- Email: [email protected]
A technique to prepare boronated B72.3 monoclonal antibody for boron neutron capture therapy
Nucl. Med. Eiol. Vol. 20, No. I, pp. 14, 1993 Printed in Great Britain
0883-2897/93 $5.00 + 0.00 Pergamon Press Ltd
A Technique to Prepare Boronated B72.3 Monoclonal Antibody for Boron Neutron Capture Therapy GIRISH N. RANADIVE’y2*, HOWARD S. ROSENZWEIG’**, MICHAEL W. EPPERLY’ and WILLIAM D. BLOOMER2 ’Department of Radiation Oncology, University of Pittsburgh, Pittsburgh, PA 15213 and *The Radiation Medicine Institute, Evanston Hospital Corporation,
Evanston, IL 60201, U.S.A.
(Received 24 June 1992)
B72.3 monoclonal antibody has been successfully boronated using mercaptoundecahydro-closo-dodecaborate (boron cage compound). The reagent was incorporated by first reacting the lysine residues of the antibody with m-maleimidobenzoyl succinimide ester (MBS), followed by Michael addition to the maleimido group by the mercapto boron cage compound to form a physiologically stable thioether linkage. Boron content of the antibody was determined by atomic absorption spectroscopy. For biodistribution studies, boronated antibody was radioiodinated with iodogen. ‘251-labeled and boronated B72.3 monoclonal antibody demonstrated clear tumor localization when administered via tail vein injections to athymic nude mice bearing LS174-T tumor xenografts. Boronated antibody was calculated to deliver lo6 boron atoms per tumor cell. Although this falls short of the specific boron content originally proposed as necessary for boron neutron capture therapy (BNCT), recent calculations suggest that far fewer atoms of “B per tumor cell would be necessary to effect successful BNCT when the boron is targeted to the tumor cell membrane.
Introduction The scientific rationale for boron neutron capture therapy (BNCT) as a potential treatment for malignant disease is based upon the nuclear reaction which occurs when ‘OBis irradiated with low energy thermal and epithermal neutrons. The resulting nuclear reaction produces alpha particles with high linear energy transfer and radiobiological effectiveness. Because the neutron capture cross section of ‘OBexceeds that of the major chemical elements in normal tissues by several orders of magnitude, competing nuclear reactions and normal tissue irradiation will be minimized if “B is selectively localized within tumors and the neutron beam focused thereon. Although the clinical potential of BNCT was recognized as early as 1936 (Lecher, 1936), it was not until the 1950s that it was actually used in the treatment of malignant tumors (Sweet and Javid, 1952; Farr et al., 1954; Godwin et al., 1955). The initial experience with glioblastoma multiforme was subsequently extended but clinical results to date
*All correspondence and reprint requests should be addressed to: Girish N. Ranadive, The Radiation Medicine Institute, 2650 Ridge Avenue, Evanston, IL 60201, U.S.A.
have been disappointing (Hatanaka et al., 1986). A major reason for these disappointing results was the lack of tumor selective boron-containing compounds, the lack of which significantly compromised therapeutic efficacy by increasing normal tisue irradiation. So-called “first” and “second” generation boron-containing compounds have generally lacked tumor specificity, resulting in low tumor-to-normal tissue uptake ratios. Current research is directed toward the design and synthesis of “third” generation boron-containing compounds which demonstrate preferential, if not specific, tumor cell binding (Fairchild and Bond, 1985). Based upon earlier calculations which assumed a homogeneous intracellular distribution of boron, a minimum boron content of approx. 3Opg/g tumor has traditionally been considered necessary for successful BNCT (Fairchild and Bond, 1985). However, by localizing boron to the tumor cell membrane as with monoclonal antibodies, the concentration of boron, calculated on a per cell or weight basis of tumor, required for effective BNCT can be markedly reduced (Humm and Cobb, 1990; Gabel et al., 1987; Fairchild, 1989). Boronated monoclonal antibodies to tumor associated antigens have previously been studied as vehicles for preferential l”B localization. Alam et al. (1987) were able to incorporate a relatively large number
2
GIRISH
N.
RANADNE
of “B atoms onto antibody but not without substantial loss of immunoreactivity. Another method for incorporating “B in antibodies, reducing existing disulfide linkages and reforming them with Na,(i”B),,H,,SH (Fairchild and Bond, 1985) led to 9-13 boron cages (108-156 boron atoms) incorporated into antithymocyte globulin (Alam et al., 1983). However, these boronated antibodies could not be studied in vivo because the disulfide linkage between the boron cage and the protein sulfhydryl group was prone to enzymatic cleavage before reacting with the antigenic sites on target lymphocytes. In this communication, we describe a method to derivatize the B72.3 monoclonal antibody with mercaptoundecahydro-close-dodecaborate, an anionic polyhedral borane compound (Soloway et al., 1967) that has been used in limited clinical trials with glioblastoma multiforme in Japan (Hatanaka et al., 1986). Our methodology to modify B72.3, forming physiologically stable thioether linkages with boron cages, leads to the incorporation of 16 cages (192 ‘OB atoms) per antibody molecule with little loss of immunoreactivity.
Materials and Methods B72.3 monoclonal antibody was a gift from Dr J. Schlom (NCI, Bethesda, Md) and was isolated from ascites fluid (Colcher et al., 1984). The LS174-T human colon adenocarcinoma was obtained from ATCC (Rockville, Md). Dicesium mercaptoundecahydro-close-dodecaborate (boron cage compound, BCC) enriched with ‘OBwas purchased from Callery Chemicals (Evans City, Pa), maleimidobenzoyl succinimidyl ester (MBS) and iodogen from Pierce (Rockford, Ill.), and Na’251 (14 mCi/mg) from Amersham Corp. (Arlington Heights, Ill.).
et al.
B72.3 monoclonal antibody (1 mg, 500 pL) in 20mM HEPES (PH 8.0) was reacted with a 50-fold molar excess of MBS and stirred for 3 h, after which excess reagent was removed by gel filtration on a Sephadex G-25 column, Antibody, recovered in the void volume was then reacted with a 50-fold molar excess of BCC and stirred at room temperature for 2 h after which excess reagent was again removed by gel filtration (Fig. 1). Protein within the void volume was concentrated on Centricon(Amicon Corp., Beverly, Mass.). Boronated B72.3 and unmodified B72.3 antibody (as a control) was radioiodinated using standard iodogen methodology (Colcher et al., 1984). Protein associated radioactivity (78% yield) was collected in the void volume using gel filtration. A tumor non-specific antibody human IgG (Jackson Laboratories, West Grove, Pa) was boronated and iodinated in analogous fashion. The purity of the boronated antibody was checked by HPLC on Superdex-75 (HR 10/30) column (Pharmacia, Piscataway, N.J.) using 20 mM Tris at pH 8.0, with a 1.OmL flow rate (Fig. 2). Immunoreactivity and biodistribution
The immunoreactivity of the radioiodinated and boronated B72.3 (RIBA) was measured using LSl74-T human colon adenocarcinoma cells obtained from xenografts grown in athymic nude mice (Harlan-Sprague-Dawley, Indianapolis, Ind.). LSl74-T cells (1 x 10’) were injected subcutaneously in the hind leg. Ten days later, the xenografts were removed and single cell suspensions prepared using Dispase (Boehringer Mannheim Corp., Indianapolis, Ind.). Single cell suspensions ranging from 0.1 to 5 x 10’ cells were placed in test tubes to which 1.0 pg of RIBA were added. Cell suspensions were
0 +
SH
A
=
Fig. 1. Reaction scheme for boronation of the antibody.
A technique to prepare boronated B72.3 monoclonal antibody
3
Fig. 2. HPLC analysis of boronated antibody on a Superdex-75 (HR 10/30) column. incubated for 1 h at 37”C, then washed twice with Hank’s balanced salt solution and the radioactivity counted (Badger et al., 1987). The biodistribution of the ‘251-labelled antibodies was determined by sacrificing the mice by nembutal overdose, and excising, washing, blotting, weighing and counting the various organs in a Packard 5650 y-counter (Packard Instrument Co., Downers Grove, Ill.). The percent injected dose per gram of tissue (Table 1 and Fig. 4) and tumor-to-tissue ratios (Table 2) were calculated.
Results Determination of specljic boron content The protein concentration of an aliquot of boronated B72.3 monoclonal antibody was measured using the Bio-Rad protein assay reagent. An aliquot (10 pL) of the same sample was also sent for boron determination by atomic absorption (Environment One Inc., Valley View, Ohio). From these analyses, we determined that an average of 16 boron cages were attached to each antibody molecule. The specific
Fig. 3. Scans of mice bearing a LS174-T colon carcinoma tumor xenograft, injected via tail vein with boronated/iodinated antibody.
GIRISH N. RANADIVE et al.
4
Table 1. Biodistribution of RIB* (B72.3 and human IgG) and “‘I-B72.3: percent injected dose/g tissue. The results are the average of 3 mice for each time point ‘*‘I-B72.3
Tissue
RIB (B72.3)
Brain Blood Liver Gut Kidney Bone Tumor
24 h post-injection 0.27 + 0.02 0.59 f 0.14 6.69 f 0.28 5.97 f 0.37 2.96 + 0.37 2.39 f 0.56 0.82+0.13 0.77 * 0.03 3.60 k 0.27 2.21 * 0.01 0.85 f 0.18 0.75 f 0.01 8.64 f 0.10 14.16 + 3.78
Brain Blood Liver Gut Kidney Bone Tumor
0.46 4.68 2.22 0.47 2.96 0.74 8.56
Brain Blood Liver Gut Kidney Bone Tumor
0.13 * 0.09 2.87 f 1.34 1.36 f 0.63 0.31 f 0.04 1.66 f 0.59 0.43 & 0.17 7.29 + 2.09
f f f f k * f
48h post-injection 0.16 0.27 + 0.16 1.27 3.80 & 0.95 0.57 1.88 * 0.55 0.08 0.43 + 0.16 0.45 1.65 k 0.48 0.14 0.50 * 0.20 1.99 11.29 + 4.55
RIB (human IgG) 0.16 f 2.54 + 1.70 f 0.87 f 1.36 f 0.56 + 2.02 f
0.04 0.24 0.07 0.12 0.01 0.01 0.55
0.13 f 0.10 2.95 + 1.97 1.67 i 0.81 0.61 i_ 0.29 1.60+ 0.91 0.51 i 0.28 1.13 * 0.47
1”
96h post-injection
0.10 + 1.54& 0.83 f 0.21 f 0.63 f 0.21 + 6.87 f
0.07 1.07 0.40 0.13 0.41 0.14 2.53
0.07 + 0.02 2.06 k 0.16 1.02 f 0.05 0.32 of:0.05 0.94 f 0.07 0.27 f 0.03 0.77 f 0.14
*RIB = radioiodinated/boronated.
boronation was calculated to be 83.2 ng/pg protein. An aliquot of boronated B72.3 was then radiolabeled with lzsI. The radioactivity of RIBA was counted on a Packard 5650 y-counter. We calculated the specific radioactvity to be 6.5 pCi/pg protein. Therefore, 1 PCi of antibody contained 12.8 ng loB. Corrections for 1251decay were made in all subsequent studies. Immunoreactivity,
imaging and biodistribution
The immunoreactivity of radioiodinated and boronated B72.3 antibody (RIBA) was found to be 75-80% compared to radioiodinated B72.3, thus we do not see a major loss of the immunoreaci
i Table 2. Biodistribution studies of RIB* (B72.3 and human IgG) and lz51-B72.3: tumor-to-tissue ratios. The results are the average of 3 mice for each time point Tissue
RIB (B72.3)
‘251-B72.3
Brain Blood Liver Gut Kidney Bone
24 h post-injection 32.0 + 4.3 24.2 + 0.7 1.3 f 0.1 2.4 k 0.5 2.9 + 0.3 5.9kO.l 11.0&2.8 18.3 + 4.0 2.4 + 0.2 6.4 k 1.2 10.0 + 1.2 18.8 * 3.3
Brain Blood Liver Gut Kidney Bone
18.6 f 1.8 f 3.9 f 18.2 f 2.9 f 12.0 f
Brain Blood Liver Gut Kidney Bone
58.0 f 22.1 2.5 + 0.7 5.3 & 0.8 24.0 + 2.7 4.4 * 0.1 17.0 f 2.7
48 h post-injection 2.0 42.1 + 9.0 0.2 3.0 * 0.4 0.1 6.0 &-0.5 2.8 26.4 k 3.4 0.3 6.8 k 0.6 1.4 22.4 & 0.3
RIB (human IgG) 12.6 k 4.3 0.8 +O.l 1.2 * 0.4 2.3 + 0.3 1.5 If-o.4 3.6 k 0.9 8.7 + 0.4 f 0.7 f 1.9 f 0.7 + 2.2 *
1.9 0.11 0.29 0.36 0.24 0.13
96h post -injection
*RIB = radioiodinated/boronated.
70.9 + 17. I 4.5 f 0.2 8.3 + 0.7 32.6 + 3.7 ll.OkO.6 33.2 f 2.6
11.2 + 3.4 0.4+0.1 0.8 + 0.1 2.4 + 0.3 0.8 + 0.1 2.8 + 0.28
Fig. and
4. Biodistribution
of RIB-B72.3,
I
RIB
human
IgG
1251-B72.3 in athymic nude mice bearing LS17Ctumor xenografts.
tivity of the antibody on boronation. This was also reflected in the biodistribution and imaging studies. Scintigraphic and biodistribution studies were performed in athymic nude mice bearing LS174-T human colon adenocarcinoma tumor xenografts of approx. 1 cm3 in volume. Potassium iodide was added to the drinking water to block thyroid uptake of free 1251.Scintiphotos were obtained at 4, 24, 48 and 96 h after tail vein injection of 20 PCi of either RIBA (Fig. 3), 1251-B72.3 or radioiodinated boronated human IgG using a pinhole collimator on a Pho Gamma V y-camera (Siemens, Hoffman Estates, Ill.) with a Scintiview image control station (Fig. 3). Tumor uptake was noticeable in mice injected with RIBA and ‘*‘I-B72.3 but not for ‘*‘I-human IgG. Uptake was not seen in any other tissues.
A technique to prepare boronated B72.3 monoclonal antibody
5
The recent development of tumor-affinic boroncontaining compounds has necessarily led to a reevaluation of the microdosimetric requirements for BNCT. Calculations based upon uniformly distributed boron are clearly inappropriate for boronated monoclonal antibodies exhibiting any kind of tumor cell specificity. A series of Monte Carlo simulations were performed where the “B(n, u)7Li reaction is assumed to occur on the cell membrane (Kalend et al., 1991). Under these circumstances, Brain
l-b
o-
0
7
24
46
n
1 96
h Fig. 5. Localization of RIB-B72.3 antibody in the tumor,
over a period of 96 h.
RIBA and ‘*‘I-B72.3 showed selective uptake in the tumor with no accumulation in any other organ, while non-specific human IgG antibody demonstrated no uptake by the tumor or any other organ over the period of 96 h. The biodistribution of RIBA and ‘25I-B72.3 was similar except in the tumor where uptake was less for RIBA, which presumably reflects the marginal loss of immunoreactivity due to the addition of the boron cages. The RIBA percent injected dose per gram tissue for tumor ranged between 7.29 and 8.64 at each time point examined between 24 and 96 h after injection. Analysis of tumor-to-tissue ratios shows progressive clearance of RIBA from all organs with increasing time (Fig. 5). Using an injected RIBA dose of 20 pCi, on average 1.322 pCi or 16.93 ng “B localized within the tumor (i.e. 1.02 x lOI atoms of “B/g of tumor). Assuming a gram of tumor contains lo9 cells, there are 1.02 x lo6 atoms of “B per tumor cell.
the minimum number of boron atoms necessary to inactivate a tumor cell nucleus is a function of the ratio of nuclear-to-cell volumes. For nuclear-to-cell volume ratios >0.2, fewer than 100 neutron capture events are required for cell inactivation. For nuclearto-cell volume ratios of GO.1, the threshold for successful BNCT is 400 neutron capture events but the requirement rises exponentially for smaller ratios. On the other hand, for a nuclear-to-cellular ratio of one, only 4 such events are required. On the basis of these calculations, BNCT using boronated monoclonal antibodies appears to be very feasible. Our methodology to introduce BCC onto B72.3 monoclonal antibody through the formation of stable thioether linkages occurs with little or no apparent loss of immunoreactivity. The experimentally determined concentration of “B in tumor (1.02 x lo6 atoms of “B/tumor cell) appears to be more than adequate to carry out successful BNCT. Acknowledgements-This work was supported by grants from the U.S. Department of Energy (DE-FGOZ89ER60869) and the Claude Worthington Benedum Foundation. By acceptance of this article, the publisher acknowledges the U.S. Government’s right to retain a non-exclusive royalty-free license in and to any copyright covering this paper.
References Discussion It is critically important in BNCT to accumulate i”B within the target tumor so as to minimize normal tissue damage. In the early clinical trials of BNCT, extensive damage to normal vasculature appears to have been largely the result of poor discrimination between tumor-to-blood distributions (Mishima et al., 1986). Recent attempts to identify agents for the selective delivery of l”B to tumors have yielded a variety of so-called “third generation” pro-drugs. Boronated amino acids, such as p-boronophenylalanine, appear to accumulate in melanoma cells (Ichihashi et al., 1982). Furthermore, boronated porphyrins (Kahl and Micca, 1986), phthalocyanines (Alam et al., 1989), purines and pyrimidines (Schinazi and Prusoff, 1978) appear to localize within tumors. Low density lipoproteins encapsulating i”B (Laster et al., 1991) and boronated monoclonal antibodies against tumor-associated antigens (Goldenberg et al., 1982) are yet other examples of “third generation” boron-containing compounds.
Alam F., Bapat B. V., Soloway A. H., Barth R. F., Mafune N. and Adams D. M. (1989) Boronated compounds for neutron capture therapy. Strahlenther. Onkol. 165, 12. Alam F., Soloway A. H. and Barth R. F. (1987) Boronation of antibodies with mercaptoundecahydro-closododecarborate (2-) anion for potential use in boron neutron capture therapy. Appl. kadiat. Isot. 38, 503. Alam F.. Solowav A. H.. Barth R. F. et al. (1983) Boronation of polyclonal and monoclonal antibodies for neutron capture. In First International Symposium on Neutron Capture Therapy, Cambridge, Mass.,. 12-14 October 1983. .DD. 229-236. Brookhaven National Laboratories, Reports BNL-51730. Badger C. C., Krohn K. A. and Bernstein I. A. (1987) In vitro measurement of avidity of radioiodinated antibodies. Nucl. Med. Biol. 14, 605610. Colcher D., Keenan A. M., Larson S. M. and Schlom J. (1984) Prolonged binding of a radiolabeled monoclonal antibody (B72.3) for the in situ radioimmunodetection of human colon carcinoma xenografts. Cancer Res. 44, 5744.
Colcher D., Zalutsky M., Kaplan W. et al. (1983) Radiolocalization of human mammary tumors in athymic mice by a monoclonal antibody. Cancer Res. 43, 736. Fairchild R. G. (1989) Dose rate and therapeutic gain. In Clinical Aspects of Neutron Capture Therapy (Edited
6
GWSH N. RANADIVE el al.
by Fairchild R. G., Bond V. P. and Woodhead A. D.), p. 1. Plenum Press, New York. Fairchild R. G. and Bond V. P. (1985) Current status of B-lOneutron capture therapy: enhancement of tumor dose via beam liltration and dose rate, and the effects of these parameters on minimum boron content: a theoretical evaluation. ht. J. Radiat. Oncol. Biol. Phys. 11, 831. Farr L. E.. Sweet W. H.. Robertson J. S. et al. (1954) Neutron ‘capture therapy with boron in the treatment of glioblastoma multiforme. Am. J. Roentgenol. 71, 279. Gabel D., Holstein H., Larsson B. et al. (1987) Quantitative neutron capture radiography for studying the biodistribution of tumor-seeking boron-containing compounds. Cancer Res. 47, 5451. Godwin J. T., Farr L. E., Sweet W. H. and Robertson J. S. (1955) Pathological study of eight patients with glioblastoma multiforme treated by neutron capture therapy using boron 10. Cancer 8, 601. Goldenberg D. M., Sharkey R. M., Primus F. J. et al. (1982) Neutron capture therapy for human cancer: in vivo results on the preparation of boron-labeled antibodies to carcinoembryonic antigen. Proc. Natl. Acad. Sci. U.S.A. 79, 3011. Hatanaka H., Kamano S., Amano K. et al. (1986) Clinical experience of boron-neutron capture therapy of gliomas -a comparison with conventional chemo-immuno-radiotherapy. In Boron Neutron Capture Therapyfor Tumors (Edited by Hatanaka H.), p. 449. Nishimura, Niigata, Japan. Human J. L. and Cobb L. M. (1990) Nonuniformity of tumor dose in radioimmunotherapy. J. Nucl. Med. 31,75. Ichihashi M., Nakanishi J. and Mishima Y. (1982) Specific killing effect of loB-para-boranophenylananine in thermal
neutron capture therapy of malignant melanoma: in vitro radiobiological evaluation. J. Invest. Dermatol. 78, 215.
Kahl S. B. and Micca P. (1986) Chemical and biological studies on boronated tetraphenyl porphyrins. In Neutron Capture Therapy (Edited by Hatanaka H.), p. 61. Nishimura, Niigata, Japan. Kalend A. M., Bloomer W. D., Epperly M. W. et al. (1991) Cellular energy and tract densities by Monte Carlo simulation of the i”B (n,a) ‘Li reaction in human cell lines assayed with monoclonal antibodies. Med. Phys. 18, 662.
Laster B. H., Kahl S. B., Popenoe E. A., Pates D. and Fairchild R. G. (1991) Biological efficacy of boronated low-density lipoprotein for boron neutron capture therapy as measured in cell culture. Cancer Res. 51,4588. Lecher G. L. (1936) Biological effects and therapeutic possibilities of neutrons. Am. J. Roentgenof. 36, 1. Mishima Y., Ichihashi M., Tsuji M. et al. (1986) Prerequisites of first clinical trials of melanoma selective thermal neutron capture therapy. In Neutron Capture Therapy (Edited by Hatanaka H.), p. 230. Nishimura, Nigata, Japan. Schinaxi R. F. and Prusoff W. H. (1978) Synthesis and properties of boron and silicon substituted uracil or Z’deoxyuridine. Tetrahed. L&t. SO, 4981. Soloway A. H., Hatanaka H. and Davis M. A. (1967) Penetration of brain and brain tumor VII. Tumor binding sulthydryl compounds. J. Med. Chem. 10, 714. Sweet W. H. and Javid M. (1952) The possible use of neutron-capturing isotopes such as boron-10 in the treatment of neoplasms. I. Intracranial tumors. J. Neurosurg. 9, 200.
0883-2897/93 $5.00 + 0.00 Pergamon Press Ltd
A Technique to Prepare Boronated B72.3 Monoclonal Antibody for Boron Neutron Capture Therapy GIRISH N. RANADIVE’y2*, HOWARD S. ROSENZWEIG’**, MICHAEL W. EPPERLY’ and WILLIAM D. BLOOMER2 ’Department of Radiation Oncology, University of Pittsburgh, Pittsburgh, PA 15213 and *The Radiation Medicine Institute, Evanston Hospital Corporation,
Evanston, IL 60201, U.S.A.
(Received 24 June 1992)
B72.3 monoclonal antibody has been successfully boronated using mercaptoundecahydro-closo-dodecaborate (boron cage compound). The reagent was incorporated by first reacting the lysine residues of the antibody with m-maleimidobenzoyl succinimide ester (MBS), followed by Michael addition to the maleimido group by the mercapto boron cage compound to form a physiologically stable thioether linkage. Boron content of the antibody was determined by atomic absorption spectroscopy. For biodistribution studies, boronated antibody was radioiodinated with iodogen. ‘251-labeled and boronated B72.3 monoclonal antibody demonstrated clear tumor localization when administered via tail vein injections to athymic nude mice bearing LS174-T tumor xenografts. Boronated antibody was calculated to deliver lo6 boron atoms per tumor cell. Although this falls short of the specific boron content originally proposed as necessary for boron neutron capture therapy (BNCT), recent calculations suggest that far fewer atoms of “B per tumor cell would be necessary to effect successful BNCT when the boron is targeted to the tumor cell membrane.
Introduction The scientific rationale for boron neutron capture therapy (BNCT) as a potential treatment for malignant disease is based upon the nuclear reaction which occurs when ‘OBis irradiated with low energy thermal and epithermal neutrons. The resulting nuclear reaction produces alpha particles with high linear energy transfer and radiobiological effectiveness. Because the neutron capture cross section of ‘OBexceeds that of the major chemical elements in normal tissues by several orders of magnitude, competing nuclear reactions and normal tissue irradiation will be minimized if “B is selectively localized within tumors and the neutron beam focused thereon. Although the clinical potential of BNCT was recognized as early as 1936 (Lecher, 1936), it was not until the 1950s that it was actually used in the treatment of malignant tumors (Sweet and Javid, 1952; Farr et al., 1954; Godwin et al., 1955). The initial experience with glioblastoma multiforme was subsequently extended but clinical results to date
*All correspondence and reprint requests should be addressed to: Girish N. Ranadive, The Radiation Medicine Institute, 2650 Ridge Avenue, Evanston, IL 60201, U.S.A.
have been disappointing (Hatanaka et al., 1986). A major reason for these disappointing results was the lack of tumor selective boron-containing compounds, the lack of which significantly compromised therapeutic efficacy by increasing normal tisue irradiation. So-called “first” and “second” generation boron-containing compounds have generally lacked tumor specificity, resulting in low tumor-to-normal tissue uptake ratios. Current research is directed toward the design and synthesis of “third” generation boron-containing compounds which demonstrate preferential, if not specific, tumor cell binding (Fairchild and Bond, 1985). Based upon earlier calculations which assumed a homogeneous intracellular distribution of boron, a minimum boron content of approx. 3Opg/g tumor has traditionally been considered necessary for successful BNCT (Fairchild and Bond, 1985). However, by localizing boron to the tumor cell membrane as with monoclonal antibodies, the concentration of boron, calculated on a per cell or weight basis of tumor, required for effective BNCT can be markedly reduced (Humm and Cobb, 1990; Gabel et al., 1987; Fairchild, 1989). Boronated monoclonal antibodies to tumor associated antigens have previously been studied as vehicles for preferential l”B localization. Alam et al. (1987) were able to incorporate a relatively large number
2
GIRISH
N.
RANADNE
of “B atoms onto antibody but not without substantial loss of immunoreactivity. Another method for incorporating “B in antibodies, reducing existing disulfide linkages and reforming them with Na,(i”B),,H,,SH (Fairchild and Bond, 1985) led to 9-13 boron cages (108-156 boron atoms) incorporated into antithymocyte globulin (Alam et al., 1983). However, these boronated antibodies could not be studied in vivo because the disulfide linkage between the boron cage and the protein sulfhydryl group was prone to enzymatic cleavage before reacting with the antigenic sites on target lymphocytes. In this communication, we describe a method to derivatize the B72.3 monoclonal antibody with mercaptoundecahydro-close-dodecaborate, an anionic polyhedral borane compound (Soloway et al., 1967) that has been used in limited clinical trials with glioblastoma multiforme in Japan (Hatanaka et al., 1986). Our methodology to modify B72.3, forming physiologically stable thioether linkages with boron cages, leads to the incorporation of 16 cages (192 ‘OB atoms) per antibody molecule with little loss of immunoreactivity.
Materials and Methods B72.3 monoclonal antibody was a gift from Dr J. Schlom (NCI, Bethesda, Md) and was isolated from ascites fluid (Colcher et al., 1984). The LS174-T human colon adenocarcinoma was obtained from ATCC (Rockville, Md). Dicesium mercaptoundecahydro-close-dodecaborate (boron cage compound, BCC) enriched with ‘OBwas purchased from Callery Chemicals (Evans City, Pa), maleimidobenzoyl succinimidyl ester (MBS) and iodogen from Pierce (Rockford, Ill.), and Na’251 (14 mCi/mg) from Amersham Corp. (Arlington Heights, Ill.).
et al.
B72.3 monoclonal antibody (1 mg, 500 pL) in 20mM HEPES (PH 8.0) was reacted with a 50-fold molar excess of MBS and stirred for 3 h, after which excess reagent was removed by gel filtration on a Sephadex G-25 column, Antibody, recovered in the void volume was then reacted with a 50-fold molar excess of BCC and stirred at room temperature for 2 h after which excess reagent was again removed by gel filtration (Fig. 1). Protein within the void volume was concentrated on Centricon(Amicon Corp., Beverly, Mass.). Boronated B72.3 and unmodified B72.3 antibody (as a control) was radioiodinated using standard iodogen methodology (Colcher et al., 1984). Protein associated radioactivity (78% yield) was collected in the void volume using gel filtration. A tumor non-specific antibody human IgG (Jackson Laboratories, West Grove, Pa) was boronated and iodinated in analogous fashion. The purity of the boronated antibody was checked by HPLC on Superdex-75 (HR 10/30) column (Pharmacia, Piscataway, N.J.) using 20 mM Tris at pH 8.0, with a 1.OmL flow rate (Fig. 2). Immunoreactivity and biodistribution
The immunoreactivity of the radioiodinated and boronated B72.3 (RIBA) was measured using LSl74-T human colon adenocarcinoma cells obtained from xenografts grown in athymic nude mice (Harlan-Sprague-Dawley, Indianapolis, Ind.). LSl74-T cells (1 x 10’) were injected subcutaneously in the hind leg. Ten days later, the xenografts were removed and single cell suspensions prepared using Dispase (Boehringer Mannheim Corp., Indianapolis, Ind.). Single cell suspensions ranging from 0.1 to 5 x 10’ cells were placed in test tubes to which 1.0 pg of RIBA were added. Cell suspensions were
0 +
SH
A
=
Fig. 1. Reaction scheme for boronation of the antibody.
A technique to prepare boronated B72.3 monoclonal antibody
3
Fig. 2. HPLC analysis of boronated antibody on a Superdex-75 (HR 10/30) column. incubated for 1 h at 37”C, then washed twice with Hank’s balanced salt solution and the radioactivity counted (Badger et al., 1987). The biodistribution of the ‘251-labelled antibodies was determined by sacrificing the mice by nembutal overdose, and excising, washing, blotting, weighing and counting the various organs in a Packard 5650 y-counter (Packard Instrument Co., Downers Grove, Ill.). The percent injected dose per gram of tissue (Table 1 and Fig. 4) and tumor-to-tissue ratios (Table 2) were calculated.
Results Determination of specljic boron content The protein concentration of an aliquot of boronated B72.3 monoclonal antibody was measured using the Bio-Rad protein assay reagent. An aliquot (10 pL) of the same sample was also sent for boron determination by atomic absorption (Environment One Inc., Valley View, Ohio). From these analyses, we determined that an average of 16 boron cages were attached to each antibody molecule. The specific
Fig. 3. Scans of mice bearing a LS174-T colon carcinoma tumor xenograft, injected via tail vein with boronated/iodinated antibody.
GIRISH N. RANADIVE et al.
4
Table 1. Biodistribution of RIB* (B72.3 and human IgG) and “‘I-B72.3: percent injected dose/g tissue. The results are the average of 3 mice for each time point ‘*‘I-B72.3
Tissue
RIB (B72.3)
Brain Blood Liver Gut Kidney Bone Tumor
24 h post-injection 0.27 + 0.02 0.59 f 0.14 6.69 f 0.28 5.97 f 0.37 2.96 + 0.37 2.39 f 0.56 0.82+0.13 0.77 * 0.03 3.60 k 0.27 2.21 * 0.01 0.85 f 0.18 0.75 f 0.01 8.64 f 0.10 14.16 + 3.78
Brain Blood Liver Gut Kidney Bone Tumor
0.46 4.68 2.22 0.47 2.96 0.74 8.56
Brain Blood Liver Gut Kidney Bone Tumor
0.13 * 0.09 2.87 f 1.34 1.36 f 0.63 0.31 f 0.04 1.66 f 0.59 0.43 & 0.17 7.29 + 2.09
f f f f k * f
48h post-injection 0.16 0.27 + 0.16 1.27 3.80 & 0.95 0.57 1.88 * 0.55 0.08 0.43 + 0.16 0.45 1.65 k 0.48 0.14 0.50 * 0.20 1.99 11.29 + 4.55
RIB (human IgG) 0.16 f 2.54 + 1.70 f 0.87 f 1.36 f 0.56 + 2.02 f
0.04 0.24 0.07 0.12 0.01 0.01 0.55
0.13 f 0.10 2.95 + 1.97 1.67 i 0.81 0.61 i_ 0.29 1.60+ 0.91 0.51 i 0.28 1.13 * 0.47
1”
96h post-injection
0.10 + 1.54& 0.83 f 0.21 f 0.63 f 0.21 + 6.87 f
0.07 1.07 0.40 0.13 0.41 0.14 2.53
0.07 + 0.02 2.06 k 0.16 1.02 f 0.05 0.32 of:0.05 0.94 f 0.07 0.27 f 0.03 0.77 f 0.14
*RIB = radioiodinated/boronated.
boronation was calculated to be 83.2 ng/pg protein. An aliquot of boronated B72.3 was then radiolabeled with lzsI. The radioactivity of RIBA was counted on a Packard 5650 y-counter. We calculated the specific radioactvity to be 6.5 pCi/pg protein. Therefore, 1 PCi of antibody contained 12.8 ng loB. Corrections for 1251decay were made in all subsequent studies. Immunoreactivity,
imaging and biodistribution
The immunoreactivity of radioiodinated and boronated B72.3 antibody (RIBA) was found to be 75-80% compared to radioiodinated B72.3, thus we do not see a major loss of the immunoreaci
i Table 2. Biodistribution studies of RIB* (B72.3 and human IgG) and lz51-B72.3: tumor-to-tissue ratios. The results are the average of 3 mice for each time point Tissue
RIB (B72.3)
‘251-B72.3
Brain Blood Liver Gut Kidney Bone
24 h post-injection 32.0 + 4.3 24.2 + 0.7 1.3 f 0.1 2.4 k 0.5 2.9 + 0.3 5.9kO.l 11.0&2.8 18.3 + 4.0 2.4 + 0.2 6.4 k 1.2 10.0 + 1.2 18.8 * 3.3
Brain Blood Liver Gut Kidney Bone
18.6 f 1.8 f 3.9 f 18.2 f 2.9 f 12.0 f
Brain Blood Liver Gut Kidney Bone
58.0 f 22.1 2.5 + 0.7 5.3 & 0.8 24.0 + 2.7 4.4 * 0.1 17.0 f 2.7
48 h post-injection 2.0 42.1 + 9.0 0.2 3.0 * 0.4 0.1 6.0 &-0.5 2.8 26.4 k 3.4 0.3 6.8 k 0.6 1.4 22.4 & 0.3
RIB (human IgG) 12.6 k 4.3 0.8 +O.l 1.2 * 0.4 2.3 + 0.3 1.5 If-o.4 3.6 k 0.9 8.7 + 0.4 f 0.7 f 1.9 f 0.7 + 2.2 *
1.9 0.11 0.29 0.36 0.24 0.13
96h post -injection
*RIB = radioiodinated/boronated.
70.9 + 17. I 4.5 f 0.2 8.3 + 0.7 32.6 + 3.7 ll.OkO.6 33.2 f 2.6
11.2 + 3.4 0.4+0.1 0.8 + 0.1 2.4 + 0.3 0.8 + 0.1 2.8 + 0.28
Fig. and
4. Biodistribution
of RIB-B72.3,
I
RIB
human
IgG
1251-B72.3 in athymic nude mice bearing LS17Ctumor xenografts.
tivity of the antibody on boronation. This was also reflected in the biodistribution and imaging studies. Scintigraphic and biodistribution studies were performed in athymic nude mice bearing LS174-T human colon adenocarcinoma tumor xenografts of approx. 1 cm3 in volume. Potassium iodide was added to the drinking water to block thyroid uptake of free 1251.Scintiphotos were obtained at 4, 24, 48 and 96 h after tail vein injection of 20 PCi of either RIBA (Fig. 3), 1251-B72.3 or radioiodinated boronated human IgG using a pinhole collimator on a Pho Gamma V y-camera (Siemens, Hoffman Estates, Ill.) with a Scintiview image control station (Fig. 3). Tumor uptake was noticeable in mice injected with RIBA and ‘*‘I-B72.3 but not for ‘*‘I-human IgG. Uptake was not seen in any other tissues.
A technique to prepare boronated B72.3 monoclonal antibody
5
The recent development of tumor-affinic boroncontaining compounds has necessarily led to a reevaluation of the microdosimetric requirements for BNCT. Calculations based upon uniformly distributed boron are clearly inappropriate for boronated monoclonal antibodies exhibiting any kind of tumor cell specificity. A series of Monte Carlo simulations were performed where the “B(n, u)7Li reaction is assumed to occur on the cell membrane (Kalend et al., 1991). Under these circumstances, Brain
l-b
o-
0
7
24
46
n
1 96
h Fig. 5. Localization of RIB-B72.3 antibody in the tumor,
over a period of 96 h.
RIBA and ‘*‘I-B72.3 showed selective uptake in the tumor with no accumulation in any other organ, while non-specific human IgG antibody demonstrated no uptake by the tumor or any other organ over the period of 96 h. The biodistribution of RIBA and ‘25I-B72.3 was similar except in the tumor where uptake was less for RIBA, which presumably reflects the marginal loss of immunoreactivity due to the addition of the boron cages. The RIBA percent injected dose per gram tissue for tumor ranged between 7.29 and 8.64 at each time point examined between 24 and 96 h after injection. Analysis of tumor-to-tissue ratios shows progressive clearance of RIBA from all organs with increasing time (Fig. 5). Using an injected RIBA dose of 20 pCi, on average 1.322 pCi or 16.93 ng “B localized within the tumor (i.e. 1.02 x lOI atoms of “B/g of tumor). Assuming a gram of tumor contains lo9 cells, there are 1.02 x lo6 atoms of “B per tumor cell.
the minimum number of boron atoms necessary to inactivate a tumor cell nucleus is a function of the ratio of nuclear-to-cell volumes. For nuclear-to-cell volume ratios >0.2, fewer than 100 neutron capture events are required for cell inactivation. For nuclearto-cell volume ratios of GO.1, the threshold for successful BNCT is 400 neutron capture events but the requirement rises exponentially for smaller ratios. On the other hand, for a nuclear-to-cellular ratio of one, only 4 such events are required. On the basis of these calculations, BNCT using boronated monoclonal antibodies appears to be very feasible. Our methodology to introduce BCC onto B72.3 monoclonal antibody through the formation of stable thioether linkages occurs with little or no apparent loss of immunoreactivity. The experimentally determined concentration of “B in tumor (1.02 x lo6 atoms of “B/tumor cell) appears to be more than adequate to carry out successful BNCT. Acknowledgements-This work was supported by grants from the U.S. Department of Energy (DE-FGOZ89ER60869) and the Claude Worthington Benedum Foundation. By acceptance of this article, the publisher acknowledges the U.S. Government’s right to retain a non-exclusive royalty-free license in and to any copyright covering this paper.
References Discussion It is critically important in BNCT to accumulate i”B within the target tumor so as to minimize normal tissue damage. In the early clinical trials of BNCT, extensive damage to normal vasculature appears to have been largely the result of poor discrimination between tumor-to-blood distributions (Mishima et al., 1986). Recent attempts to identify agents for the selective delivery of l”B to tumors have yielded a variety of so-called “third generation” pro-drugs. Boronated amino acids, such as p-boronophenylalanine, appear to accumulate in melanoma cells (Ichihashi et al., 1982). Furthermore, boronated porphyrins (Kahl and Micca, 1986), phthalocyanines (Alam et al., 1989), purines and pyrimidines (Schinazi and Prusoff, 1978) appear to localize within tumors. Low density lipoproteins encapsulating i”B (Laster et al., 1991) and boronated monoclonal antibodies against tumor-associated antigens (Goldenberg et al., 1982) are yet other examples of “third generation” boron-containing compounds.
Alam F., Bapat B. V., Soloway A. H., Barth R. F., Mafune N. and Adams D. M. (1989) Boronated compounds for neutron capture therapy. Strahlenther. Onkol. 165, 12. Alam F., Soloway A. H. and Barth R. F. (1987) Boronation of antibodies with mercaptoundecahydro-closododecarborate (2-) anion for potential use in boron neutron capture therapy. Appl. kadiat. Isot. 38, 503. Alam F.. Solowav A. H.. Barth R. F. et al. (1983) Boronation of polyclonal and monoclonal antibodies for neutron capture. In First International Symposium on Neutron Capture Therapy, Cambridge, Mass.,. 12-14 October 1983. .DD. 229-236. Brookhaven National Laboratories, Reports BNL-51730. Badger C. C., Krohn K. A. and Bernstein I. A. (1987) In vitro measurement of avidity of radioiodinated antibodies. Nucl. Med. Biol. 14, 605610. Colcher D., Keenan A. M., Larson S. M. and Schlom J. (1984) Prolonged binding of a radiolabeled monoclonal antibody (B72.3) for the in situ radioimmunodetection of human colon carcinoma xenografts. Cancer Res. 44, 5744.
Colcher D., Zalutsky M., Kaplan W. et al. (1983) Radiolocalization of human mammary tumors in athymic mice by a monoclonal antibody. Cancer Res. 43, 736. Fairchild R. G. (1989) Dose rate and therapeutic gain. In Clinical Aspects of Neutron Capture Therapy (Edited
6
GWSH N. RANADIVE el al.
by Fairchild R. G., Bond V. P. and Woodhead A. D.), p. 1. Plenum Press, New York. Fairchild R. G. and Bond V. P. (1985) Current status of B-lOneutron capture therapy: enhancement of tumor dose via beam liltration and dose rate, and the effects of these parameters on minimum boron content: a theoretical evaluation. ht. J. Radiat. Oncol. Biol. Phys. 11, 831. Farr L. E.. Sweet W. H.. Robertson J. S. et al. (1954) Neutron ‘capture therapy with boron in the treatment of glioblastoma multiforme. Am. J. Roentgenol. 71, 279. Gabel D., Holstein H., Larsson B. et al. (1987) Quantitative neutron capture radiography for studying the biodistribution of tumor-seeking boron-containing compounds. Cancer Res. 47, 5451. Godwin J. T., Farr L. E., Sweet W. H. and Robertson J. S. (1955) Pathological study of eight patients with glioblastoma multiforme treated by neutron capture therapy using boron 10. Cancer 8, 601. Goldenberg D. M., Sharkey R. M., Primus F. J. et al. (1982) Neutron capture therapy for human cancer: in vivo results on the preparation of boron-labeled antibodies to carcinoembryonic antigen. Proc. Natl. Acad. Sci. U.S.A. 79, 3011. Hatanaka H., Kamano S., Amano K. et al. (1986) Clinical experience of boron-neutron capture therapy of gliomas -a comparison with conventional chemo-immuno-radiotherapy. In Boron Neutron Capture Therapyfor Tumors (Edited by Hatanaka H.), p. 449. Nishimura, Niigata, Japan. Human J. L. and Cobb L. M. (1990) Nonuniformity of tumor dose in radioimmunotherapy. J. Nucl. Med. 31,75. Ichihashi M., Nakanishi J. and Mishima Y. (1982) Specific killing effect of loB-para-boranophenylananine in thermal
neutron capture therapy of malignant melanoma: in vitro radiobiological evaluation. J. Invest. Dermatol. 78, 215.
Kahl S. B. and Micca P. (1986) Chemical and biological studies on boronated tetraphenyl porphyrins. In Neutron Capture Therapy (Edited by Hatanaka H.), p. 61. Nishimura, Niigata, Japan. Kalend A. M., Bloomer W. D., Epperly M. W. et al. (1991) Cellular energy and tract densities by Monte Carlo simulation of the i”B (n,a) ‘Li reaction in human cell lines assayed with monoclonal antibodies. Med. Phys. 18, 662.
Laster B. H., Kahl S. B., Popenoe E. A., Pates D. and Fairchild R. G. (1991) Biological efficacy of boronated low-density lipoprotein for boron neutron capture therapy as measured in cell culture. Cancer Res. 51,4588. Lecher G. L. (1936) Biological effects and therapeutic possibilities of neutrons. Am. J. Roentgenof. 36, 1. Mishima Y., Ichihashi M., Tsuji M. et al. (1986) Prerequisites of first clinical trials of melanoma selective thermal neutron capture therapy. In Neutron Capture Therapy (Edited by Hatanaka H.), p. 230. Nishimura, Nigata, Japan. Schinaxi R. F. and Prusoff W. H. (1978) Synthesis and properties of boron and silicon substituted uracil or Z’deoxyuridine. Tetrahed. L&t. SO, 4981. Soloway A. H., Hatanaka H. and Davis M. A. (1967) Penetration of brain and brain tumor VII. Tumor binding sulthydryl compounds. J. Med. Chem. 10, 714. Sweet W. H. and Javid M. (1952) The possible use of neutron-capturing isotopes such as boron-10 in the treatment of neoplasms. I. Intracranial tumors. J. Neurosurg. 9, 200.