Bioconjugate Chem. 1997, 8, 595−604
595
Development of a Streptavidin-Anti-Carcinoembryonic Antigen Antibody, Radiolabeled Biotin Pretargeting Method for Radioimmunotherapy of Colorectal Cancer. Studies in a Human Colon Cancer Xenograft Model† Robert M. Sharkey,*,‡ Habibe Karacay,§ Gary L. Griffiths,§ Thomas M. Behr,‡ Rosalyn D. Blumenthal,‡ M. Jules Mattes,‡ Hans J. Hansen,§ and David M. Goldenberg‡ Garden State Cancer Center, Belleville, New Jersey 07109, and Immunomedics, Inc., Morris Plains, New Jersey 07950. Received February 25, 1997X
Pretargeting methodologies can produce high tumor:blood ratios, but their role in cancer radioimmunotherapy (RAIT) is uncertain. A pretargeting method was developed using a streptavidin (StAv) conjugate of MN-14 IgG, an anti-carcinoembryonic antigen (CEA) murine monoclonal antibody (mab) as the primary targeting agent, an anti-idiotype antibody (WI2 IgG) as a clearing agent, and DTPAor DOTA-conjugated biotin as the radiolabeled targeting agent. A variety of reagents and conditions were examined to optimize this method. At 3 h, 111In-DTPA-peptide-biotin tumor uptake was 3.9 ( 0.8 % per gram and tumor:blood ratios were >11:1. By 24 h, this ratio was 178:1, but tumor accretion declined in accordance with the gradual loss of StAv-MN-14 from the tumor. Tissue retention was highest in the liver and kidneys, but their tumor:organ ratios were >2:1. Dosimetry predicted that radiolabeled MN-14 alone would deliver higher tumor doses than this pretargeting method. Increasing the specific activity and using DOTA-biotin in place of DTPA increased tumor uptake nearly 2-fold, but analysis of StAv-MN-14’s biotin-binding capacity indicated over 90% of the initial biotin-binding sites were blocked within 24 h. Animals fed a biotin-deficient diet had 2-fold higher 111In-DOTAbiotin uptake in the tumor, but higher uptake also was observed in all normal tissues. Although exceptionally adept at achieving high tumor:blood ratios rapidly, the tumor uptake of radiolabeled biotin with this pretargeting method is significantly (p < 0.0001) lower than that with a radiolabeled antibody. Endogenous biotin and enhanced liver and kidney uptake may limit the application of this method to RAIT, especially when evaluating the method in animals, but with strategies to overcome these limitations, this pretargeting method could be an effective therapeutic alternative.
INTRODUCTION
Improving the therapeutic window for RAIT requires either increasing the amount and duration of radiolabeled antibody delivered to the tumor or decreasing the amount and residence time in normal tissues. Maximum tumor accretion is observed with whole IgG, but with a blood clearance time of 2-3 days in patients, the red marrow also is exposed to a substantial radiation dose. Bivalent and monovalent antibody fragments are cleared more rapidly from the blood with higher tumor:blood ratios, and thus, a higher radiation dose can be given. For example, nude mice can survive 275 µCi of 131Ilabeled IgG, but with F(ab′)2 fragments, the dose can be escalated to 1.0-1.2 mCi (1). However, antibody fragments have lower tumor uptake and a shorter half-life in the tumor than IgG (2). There have been reports in some animal models of improved antitumor effects with F(ab′)2 fragments over whole IgG (3), but we could not observe an appreciable advantage of a single injection of an antibody fragment in comparison to whole IgG in animals (4). Fab and even single-chain antibodies (scFv) have not been widely used for therapy due to their low tumor and high renal uptake. Renal accretion of fragments can be reduced substantially by basic amino acids † Presented at the Sixth Conference on Radioimmunodetection and Radioimmunotherapy of Cancer in Princeton, NJ, on October 10-12, 1996. * Address for correspondence and reprints: Robert M. Sharkey, Ph.D., GSCC, 520 Belleville Ave., Belleville, NJ 07109. ‡ Garden State Cancer Center. § Immunomedics, Inc. X Abstract published in Advance ACS Abstracts, July 1, 1997.
S1043-1802(97)00101-8 CCC: $14.00
(5-7), and Behr et al. (8) showed that 90Y-labeled Fab fragments have antitumor effects rivaling that of whole IgG. Thus, with their smaller size enabling them to penetrate tumors more uniformly than whole IgG (9, 10), antibody fragments may yet become highly effective radiotherapeutics. Other approaches to retaining high tumor uptake while reducing blood concentrations of radiolabeled antibody have been explored. Our studies (11, 12), as well as those of others (13-15), showed that the administration of antiantibodies can reduce blood concentrations of radiolabeled antibodies very quickly. The rapid clearance of the antibody from the blood also reduces the level of antibody in the tumor, but it usually take 1-2 days before tumor levels decrease substantially. Early indications suggest that 131I-NP-4 anti-carcinoembryonic antigen (CEA)1 IgG combined with an anti-idiotype antibody given at 24 h improves antitumor activity in tumor-bearing nude mice (16). Since the anti-antibody and radiolabeled antibody complexes are cleared predominantly by the liver, this procedure is ideally suited only for radioiodinated and not radiometal-labeled (i.e., 90Y) antibodies. Extracorporeal removal of radiolabeled antibodies from the blood is also possible (17, 18). This procedure allows antibodies with any radiolabel to be used. One clinical study suggested that this method reduced tumor uptake by only 10% while decreasing blood pool by 59%, but other studies have shown a 20-25% loss from the tumor (17). 1 Abbreviations: Av, avidin; CEA, carcinoembryonic antigen; gal, galactose; DTPA-peptide-biotin, DTPA-D-Phe-D-Lys-biotin; i.v., intravenous; % ID/g, percent injected dose per gram; MIRD, medical internal radiation dose; StAv, streptavidin.
© 1997 American Chemical Society
596 Bioconjugate Chem., Vol. 8, No. 4, 1997
Pretargeting approaches dissociate the radiolabel from the large, slow-clearing antibody to a small, extremely fast-clearing compound. The radiolabeled carrier also is designed to pass freely from the kidneys, thereby avoiding high renal retention. Several pretargeting approaches have been described (19), and each reportedly produces high tumor:blood ratios within hours. The avidin (or streptavidin)-biotin procedures are of particular interest due to the exceptionally high affinity of biotin for avidin, and with up to four potential biotinbinding sites per avidin molecule, the amount of radioactivity targeted to the tumor may be amplified. Several approaches have been used: pretarget with biotinconjugated mab, followed by radiolabeled avidin (Av) or more commonly streptavidin (StAv); pretarget with biotin-conjugated mab followed by StAv or Av clearance/ bridging, and then target with radiolabeled biotin; or pretarget with StAv-conjugated mab followed by targeting with radiolabeled biotin. Although sometimes referred to as two- or three-step procedures, these methods commonly require a minimum of three, and as many as five, separate injections, each one being precisely timed and with carefully determined dosages. On the basis of the findings of Axworthy et al. (20), who showed similar tumor accretion of labeled biotin in comparison to a whole IgG, we decided to investigate a streptavidin mab, radiolabeled biotin pretargeting method. Our purpose in developing this technology was to determine if a pretargeting approach could be an effective therapeutic method. This report, and that of Karacay et al. (21), describe our initial investigation of this approach and examine the prospects for its development as a viable approach for the therapy of cancer. MATERIALS AND METHODS
Reagents. All reagents used in these studies were described in detail by Karacay et al. (21). Briefly, the studies were conducted with a StAv-MN-14 anti-CEA IgG conjugate prepared by mixing StAv (Pierce, Rockford, IL) that was modified with sulfosuccinimidyl 4-(N-maleimidomethyl)cyclohexane-1-carboylate and mercaptoethylamine (SMCC)-reduced murine MN-14 anti-CEA IgG mab (22) (method 2 in Karacay et al. (21)]. In most instances, the conjugate administered to the animals contained a tracer quantity (36 galactose per WI2. In order to determine a suitable protein dose for the WI2 clearing agent, clearance studies were first
Bioconjugate Chem., Vol. 8, No. 4, 1997 597
Figure 2. Assessment of primary antibody clearance by the WI2 or gal-WI2 anti-MN-14 idiotype IgG clearing agent in BALB/c mice. Animals were injected intravenously with 25 µCi (2 µg) of 131I-MN-14 IgG. Three days later, groups of animals (n ) 3) were given increasing doses of either WI2 or gal-WI2 IgG, starting at 5 times the molar amount of MN-14 remaining in the blood at the time of WI2 injection, and increasing 10, 50, and 100 times. The clearing agent contained 125I-WI2 or galWI2 so the distribution of the clearing agent could be monitored. The left panels show the distribution of 133I-MN-14 at the designated times after the injection of the clearing agent, while the right panels show the corresponding WI2 or gal-WI2 distribution in these same animals. Data are shown only for the group of animals given 5 times the amount of WI2 or gal-WI2, since no differences were seen between these groups and those with the higher concentrations of clearing agents. Values represent the means, and the bars are the SD for each observation.
conducted in BALB/c mice injected with 131I-MN-14 IgG (2 µg), followed 72 h later by WI2 or gal-WI2. The amount of WI2 was adjusted to prescribed molar excesses in relationship to the amount of MN-14 IgG remaining in the entire blood pool, on the basis of the specific activity of the radiolabeled MN-14 conjugate. WI2 or gal-WI2 in molar excesses of 5-, 10-, 50-, and 100-fold were tested. 125I-labeled WI2 or gal-WI2 was added to the WI2 or gal-WI2, respectively, as a tracer dose. The blood concentration of 131I-StAv-MN-14 IgG was reduced as effectively with a 5-fold molar excess as with higher molar concentrations of either WI2 or gal-WI2 (data not shown), indicating that this lower molar excess was sufficient for clearing StAv-MN-14 from the blood for either agent. As seen in Figure 2 (left panel), the StAv-MN-14 blood concentration was reduced from 23.3 ( 2.6 % per gram to between 2 and 4 % within just 15 min of the WI2 or gal-WI2 injection, respectively. There appeared to be a more rapid blood clearance of the MN14 with the WI2, but within 4 h of their injection, no differences were seen between the concentrations of MN14 remaining in the blood, regardless of which clearing agent was used. As expected, the gal-WI2, as reflected by the 125I-labeled tracer (Figure 2, right panel), was itself cleared from the blood very quickly, with 2.0 ( 0.6% of the injected dose per gram in the blood within 15 min. A separate group of BALB/c mice which were coinjected with an identical concentration of only radiolabeled galWI2 and WI2 (i.e., no preinjection of MN-14) showed identical blood clearance kinetics for gal-WI2 and WI2 (not shown). This further illustrated the fact that there was an abundant molar excess of gal-WI2 and WI2 to affect the clearance of MN-14 without affecting their own clearance. Regardless of the clearing agent used, MN-14 was cleared primarily into the liver, with the liver concentration of MN-14 increasing from 3.2 % injected dose per gram to between 30 and 40% within 15 min of the WI2 or gal-WI2 injection, respectively (Figure 2, left panel). An interesting difference was found between gal-WI2
598 Bioconjugate Chem., Vol. 8, No. 4, 1997
and WI2 in relationship to the amount of radiolabeled MN-14 remaining in the liver. Within 24 h of the unconjugated WI2 injection, the concentration of MN-14 in the liver had decreased to 2.4 ( 0.5%, whereas the amount of MN-14 in the liver of animals who received gal-WI2 had decreased to only 22.0 ( 2.9%. As shown previously (27), galactosylated proteins are taken up in the liver, but unlike antigen-antibody complexes that are removed predominantly by the reticuloendothelial (RE) cells in the liver, galactosylated proteins are removed by galactose receptors on hepatocytes. This was observed in a separate group of animals (three animals per time point) that received a mixture of 131I- and 125Ilabeled gal-WI2 and unconjugated WI2 (not shown). In these animals, 63.3 ( 3.0 % of the gal-WI2 was in the liver within 15 min, with only 1.2 ( 0.1 % remaining in the blood. This compares to 20.0 ( 1.6% in the liver for unconjugated WI2 alone with a blood concentration of 28.8 ( 5.3% at the same time point. The level of galWI2 in the liver even in these animals remained significantly higher than that of WI2 alone for the 24 h this study was monitored (at 24 h, 19.2 ( 7.0 vs 6.2 ( 0.5% for gal-WI2 vs WI2, respectively; n ) 3, p ) 0.033). Thus, the slower rate of MN-14 clearance from the liver in animals given gal-WI2 as the clearing agent was probably due to its association with gal-WI2. The clearance properties of StAv-MN-14 with either unconjugated WI2 or gal-WI2 was identical to that described above for MN-14, regardless of the StAv-MN14 protein dose administered, as long as the dose of the WI2 was given in the prescribed 5-fold molar excess. Despite the differences in how gal-WI2 and WI2 are distributed, the most important finding was the fact that the MN-14 cleared by the gal-WI2 was retained in the liver at a higher level than that seen when WI2 alone was used. This suggests that agents cleared via the hepatocytes may not be catabolized from the liver as rapidly as when they are cleared by the RE cells. Studies in nude mice bearing GW-39 xenografts showed more WI2 in the tumor (3.0 ( 1.0 and 4.2 ( 0.3% at 4 and 24 h after WI2 injection, respectively) than gal-WI2 (2.0 ( 0.9 and 0.03 ( 0.1% at 4 and 24 h after gal-WI2 injection, respectively). Even though earlier studies had discounted the possibility that an anti-idiotypic antibody was the cause of reduced levels of radiolabeled primary antibody in a tumor (12), gal-WI2 was selected as a clearing agent in the ensuing studies. The next important step in the development of the twostep pretargeting approach was determining an optimal time for administering the radiolabeled biotin. A major consideration here is the fact that the concentration of StAv-MN-14 in the blood must be reduced sufficiently to avoid binding the radiolabeled biotin to the StAv in the blood before it has an opportunity to localize to the StAv targeted to the tumor via the MN-14. Experiments were conducted in tumor-bearing nude mice to evaluate administration of 111In-DTPA-peptide-biotin given 2 and 24 h after gal-WI2 injection. Figures 3 and 4 show the results from the group that received 111In-DTPApeptide-biotin 2 h after the gal-WI2 injection. In comparison, animals given radiolabeled biotin 24 h after the gal-WI2 had identical tumor:nontumor ratios, but the percentage of labeled biotin in all of the tissues, and most importantly in the tumor, was reduced about 5-fold (not shown). This result may be related to the lower amount of StAv-MN14 in the various tissues between the 2 and 24 h periods. Since successful therapy depends on not only the tumor:nontumor ratio but also the absolute amount of radioactivity in the tumor, it was
Sharkey et al.
Figure 3. GW-39 uptake of pretargeted 111In-DTPA-peptide biotin and 88Y-StAv-MN-14 IgG with or without clearance with gal-WI2. All animals were given 200 µg of StAv-MN-14 IgG intravenously containing 2 µCi of 88Y-StAv-MN-14 IgG. One group of animals received no further injections and were necropsied 3 h and 1, 2, 3 and 7 days later (n ) 5/time) to determine tumor uptake and organ distribution. The remaining animals received i.v. injections of gal-WI2 (5 times the molar concentration of StAv-MN-14 in the blood) 48 h after the StAvMN-14 injection. Two hours later, all animals were injected with 111In-DTPA-peptide-biotin [6 µg (5.7 × 10-9 mol)/40 µCi] 2 h after the gal-WI2 injection. At 3, 24, 48, 72, and 168 h after the 111In-DTPA-peptide-biotin injection, five animals were necropsied. Average tumor sizes ((SD) were 0.052 ( 0.009, 0.055 ( 0.016, 0.086 ( 0.019, 0.109 ( 0.028, and 0.199 ( 0.029 g for the first group at 3 h and 1, 2, 3, and 7 days, respectively. The second group had values of 0.092 ( 0.023, 0.102 ( 0.02, 0.106 ( 0.041, 0.137, 0.038, and 0.184 ( 0.038 g at 3 h and 1, 2, 3, and 7 days after the 111In-DTPA-peptide-biotin injection (i.e., 51, 72, 96, 120, and 216 h after 88Y-StAv-MN-14 injection). The arrows indicate the mean percent of injected dose per gram in the tumor at the times shown.
Figure 4. Tumor:nontumor ratios from the data shown in Figure 2 for the blood, liver, and kidneys.
determined that the earlier administration of the radiolabeled biotin was favored over delaying its injection. Figures 3 and 4 also show the biodistribution of 88YStAv-MN-14 IgG used as a tracer in the pretargeting method, as well as in a separate group of animals that received only the 88Y-labeled StAv-MN-14 IgG. These three situations illustrate the results that may be obtained if 90Y-labeled MN-14 was used alone (88YStAv-MN-14 alone) or in combination with a second antibody (88Y-StAv-MN-14 cleared by gal-WI2 in the pretargeting approach) or the pretargeting of 90Y-DTPApeptide-biotin was used (as modeled by the 111In-DTPApeptide-biotin in these biodistribution studies). As seen in Figure 3, the percent injected dose of 88Y-StAv-MN14 alone is sustained at a level of 50-60% for 5 days after taking 2 days to reach the maximum tumor uptake level. Administration of gal-WI2 causes a slow but steady decline of the 88Y-StAv-MN-14 in the tumor, taking 48 h before the level in the tumor was lower than the 88Y-
Bioconjugate Chem., Vol. 8, No. 4, 1997 599
Avidin−Biotin Pretargeting for RAIT
Table 1. Prediction of Maximum Radiation Dose Delivered to the Tumor Based on the Biodistribution of Each of the Listed Radiolabeled Materialsa method 131I-MN-14 131I-MN-14
IgG alone IgG/gal-WI2 at 24 h
90Y-StAv-MN-14
IgG alone pretargeting/90Y-biotin
a 111In-DTPA-peptide-biotin
maximum tumor dose (cGy)
MTD (mCi)
16716 20824
0.275 0.877
4264 2349
0.126 1.0
was used to predict the behavior of
Figure 5. Comparison of 111In-DTPA-peptide-biotin uptake in GW-39 with and without pretargeting. The data for the StAv-MN-14 pretargeting group were taken from Figure 2. A separate group of nude mice (n ) 5/observation) bearing small GW-39 tumors (0.03 ( 0.01, 0.044 ( 0.017, 0.057 ( 0.015, and 0.048 ( 0.02 g at 1, 3, 24, and 48 h, respectively) were injected intravenously with 40 µCi/6 µg of 111In-DTPA-peptide-biotin. Due to the very rapid clearance of this agent, monitoring was only possible over 48 h given the injected activity.
StAv-MN-14 alone. For pretargeting, 200 µg of StAvMN-14 IgG was given, followed 48 h later with gal-WI2 (5-fold molar excess), and then 2 h later, 40 µCi (5.7 × 10-9 mol) was given. Three hours after 111In-DTPApeptide-biotin injection, tumor uptake was 3.9 ( 0.8% per gram, but it too declined over time at a rate analogous to the decline of StAv-MN-14 in the tumor. Tumor:blood ratios (Figure 4) for the pretargeting method were excellent, reaching 11.3 ( 3.7 just 3 h after the injection of the 111In-DTPA-peptide-biotin to 178 ( 42 within 1 day. The level of 111In-DTPA-peptide-biotin in the blood became too low to monitor accurately after 48 h, confirming the very rapid blood clearance of this small molecule. Tumor:blood ratios for 88Y-StAv-MN-14 were less than 2:1 for a period of 3 days. However, if cleared by galWI2, tumor:blood ratios for 88Y-StAv-MN-14 increased to 59 ( 19 within just 5 h of the gal-WI2 injection. In the pretargeting method, the liver and kidney had the highest uptake of 111In-DTPA-peptide-biotin and the lowest tumor:nontumor ratios of all of the times tested. Tumor:liver ratios for the 111In-DTPA-peptide-biotin started at a level of 6.0 ( 0.6 but within 3 days had decreased to 2:1. Tumor:kidney ratios were 1.9 ( 0.04 just 3 h after 111In-DTPA-peptide-biotin injection, remaining greater than 1:1 for 48 h, but by 72 h, it decreased to 0.8 ( 0.3. Tumor:lung and spleen ratios were between 15 and 30:1 over the first 72 h (not shown). In order to affirm the fact that the tumor uptake of the 111In-DTPA-peptide-biotin was due to specific binding by the StAv localized in the tumor, the biodistribution of 111In-DTPA-peptide-biotin alone was evaluated. As shown in Figure 5, the percent uptake of 111In-DTPApeptide-biotin alone was only a fraction of that achieved if the tumor had been pretargeted with the StAv-MN14. Three hours after injection, only 0.05 ( 0.016 of the 111 In-DTPA-peptide-biotin was in the tumor. This was nearly 80-fold less than that achieved with the pretargeting method shown earlier, supporting the specific
dose-limiting organ (absorbed dose in cGy) red marrow (2500) red marrow (2400) liver (7000) red marrow (2500) liver (7000) red marrow (176) kidney (2574)
90Y.
targeting of 111In-DTPA-peptide-biotin to the StAv in the pretargeting approach. Dosimetry. In order to predict the therapeutic potential for pretargeting, these extended biodistribution studies were used to calculate radiation-absorbed doses to the tissues and tumors if 90Y-labeled biotin was used. Table 1 shows the maximum tumor dose projected to be delivered at the maximum tolerated dose (MTD). Included in the table also are projections for 90Y-MN-14 IgG without second antibody clearance, using the data from the 88Y-StAv-MN-14. In addition, radiation dose estimates are provided from biodistribution studies conducted in nude mice bearing size-matched GW-39 tumors that were given 131I-MN-14 IgG with and without blood clearance by gal-WI2 given at 24 h after the 131I-MN14 was administered. The MTD for 90Y-MN-14 IgG and 131 I-MN-14 IgG without second antibody clearance has been determined empirically to be 120 and 275 µCi, respectively (4). However, without an empirical determination of the MTD for the other procedures, the following assumptions were made to estimate a possible MTD. For 131 I-MN-14 IgG cleared by a second antibody, tumor doses were normalized to a maximum blood dose of 2500 cGy, which represented the blood dose delivered at the MTD for each agent alone. Secondarily, the absorbed dose could not exceed 7000 cGy to either the liver or kidney. This limit is based on studies performed in nude mice with 90Y-labeled MN-14 Fab fragments, which showed animals tolerating this absorbed dose level (8). The data indicate that 131I-MN-14 IgG cleared with the gal-WI2 at 24 h would achieve the highest tumor dose at the projected MTD, followed by 131I-MN-14 IgG alone. Since the gal-WI2 clearance mechanism resulted in sustained liver retention of the 131I-MN-14 IgG, liver doses were determined to be dose-limiting before the 2500 cGy dose to the red marrow (as estimated from the blood doses) could be achieved. Doses to the red marrow for the pretargeting approach were the lowest of all methods, being only 176 cGy. However, projected maximum tumor doses for the pretargeting approach were nearly 8-fold lower than that using 131I-MN-14 alone. Some of the differences in tumor dose can be attributed to the very small size of these tumors (i.e., ∼0.1 g), and thus, a large portion of the radiation dose for a long-range β-emitting particle such as 90Y would be lost to the surrounding tissue. In addition, there are other inherent difficulties in determining dosimetry for internally administered radionuclides, and there is incomplete information regarding the relationship of toxicity to these radiation dose estimates. Therefore, these dose estimates should be interpreted cautiously. Nevertheless, the data suggest that, in order for the pretargeting approach to compete with a radiolabeled antibody with or without the second antibody method for blood clearance, either the amount of radiolabel delivered to the tumor needs to be increased, doses to the normal tissues (notably liver and then
600 Bioconjugate Chem., Vol. 8, No. 4, 1997
Sharkey et al.
Figure 6. Effect of preinjection of unlabeled biotin on the biodistribution of pretargeted 111In-DTPA-peptide-biotin. Animals bearing GW-39 tumors (0.087 ( 0.038 and 0.098 ( 0.03 g for 10 µg of unlabeled biotin and none, respectively) were given 400 µg of StAv-MN-14 IgG (contains the 88Y-StAv-MN-14 tracer). Two days later, 5× gal-WI2 was given. Two hours thereafter, one group of animals (n ) 5) was injected intravenously with 10 µg of unlabeled biotin and then 5 min later received 111In-DTPA-peptide-biotin (6 µg/40 µCi), while the other group received only the radiolabeled biotin (n ) 5). All animals were necropsied 3 h after the 111In-biotin injection. Note the scale difference for the blood data.
Figure 7. Evaluation of galactosylated, biotinylated HSA as a clearing agent for the pretargeting approach. Nude mice bearing GW-39 tumors (0.098 ( 0.03 and 0.118 ( 0.03 g at 3 and 24 h for the gal-WI2 group and 0.078 ( 0.023 and 0.145 ( 0.079 g for the HSA group, respectively) were injected intravenously with 400 µg of StAv-MN-14 IgG (containing 88Y-StAvMN-14). Forty-eight hours later, 200 µg of the gal-HSA-biotin was given intravenously, and then 2 h later, the 111In-DTPApeptide-biotin was injected. Animals (n ) 5/group) were necropsied 3 and 24 h after the radiolabeled biotin injection.
kidney) need to decrease further, or a combination of both needs to be used. Attempts To Optimize the Pretargeting Approach. Initially, strategies to optimize the pretargeting approach focused on reducing uptake in the liver. Due to its accessibility to blood-borne agents, it was possible that the liver may be acting as a sink for the 111InDTPA-peptide-biotin, thereby reducing the supply of biotin able to reach the tumor. Thus, the first approach that was evaluated involved a preinjection of unlabeled biotin, the rationale being that a brief exposure of unlabeled biotin may bind the StAv-biotin binding sites in the liver, giving the radiolabeled biotin more opportunity to localize to StAv in the tumor. Nude mice bearing GW-39 tumors were first pretargeted with 400 µg of StAv-MN-14 and cleared with gal-WI2 48 h later. Two hours after the gal-WI2 injection, animals were injected intravenously with 10 µg of unlabeled biotin, followed 5 min later with the radiolabeled biotin. A separate group was given the radiolabeled biotin without preinjection of unlabeled biotin. Preinjection of unlabeled biotin had the desired effect of reducing liver uptake of the radiolabeled biotin, but there was also a significant decrease in the amount of radiolabeled biotin targeted to the tumor compared to tumor uptake seen when unlabeled biotin was not preadministered (p ) 0.005, Figure 6). The end result was that most tumor:nontumor ratios remained unchanged (except tumor:kidney which decreased). This study is important also because it shows that binding of the radiolabeled biotin to the tumor can be inhibited, thereby further supporting the possibility that a specific interaction between biotin and StAv in the tumor is occurring. In a separate study, animals were pretargeted in an identical fashion but were given 1, 10, or 100 µg of unlabeled biotin 5 min before the radiolabeled biotin (not shown). This study indicated that 10 µg of biotin had sufficiently inhibited the biotin-binding sites in the tissues, since no further reduction in normal tissue uptake (or tumor) was seen with the 100 µg of unlabeled biotin dose. However, since this procedure did not enhance tumor uptake or tumor:nontumor ratios, other strategies were pursued. The next attempt was to use a different clearing agent. Axworthy et al. (20) had used galactosylated, biotinylated human serum albumin (HSA) as their clearing agent. Although using an inert carrier, such as HSA, requires biotinylation for the clearing agent to form a complex
with the circulating StAv-mab conjugate, this conjugate configuration can also have a dual role by blocking StAvbiotin binding sites in the tissues prior to the addition of the radiolabeled biotin. Two groups of nude mice bearing GW-39 tumors were pretargeted with 400 µg of StAvMN-14 IgG, and after 48 h, gal-HSA-biotin3 (81% of the available lysines were modified with galactose) or galWI2 was given at a 5-fold molar excess in relationship to the amount of StAv-MN-14 in the blood. Two hours later, 111In-DTPA-peptide-biotin was given. These dosages and timing were consistent with that reported by Axworthy et al. (20). Unlike the StAv-MN-14 that was easily cleared by gal-WI2 to 0.5 ( 0.08% in the blood within 5 h of the gal-WI2 injection, there was a negligible effect on the clearance of StAv-MN-14 from the blood with the gal-HSA-biotin, with 88Y-StAvMN-14 IgG blood concentrations at 16.3 ( 2.5% at 5 h after the biotinylated HSA clearing agent was given (Figure 7). Tumor uptake and tumor:blood ratios for 111 In-DTPA-peptide-biotin were not improved and in fact were much lower than that achieved using gal-WI2 as the clearing agent. Low tumor:blood ratios for the 111In-DTPA-peptide-biotin (only 2.8 ( 0.6 at 24 h) were caused primarily by the higher concentration of 111InDTPA-peptide-biotin in the blood compared to the levels seen in the pretargeting approach, when gal-WI2 was used as a clearing agent (0.7 ( 0.08 vs 0.03 ( 0.007% for gal-HSA-biotin vs gal-WI2 at 24 h, respectively). The higher concentration of 111In-biotin in the blood is most likely due to its association with residual StAvMN-14. Another experiment was performed with a 10fold excess of gal-HSA-biotin with similar results. The failure of the HSA clearing agent to remove the StAvMN-14 conjugate could be explained by three possibilities: the biotin on the HSA was unable to bind to the StAv-MN-14 conjugate, it was cleaved by biotinidases, or the StAv biotin-binding sites on the MN-14 conjugate were already bound with biotin from endogenous supplies. In vitro and in vivo data showed quantitative complexation of StAv-MN-14 with gal-HSA-biotin3 or HSA-biotin conjugates prepared for these studies (21), suggesting the biotin was accessible. Thus, it was likely that the StAv-MN-14 already had a portion of its biotin binding sites occupied due to endogenous biotin. Given the implications that endogenous biotin would have on this type of pretargeting approach, additional studies were designed to elucidate the nature of this problem.
Bioconjugate Chem., Vol. 8, No. 4, 1997 601
Avidin−Biotin Pretargeting for RAIT Table 2. Effect of Endogenous Biotin on the Biotin-Binding Ability of StAv-MN-14 after i.v. Injection in Tumor-Bearing Nude Micea percent free biotin-binding sites remaining on StAv-MN-14
condition standard lab chow StAv predose biotin-deficient diet
1 h after 24 h after 48 h after StAv-Mn-14 StAv-Mn-14 StAv-Mn-14 injection injection injection 100 NDb ND
2 ND ND
2.4-2.9 3.3 100
a Animals were injected intravenously with 200 µg of StAvMN-14 IgG. At the times indicated, two mice were bled and the serum was either pooled or evaluated separately. Animals predosed with StAv had received a single injection of 500 µg of StAv 1 day prior to StAv-MN-14 injection, whereas animals which were placed on a biotin-deficient diet were fed this diet for 7 days prior to the StAv-MN-14 IgG injection. b ND, not determined.
In order to investigate this possibility, nude mice were injected with 200 µg of StAv-MN-14 IgG containing a trace of 88Y-StAv-MN-14. At 1, 24, and 48 h later, the mice were bled. Serum was collected and then mixed with varying amounts of the radiolabeled DTPA-peptidebiotin. After incubation for 5 min at room temperature, the samples were run over a size-exclusion HPLC column, and profiles were evaluated for the amount of radiolabeled biotin associated with the high-molecular weight StAv-MN-14 peak. Knowing the specific activity of the StAv-MN-14 and the known amount of biotin added to each sample, the number of biotin binding sites remaining on the StAv-MN-14 was determined. Values were compared to those of the same batch of StAv-MN14 that was not injected in vivo. The results, shown in Table 2, indicate that StAv-MN-14 taken at 1 h retained nearly 100% of its original biotin-binding capacity, whereas StAv-MN-14 taken from animals after 24 and 48 h had lost over 90% of its biotin-binding capacity. Since our earlier findings showed complete retention of StAvMN-14 in binding biotin when incubated in vitro in serum over a 24 h period (21), these in vivo results suggest that continued exposure to body fluids provides greater opportunity for blockade of the biotin-binding sites. Indeed, considering that the published serum level of biotin in rats is 3.1 × 10-8 M (28) (4.5 × 10-11 mol in a mouse assuming a 1.5 mL blood volume) and that 200 µg of StAv-MN-14 contains 4.3 × 10-10 mol of StAv with its full biotin-binding capacity (i.e., four binding sites/molecule), there should have been sufficient biotin-binding capacity to avoid this problem. Thus, with evidence that a majority of the biotin-binding sites were blocked, studies were designed to investigate the pretargeting method in animals with diminished levels of biotin. Rosebrough and Hartley (28) showed that endogenous stores of biotin could be reduced in rabbits and dogs by administering avidin. In order to investigate this procedure, nude mice bearing GW-39 tumors were given i.v. injections of 20, 100, or 500 µg of StAv 24 h before the administration of 200 µg of StAv-MN-14 IgG. Another group of animals received 100 µg of StAv at the same time that the StAv-MN-14 was given. After waiting 48 h, gal-WI2 was given, followed 2 h later with 111InDTPA-peptide-biotin. The animals were then necropsied 3 h later. The results shown in Figure 8 indicate no improvement in tumor retention compared to the results from animals that did not receive additional StAv, and in fact, at the higher StAv doses, increased uptake of 111In-DTPA-peptide-biotin was seen in the liver and blood, but more notably, in the kidneys, the major organ where StAv is known to accrete (30). This also occurred in animals that were coinjected with 100 µg of StAv on
Figure 8. Effect of StAv pretreatment to reduce endogenous biotin levels on the biodistribution of 111In-DTPA-peptidebiotin in nude mice. Animals in groups 1-4 were given varying amounts of StAv and 1 day later given 200 µg of StAv-MN-14. Group 5 received 100 µg of StAv on the same day the StAvMN-14 was given. Two days later, the StAv-Mn-14 was cleared with gal-WI2; 2 h later, the 111In-DTPA-peptide-biotin was given, followed by necropsy at 3 h. Each group of animals contained three or four animals with an average tumor size of 0.042 ( 0.027, 0.038 ( 0.017, 0.068 ( 0.012, 0.035 ( 0.010, and 0.036 ( 0.01 g for groups 1-5, respectively.
the day of the StAv-MN-14 injection. Measurement of StAv-MN-14’s ability in the blood at 48 h to bind biotin showed that the StAv pretreatment did not reduce endogenous biotin blockade of the conjugate (Table 2). Prior to examination the use of a biotin-deficient diet to reduce endogenous biotin stores, studies were conducted to determine how the specific activity of the labeled biotin affected its biodistribution, and to assess if DOTA would provide an advantage over DTPA. In vivo studies with 111In-DOTA-peptide-biotin suggested that this conjugate has slower blood clearance than DTPApeptide-biotin (21) and that perhaps slowing the blood clearance of the labeled biotin might provide a greater opportunity for tumor binding. Four separate groups of tumor-bearing nude mice were selected. Each animal received 200 µg of StAv-MN-14 followed 2 days later with gal-WI2. Two hours later, the animals received an i.v. injection of either DTPA- or DOTA-conjugated biotin labeled at either a high (1.2 × 10-9 mol of the labeled biotin) or low (5.7 × 10-9 mol of the labeled biotin) specific activity. Only the high-specific activity DOTA showed a significant increase in tumor uptake, and this was apparent only 3 h after injection (p < 0.01, Figure 9). By 24 h after the biotin injection, all of the groups had similar amounts of labeled biotin in the tumor. The high-specific activity DOTA also had a higher level of radioactivity in the blood, compared to the lower specific activity DOTA or the DTPA-conjugated biotin. These results were also most pronounced at 3 rather than at 24 h. However, since increasing the specific activity and changing the chelate improved tumor uptake from 4 to approximately 8% of the injected dose per gram, this combination was used in the next study to determine the effects of biotin-deficient diets. Two groups of animals were initiated on either biotindeficient or standard lab chow diets immediately upon receipt from the vendor. Within 2 days, a GW-39 tumor suspension was given to initiate tumor growth, and then on day 7, all animals were given 200 µg of StAv-MN14. Two days later, gal-WI2 was given, and 2 h later, 111 In-DOTA-peptide-biotin (1.2 × 10-9 mol) was ad-
602 Bioconjugate Chem., Vol. 8, No. 4, 1997
Sharkey et al.
Figure 11. Tumor:nontumor ratios for the study described in Figure 10. Statistical evaluation was performed by a general Student’s t test with p values given in the figure. N.S. means not significant at a p < 0.05 level.
Figure 9. Comparison of DTPA-peptide-biotin and DOTApeptide-biotin at low and high specific activities in the two-step pretargeting approach. Nude mice bearing GW-39 tumors were divided into four groups as indicated (tumor sizes were 0.11 ( 0.019 and 0.127 ( 0.082, 0.104 ( 0.022 and 0.137 ( 0.056, 0.113 ( 0.035 and 0.147 ( 0.027, and 0.122 ( 0.031 and 0.101 ( 0.064 g for groups 1-4 at 3 and 24 h, respectively). All animals received an initial injection of 200 µg of StAv-Mn-14 IgG followed 2 days later with clearance by gal-WI2, and then 2 h later, the radiolabeled biotin was given. The low specific activity contained 5.7 × 10-9 mol of radiolabeled chelate-peptide-biotin (approximately 6 µg each), whereas the high specific activity contained only 1.2 × 10-9 mol of radiolabeled chelate-peptidebiotin. Animals were necropsied 3 and 24 h after the biotin injection.
Figure 10. Effect of biotin-deficient diet on the biodistribution of pretargeted 111In-DOTA-peptide-biotin. Nude mice were fed standard autoclavable lab chow or a biotin-deficient diet (supplemented with 30% egg solids). On day 2, animals were inoculated with GW-39 tumors, and on day 7, 200 µg of StAv-MN-14 was injected. Gal-WI2 was given 2 days later, with 111In-DOTApeptide-biotin (1.2 × 10-9 mol) injected intravenously 2 h thereafter. Animals were necropsied 3 and 24 h after the radiolabeled biotin was given (n ) 5/group). Tumor sizes at 3 and 24 h were 0.052 ( 0.011 and 0.072 ( 0.016 g, respectively, for the animals given the biotin-deficient diet and 0.71 ( 0.018 and 0.88 ( 0.033 g for the animals fed the normal rodent diet, respectively.
ministered. The animals were necropsied 3 and 24 h after the radiolabeled biotin injection. Under these conditions, tumors taken from animals given biotin-deficient diets had a nearly 2-fold higher (16.9 ( 4.6) tumor uptake than animals fed the standard lab chow (Figure 10). This level of 111In-DOTA-peptide-biotin in the tumor was sustained over the 24 h monitoring period. Although tumor uptake was increased, uptake in all normal tissues also increased. Interestingly, tumor:liver and kidney
ratios in animals given the biotin-deficient diet were not significantly different at 24 h (Figure 11). Although tumor:blood ratios were significantly lower in animals given the biotin-deficient diet on day 1, the actual ratios exceeded 50:1. Thus, even under these conditions, the pretargeting method maintained its tumor:blood advantage compared to a radiolabeled antibody alone. Analysis of serum taken from these mice 48 h prior to administration of gal-WI2 revealed the conjugate had retained 100% of its original bitoin-binding capacity (Table 2). DISCUSSION
Pretargeting approaches are very attractive for use with radiolabeled antibodies because they offer exceptional tumor:blood ratios. This property is highly desirable for imaging and therapeutic applications. For imaging, the simple use of an antibody fragment may be preferred over the multiple steps involved in a pretargeting approach. The same can be said for therapy, where the simplest approach will prevail over those requiring multiple steps, unless these alternative approaches improve the therapeutic outcome. The work presented herein, and in a companion article (21), illustrates some of the complexities involved in developing a two-step, streptavidin-biotin pretargeting method. Multiple agents need to be prepared and tested, dosages optimized, and appropriate dosing schedules evaluated. After evaluation of a number of different agents and parameters, a successful two-step pretargeting approach was established that produced excellent tumor:blood ratios within a few hours. The liver and kidney were the major organs for retention of the radiolabeled biotin, but even this uptake produced tumor: nontumor ratios in excess of 2:1, which is acceptable for imaging applications but may need to be improved for therapy. A major concern for the pretargeting approach was the fact that the percentage of radiolabeled biotin targeted to the tumor was much lower than that observed with a radiolabeled antibody alone. This may present a problem for therapeutic applications that will require a critical concentration of radioactivity in order to effect changes in tumor growth. However, the primary issue is not the percent of injected radioactivity delivered to the tumor but the absolute amount of radioactivity that can be delivered at the maximum tolerated dose for each procedure in combination with the longevity of tumor binding. Dosimetry estimates suggest that, with the current pretargeting approach and 4% of the injected dose per gram tumor, approximately 40 µCi could be delivered (assuming 1 mCi is the MTD). At a 60% injected dose per gram, an antibody labeled with 90Y could deliver 72 µCi/g to the tumor (MTD of 120 µCi), whereas an 131I-labeled antibody could deliver 165 µCi/g. From this perspective, the pretargeting approach would need to increase the percentage delivered to the tumor at least
Avidin−Biotin Pretargeting for RAIT
2-fold, or if liver and kidney were reduced at least 2-fold and this allowed a doubling in the administered dose, then the pretargeting approach would be more competitive with other forms of RAIT. By itself, the labeled biotin is cleared very quickly from all tissues. Our studies showed that, within 1 day, only 0.004 and 0.01% per gram of 111In-DTPA-peptide and DOTA-peptide-biotin, respectively, remained in the blood (21). The kidney was the organ with the highest uptake, but it was as low as 0.22 and 0.17% per gram 1 day after the biotin injection. In animals given StAvMN-14 IgG, there were higher levels of radiolabeled biotin in all the major organs. Thus, the altered distribution was due to the radiolabeled biotin binding to the StAv contained in these normal tissues. Indeed, it was possible to show that, within just 5 min, biotin-binding sites could be saturated with a preinjection of 10 µg of unlabeled biotin. Several approaches were examined to increase tumor uptake or to reduce normal tissue uptake. Tumor uptake was increased somewhat by raising the specific activity of the labeled biotin and by using DOTA rather than DTPA. Trying to inhibit biotin binding in tissues selectively is difficult. In most instances, inhibition of residual StAv biotin-binding sites caused a concomitant reduction in the level of radiolabeled biotin targeted to the tumor, thereby negating the value of the procedure. Another important finding was the fact that endogenous biotin levels in the mouse seriously inhibit this approach. Hnatowich et al. (30) had brought this problem to attention earlier, but it has not been studied in detail in the literature. On the basis of published levels of biotin in mouse serum, and our in vitro studies showing the full biotin-binding capability of the conjugates, it was assumed that there would be abundant biotin-binding sites on the StAv with the doses administered. It was interesting to discover that problems with endogenous biotin could not be appreciated by incubating the conjugate in serum in vitro, but rather, the conjugate needed to be injected in the mice and then assayed over time. Our experience showed that, within 1 day, the vast majority of biotin-binding sites on the conjugate were blocked. Since tumor localization in nude mice generally requires 2 days to achieve optimal tumor accretion, this is a serious disadvantage for assessing this methodology preclinically. The fact that 111In-biotin uptake in the tumor and normal tissues could be reduced by a preinjection of unlabeled biotin clearly showed that not all biotin-binding sites were blocked by endogenous biotin. Fortunately, feeding animals a biotin-free diet can reduce endogenous biotin levels, and temporary reductions in biotin do not have a serious impact on the health of the animals. Using this approach, unfortunately, exacerbates the problem of residual StAv in the tissues capable of binding the radiolabeled biotin. Although biotin levels in humans are reported to be more than 10-fold lower than those in mice (28), it is still uncertain whether endogenous biotin will play an equally important role in humans because these measurements only account for serum levels and do not account for other storage sites. Thus, it is not merely a question of serum biotin levels but a test subject’s capacity to restore serum biotin homeostasis, either by liberation of biotin from intracellular stores or by an increase in absorption of dietary biotin. Kalofonos et al. (31) showed that, 3 days after the injection of only 1.0 mg of StAv-conjugated antibody, the antibody in the blood of patients still retained its ability to bind biotin. However, the precise number of biotin-binding sites remaining was not reported. Nevertheless, this result is encouraging, especially since
Bioconjugate Chem., Vol. 8, No. 4, 1997 603
higher conjugate dosages will further reduce the influence of endogenous biotin in patients. At least in animal models, which may predict whether human studies are warranted, endogenous biotin is a serious problem. We elected to evaluate the StAv-antibody, radiolabeled biotin approach on the basis of early reports by Axworthy et al. (20), who stated that they were able to achieve > 20% injected dose per gram 111In-biotin tumor uptake, similar to the level obtained with their radiolabeled antibody alone. If identical tumor uptake with markedly improved tumor:blood ratios is possible with this pretargeting approach, as compared to the radiolabeled antibody, then improved antitumor effects should also be possible. Indeed, Axworthy et al. (32) have reported improved therapeutic effects in animals with this pretargeting approach. If a tumor uptake can be achieved with this pretargeting approach identical to that seen with the radiolabeled antibody, it would be most encouraging. As pointed out by O’Connor and Bale (33), only a small portion of an antibody will have an opportunity to pass through a tumor. With an agent that is rapidly cleared from the blood, the portion passing through the tumor would be substantially lower than a substance that is cleared from the blood more slowly. This relationship is evident when comparing tumor uptake of whole IgG to antibody fragments; the uptake is always substantially higher with the whole IgG. However, the pretargeting approach offers other advantages that could alter this relationship, namely an increase in the number (maximum of four biotins/StAv) and affinity of the binding sites in the tumor (i.e., StAv-biotin vs antibodyantigen). Combined with the fact that, as a small molecule, biotin is able to be distributed more quickly and uniformly throughout the tumor than a large macromolecule, this pretargeting approach has the potential to compete effectively with directly radiolabeled antibodies. However, the complexity of these procedures in comparison to using radiolabeled antibodies, coupled with the possibility of increased immunogenicity of StAv-mab conjugates, makes it necessary that these procedures exceed the therapeutic ability of the simpler approach. Clinical trials will ultimately be required to make this determination. Despite some of the problems cited herein, we have yet to investigate further modifications that could either improve the total amount of radioactivity delivered to the tumor or reduce the level in normal tissues. For example, using WI2 in place of gal-WI2 may reduce the quantity of StAv-MN-14 available in the liver, and perhaps waiting 4-8 h between the injection of the clearing agent until the labeled biotin is given will further reduce the potential for high liver accretion. Once fully optimized, it will also be important to establish empirically a maximum tolerated dose level for the pretargeting approach rather than depending on dosimetry to decide the fate of this promising procedure. Given the relatively surprising finding that a single injection of 90Y-labeled MN-14 Fab may have a therapeutic benefit comparable to that of 90Y-IgG (8), we are encouraged to further explore the potential therapeutic benefit that may be afforded by such pretargeting approaches. ACKNOWLEDGMENT
The authors thank Ms. Rosarito Aninipot for her technical assistance in performance of the animal studies and Mr. Mark Prysbylowski for the purification of the antibodies. This study was supported in part by U.S. Public Health Service Grants CA-37895 and CA-39841.
604 Bioconjugate Chem., Vol. 8, No. 4, 1997 LITERATURE CITED (1) Blumenthal, R. D., Sharkey, R. M., Haywood, L., Natale, A. M., Wong, G. Y., Siegel, J. A., Kennel, S. J., and Goldenberg, D. M. (1992) Targeted therapy of athymic mice bearing GW-39 human colonic cancer micrometastases with 131Ilabeled monoclonal antibodies. Cancer Res. 52, 6036-6044. (2) Sharkey, R. M., Blumenthal, R. D., Hansen, H. J., and Goldenberg, D. M. (1990) Biological considerations for radioimmunotherapy. Cancer Res. 50 (Suppl.) 964s-969s. (3) Buchegger, F., Pe`legrin, A., Delaloye, B., Bischof-Delaloye, A., and Mach, J.-P. (1990) Iodine-131-labeled MAb F(ab′)2 fragments are more efficient and less toxic than intact antiCEA antibodies in radioimmunotherapy of large colon carcinoma grafted in nude mice. J. Nucl. Med. 31, 1035-1044. (4) Sharkey, R. M., Blumenthal, R. D., Behr, T. M., Wong, G., Haywood, L., Forman, D., Griffiths, G., and Goldenberg, D. M. (1997) Selection of radioimmunconjugates for the therapy of well-established or micrometastatic colon carcinoma. Int. J. Cancer (in press). (5) Behr, T. M., Sharkey, R. M., Juweid, M. E., Blumenthal, R. D., Dunn, R. M., Griffiths, G. L., Wolf, F. G., Becker, W. S., and Goldenberg, D. M. (1995) Reduction of the renal uptake of radiolabeled monoclonal antibody fragments by cationic amino acids and their derivatives. Cancer Res. 55, 3825-3834. (6) DePalatis, L. R., Frazier, K. A., Cheng, R. C., and Kotite, N. J. (1995) Lysine reduces renal accumulation of radioactivity associated with injection of the [177Lu]R-[2-(4-aminophenyl)ethyl]-1,4,7,10-tetraaza-cyclodecane-1,4,7,10-tetraacetic acid-CC49 Fab radioimmunoconjugate. Cancer Res. 55, 5288-5295. (7) Kobayashi, H., Yoo, T. M., Kim, I. S., Kim, M.-K., Le, N., Webber, K. O., Pastan, I., Paik, C. H., Eckelman, W. C., and Carrasquillo, J. A. (1996) L-lysine effectively blocks renal uptake of 125I- or 99mTc-labeled anti-Tac disulfide-stabilized Fv fragment. Cancer Res. 56, 3788-3795. (8) Behr, T. M., Sharkey, R. M., Sgouros, G., Blumenthal, R. D., Dunn, R. M., Kolbert, K., Griffiths, G. L., Siegel, J. A., Becker, W. S., and Goldenberg, D. M. (1997) Improved cancer therapy by overcoming nephrotoxicity of radiometal-labeled immunoconjugates and peptides: biological effects in relation to internal radiation dosimetry in a human colon cancer model. Cancer (in press). (9) Blumenthal, R. D., Fand, I., Sharkey, R. M., Boerman, O. C., Kashi, R., and Goldenberg, D. M. (1991) The effect of antibody protein dose of tumor distribution of radiolabeled antibodies: an autoradiographic study. Cancer Immunol. Immunother. 33, 351-358. (10) Fujimori, K., Covell, D. G., Fletcher, J. E., and Weinstein, J. N. (1989) Modeling analysis of the global and microscopic distribution of immunoglobulin G, F(ab′)2, and Fab in tumors. Cancer Res. 49, 5656-5663. (11) Sharkey, R. M., Primus, F. J., and Goldenberg, D. M. (1984) Second antibody clearance of radiolabeled antibody in cancer radioimmunodetection. Proc. Natl. Acad. Sci. U.S.A. 81, 2843-2846. (12) Sharkey, R. M., Boerman, O. C., Natale, A., Pawlyk, D., Monestier, M., Losman, M. J., and Goldenberg, D. M. (1992) Enhanced clearance of radiolabeled murine monoclonal antibody by a syngeneic anti-idiotype antibody in tumor-bearing nude mice. Int. J. Cancer 51, 266-273. (13) Goodwin, D., Meares, C., Diamanti, C., McCall, M., Lai, C., Torti, F., McTigue, M., and Martin, B. (1984) Use of specific antibody for rapid clearance of circulating blood background from radiolabeled tumor imaging proteins. Eur. J. Nucl. Med. 29, 226-243. (14) Begent, R. H. J., Ledermann, J. A., Green, A. J., Bagshawe, K. D., Riggs, S. J., Searle, F., Keep, P. A., Adam, T., Dale, R. G., and Glaser, M. G. (1989) Antibody distribution and dosimetry in patients receiving radiolabelled antibody therapy for colorectal cancer. Br. J. Cancer 60, 406-412. (15) Pedley, R. B., Dale, R., Boden, J. A., Begent, R. H. J., Keep, P. A., and Green, A. J. (1989) The effect of second antibody clearance on the distribution and dosimetry of radiolabelled anti-CEA antibody in a human colonic tumor xenograft model. Int. J. Cancer 43, 713-718. (16) Blumenthal, R. D., Sharkey, R. M., and Goldenberg, D. M. (1995) Overcoming Dose-Limiting, Radioantibody-Induced
Sharkey et al. Myelotoxicity. In Cancer Therapy with Radiolabeled Antibodies (D. M. Goldenberg, Ed.) pp. 295-314, CRC Press, Boca Raton, FL. (17) Norrgren, K., Strand, S.-E., Nilsson, R., Lindgren, L., and Sjo¨gren, H.-O. (1993) A general, extracorporeal immunoadsorption method to increase the tumor-to-normal tissue ratio in radioimaging and radioimmunotherapy. J. Nucl. Med. 34, 448-454. (18) Lear, J. L., Kasliwal, R. K., Feyerabend, A. J., Pratt, J. P., Bunn, P. A., Dienhar, D. G., Gonsaly, R., Johnson, T. K., Bloedow, D. C., Maddock, S. W., and Glenn, S. D. (1991) Improved tumor imaging with radiolabeled monoclonal antibodies by plasma clearance of unbound antibody with antiantibody column. Radiology 179, 509-512. (19) Goodwin, D. A., and Meares, C. F. (1997) Pretargeting: General principles. Cancer (in press). (20) Axworthy, D. B., Fritzberg, A. R., Hylarides, M. D., Mallet, R. W., Theodore, L. J., Gustavson, L. M., Su, F.-M., Beaumier, P. L., and Reno, J. M. (1994) Preclinical evaluation of an antitumor monoclonal antibody/streptavidin conjugate for pretargeted 90Y radioimmunotherapy in a mouse xenograft model. J. Immunother. 16, 158 (abstract). (21) Karacay, H., Sharkey, R. M., Griffiths, G. L., Govindan, S. V., McBride, W. J., Goldenberg, D. M., and Hansen, H. J. (1997) Development of a streptavidin-anti-carcinoembryonic antigen antibody, Radiolabeled Biotion pretargeting method for radioimmunotherapy of colorectal cancer. Reagent Development. Bioconjugate Chem. 8, 585-594. (22) Hansen, H. J., Goldenberg, D. M., Newman, E. S., Grebenau, R., and Sharkey, R. M. (1993) Characterization of secondgeneration monoclonal antibodies against carcinoembryonic antigen. Cancer 71, 3478-3485. (23) Goldenberg, D. M., and Hansen, H. J. (1972) Carcinoembryonic antigen present in human colonic neoplasms serially propaged in hamsters. Science 175, 1117-1118. (24) Sharkey, R. M., Motta-Hennessy, C., Pawlyk, D., Siege, J. A., and Goldenberg, D. M. (1990) Biodistribution and radiation dose estimates for yttrium- and iodine-labeled monoclonal antibody IgG and fragments in nude mice bearing human colonic tumor xeongrafts. Cancer Res. 50, 23302336. (25) Sharkey, R. M., Primus, F. J., and Goldenberg, D. M. (1987) Antibody protein dose and radioimmunodetection of GW-39 human colon tumor xenografts. Int. J. Cancer 39, 611-617. (26) Boerman, O. C., Sharkey, R. M., Wong, G. Y., Blumenthal, R. D., Aninipot, R. L., and Goldenberg, D. M. (1992) Influence of antibody protein dose on therapeutic efficacy of radioiodinated antibodies in nude mice bearing GW-39 human tumors. Cancer Immunol. Immunother. 35, 127-134. (27) Mattes, M. J. (1987) Biodistribution of antibodies after intraperitoneal or intravenous injection and effect of carbohydrate modification. J. Natl. Cancer Inst. 79, 855-863. (28) Rosebrough, S. F., and Hartley, D. F. (1995) Quantification and lowering of serum biotin. Lab. Anim. Sci. 45, 554-557. (29) Rosebrough, S. F. (1993) Pharmacokinetics and biodistribution of radiolabeled avidin, streptavidin, and biotin. Nucl. Med. Biol. 20, 663-668. (30) Hnatowich, D. J. (1994) The in vivo uses of streptavidin and biotin: a short progress report. Nucl. Med. Commun. 15, 575-577. (31) Kalofonos, H. P., Rusckowski, M., Siebecker, D. A., Sivolapenko, G. B., Snook, D., Lavender, J. P., Epenetos, A. A., and Hnatowich, D. J. (1990) Imaging of tumor in patients with indium-111-labeled biotin and streptavidin-conjugated antibodies: Preliminary communication. J. Nucl. Med. 31, 1791-1796. (32) Axworthy, D. B., Beaumier, P. L., Bottino, S., Goshorn, R. W., Mallett, R. W., Stone, D. M., Su, F.-M., Theodore, L. J., Yau, E. K., and Reno, J. M. (1996) Preclinical optimization of pretargeted radioimmunotherapy components: High efficiency, curative 90Y delivery to mouse tumor xenografts. Tumor Targeting 2, 156-157 (abstract). (33) O’Connor, S. W., and Bale, W. F. (1984) Accessibility of circulating immunoglobulin G to the extravascular compartment of solid rat tumors. Cancer Res. 44, 3719-3723.
BC970101V