclei. The mean specific energy with astatine-211 within the follicle lumen was 0.17 1.1 and 2.0 mGy, for human, rat and mouse, respectively, with corresponding single-hit mean specific energy 370, 610 and 1200 mGy. Approximately 45% of the alpha particles in a astatine-211 decay deposit all the kinetic energy within the follicle lumen for the human model, compared to 7% and less than 0.5% for the rat and mouse model, respectively. The results demonstrate the importance of taking the range of the alpha particles and the spatial distribution of astatine-211 into account for dosimetric evaluation of thyroid exposure. The dosimetry for 211At in the thyroid cannot readily be translated between the species.
Radiobiological effects in major risk organs
Thyroid toxicity of astatine-211 and iodine-131 in mice studied by gene expression analysis
Female BALB/c mice were i.v. injected with astatine-211 or iodine-131 and the absorbed dose to the thyroid was 0.05-32 Gy and 0.85-17 Gy, respectively. Gene expression profiling (Illumina platform and Nexus) was performed using total RNA from thyroids removed 24 h after injection.
Analysis of astatine-211 irradiated thyroids revealed distinct gene expression profiles compared to non-irradiated controls, with more differentially expressed transcripts at lower absorbed doses (1225 and 1636 transcripts for 0.05 and 0.5 Gy, respectively). Down-regulation of gene expression was more frequent at lower and up-regulation at higher absorbed doses, with an intermediate phase around 1.4 Gy. A high amount of differentially expressed transcripts was unique for 0.5 and 1.4 Gy. Less than 7.5% of the transcripts detected at 0.05 11 and 32 Gy were unique. Of the 130 transcripts differentially expressed in all irradiated groups, 43 transcripts were up-regulated for 1.4 Gy, but down-regulated in all other groups. Biological processes triggered for all absorbed doses were related to transport, metabolism and muscle contraction. Briefly, high absorbed doses affected genes involved in cytoskeleton reorganisation, while low doses produced radiation-induced changes involved in cellular maintenance and metabolism. Interestingly, the type of cellular response to astatine-211 exposure marked a transition phase at 1.4 Gy.
Corresponding results after iodine-131 exposure: Compared with non-irradiated controls, 497 transcripts were altered for 0.85 Gy (more down), 546 transcripts for 8.5 Gy and 90 transcripts for 17 Gy (more up). Of these, 17 transcripts were affected in all irradiated groups. Totally, 36, 31 and 41 biological processes were affected for 0.85 8.5 and 17 Gy, respectively; only one was found in all groups. The cellular response to radiation strongly differs with absorbed dose. For 0.85 Gy, anti-apoptosis, signal transduction and transcription were affected, while 17 Gy affected metabolic functions, immune response and chemotaxis. Inflammatory response was seen for both 8.5 and 17 Gy. Effects on cell cycle arrest, muscle contraction and biogenesis were seen for 8.5 Gy alone. Influence on cell growth, cell adhesion and negative regulation of apoptosis were seen for both 0.85 and 8.5 Gy. Briefly, low absorbed dose activated cellular repair mechanisms, whereas higher doses activated cellular protection mechanisms such as repair and apoptosis.
The response was very different for astatine-211 and iodine-131. In general, irradiation by astatine-211 affected the expression of more genes than by iodine-131. Iodine-131 affected genes involved in cell cycle functions and inflammatory processes, while astatine-211 affected genes involved in transport processes. The dose response relationships were also clearly different between the radionuclides. In conclusion, the results indicate that astatine-211 and iodine-131 irradiation of the thyroid affect the cell signalling pathways differently, with more dramatic effects obtained from astatine-211 than from iodine-131.
Radiobiological effects of astatine-211 and iodine-131 on kidney, liver, spleen and lung in mice studied by gene expression analysis
In a similar way as described above BALB/c mice were i.v. injected with astatine-211 or iodine-131. Kidneys, liver, lungs and spleen were removed after 24 h and total RNA was extracted and analysed. Results for iodine-131: A surprisingly strong response was observed despite the low absorbed doses delivered (0.1-9.7 mGy). The number of transcripts affected was lowest for kidney cortex (260) and highest for lung (857). Most affected transcripts were specific for the different absorbed doses and few were found in more than one tissue. For the transcripts affected at all dose levels the response was in general independent of dose and only a few transcripts showed increasing or decreasing regulation with increasing absorbed dose. Affected biological processes were primarily associated with the normal functions of the each tissue. Immune response only was affected in all tissues and processes related to metabolism were affected in several tissue types. Evaluation of the results for astatine-211 is on-going. Briefly, the response was very different for the two radionuclides. In general, astatine-211 affected the expression of more genes than iodine-131 compared to non-exposed mice. Furthermore, the two radionuclides affected expression of different genes involved in different biological processes.
The relative biological effects and effectiveness of alpha emitters
Studies on relative biological effectiveness (RBE) should compare two radiation types while keeping parameters like time of exposure and dose rate constant. Such studies are very difficult to perform with radiopharmaceuticals due to the limited number of radionuclides emitting high and low-LET radiation, radiolabelling techniques and compounds available. Preferably, the same compound should be used and labelled with two radionuclides with similar half-life and labelling technique, while the tumour binding properties are retained. Furthermore, different end-points will give different results.
The focus has been on studies with alpha-emitting radionuclides and to get data on therapeutic effects and toxicity. Few corresponding studies on reference radionuclides were done, partly due to lack of suitable compounds. In the future, other less strict RBE-like parameters need to be defined for appreciation of the higher biological effects of alpha-emitting radionuclides.
Radiobiological effects of bismuth-213-MAb on cancer cells studied by quantification of gamma-H2AX
Cytotoxicity of alpha-particles is particularly based on induction of DNA double-strand breaks (DSBs). Methods detecting DSBs might be useful to estimate absorbed doses in tumour and normal tissue. At sites of DSBs, histones H2AX are phosphorylated to gamma-H2AX that can be detected by immunofluorescence, flow cytometry and Western blotting. Cultured EJ28 bladder cancer cells were exposed to different activity concentrations of 213Bi-anti-EGFR-MAb for three h and gamma-H2AX was quantified at different times after exposure. The number of gamma-H2AX foci per cell and the number of gamma-H2AX foci positive cells increased with increasing 213Bi activity concentration. Incubation at either 4°C or 37°C unexpectedly gave similar results. The number of gamma-H2AX foci only gradually decreased with time after incubation. Because gamma-H2AX directs repair factors to the sites of DNA-DSBs and rapidly becomes dephosphorylated after repair, the results suggest that DSB repair is impaired in EJ28 cells. Quantification of DSBs via gamma-H2AX foci might allow for estimation of radiation dose and cytotoxicity.
Radiobiological effects of bismuth-213-MAb on cancer cells in vitro determined by gene expression analysis
Whole genome gene expression profiling was performed at 6, 24 and 48 h after incubation of HSC45-M2 gastric cancer cells with bismuth-213-d9MAb. Irradiation induced both up- and down-regulation of approximately 1,000 genes at each time point analysed. Eight genes appeared up-regulated and 12 genes down-regulated throughout. Among the genes showing continuous up-regulation, COL4A2, NEDD9 and C3 have not been associated with cellular response to high LET radiation so far. The same holds true for WWP2, RFX3, HIST4H4 and JADE1 that showed continuous down-regulation. The consistently up-regulated (ITM2C, FLJ11000, MSMB) and down-regulated (HCG9, GAS2L3, FLJ21439) genes have not been associated with any biological process or molecular function so far. Thus, these findings revealed interesting new targets for selective elimination of tumour cells and new insights regarding response of tumour cells to alpha-emitter exposure.
Effects of dose-rate and hypoxia
Hypoxic cells within tumours are radioresistant, but high LET alpha-particles are assumed to damage cells independently of their oxygenation status. Therefore, the cytotoxicity of bismuth-213-anti-EGFR-MAb was compared with that of photon irradiation at both normoxic and hypoxic conditions in squamous epithelium carcinoma CAL33 cells. The hypoxia marker HIF-1alpha was assayed via Western blotting at different time points after incubation in the hypoxia chamber. CAL33 cells were incubated with bismuth-213-anti-EGFR-MAb (0-1.48 MBq/ml) or irradiated with photons (0-12 Gy) under normoxic or hypoxic conditions. Cell survival, analysed by clonogenic assay and WST colorimetric assay, decreased with increasing activity concentration (bismuth-213) or dose (photons). After photon irradiation survival of hypoxic cells was significantly higher than of normoxic cells. In contrast, similar survival was found in normoxic and hypoxic cells after incubation with 213Bi-anti-EGFR-MAb. Thus, alpha-emitters such as bismuth-213 are effective tools for eradication of hypoxic tumour cells.
Maximum tolerated activities for alpha-emitting vectors
Toxic effects were obtained in the liver using bismuth-213-labelled antibodies given i.v. while other bismuth-213-labelled antibodies gave bone marrow toxicity after i.p. administration. For smaller peptides toxicity was found for the kidneys and bone marrow, with MTAs of 25 MBq of bismuth-213 and 20-30 kBq of actinium-225. With a time period allowing recovery of side effects more fractions could be given without acute toxicity.
Optimal therapy strategies and protocols
Based on the results of the TARCC project some strategies for finding optimal treatment protocols can be drawn. The biodistribution of astatine-211-, bismuth-213- and actinium-225-labelled radiopharmaceuticals studied were very different and the critical organs should be defined for each radiopharmaceutical. The dosimetry for the critical organs must be further studied, if possible on the cellular level for some organs and the maximum tolerated activity or tolerance dose should be found. Based on data from the project we can conclude that at least the following critical organs are found for i.v. or i.p. administration: thyroid and bone marrow for astatine-211-labelled compounds, liver and bone marrow for bismuth-213-labelled proteins and peptides, kidneys for actinium-225-labelled proteins and kidneys for bismuth-213- and actinium-225-labelled peptides. For intracavitary administration other critical organs should also be considered. Thus, techniques blocking thyroid and kidney uptake should be used when appropriate. Furthermore, the promising results showing that alpha particles can eradicate also hypoxic tumour cells indicate that therapy with alpha emitting radiopharmaceuticals should be considered also for patients with reduced tumour perfusion and hypoxia.
Conclusions:
From the results obtained in the TARCC project, it could be derived that alpha-radionuclide conjugates are superior to beta-radionuclide conjugates in treatment of minimal disease and disseminated tumour cells. Many different models have been explored, in both preclinical experiments to demonstrate anti-tumour efficacy and toxicity and in in vitro models that explored more fundamental aspects of alpha radionuclide therapy. Efficacy was clearly established in models of minimal disease, as expected, with minimal toxicity. It was also confirmed that the cytotoxic effect of alpha particles is independent of hypoxia and that response of the cells to this particular kind of irradiation differs from that observed with low linear energy transfer radiations (photons, electrons). These results clearly demonstrate that alpha radionuclide therapy has the potential of killing even the last tumour cell, including cancer stem cells, which are known to be resistant to chemotherapy and classical radiotherapy. This aspect will certainly be the objective of future research.
Future clinical trials could be developed with a maximum safety, despite of the high intrinsic toxicity of alpha particles.
Ultimately, targeted alpha-radionuclide therapy appears as a promising approach that could reduce side-effects resulting from current targeted radionuclide therapy (TRT), a treatment option that is generally well tolerated giving patients a fast and clear improvement in quality of life.
Project website: http://www.tarcc.org
Email: contact@tarcc.org