NW/. Med. Bid. Vol. 17, No. 8. pp. 819-827, Inr. J. Radiat. Appl. Instrum. Parr B Printed in Great Britain. All rights reserved

0883-2897/90 $3.00 + 0.00 Copyright 0 1990 Pcrgamon Press plc

1990

Design of Compounds Having an Enhanced Tumour Uptake, Using Serum Albumin as a Carrier. Part I H. SINN,*

H. H. SCHRENK, E. A. FRIEDRICH, and W. MAIER-BORST

U. SCHILLING

lnstitut fiir Radiologie und Pathophysiologie, Deutsches Krebsforschungszentrum, Im Neuenheimer Feld 280, D-6900 Heidelberg, F.R.G. (Received

6 September

1989; in revised form

20 February

1990)

The search for a radioiodinated “cumulative” protein label, stored within cells following intracellular protein degradation, suggested that plasma protein turnover of tumours might be of use. While earlier investigators were primarily interested in metabolism and utilization of plasma proteins by tumours, we tried to utilize the tumour protein turnover to channel radioiodine lab&d compounds, covalently bound to serum albumin, into neoplastic tissues. To identify those parameters which influence the tumour uptake and storage, we investigated a series of compounds having different chemical and physicochemicaf properties. Unbound, small molecular weight compounds were rapidly eliminated from the circulatory system. They had a prolonged biological half life if linked to serum albumin (SA), especially when derivatized with deoxysorbitol. Parallel with the prolongation of the biological half-life we noted a remarkable increase in tumour uptake, which was not accompanied by increased liver activity. Furthermore, without thyroid blockade, we failed to detect significant radioiodine uptake in this organ after 24 or 72 h. This is due to the particular coupling mechanism, which may be relevant for other (radio)iodinated pharmaceuticals used in medicine. Glucose and aromatic amines, as well as aromatic aldehydes and glucamine react to form deoxysorbitol derivates, which then have similar biokinetics after linkage to serum albumin. This indicates that a new approach in tumour detection and possibly in tumour therapy may be possible when SA is used as a carrier molecule, using the described labelling procedure.

Introduction New information on the pathophysiology of tumours (Jain, 1987, 1989), in combination with published findings on protein turnover in tumour-bearing animals and patients (Babson and Winnick, 1954; Bush and Greene, 1955; Busch er al., 1961; von Euler et al., 1963; Hradec, 1958; LePage et al., 1952; Rossing, 1968), stimulated us to examine the criteria which influence the accumulation of substances by tumours. Small molecular weight substances of very different chemical and physicochemical characters e.g. Sfluoro-uracii (5FU), tetracycline derivates, porphyrines and x-ray contrast media show very low tumour accumulation even though they are useful in tumour diagnosis and therapy. These results appear to confirm Jain’s concept on pathophysiology of tumours (Jain, 1987, 1989). Our concepts were given the direction by Pittmans “cumulative” protein label (Pittman, et al., 1983), and the suspected high rate of serum albumin (SA) turnover in tumours. We began a series of chemical syntheses and animal experiments

using different compounds as serum albumin labels. The experiments were to provide answers to the following questions:

*Author for correspondence. 819

(1) Is SA turnover in tumours high enough to be used for positive tumour imaging, when a labelled SA is used, which shows cumulative uptake? (2) Which structure does a protein-linked compound need to remain in circulation and encounter tumour accumulation? (3) What kind of label persists in a tumour, or in tumour cells, after intracellular protein degradation? (4) Is the biological behaviour of SA changed when it is labelled with a foreign compound? (5) Up to what molar ratio can SA be loaded with foreign compounds, without losing its normal properties? At first we synthesized tyramine cellobiose (TCB), using Pittman’s protocol, and labelled this compound with radioiodine. The *I-TCB was linked to rat serum albumin (RSA) and was injected intravenously

820

H. SINNer al.

to tumour bearing rats. The activity distribution was registered non-invasively at different time intervals, by means of a y camera. The tracer technique enabled comparison of results achieved with different compounds. We performed a series of in vitro trials to obtain information on the coupling mechanism of TCB and cyanuric chloride. Since we were able to couple the radioiodinated derivates of: phenol, enamines of tyramine, 4-hydroxy-benzaldehyde, and tetrakis-4hydroxyphenyl-porphine-sulphonate very efficiently, we concluded that cyanuric chloride reacts with the hydroxylic group of the phenyl ring. In the following we investigated the biological behaviour of radioiodine labelled compounds such as: phenol, tyramine, tyramine-l\‘-I’-deoxysorbitol (TDS), 4-hydroxybenzylamine-N-l ‘-deoxysorbitol (HBADS), 4-amino-1,8naphthalic-acid-tyrimid (ANT), Camino-1,8-naphthalic-acid-tyrimide- N- l’-deoxysorbitol (ANTDS) as well as the tetrakis-4-hydroxyphenyl-porphinesulphonate (THPPS) and the corresponding SA linked substances as well. Since ANTDS and THPPS are compounds which are potentially useful bifunctional protein labels, or vice versa, and since SA can be a valuable carrier for THPPS, a substance to which growing interest is directed for photodynamic therapy (PDT), we paid special attention to these two compounds. Whereas free, small molecular weight compounds in general were trapped very rapidly by the liver but showed only a moderate tumour uptake, the protein linked compounds stayed longer in the circulatory system and showed higher tumour accumulation rates. In particular, sugar or deoxysorbitolderivatized compounds, when linked to SA, showed no alteration in distribution or biological half-life of the carrier protein, and exhibited altogether considerable tumour accumulation rates, in some cases more than 25% of total body activity.

Materials Phenol, tyramine, tyramine -HCl and D-cellobiose were purchased from Serva, Heidelberg, F.R.G. D-Glucose and sodium cyanoborhydrid (NaBH,CN) were delivered by E. Merck, Darmstadt, F.R.G. Ethylene glycol, cyanuric chloride and Caminonaphthalic-acid-anhydride were bought from Aldrich, Steinheim, F.R.G. Rat serum albumin (RSA) was purchased from Sigma, Deisenhofen, F.R.G. and 4-hydroxy-benzaldehyde was from Fluka, Neu-Ulm, F.R.G. Glucamine was a generous gift from Professor Dr H. Elias, Technical Highschool of Darmstadt, F.R.G. Tetrakis-4-hydroxyphenyl-porphine-sulphonate (THPPS) was made available by Professor Dr D. Wohrle, University of Bremen, F.R.G. i3’I was purchased from Amersham Buchler, Braunschweig, F.R.G. Animals

BD IX and Wistar rats were delivered from Zentrale Versuchstieranstalt, Hannover, F.R.G.

Tumour

Ovarian carcinoma O-342 was made available by Dr Zeller, German Cancer Research Center, Heidelberg, F.R.G.

Methods (1) Synthesis of tyramine-cellobiose (TCB)

TCB was synthesized in two different ways: (a) following the protocol of Pittman and (b) according to an own elaborated procedure, which works much faster, gives better yields and is applicable to syntheses of other compounds, of which part of the initial materials are water insoluble: 34.8 mg tyramine. HCI (0.2 mM) and 85 mg cellobiose (0.25 mM) are dissolved in 2 mL ethylene glycol by heating up to 120°C in a glycerine bath. Subsequently 60mg NaBH,CN (0.9mM) are added and heating is continued for further 10min. After cooling down to room temperature the pale yellow solution is diluted by a factor of ten with distilled water. The tyraminecellobiose content is directly analysed and separated in semi-preparative amounts by high performance liquid chromatography (HPLC). (a) HPLC apparatus for analytical use. Two high pressure pumps: P400, a Gradient programmer, a variable u.v./vis.-monitor, a Rheodyne injection valve; u.v.-detector: 280 nm; range: 2.0 arbitrary units; recorder: Shimadzu C-RSA. (6) HPLC apparatus for semi-preparative use. The HPLC apparatus was provided additionally with an auto-sampling system consisting of three motor valves (two HMV P and one HMV 6), a magnetic valve (SAH 24/30 T) and a time programmer (alphatronic PT 810 S). The complete HPLC system including the progressive installation for substance separation in semipreparative amounts was manufactured by Latek, Heidelberg, F.R.G. The separation conditions were as follows: Sample vol.: 100 p L; pre-column: 50 x 4 mm C-18 20 pm (Latek, Heidelberg, F.R.G.); column: 250 x 4 mm, Nucleosil5-SA 5 pm (Latek, Heidelberg); Eluent: (a) 0.1 M NH,acetate, pH 6.0, 90%, (b) acetonitrile, 10%; flowrate: 1 mL/min; Retention time: (a) tyramine-cellobiose, 10.76 min (b) tyramine, 15.50 min; yield: ca 73%. Independently from the applied method (Pittman or our own) TCB samples showed identical retention times on HPLC, and the corresponding fractions had identical u.v.-spectra. (2) Synthesis of tyramine-N-I’-deoxysorbitol

(TDS)

A quantity of 34.8 mg tyramine*HCI (0.2 mM) and 200 mg D( + )glucose (1 .OmM) are dissolved together in 1 mL ethylene glycol by gentle warming up. Subsequently 120 mg NaBH, CN (1.8 mM) are added and the mixture is then heated up to 120°C in a glycerine bath for 20min. After cooling to room temperature

Compounds with an enhanced tumour uptake

the clear pale yellow ethylene glycol solution is diluted by a factor of 10 with distilled water. This diluted solution can be used directly for analyses or semi-preparative separation of the components on HPLC under conditions similar to TCB, as described previously. Retention time: (a) TDS, 13.2min (b) tyramine, 15.5 min; yield: ca 46%. (3) Synthesis of 4-hydroxy-benzylamine-N-l’-deoxysorbitol (HBADS) A quantity of 36 mg 4-hydroxy-benzaldehyde (0.3 mM) are mixed together with 54mg glucamine (0.3 mM) in 1 mL ethylene glycol under warming up, with subsequent addition of 30 PL concentration HCI and I55 mg NaBH,CN (2.5 mM). The reaction temperature is kept at 100°C in a glycerine bath for 1 h. After cooling down to room temperature the mixture is diluted by a factor of ten with distilled water. HBADS is purified and separated by HPLC under similar conditions as described earlier. Retention time: (a) 4-hydroxy-benzaldehyde, 7.8 min (b) HBADS 10.3 min; yield: ca 45%. (4) Synthesis

of

4-amino-naphthalic-acid-tyrimide

(ANT)

A quantity of 426 mg 4-amino-l &naphthalic acid anhydride (2 mM) are mixed together with 1,37 g tyramine (10 mM) in 30 mL dimethylformamide and heated up to 140°C in a glycerine bath for 30min. Following this, the dimethylformamide is withdrawn with an oil pump vacuum and the residue is refluxed in 100 ml 6 N HCl for 2.5 h. After cooling down to room temperature the yellow-brown sediment is separated by filtration, redissolved in 100 mL hot alcohol (methanol :ethanol 1: 1, v/v) and precipitated again by adding 300 mL 6 N HCI. To complete the precipitation the solution has to be stored at +4”C for 2 h. The resulting sediment is separated by filtration and dried in a desiccator. Yield: 498 mg (about 75% of theory). ANT exhibits, when dissolved in 1,Cdioxane or comparable organic solvents, an intensive yellowgreen fluorescence and has an absorption maximum at 445 nm (in 1,4-dioxane). HPLC conditions for analytic and semi-preparative substance separation are as follows: sample vol.: 100 pL; pre-column: 50 x 4 mm: Cl8 30 pm (Latek, Heidelberg, F.R.G.); column: 250 x 10 mm; Cl8 5 pm GO (Latek, Heidelberg, F.R.G.); eluent: (a) 0.1% formic acid in water, 40% (b) methanol, 60%; gradient: 60-100% methanol; gradient time: I2 min; exponent: 2; delay time: 2 min; flow rate: 4mL/min; u.v.-detector: 445 nm; range: 0.5 arbitrary units; retention time: 11.75 min. (5) Synthesis of’4-amino-I$-naphthalicacid-tyrimideN-I’-deoxysorbitol

(ANTDS)

An amount of 10 mg ANT (30 PM) are mixed together with 200 mg glucose (about 1.l mM) in 1 mL

821

ethylene glycol and 0.2mL 25% acetic acid. This mixture is warmed up until all components are dissolved and 68 mg of NaBH, CN (about 1.1 mM) are added. Subsequently this reaction mixture is heated for 2 h at 100°C in a glycerine bath. After cooling down to room temperature the mixture is diluted in a ratio of l/10 with a solution of methanol/dioxan/ water (3 : 3 : 4 v/v/v). HPLC conditions. Similar to ANT; Retention time: (a) ANT-N-I’-deoxysorbitol, 8.60 min; (b) ANT, 11.75 min; Yield: about 20% of theory. After HPLC separation, the eluent is removed by vacuum distillation. The solid yellow residue is dissolved in a mixture of water and dioxane (7:3 v/v), to a concentration of 0.5 mg/mL. ANTS shows the same intensive yellow-green fluorescence in aqueous solutions as ANT does in 1,Cdioxane.

Radioiodination of Synthesized Compounds In this reaction, species, amount, concentration and pH of the oxidizing agent is crucial, since the agent is required for generating IO- which transfers cationic iodine (I+) onto the aromatic ring (Scheme 1). R u

" -

OH t*IO--

/\ -

R q

OH + HOI'

After a series of pilot experiments we found that elemental bromine, dissolved in phosphate buffer (PPB, 0.13 M, pH 7.4) revealed the best properties of all investigated oxidizing substances. The transformation from I- to IO- is very fast and the reaction of IO- with a hydroxyphenyl residue is faster than that of BrC-. In general 1.0-1.2 pg bromine/mCi (37 MBq) radioiodine are sufficient to incorporate 90% of the activity into the hydroxyphenyl residue. Additionally there is no interaction between bromine and cyanuric chloride, necessary for coupling the radioiodinated compound. To about 0.1 PM of the appropriate compound (phenol: 9.5 pg, tyramine: 14bg, ANT: 33 pg, HBADS: 30 pg, TDS: 30 pg, TCB: 46 pg, ANTDS 5Opg and TPPS: 125pg), dissolved in 50-IOOpL phosphate buffer (PPB: 0.13 M, pH 7.4) (or in a mixture of PPB and dimethylacetamide or dioxane (1: 1, v/v), required in the case of ANT and ANTDS) l-5 mCi (37-185 MBq) of radioiodine were added. With 1.0-l .2 gg bromine/mCi (37 MBq) dissolved in PPB with a concentration of 1 mg/mL, iodide is converted into the required iodonium form to replace a hydrogen atom in the ortho position to the hydroxylic group. In general the iodination reaction does not need more than 5 min and the labelling yield is about 90%. Control of labelling efficiency is done preferably by thin layer chromatography (TLC).

H. SINNeral.

822

Plates: Silicagel 60 (5 x 20 cm) with fluorescence indicator (E. Merck, Darmstadt, F.R.G.); developing solvent: acetone butanol-1 25% ammonia water, 60 : 25 : 10 : 5 (v/v); running distance: 10 cm; radioanalyzer: Linear Analyzer LB 28 20-l Berthold, Wildbad, F.R.G.; the Rf values are shown below: Comoound

Oria. camp.

*I-camp.

*Iodide

Phenol+ Tyfamine TCB HBADS TDS ANT ANTDS THPPS

0.33 0.76 0.15 0.30 0.42 0.85 0.44 0.11

0.36 0.80 0.20 0.32 0.44 0.85 0.48 0.13

0.0 0.93 0.93 0.93 0.93 0.93 0.93 0.93

+Since phenol, *I-phenol and *iodide were running in the standard developing solvent in the front we used in this case a mixture of: chloroform/methanol: 97/3 (v/v).

From this SA solution an equimolar amount (0.1 PM corresponds to 7mg SA) is added directly to the solution of the dichloro-triazine derivatized radioiodine labelled compound. After a reaction time of at least 30min the reaction volume is filled up with PPB (0.13 M, pH 7.4) to 2.0 mL and transferred to a Centricon C 30 ultra-filtration unit (Amicon, Witten, F.R.G.). Unbound radioiodine, dioxane and in some cases dimethylacetamide are separated by centrifugation in a 45” fixed angle rotor (average diameter: 20 cm) at 6000 rpm, for about 15 min. Subsequently the reduced volume (from 2.0 mL to 70-100 pL) is directly injected into an HPLC apparatus suitable for protein analysis and separation of the monomer fraction. HPLC apparatus. High pressure pump: P 400 (Latek, Heidelberg, F.R.G.); Rheodyne injection valve; u.v.-monitor: Single Path Mon. UV-1 (Phar-

Cyanuric chloride coupling of radioiodinated compounds

In this reaction the pH range and reaction time are of importance. We found that the pH range of 7.4-7.6 is not only ideal for iodination of hydroxy-phenyl residues but also for the first coupling step. To guarantee high efficiency in this reaction it is necessary to prepare a fresh solution of freshly recrystallized cyanuric chloride in dry 1&dioxane at a concentration of lOmg/mL. Related to the mass of the iodinated compounds (we labelled in general 0.1 p M) a surplus of lo-20% cyanuric chloride corresponding to 0.11-O. 12 PM (20-22 ng or 2.0-2.2 pL of the dioxane solution) is required. The dioxane solution of cyanuric chloride is added directly to the solution of the iodination reaction and vortexed for some seconds. The required reaction time is about 5 min. Subsequently SA solution can be added to realize protein coupling.

macia, Freiburg, F.R.G.); y-detector: NaI 1.5 in. x 2 in. with Berthold electronics (Berthold, Wildbad, F.R.G.); recorder: Shimadzu, Chromatopac CRlA (Latek, Heidelberg); motor valve: HMV P (Latek, Heidelberg, F.R.G.). HPLC conditions. Sample vol.: 100 p L; pre-column: 50 x 4 mm, Zorbax Diol (DuPont, Dreieich, F.R.G.); column I: Zorbax GF 450 (DuPont, Dreieich, F.R.G.); column II: Zorbax GF 250 (DuPont, Dreieich, F.R.G.); eluent: 0.2 M phosphate buffer, pH 7.6, with 0.05% NaN,; flow: 1 mL/min; u.v./vis-monitor: 280 nm, range 1.0 A.U.; retention time: 19.8 min (monomer fraction of RSA). Before the labelled and separated monomer fraction of SA can be injected into an animal, the protein must be purified from toxic NaN, by centrifugation in a C-30 ultra-filtration unit.

Coupling of the dichloro-triazine derivatives to SA CL I

R

/

/\ --Q-

\

I* ,T;

/ Protein

To realize a high protein labelled dichloro-triazine SA has to be dissolved buffer (PPB 0.13 M, pH lOma/mL. The resulting

binding of the radioiodine derivatized compound, the in an alkaline phosphate 9.3) at a concentration of oH of the solution is 8.8. _.

Animal Experiments Since this paper would have become too comprehensive had we included all biological experiments, we decided to publish our results in two parts.

823

Compounds with an enhanced tumour uptake

b

ul 2 .

Fig. I. Two typical chromatograms of an “‘I-TDS labelled RSA, analyzed by HPLC and registered by y-detection. (a) Original batch, applied to HPLC after concentration and simultaneous separation of low molecular weight impurities on a centricon C-30 ultra-filtration unit. The content of oligomer (16.66min) and dimer (17.54min) components is approx. 25%. (b) Analytical control run of a HPLC purified and separated monomer fraction of “‘I-TDS-RSA. The content of dimer and oligomer impurities is now below 0.25%.

Part I: with emphasis on chemical aspects such as synthesis, purification, analyses, radioiodination and coupling to the protein including one typical animal experiment. Part II: comprising the results of all biological investigations, performed with different rat species, different tumour models, different ways of tumour transplantation, different methods of activity application, isolation of tumour cells etc., with special respect to tumour physiology and consequences of these data. Camera system: Searle, Pho-Gamma IV, FLOV (Siemens, Erlangen, F.R.G.); data recording system: Medworker N 4. frame mode; matrix: 128 x 128 (Gaede. Freiburg, F.R.G.) animal model: rats, BD IX. female, 8 weeks old, 200-220 g; tumour: ovarian carcinoma (O-342, Zeller, DKFZ); implanted i.m. into the left hind-leg; inoculated cell number: 5 x 106;start of the experiment: 10 days after tumour implantation; tumour diameter at the starting point: IO x 12 mm; tumour diameter at the end of the experiment: 29 x 24 mm; method of application: i.v., lateral tail vein; time duration of the experiment: 72 h; Number of animals used in this experiment: 4; administered activity: 300 PCi (Il. 1 MBq) “‘I; administered volume: 0.3 mL; Sp. act.: 1 mCi (37 MBq)/S pg TDS x 1 mg RSA x 1 mL PPB pH 7.4; Time points of data acquisition: IO min, 1,2,4, 24,48 and 72 h post-injection. Time-dependent activity distribution of ‘3’1TDS-RSA is representative for the other compounds derivatized with deoxysorbitol and linked to SA.

Rt?SUltS Pilot studies with different small molecular weight compounds labelled with radioiodine and bound to SA showed promising tumour accumulation: (a) Radioiodine labelled compounds such as phenol, 4-amino-phenol, tyramine and thyrosine failed to show significant accumulation in tumours (