Catabolism of radioiodinated murine monoclonal antibody F(ab')2

Jul 1, 1995 - ... Keith O. Webber, Ira Pastan, Otto Gansow, and William C. Eckelman ... Catherine F. Foulon, Kevin L. Alston, and Michael R. Zalutsky...
0 downloads 0 Views 1MB Size
Bioconjugate Chem. 1995, 6,493-501

493

Catabolism of Radioiodinated Murine Monoclonal Antibody F(ab’)2 Fragment Labeled Using N-Succinimidyl3-Iodobenzoateand Iodogen Methods Pradeep K. Garg, Kevin L. Alston, and Michael R. Zalutsky* Department of Radiology, Duke University Medical Center, Durham, North Carolina 27710. Received February 7, 1995@

The F(ab’)z fragment of monoclonal antibody (MAb) Mel-14 was labeled with lZ5Iusing the Iodogen method and by reaction with N-succinimidyl 3-[12511iodobenzoate(SIB). The labeled catabolites generated after exposure to tissue homogenates i n vitro and following administration of labeled F(ab’h into normal mice were investigated by size-exclusion HPLC, gel electrophoresis, and reverse-phase HPLC. Rapid conversion of F(ab‘)z to Fab was observed with both labeling methods. With F(ab‘)z labeled using the Iodogen method, the primary low molecular weight catabolites appeared to be [12511iodideand, to a lesser extent, mono[12511iodotyrosine.With SIB, [12511iodideand [12511iodobenzoic acid (IBA) as well as the glycine and lysine conjugates of IBA were all observed. Differences in low molecular weight catabolic products could explain the more rapid normal tissue clearance with MAbs and MAb fragments labeled with SIB compared with those labeled using iodogen.

INTRODUCTION

The impact of radiolabeled monoclonal antibodies (MAbs) on the clinical management of cancer remains limited. This has led to numerous strategies for optimization of this approach for tumor diagnosis and therapy including the development of improved radiolabeling methodology. A key criterion for evaluating the suitability of a protein labeling method for in vivo applications is the extent to which normal tissue uptake can be avoided. While tissue distribution measurements can provide valuable information pertinent to this problem, an understanding of the nature of the labeled species created in the catabolism of labeled MAbs may be even more instructive, particularly in facilitating the design of next generation acylation agents. The catabolism of labeled MAbs has been investigated directly in only a few publications, and these have dealt with intact IgG and F(ab’Iz fragments labeled with lllIn, and 99mTc. These studies have utilized a variety of analytical techniques including HPLC, SDSPAGE, and thin layer chromatography to characterize the labeled species created following i n vitro exposure to liver homogenates (1) and i n vivo administration in mice (2-5). These studies have documented the existence of multiple labeled catabolites of both high and low molecular weight and differences in stability between different labeling methods (3,5) and radionuclides (1 and attempted to explain the mechanisms responsible for increased retention of radiometals compared with radioiodine in normal tissues. The more rapid normal tissue clearance of radioiodinated MAbs and fragments compared with radiometals is one factor that has led to the continued use of 1311for MAb labeling in spite of its less than ideal nuclear decay properties. Iodine-131remains the most commonly used radionuclide in clinical radioimmunotherapy trials, and encouraging responses have been observed in certain

* Correspondence should be addressed to this author at the Department of Radiology, Box 3808, Duke University Medical Center, Durham, NC 27710. Phone: (919) 684-7708; FAX: (919) 684-7121. Abstract published in Advance ACS Abstracts, July 1,1995. @

patient populations (6-9). Clinical investigations have utilized MAbs labeled by direct methods which result predominantly in the formation of an iodinated tyrosine residue on the protein (lo), and indirect evidence, such as thyroid uptake, has suggested that MAbs labeled using these methods undergo dehalogenation in vivo (11, 12). To circumvent this problem, alternative radioiodination strategies have been developed, the most successful of which involves the synthesis of N-succinimidyl 3- or 4-[13111iodobenzoate(SIB)by iododestannylation, followed by reaction of SIB with the MAb (13-17). Significant reduction in thyroid uptake compared with MAbs labeled using direct methods generally has been observed, and ’’ in one report, radioiodination of a MAb using SIB increased its effectiveness for the treatment of human tumor xenografts in athymic mice (18). The nature of the catabolites generated from radioiodinated MAbs has not been investigated extensively. With conventionally labeled MAbs, loss of label generally is assumed to occur via deiodination. However, the production of free iodide has only been inferred indirectly, through the accumulation of activity in the thyroid and stomach (13). A better understanding the nature of the labeled species generated from the catabolism of radioiodinated MAbs is clearly needed. In this investigation, we report our preliminary observations on the catabolic products generated from a radioiodinated F(ab’Iz fragment labeled using the iodogen and SIB methods. EXPERIMENTAL PROCEDURES

Materials. Sodium [1251]iodidein pH 7-11 NaOH was obtained from DuPont-New England Nuclear. Iodogen was purchased from Pierce Chemical Co. Mel-14 F(ab‘Iz was obtained as a gift from Dr. Dare11 Bigner of Department of Pathology, Duke University Medical Center. This MAb reacts with the chondroitin sulfate proteoglycan present on human gliomas and melanomas (19). Tumor targeting of Mel-14 F(ab’)z radioiodinated using the iodogen (20) and SIB methods (21,22) has been documented in athymic mouse xenograft models, and I3lI-labeled Mel-14 F(ab’Iz is currently being evaluated in clinical radioimmunotherapy trials (9). The purifica-

1 043-1802/95/2906-0493$09.00/0 0 1995 American Chemical Society

494 Bioconjugate Chem., Vol. 6,No. 4, 1995

tion and fragmentation of this murine IgGza F(ab'Iz fragment have been described in a previous publication (20). All other reagents were purchased from Aldrich Chemical Co. General. NMR spectra were recorded on a General Electric Midfield GN-300 spectrometer. Chemical shifts for protons are reported in ppm downfield from an internal tetramethylsilane standard (0.00 ppm). Mass spectral data and elemental analyses were provided by Oneida Research Services (Whitesboro, NY). Melting points were taken on a Haake-Buchler variable heat apparatus and are uncorrected. Radioiodination of Mel-14 F(ab)B Using Iodogen. Mel-14 F(ab')z (2.5 mg/mL in 100 pL of PBS) and lZ5I(350 pCi) were added to a glass tube coated with 10 pg of iodogen and allowed to incubate at room temperature for 10 min. The labeled MAb fragment was isolated in 80% yield by performing gel-filtration chromatography using a Sephadex G-25 column eluted with 100 mM PBS. Protein-associated activity, determined by trichloroacetic acid precipitation, was 97%. Radioiodination of Mel-14 F(abIz Using SIB. Synthesis of N-succinimidyl 3-(tri-n-butylstannyl)benzoate and its subsequent reaction with radioiodine to produce SIB has been described (13, 23). Briefly, SIB was prepared by adding 10 pL of 5% acetic acid, 20 pmol of tert-butyl hydroperoxide, and 5 pmol of N-succinimidyl 3-(tri-n-butylstannyl)benzoateto the sodium [12511iodide solution. After a 10 min reaction a t room temperature, SIB was isolated by HPLC using a silica column (Alltech, Adsorbosphere-10, 10 pm, 250 x 4.6 mm) eluted with ethyl acetate/hexane/acetic acid (30:70:0.2). After solvent evaporation, Mel-14 F(ab')z (250 pg; 5 mg/mL) was added to the SIB residue and incubated for 15 min at room temperature. After termination of the reaction by addition of 200 pL of 0.2 M glycine in borate buffer (pH 8.51, the lZ5I-labeledMAb fragment was isolated using a Sephadex G-25 gel filtration column eluted with 100 mM PBS. Protein-associated activity, determined by trichloroacetic acid precipitation, was 99%. Preparation of Standards for HPLC Analysis. To facilitate interpretation of the reverse-phase HPLC analyses of lower molecular weight catabolites, the glycine and lysine conjugates of 3-iodobenzoic acid (IBA) were synthesized. SIB (unlabeled) was first prepared by dissolving IBA (2.0 g, 0.01 M) in 100 mL of THF followed by the addition of 2.06 g of dicyclohexylcarbodiimide and 1.15 g of N-hydroxysuccinimide. After stirring for 6 h a t room temperature, the precipitated dicyclohexylurea was filtered off and the solvent was evaporated on a rotary evaporator. The desired compound was obtained as a white crystalline compound, mp 154-155 "C, by purifying the residue on a silica column eluted with hexane, with 10% ethyl acetate in hexane, and finally, with 30% ethyl acetate in hexane. To produce 2-[N-(3-iodobenzamido)lacetic acid, the glycine conjugate of IBA (IBA-Gly),glycine (25 mg), was dissolved in 100 pL of a (50:50) DMFhorate buffer, pH 8.5, mixture and added to a solution of SIB (115 mg) in 200 pL of THF. The mixture was stirred for 2 h a t room temperature, and the required compound was separated as a white crystalline compound (mp 140-141°C) from the reaction mixture using a silica column. Mass spectra, m l z (E1 mode) 306, 288, 261,231, 225. Anal. Calcd for CgHsNOJ: C, 35.43; H, 2.64; N, 4.59. Found: C, 35.90; H, 2.74; N, 4.90. The lysine conjugate of IBA (IBA-Lys), 6-[N-(3-iodobenzamido)l-2-aminocaproicacid, was prepared by first reacting a solution of SIB (345 mg) in anhydrous THF with N a-t-Boc-lysine (246 mg) dissolved in the borate buffer (200 pL; pH 8.5). After stirring at

Garg et al. room temperature for 4 h, the t-Boc derivative was isolated in 65% yield by employing silica gel flash column chromatography. Anal. Calcd for CISHZ~NZO~I: C, 45.38; H, 5.25; N, 5.88. Found: C, 45.14; H, 5.56; N, 5.45. The t-Boc protective group was then removed by treating with trifluoroacetic acid in methylene chloride at 60 "C for 30 min. The solvent was evaporated off, and the desired compound was isolated in 55% radiochemical yield as a white compound (mp 261-263 "C). Mass spectra, m l z (E1 mode) 377, 333, 314, 265, 225. Anal. Calcd for C I ~ H ~ ~ N ZC, O41.49; ~ I : H, 4.52; N, 7.45. Found: C, 41.81; H, 4.62; N, 7.29. HPLC Analysis. Radiolabeled catabolite analyses by HPLC were performed on a Beckman System Gold package which included a Model 126 programmable solvent module, a Model 168 diode array detector, and a Model 170 radioisotope detector. Data analysis was accomplished using Beckman Gold software V 7.11 on an IBM computer. Size-exclusion chromatography was performed using a Bio-Si1 SEC 250 gel filtration column (Bio-Rad, 600 x 7.5 mm) eluted with PBS at a flow rate of 1 mumin. Fractions of 0.5 mL were collected and counted using an automated y counter. Molecular weight assignments were made based on comparison with gel filtration molecular weight standards (Bio-Rad) which were run under identical conditions. Compounds with molecular masses of less than 10 kDa (LMW) were separated from the tissue homogenate supernatants using Centricon-10 filtering cartridges (Amicon) and were analyzed by HPLC using a reversephase column (Alltech, Adsorbosphere (2-18 10 pm, 250 x 4.6 mm) eluted in isocratic mode with MeOWHzO/ AcOH (45:55:0.2) at a flow rate of 1 mumin. One milliliter fractions were collected and counted for lZ5I activity using an automated y counter. Cold compounds, including IBA, IBA-Gly, and IBA-Lys, were analyzed by HPLC to determine the retention time for these potential catabolic products. Because of the potential for variability in retention time with column age and minor change in buffer preparation for the reverse-phase column, these HPLC standards were run before each series of tissue analyses. SDS-PAGE. Iodine-125-labeled Mel-14 F(ab')z and aliquots of tissue supernatants were analyzed by SDSPAGE using 4-20% gradient gels (Bio-Rad) under nonreducing conditions. Dried gels were exposed to X-ray film (Ektascan MEM-1, Kodak, Rochester, NY),and the distribution of radioactivity among the different bands was analyzed using a Bio-Rad GS-670 imaging densitometer. Ex Vivo Studies. Urine and blood were collected from 2-3 normal Balb/c mice. Additional groups of 2-3 mice were killed by halothane overdose, and the liver, spleen, and kidneys were excised, washed with saline, chopped coarsely with scissors, and homogenized in a hand-held glass tissue homogenizer. Depending on organ weight, between 0.2 and 1.0 g of each tissue homogenate, as well as 0.2-0.5 mL of urine and blood, was incubated a t 37 "C with 5-10 pCi of 1251-labeledMel-14 F(ab')z. The generalized scheme for sample analysis is summarized in Figure 1. After 3 and 24 h, homogenates and biological fluids were centrifuged at 1300g for 15 min, and the pellet was washed twice with 300 pL of PBS, pH 7.4. Aliquots of both the combined supernatants and pellets were assayed for lz5Ito determine the fraction of the activity remaining in the pellet. The supernatant was passed through a 0.45 pm filter and analyzed using sizeexclusion HPLC as well as SDS-PAGE. A portion of the supernatant was centrifuged through a Centricon-10 cartridge (Millipore, Bedford, MA) to isolate catabolites

Bioconjugate Chem., Vol. 6,No. 4, 1995 495

Catabolism of Radioiodinated F(ab’)? Fragment

1

Mel-14 F(ab’)2

[Tissue Eomogenate CentFLluge 1300g, 15 min

r

i

Pellet

I

I I

20

30

IODOGEN

I

I I ’

Supernatant

t

t

Count

+ + +

Count Size exclusion 1

1-

SIB

1

0.4! I

amfilter

I Filtrate , Centricon 10

Reverse phase HPLC

Count

I

Figure 1. Scheme used for sample preparation and analysis of catabolites from radioiodinated Mel- 14 F(ab’)z. Table 1. Tissue Distribution of Radioiodine Following Injection of 12%LabeledMel-14 F(ab)2 in Normal Mice Prepared Using Iodogen and SIB Methods % injected dose per organR

3 h postinjection tissue SIB iodogen liver 5.37 f 0.66 7.63 f 0.43 spleen 0.34 i 0.03 0.46 f 0.05 lungs 1.79 i 0.11 2.87 f 0.50 kidneys 10.60 f 0.10 14.30 f 0.50 blood 21.90 f 2.40 29.60 f 1.80 a

24 h postinjection SIB 0.51 i 0.05 0.04 i 0.01 0.14 i 0.01 0.52 f 0.03 1.21 f 0.07

iodogen 1.12 f 0.13 0.08 f 0.01 0.19 f 0.01 1.22 f 0.41 2.53 i 0.20

Mean f standard deviation.

with molecular weights less than 10 kDa. These were analyzed by reverse-phase HPLC. Analysis of Labeled Catabolites Generated in Vivo. Experiments were performed using 1251-labeled Mel-14 F(ab’)z prepared by both the iodogen and SIB methods. Balb/c mice were injected intravenously with 10 pCi (5-7 pg) of radioiodinated Mel-14 F(ab‘)z, and groups of 3 animals were killed by halothane overdose after 3 and 24 h. Urine was collected before killing the mice. Liver, spleen, lung, kidneys, and blood were removed and washed with saline. After counting the tissues for lz5Iactivity, analysis of labeled catabolites was performed using the same procedures described for the in vitro studies. RESULTS

The tissue distribution of lz5I activity was measured to ensure that differences in normal tissue uptake observed in prior studies existed in the animals used in the catabolism experiments. As summarized in Table 1, accumulation of lZ5Iactivity in normal tissues following injection of lZ5I-labeledMel-14 F(ab‘)z was higher when the iodogen method was used for radioiodination. Even at 3 h, significantly lower normal tissue levels were seen with F(ab’Izlabeled using SIB (P