Synthesis and characterization of nido-carborane-cobalamin conjugates

Nuclear Medicine & Biology, Vol. 27, pp. 89 –92, 2000 Copyright © 2000 Elsevier Science Inc. All rights reserved.

ISSN 0969-8051/00/$–see front matter PII S0969-8051(99)00081-5

Synthesis and Characterization of nido-Carborane-Cobalamin Conjugates Harry P. C. Hogenkamp,1 Douglas A. Collins,2 David Live,1 Linda M. Benson3 and Stephen Naylor3 1

DEPARTMENT OF BIOCHEMISTRY, UNIVERSITY OF MINNESOTA, MINNEAPOLIS, MINNESOTA, USA; 2DEPARTMENT OF

DIAGNOSTIC RADIOLOGY, MAYO CLINIC, ROCHESTER, MINNESOTA, USA; AND 3BIOMEDICAL MASS SPECTROMETRY FACILITY, MAYO CLINIC, ROCHESTER, MINNESOTA, USA

ABSTRACT. Three vitamin B12 (cyanocobalamin) conjugates bearing one nido-carborane molecule or two nido-carborane molecules linked to the propionamide side chains via a four carbon linker have been synthesized. Reaction of o-carboranoylchloride with 1,4-diaminobutane in pyridine produced nido-carboranoyl(4-amidobutyl)amine, which was linked to the b- and d-monocarboxylic acids and the b,d-dicarboxylic acid of cyanocobalamin. Mass spectrometry analysis as well as 11B nuclear magnetic resonance demonstrated that during the reaction of o-carboranonylchloride with diaminobutane one of the boron atoms was eliminated. In vitro biological activity of the cyanocobalamin-nido-carborane conjugates was assessed by the unsaturated vitamin B12 binding capacity assay. When compared with 57Co cyanocobalamin, the biological activity of cyanocobalamin-b-nido-carborane, cyanocobalamin-d-nido-carborane, and cyanocobalamin-b-dbis-nido-carborane conjugates were 92.93%, 35.75%, and 37.02%, respectively. These findings suggest that the 10B cobalamin conjugates might be useful agents in treating malignant tumors via neutron capture therapy. NUCL MED BIOL 27;1:89 –92, 2000. © 2000 Elsevier Science Inc. All rights reserved. KEY WORDS: Boron neutron capture therapy (BNCT), Vitamin B12, Cyanocobalamin-nido-carborane conjugates, Boronated vitamin B12 INTRODUCTION The limitation of radiation therapy, and indeed all oncologic therapies, is the specific delivery of lethal agents to malignant cells with minimal effect on surrounding normal tissue. The potential of neutron capture therapy (NCT) is to selectively irradiate tumors with thermal or epithermal neutrons to minimize radiation damage to nonmalignant tissue. However, for NCT to be successful, a sufficient number of 10B or 157Gd atoms, or possibly a combination of the two atoms, must be delivered to the tumor. Once the NCT agent concentrates within neoplastic tissue, the potential for delivering a localized dose of therapeutic radiation arises. The reaction of 10B atoms with thermal or epithermal neutrons results in the formation of excited 11B nuclei, which undergo fission to yield highly energetic 4He and 7Li ions (6, 11): He ⫹ 7Li ⫹ 2.79 MeV (6.3%)

4

B ⫹ 1n 3 [11B]m n

10

He ⫹ 7Li ⫹ ␥0.48 MeV ⫹2.31 MeV (93.7%)

4

The lithium and helium atoms produced during NCT give up their disintegration energy within 5 and 9 ␮m, respectively, of the 10B atom. Therefore, the deposit of therapeutic radiation should remain within the individual tumor cells. Conversely, the gamma energy released during the 157Gd (n ␥) reaction (0 –7.9 MeV) extends beyond single cell boundaries, thus delivering radiation to the surrounding normal tissue. However, this Address correspondence to: H. P. C. Hogenkamp, Ph.D., Department of Biochemistry, University of Minnesota Medical School, Minneapolis, MN 55455, USA; e-mail: [email protected] Received 7 June 1999. Accepted 1 September 1999.

increases the chance of hitting all the cells within the tumor. There are intercellular, internal conversion, and Auger electrons emitted during the 157Gd neutron capture reaction. Unfortunately, for these high linear energy transfer (LET) particles to be efficacious they must be directly in contact with DNA. Because 157Gd has the greatest cross-sectional diameter (255,000 b), which is approximately 66 times greater than 10B (3,838 b) of the stable nuclides, the chance and rate of neutron capture reaction increases linearly. Numerous boron-containing analogs have been synthesized and some of them have shown promise in initial biological tests (11). For instance, Hawthorne and Shelly (7) used liposomes to deliver a boron-rich nido-carborane to tumors and Kahl et al. (9) prepared a tetraphenylporphyrin bearing four nido-carborane molecules. Very recently, Nakanishi et al. (10) prepared boron-rich oligomeric phosphate diesters and demonstrated that they accumulate in the cell nucleus. It has been well documented that 57Co cyanocobalamin has increased uptake in rapidly dividing tissues (5). Our previous studies have demonstrated that Tc-99m and In-111 radiolabeled diethylenetriaminepentaacetate adenosylcobalamin (DAC) conjugate is recognized by transcobalamin II proteins in vivo and is an effective tumor imaging agent (3, 4, Collins and Hogenkamp, unpublished data). Because the cobalamin-DTPA conjugate accumulates in proliferating tumors, the cobalamin complex radiolabeled with alpha or beta emitters may be an effective therapeutic radiopharmaceutical. We are currently exploring the potential therapeutic effect of radiolabeled DAC. We have also shown that 157Gd-DAC is a promising paramagnetic contrast agent, with approximately two times the relaxivity of the currently commercially available agents (8). Although 157Gd labeled cobalamin compounds may be useful NCT agents, we thought it would be worthwhile to label vitamin B12 with 10B to

90

H. P. C. Hogenkamp et al.

Keystone Scientific, Inc., Bellefonte, PA USA). The mobile phase consisted of water:methanol (98:2, v:v) in pump A and water: methanol (2:98, v:v) in pump B. A linear gradient was used from 5% B to 30% B over 10 min and was held at 30% B for 20 min before returning to the initial conditions (5). The purification of (6) required a longer gradient using the same mobile phases: 30% B to 70% B over 25 min and held at 70% B for 10 min before returning to the initial conditions. The separations were monitored by UV absorption at 214 or 254 nm. The flow rate was 1.0 mL/min and was split postcolumn allowing approximately 10 ␮L to flow into the mass spectrometer. Mass spectral data were collected using electrospray ionization in positive mode over a mass range of 200 to 2,100 AMU at a dwell time of 0.15 ms/0.2 AMU. The samples were prepared at 10 mg/mL in pump A mobile phase and an aliquot injected into the HPLC (1–5 ␮L). Homogeneous preparations were obtained by pooling fractions from several injections. They were dried to orange-red powders. A portion of the purified product was dissolved in water:methanol (1:1) and reanalyzed by HPLC-mass spectrometry (MS) to ascertain purity. Retention times for (5) were 13.3 min and for (6) 10.5 min.

Synthesis of Cyanocobalamin-nido-carborane Conjugates (Figure 2) o-Carborane Carboxylic Acid (2) FIG. 1. Structure of vitamin B12 (cyanocobalamin) showing the location of the b, d, and e carboxylic acids. potentially deliver localized radiation within tumors during boron NCT (BNCT). The present article describes the chemical synthesis and characterization of cyanocobalamin conjugates containing the nido-carborane nucleus, their biological activity, and the potential use of the nido-carborane complexes and 157Gd conjugates as therapeutic agents in NCT. MATERIALS AND METHODS o-Carborane, butyllithium (1.6 M solution in hexanes), and putrescine were purchased from Aldrich Chemical Company (Milwaukee, WI USA). The water-soluble carbodiimide 1-ethyl-3(3⬘dimethylaminopropyl) carbodiimide and 1-hydroxybenzotriazole were from Sigma Chemical Co. (St. Louis, MO USA). Thin layer chromatography (TLC) silica gel plates were obtained from Eastman Kodak Company (Rochester, NY USA). The cyanocobalamin b and d monocarboxylic acids and the b,d-dicarboxylic acid were prepared as described previously (1) (Fig. 1). Ultraviolet (UV)-visible spectra were recorded with a diode array spectrophotometer. 11B nuclear magnetic resonance (NMR) spectra at 192.656 MHz with 1H decoupling were recorded on a Varian (Palo Alto, CA USA) 1NOVA 600 MHz spectrometer with an 8 mm broad band probe. Approximately 10 mg of the cobalamin conjugate were dissolved in 1 mL pyridine d6 in a 5 mm NMR tube. Five hundred scans were collected with an acquisition time of 0.133 s and a relaxation delay of 1 s. Chemical shifts are given relative to BF3:(CH3CH2)2O. Mass spectra were obtained on a Sciex API 365 triple quadruple mass spectrometer system (Toronto, Ontario, Canada). Separations were accomplished using a Shimadzu high performance liquid chromatography (HPLC) system consisting of two LC-10AD pumps and an SCL-10A controller (Shimadzu Scientific Instruments, Columbia, MO USA). Analytes were monitored by UV with an ABI 785 A detector. HPLC separations were achieved using a BDS-Hypersil C8 column (150 ⫻ 4.6 mm; 120A,

A solution of o-carborane (5.0 g, 34.7 mmol) in 500 mL dry ether in a 1 L round bottom flask was cooled to ⫺78°C in a dry ice-acetone bath. The solution was flushed with argon and the flask sealed with a serum stopper. n-Butyllithium (24 mL, 1.6 M in hexanes) was slowly injected over approximately 20 min and the reaction was stirred for an additional 30 min. Crushed dry ice (10 –15 g) was then added and the mixture stirred for 1 h. The dry ice-acetone bath was removed and the reaction allowed to come to room temperature. The ether and hexanes were removed on a rotary evaporator, water (150 mL) was added, and unreacted o-carborane was removed by extraction with hexanes (2 ⫻ 100 mL). The aqueous phase was acidified with concentrated HCl and the desired product extracted with hexanes (4 ⫻ 100 mL). The combined extracts were dried over Na2SO4 and evaporated to dryness to give 5.8 g (88.7%) of o-carborane carboxylic acid, which was used without further purification.

nido-Carboranoyl(4-aminobutyl) Amide (4) o-Carborane carboxylic acid (2.0 g, 10.6 mmol), dried over P2O5, was dissolved in 30 mL thionylchloride and heated under reflux for 3 h. The solution was cooled to room temperature, evaporated to dryness, and dried over P2O5 (2.05 g, 9.9 mmol, 93%) (2). The acylchloride was dissolved in 15 mL dry pyridine, and diaminobutane (1.12 g, 12.7 mmol) was added. The reaction mixture was heated under reflux for 3 h, cooled to room temperature, and concentrated. Water (10 mL) was added and the suspension acidified with 3 M HCl. The desired product was extracted with ethylacetate (4 ⫻ 50 mL), the combined organic layers were washed once with water, dried over Na2SO4, and evaporated to dryness to yield 2.35 g (8.0 mmol, 75%) of (4). Thus far, the product has resisted crystallization from a variety of solvent mixtures; however, TLC on silica gel plates (2-propanol-NH4OH-H2O; 7:1:2) showed only one ninhydrin positive compound distinct from the diamine. 11 B-NMR (pyridine-d5, decoupled) ⫺6.90, ⫺10.25, ⫺12.72, ⫺13.66, ⫺18.16, ⫺29.07, and ⫺33.61 ppm. Isotopic ratios calculated for C7H23B9ON2 (MH⫹): 245.3 (20%), 246.3 (54%), 247.3 (96%), 248.3 (100%), 249.3 (50%), 250.3 (4%), found for (4)

nido-Carborane-Cobalamin Conjugates

91

Figure 2. Synthetic route(s) to the cyanocobalamin-nido-carborane conjugates (5 and 6).

245.2 (26%), 246.3 (54%), 247.3 (97%), 248.4 (100%), 249.3 (85%), 250.3 (64%), and 251.2 (37%).

Cyanocobalamin-nido-carborane conjugates (5) Separate reaction mixtures containing 1 g (approximately 0.66 mmol) of the b- or d-monocarboxylic acids, hydroxybenzotriazole (810 mg, 6.0 mmol), 1-ethyl-3(3-dimetylaminopropyl) carbodiimide (1.14 g, 6.0 mmol), and (4) (600 mg, 2.0 mmol) for the monocarboxylic acids in 100 mL of a water-acetone mixture (2:1) were adjusted to pH 6.9 with 1 N NaOH. The reactions were stirred at room temperature and their progress monitored by TLC. After 3 h the mixtures were concentrated to remove acetone and the resulting aqueous suspensions were extracted into 92% aqueous phenol. The phenol phases were washed with water to remove the water soluble reactants. One volume of acetone and three volumes of ether were added to each of the phenol phases and the desired cyanocobalamin-nido-carborane conjugates were back extracted into water. The aqueous layers were extracted three times with ether to remove residual phenol and unreacted (4). Finally the aqueous solutions were evaporated to dryness. The residues were triturated with acetone and the desired conjugates isolated as orange-red powders (yields 90 –95% based on the cyanocobalamin-

carboxylic acids). 11B-NMR (pyridine-d5, decoupled) ⫺7.38, ⫺12.17, ⫺14.98, ⫺18.56, ⫺20.53, ⫺30.28, and ⫺33.50 ppm. Isotope ratios calculated for C70H109CoB9O15PN15 (MH⫹): 1,584.8 (12%), 1,585.8 (37%), 1,586.8 (74%), 1,587.8 (100%), 1,588.8 (84%), 1,589.8 (44%), 1,590.8 (16%), and 1,591.8 (4%) found for (5) 1,584.8 (14%), 1,585.8 (37%), 1,586.8 (74%), 1,587.8 (100%), 1,588.8 (88%), 1,589.9 (48%), 1,590.8 (19%), 1,591.8 (6%) found for (5) 1,584.8 (15%), 1,585.8 (40%), 1,586.8 (78%), 1,587.8 (100%), 1,588.8 (86%), 1,589.8 (44%), 1,590.8 (17%), 1,591.8 (7%). UV-visible absorption maxima at 280, 306, 322, 362, 408, 520 and 548 nm.

Cyanocobalamin-bis-nido-carborane conjugate (6) A reaction mixture containing 1 g (approximately 0.66 mmol) of the b,d-dicarboxylic acid of cyanocobalamin, hydroxybenzotriazole (863 mg, 6.4 mmol), 1-ethyl-3(3-dimethylaminopropyl) carbodiimide (1.16 g, 6.0 mmol), and (4) (1.05 g, 4.3 mmol) in 100 mL of a water-acetone mixture (2:1) was adjusted to pH 6.9 with 1 N NaOH. After 3 h incubation at room temperature, the reaction was worked up as described for the synthesis of (5). Yield was 1.29 g, isolated as an orange-red power. 11B-NMR (pyridine-d5, decoupled) ⫺7.23, ⫺11.93, ⫺14.78, ⫺18.41, ⫺20.34, ⫺30.04, and ⫺33.23

H. P. C. Hogenkamp et al.

92 ppm. Isotope ratios calculated for C77H129C0B18O16PN16 (MH⫹): 1,815.1 (9%), 1,816.1 (23%), 1,817.1 (47%), 1,818.1 (77%), 1,819.1 (99%), 1,820.1 (100%), 1,821.1 (77%), 1,822.1 (44%), 1,823.1 (10%), 1,824.1 (6%) found for (6) 1,815.1 (15%), 1,816.1 (27%), 1,817.1 (51%), 1,818.1 (79%), 1,819.1 (100%), 1,820.1 (100%), 1,821.1 (78%), 1,822.1 (49%), 1,823 (25%), 1,824.1 (12%). UV-visible absorption maxima at 280, 308, 320, 362, 400, 518, and 550 nm.

In Vitro Biological Activity of the Carborane Cyanocobalamin Analogs To assess in vitro binding of the nido-carborane-cobalamin complexes (CCC) to transcobalamin proteins (TP), the unsaturated vitamin B12 binding capacity (UBBC) assay was performed as previously described (3). Serum was obtained from five patients being evaluated for pernicious anemia at the Mayo Clinic. The patients’ serum first underwent a routine clinical UBBC run. To determine if the CCC analogs would inhibit 57Co cyanocobalamin from binding to TP, the excess serum from the five patients underwent a modified UBBC. Specifically, 0.4 mL serum was treated with 4 ␮L (concentration 10 ␮g analog/mL normal saline) of the three CCC analogs: cyanocobalamin-b-nido-carborane (CCbC), cyanocobalamin-dnido-carborane (CCdC), and cyanocobalamin-b-d-bis-nido-carborane (CCbdC). The analogs were first incubated for 20 min at room temperature with the patient’s serum. Then both the clinical run and the analog treated UBBC samples were assayed as usual under dim light. RESULTS AND DISCUSSION Scheme 1 shows the four step synthesis used to prepare the three cyanocobalamin-nido-carborane conjugates (5) and (6). MS analysis as well as 11B NMR demonstrated that during the conversion of (3) to (4) the o-carborane nucleus lost a boron atom to yield the nido-carborane derivative (4). The mass spectra of the three conjugates showed the characteristic profiles reflecting the isotope composition of boron. UV-visible spectroscopy of the final products showed the maxima typical of cyanocobalamin, indicating that the vitamin B12 nucleus of the conjugates had not been modified. Although the yields of (5) are almost quantitative based on the cyanocobalamin-monocarboxylic acids, the synthetic route to (6) yields a mixture of the bis-nido-carborane conjugate and mononido-carborane complexes. Indeed, HPLC is required to isolate homogeneous (6). These mono substituted complexes may actually be very useful starting materials for the preparation of cyanocobalamin conjugates containing a nido-carborane as well as the chelator diethylenetriaminepentaacetate. Such bifunctional conjugates could be used both for simultaneous tumor imaging and BNCT. The analogs competitively blocked 57Co cyanocobalamin from binding to the transcobalamin proteins. Therefore the counts per minute of radioactivity of the modified UBBC assay was significantly lower than that of the clinical runs. The percent binding (PB) of the analogs to transcobalamin proteins was calculated as follows (PB ⫽ 100 ⫺ analog UBBC cpm o¨ clinical UBBC cpm ⫻ 100). The average PB of the five solutions (N ⫽ 10 for each

solution; i.e., two modified UBBC assays per patient) for the CCdC, CCbC, and CCbdC were 35.7%, 92.93%, and 37.02%, respectively. The decreased biological activity of the CCbC and CCbdC was not expected. Previously, the b-, d-, and e-mono DTPA-cobalamin (11) derivatives consistently interacted with TP at 90 –95% compared with cyanocobalamin (CNCBL). Therefore, the smaller nido-carborane cage (approximately one third the size of DTPA) should not have sterically hindered the TP-CCC binding. However, the addition of the nido-carborane renders the conjugate more hydrophobic and thus may interfere with the interaction of the transport proteins. Indeed, the B12-nidocarborane conjugates are less soluble in aqueous solvents than cyanocobalamin. The increased hydrophobicity may facilitate the uptake of the conjugates in certain tumor types in vivo. The avidity of these conjugates in implanted human tumors in nude mice is presently being studied. Finally, the methyl- and adenosylcobalamin-nido-carborane complexes may improve interaction with the transcobalamins when compared with the cyano derivatives. This was the case with the DTPA-cobalamin conjugates (3). The synthesis and characterization of the methyl- and adenosylcobalamin-nido-carborane complexes also are currently under investigation.

We are indebted to Dr. S. B. Kahl for making available the details of the synthesis of o-carborane monocarboxylic acid.

References 1. Anton D. L., Hogenkamp H. P. C., Walker T. E. and Matwiyoff N. A. (1980) Carbon-13 nuclear magnetic resonance studies of the monocarboxylic acids of cyanocobalamin. Assignments of the b-, d-, and e-monocarboxylic acids. J. Am. Chem. Soc. 102, 2215–2219. 2. Barth R. F., Soloway A. H. and Fairchild R. G. (1997) Boron neutron capture therapy of cancer. Cancer Res. 50, 1061–1070. 3. Collins D. A. and Hogenkamp H. P. C. (1997) Transcobalamin II receptor imaging via radiolabeled diethylenetriaminepentaacetate cobalamin analogs. J. Nucl. Med. 38, 717–723. 4. Collins D. A., Hogenkamp H. P. C. and Gebhard M. (1999) Tumor imaging via In-111-DTPA-adenosylcobalamin. Mayo Clin. Proc. 74(7), 687– 691. 5. Flodh H. (1968) Distribution and kinetics of labelled vitamin B12. Acta Radiologica Supplementum 284, 3– 80. 6. Hawthorne M. F. (1993) The role of chemistry in the development of boron neutron capture therapy of cancer. Angew. Chem. Int. Ed. Engl. 32, 950 –984. 7. Hawthorne M. F. and Shelly K. (1997) Liposomes as delivery vehicles for boron agents. J. Neurooncol. 33, 53–58. 8. Hogenkamp H. P. C. and Collins D. A. (1996) The synthesis and in vitro biological activity of cobalamin-diethylenetriaminepentaacetate complexes. Biofactors 5, 180 –182. 9. Kahl S. B., Joel D. D., Nawrocky M. M., Micca P. L., Tran K. P., Finkel G. C. and Slatkin D. N. (1990) Uptake of a nido-carboranyl porphyrin by human glioma xenografts in athymic nude mice and by syngeneic ovarian carcinomas in immunocompetent mice. Proc. Natl. Acad. Sci. USA 87, 7265–7269. 10. Nakanishi A., Guan L., Kane R. R., Kasamatsu H. and Hawthorne M. F. (1999) Toward a cancer therapy with boron-rich oligomeric phosphate diesters that target the cell nucleus. Proc. Natl. Acad. Sci. USA 96, 238 –241. 11. Sjo¨berg S., Carlsson J., Chaneolhosseini H., Gedda L., Hartmat T., Malquiest J., Naeslund C., Olsson P. and Tjarks W. (1997) Chemistry and biology of some low molecular weight boron compounds for boron neutron capture therapy. J. Neurooncol. 33, 41–52.