We investigated the biodistribution of 88Y/111In-labeled 1,4,7,10-tetraazacyclododecane-N,N′,N″,N‴-tetraacetic acid (DOTA)-biotin and therapy with 90Y-labeled DOTA-biotin in tumor-bearing mice after B3-streptavidin antibody conjugate (B3-SA) pretargeting. B3 antibody, recognizing Lewisy antigen, was conjugated to streptavidin (B3-SA). For pretargeting, 400 μg of the B3-SA was injected i.v. into mice bearing A431 tumor xenografts. After tumor localization of B3-SA, 100 μg of synthetic clearing agent was injected i.v. to clear the unbound B3-SA from the circulation. Four h later, 1 μg of radiolabeled DOTA-biotin was injected i.v. Radioimmunotherapy was performed with doses of 9.25 to 37 MBq of 90Y-labeled DOTA-biotin. As a result, radiolabeled DOTA-biotin cleared rapidly. All of the normal tissues had <2.6% of the injected dose per gram, whereas tumor uptake reached ∼15% ID/g. The total tumor uptake of radioactivity remained similar for 96 h or longer. In the first study, the median survival of the control group was 8 days, whereas it increased to >163 days in the 37 MBq 90Y group (P < 0.005). In a second therapy group, 7 of 10 mice receiving 37 MBq of 90Y and B3-SA were cured, and remained healthy for >180 days after therapy, compared with control groups, with ≤29.2 days mean survival time (P < 0.001). Tumor pretargeting with B3-SA and radiolabeled DOTA-biotin has shown favorable, specific, and fast targeting that has resulted in good tumor responses and, thus, serves as a rationale for human studies with the B3-SA pretargeting approach.
Clinical studies using radiolabeled MAbs2 for radioimmunotherapy of lymphomas have shown promising results (1). In contrast, studies performed with MAbs directed against epithelial tumors have rarely shown partial or complete remissions (1). Various physiological factors contribute to the limited therapeutic response of this targeting approach using intact MAb (1, 2, 3). Approaches are needed that result in faster delivery of higher concentrations of radiolabeled MAbs to tumor with relative sparing of normal tissues.
Several pretargeting methods that dissociate the targeting step from the radioactivity delivery step have been proposed (1, 4, 5, 6). Some of the methods are based on the high affinity binding of avidin or streptavidin to biotin (7). A promising pretargeting approach consists of three steps (8, 9): (a) the SA conjugate is allowed to target and accumulate in tumor, thus carrying the SA receptor that will later bind the radiolabeled biotin; (b) the nontumor bound MAb-SA is cleared from the circulation to prevent it from binding the radiolabeled biotin; (c) the radiolabeled biotin is administered i.v. and because of its small size, the radiolabeled biotin extravasates out of the circulation quickly and can bind to the SA on the MAb-SA conjugate that has localized in tumor (8, 9, 10).
B3 is an excellent antibody to consider for pretargeting applications (11). Biodistribution and radioimmunotherapy of 111In/90Y-radiolabeled B3 antibody has shown good tumor localization, although at the maximal tolerated doses tumor responses were not observed (12). Using the pretargeting approach, it is likely that higher doses of 90Y could be delivered to tumor compared with normal tissues and thus increase the therapeutic index. In this report, we performed a biodistribution study using 111In- and 88Y-labeled DOTA-biotin after pretargeting with B3-SA, as well as a radioimmunotherapy trial with 90Y-labeled DOTA-biotin.
MATERIALS AND METHODS
B3 is a murine IgG1k MAb directed at Lewisy and polyfucosylated-Lewisx (11). B3 was conjugated to SA by NeoRx by incubating for ∼1 h with succinimidyl 4-(N-maleimido-methyl) cyclohexane-1-carboxylate (10). To confirm that the conjugated SA to B3 was still functional, excess B3-SA was incubated with 111In-labeled DOTA-biotin, and instant thin-layer chromatography, using paper chromatography developed with normal saline, was performed and showed a bound fraction >99%.
The sCA provided by NeoRx consists of a bifunctional moiety with multiple N-acetyl-galactosamine residues linked to biotin (Mr 8651; Ref. 13). When sCA was injected to mice (n = 5) that had received 125I B3-SA, a mean drop of 90% of the circulating 125I was observed.
Biotinidase-resistant DOTA-biotin (Mr 900) was prepared by NeoRx corporation as described previously and labeled with either 111In, 88Y, or 90Y (14).
Unmodified B3 and B3-SA were labeled with 4-[125I]iodobenzoate (15) at a specific activity of 37 kBq/μg for biodistribution studies. B3 conjugated with 2-(p-isothiocyanato benzyl)-cyclohexyl-diethylenetriaminepentaacetic acid (CHX-A″-referred to as CHX) was also labeled with 111In at a specific activity of 37 kBq/μg for biodistribution study as described previously (16). Biotinidase-resistant DOTA-biotin was labeled with 111In or 88Y at a specific activity of 0.37 to 1.11 MBq/μg for biodistribution and with 90Y at a specific activity of 37 MBq/μg for therapy (10). These radiolabeled DOTA-biotin reagents consistently bound >99% to B3-SA.
A431, a human epidermoid carcinoma cell line that expresses the antigen recognized by B3, was used for immunoreactivity determination (16). The immunoreactivity of 125I B3-SA (68–73%) was in the range of that of 125I B3 (70–77%). A competition assay performed using 105 A431 cells incubated with 125I B3 and increasing doses of unlabeled B3 or B3-SA showed overlapping curves (data not shown). The IC50 was 89 nm and 95 nm for B3 and B3-SA, respectively (GraphPad Prism, San Diego, CA).
Female athymic mice (nu/nu) were inoculated s.c. with A431 cells. Animal experiments were performed under an NIH Animal Care and Use Committee approved protocol. Biodistribution or therapy studies were performed when xenografted tumors typically reached a maximum diameter of ∼0.5 cm (0.11–0.31 g). Mice were killed when tumor size reached >2 cm in the longest diameter, tumor was ulcerated, or excessive weight loss (>25%) was noted, according to the protocol guidelines. Seven days before pretargeting, the mice were given a biotin-free diet to reduce their endogenous biotin level (Biotin Deficient Purina Diet 5836C; Purina Mills, Richmond, IN). One week of this diet in BALB/c mice (n = 3 per group) caused their biotin level in serum to drop from a baseline of 15.0 ± 5.2 ng/ml to 4.77 ± 1.12 ng/ml (Vitamin Diagnostics, Inc., Cliffwood Beach, NJ).
Optimization of B3-SA Dose.
To evaluate the effect of MAb mass on tumor targeting and thus determine the optimal MAb dose for pretargeting, biodistribution studies were performed with 111In-CHX-B3. B3 was used rather than B3-SA because studies in tumor bearing mice showed a similar biodistribution in most tissues of 111In-B3 and B3-SA (data not shown), and because the quantity of B3-SA was limited. Tumor-bearing mice were injected i.v. with 0.185 MBq of 111In-CHX-B3. The amount of B3 was adjusted by adding unlabeled, unconjugated B3 in doses of 5, 50, 100, 200, 400, and 800 μg. At each dose level 5 mice were sacrificed 48 h after injection. Two additional groups were sacrificed at 24 and 72 h after receiving 400 μg to better characterize the kinetics of the dose finally selected for the therapy trial.
To evaluate the effect of dose on gross heterogeneity of distribution within the tumor, autoradiography was performed after injection of the 111In-CHX-B3. The autoradiography was performed using a BAS-1500 Fujifilm Bio-imaging analyzer (Fuji Medical Systems, Inc.). The slides were exposed to the imaging plate, and developed and analyzed according to the manufacturer’s procedure. The autoradiograms were inspected visually, and areas of interest were drawn around the individual tumor autoradiogram, and the mean and SD of the counts were determined in the manufacture’s PSLs.
111In-and/or 88Y-labeled DOTA-Biotin Biodistribution.
Tumor-bearing mice were injected i.v. with 400 μg (1.91 nmol) of B3-SA for pretargeting. After 24 h were allowed for distribution, 100 μg (11.56 nmol) of sCA was injected i.v. to clear circulating B3-SA from the blood. Four h after injection of the sCA, 1 μg (1.11 nmol) of 111In-labeled DOTA-biotin was injected i.v. Groups of 5 mice were sacrificed at 2, 24, 48, 96, and 168 h after injection. In a separate experiment, dual-isotope injection of the 111In- and 88Y-labeled DOTA-biotin was given to mice to allow direct comparison of both reagents at 2 and 96 h.
The mean %ID/g of 111In in tumor and tissues at the various time points up to 7 days was used to obtain the area under the curve of the exposure for 90Y using trapezoidal integration. From that point on, the activity in tumor was extrapolated, assuming it cleared with the estimated terminal half-life of the clearance curve (GraphPad Prism). 90Y radiation dosimetry was then calculated based on biodistribution data using the medical internal radiation dose method (17) and adjusted for organ size using the “Nodule Module” of MIRDOSE 3 (18). The data from our previous study of 111In B3 (16) was analyzed in the same way and compared with that of 90Y-labeled DOTA-biotin.
Radioimmunotherapy was performed in tumor-bearing mice by pretargeting 90Y-labeled DOTA-biotin after allowing for localization of B3-SA and clearing of the blood as described above. Groups of mice (n = 5) were treated in a dose escalation trial with 9.25, 18.5, and 37 MBq of 90Y-labeled DOTA-biotin and compared with a no-treatment control group. In an additional experiment, a group of 10 mice was treated with 37 MBq 90Y-labeled DOTA-biotin after the same pretargeting approach with B3-SA and compared with 3 control groups: (a) no treatment; (b) radioactivity-free group receiving B3-SA, clearing agent, and nonradioactive DOTA-biotin; and (c) a nonspecific MAb consisting of HAT-SA, directed against the interleukin 2 receptor, which is not expressed on A431 cells and was pretargeted followed by 100 μg sCA and 37 MBq 90Y-labeled DOTA-biotin (19).
Response to Therapy and Toxicity Assessment.
Tumor growth was monitored twice a week for 3 weeks after treatment and then once per week. A digital caliper was used to measure the tumor in two orthogonal dimensions. The volume was calculated using the formula
To determine hematological toxicity, 20 μl of blood was collected by bleeding the tail veins using an EDTA-coated capillary tube and diluted to 250 μl with PBS (n = 5/group). WBC, lymphocytes, RBC, and platelet count were performed before therapy and at weekly intervals thereafter. Serum was obtained at the time of sacrifice, and liver (total bilirubin, alk phosphatase, and albumin) and kidney (blood urea nitrogen and creatinine) function test were done.
Histological analysis was performed in 3 mice in each of the groups (control, 9.2 MBq, 18.5 MBq, and 37 MBq of the 90Y-labeled DOTA-biotin-B3-SA, and the 37 MBq 90Y-labeled DOTA-biotin-HAT-SA group). Tissues analyzed included the kidney and liver, and in some animals bone and bone marrow were also fixed in 10% buffered formalin for pathological studies by a veterinary pathologist.
Statistical analysis was performed using ANOVA for multiple groups and Student’s t test for unpaired data between two groups. Survival analysis was based on the Kaplan-Meier product limit, and groups were compared using the log-rank test (GraphPad Prism).
The concentration of 111In-CHX-B3 and 111In-B3-SA in tumor was similar, with concentrations of 13.89 ± 3.16 and 10.33 ± 2.20%ID/g at 24 h, respectively (P = 0.077). The fraction of administered 111In-CHX-B3 (%ID/g) in tumor remained similar, with doses ranging from 5 to 400 μg (Fig. 1,A). Although the %ID/g of radioactivity dropped with the 800 μg dose, the absolute tumor-bound B3 still increased. The tumor uptake of 400 μg was near maximal at 24 h (21.8 ± 2.1%ID/g) and remained grossly unchanged for 24 h through 72 h after injection (P = 0.74; Fig. 1 B). On autoradiography no gross differences in tracer heterogeneity were noted visually in the tumors. The mean activity value and SD obtained from the region of interest analyses over the autoradiographs of tumors, which are a measure of the heterogeneity of the intensity per pixel, were 120 ± 23 PSL units, 164 ± 41 PSL units, 165 ± 62 PSL units, 140 ± 36 PSL units, 116 ± 19 PSL units, and 158 ± 37 PSL units for the 5, 50, 100, 200, 400, and 800 μg antibody doses, respectively (ANOVA, P = 0.197).
111In- and 88Y-labeled DOTA-biotin cleared rapidly with >70% of the administered dose excreted by 2 h after injection. The biodistribution of 111In-labeled DOTA-biotin showed rapid and high accumulation in tumor, with 15.8 ± 0.4%ID/g at 2 h after injection, whereas all of the other organs had <2.6%ID/g (Fig. 2). Therefore, the tumor:nontumor targeting ratios at 2 h were high: 6.8 ± 2.8, 12.2 ± 4.5, 8.1 ± 0.8, and 25.9 ± 11.5 for blood, liver, kidney, and bone, respectively. At 24 h the tumor:nontumor ratios were 8.5 ± 3.7, 11.1 ± 4.8, and 5.5 ± 0.5 for blood, liver, and kidney, respectively, whereas they were only 0.88 ± 0.11, 0.34 ± 0.04, and 0.37 ± 0.04 for the non-pretargeted MAb. The %ID/g in tumor peaked at the earliest time point sampled and then decreased over time (Fig. 2). This decrease in %ID/g was almost entirely accounted for by the increasing tumor size from a mean of 0.21 g to 1.03 g over the 7 days (Fig. 2, inset).
The biodistribution of the coinfused 88Y- and 111In-labeled DOTA-biotin in tumor-bearing mice showed a similar pattern at 2 h and 96 h postinjection except in the lung, spleen, and bone at 2 h, where the 111In-labeled DOTA-biotin was found in higher concentrations (data not shown).
On the basis of 111In-CHX-B3 and 111In-labeled DOTA-biotin biodistribution data 90Y-CHX-B3 and pretargeted 90Y-labeled DOTA-biotin radiation dosimetry for tumor was calculated as 196 Gy/37MBq and 72 Gy/37MBq, respectively, assuming a 1-g tumor that does not increase in size during the therapy and correcting for absorption fraction. The radiation dose to tumor was higher for directly labeled B3. Nevertheless, the expected maximum tolerated dose for 90Y B3 (nonpretargeted) is likely to be in the range of 7.4–11.1 MBq based on prior studies (10, 20). Whereas 37 MBq 90Y could generally be delivered with the pretargeting. The tumor:blood ratios were 2.4 for 111In B3 and 9.3 for pretargeted 111In-labeled DOTA-biotin, respectively.
Effective and specific therapeutic responses were observed in tumor-bearing mice treated with 90Y-labeled DOTA-biotin after pretargeting with B3-SA. A431 tumors in the no-treatment control group grew rapidly, from 0.2 cm3 to ∼2 cm3 in <2 weeks (Fig. 3, Group 1), and these animals were sacrificed according to our animal protocol. Escalating doses of 9.25, 18.5, and 37 MBq 90Y-labeled DOTA-biotin with B3-SA pretargeting showed a dose-related response in tumor size (Fig. 3, Group 2, 3, 4, respectively). The median duration of survival of the control group was 7 days; it increased to 18, 28, and >180 days in the groups receiving 9.25, 18.5, and 37 MBq of 90Y-labeled DOTA-biotin, respectively. Four of the 5 mice in group 4 and 1 in group 3 were cured and remained healthy for >180 days after therapy, at which time they were sacrificed electively for histology, CBC, liver, and kidney chemistries.
In the second therapy experiment (Fig. 4), a group of mice that received all of the pretargeting steps but no radioactivity showed no response; their tumor growth was similar to the no-treatment control (10 versus 12 days median survival, respectively). The mice receiving nonspecific HAT-SA pretargeting and 37 MBq of 90Y-labeled DOTA-biotin showed delayed tumor growth and significantly prolonged survival (24 days) compared with the control group, but complete resolution of tumor was never observed. The mice died or were sacrificed with residual tumor. In the group treated with 90Y-labeled DOTA-biotin after pretargeting with B3-SA, tumor growth was delayed, and 8 of the 10 mice became tumor-free. Median survival was prolonged to >180 days, significantly longer than the other three groups (P < 0.0001). Seven of the 10 mice from this group remained healthy for >180 days after therapy and at sacrifice had no visual evidence of tumor. Whereas, of the 3 mice that died, 1 mouse was sacrificed because of tumor progression, 1 mouse with tumor and 1 mouse without tumor died unexpectedly of unknown cause. No autopsy was available on these mice.
All of the mice receiving up to 37 MBq of 90Y therapy survived without radiation-related lethality for 2 weeks. One mouse receiving 18.5 MBq of 90Y during the first therapy study died of a skin disease that had been present in the colony. After therapy, body weights of the nonradioactive control groups increased up to the time of sacrifice because of their large tumors. The weight of the group treated with B3-SA and 90Y-labeled DOTA-biotin showed a mean drop of 6.6% but returned to baseline by 2 weeks. In contrast, the 90Y-labeled DOTA-biotin group pretargeted with nonspecific HAT-SA had an initial mean drop of 9.6% of their weight and recovered by 21 days, but thereafter the mice progressively weakened and continued to lose weight until all had died or were sacrificed by 41 days. The drops in weight for both 90Y-treated groups were significantly different from each other (P < 0.02).
Although only a few animals (n = 3) had CBC at each therapy dose, comparison of hematology results in the therapy groups showed that there were no significant differences in platelet and RBC counts between the different groups (P > 0.12), with the nadirs occurring at ∼2 weeks. Nevertheless, a significant decrease in WBC and lymphocyte counts was observed 1–2 weeks after therapy in the 37 MBq-treated groups (P < 0.007), with the nadir occurring at ∼2 weeks and recovery by 3 weeks. This represented decreases of 79% and 94% for WBC and lymphocyte count, respectively. Liver and kidney function chemistry performed at the time of sacrifice was no different between nonradioactivity receiving controls or 90Y pretargeted groups.
Histological analysis of the liver and kidney of mice receiving the 37 MBq of 90Y-labeled DOTA-biotin-B3-SA showed similar finding as those of the control group that did not receive radioactivity. In addition, the bone and bone marrow of the mice receiving 37 MBq of 90Y-labeled DOTA-biotin-B3-SA were histologically normal (performed 180 days after therapy). The mice receiving 37 MBq of 90Y-labeled DOTA-biotin-HAT-SA showed similar findings as the control animals and did not have any histological findings in the kidney or liver to explain their toxicity.
The pretargeting approach based on SA-biotin depends on the successful delivery and function of three reagents. In this study we showed sufficient concentration of B3-SA to serve as an effective binding receptor for the 111In/90Y-labeled DOTA-biotin. In this study, the %ID/g of the MAb in tumor was shown to be similar for doses of 5–400 μg, whereas the total amount delivered continued to increase. Autoradiography showed that heterogeneity in tumor distribution was not grossly decreased at the higher antibody dose levels. There was some evidence of partial saturation of uptake at the 800 μg dose, as evidenced by the lack of uptake linearity at this higher dose. Our decision to use 400 μg was based on: (a) some evidence of saturation at the higher doses that resulted in lower tumor:normal organ uptake ratios; (b) very large doses of B3-SA that may be more difficult to clear from the blood and may result in higher tissue background; and (c) the results of these experiments, which indicated that the 400 μg dose level resulted in good tumor targeting.
The clearing step and administration of radiolabeled biotin at 24 h after giving B3-SA was chosen because near-peak antibody tumor levels were achieved by this time (Fig. 1,B). In addition, studies reported previously indicated that despite the low biotin diet, endogenous biotin fills additional binding sites of MAb-SA conjugate at 48 h as compared with 24 h, which decreases the binding sites for the subsequently administered radiolabeled biotin. The doses of sCA and DOTA-biotin we used were effective, as gauged by the high tumor concentration of radiolabeled DOTA-biotin and its fast clearance from the blood (Figs. 2). The dose of these two reagents was based on that shown previously to be effective in other studies (10). These two reagents are “universal” or independent of what MAb-SA conjugate is used, and the doses required for therapy are unlikely to be greatly dependent on what tumor antibody system is used.
Our studies showed that 111In- or 88Y-labeled DOTA-biotin was delivered in similar concentrations to tumor and normal organs. Thus, this similarity allowed us to make dosimetry estimates for 90Y based on 111In biodistribution. The pretargeting approach resulted in rapid delivery of high doses of radiation to the tumor. The calculated dose delivered to the tumor is likely to be an underestimate, because in our pretargeted animals, the tumor size increased, resulting in decreasing %ID/g. If these animals had received therapy, their tumors would not have continued to grow and the radiation dose would have been much higher. 90Y B3 dosimetry extrapolated from previous biodistribution studies suggested that high radiation doses to tumor could be delivered and that much higher doses to normal tissues on an equivalent MBq basis would also be delivered compared with the pretargeting approach. Although the dose-limiting toxicity from 90Y B3 (not pretargeted) was not evaluated in this study, works by others suggest that the maximal tolerated dose by nude mice was in the range of 7.4–11.1 MBq (10, 20). These lower tolerated doses would significantly limit the maximal radiation dose to tumor as compared with the pretargeted approach, where higher pretargeted amounts of 90Y were possible. Our dosimetry estimates for 90Y B3 on a per MBq basis were in the same range reported by Axworthy et al. (10) for nonpretargeted MAb NR-LU-10. This is because of similar peak concentration and kinetics in tumor for their MAb (29.9%ID/g) and ours (26.3%ID/g) at 24 h. In contrast, the peak dose observed in this study from pretargeted radiolabeled DOTA-biotin was lower than the 34.9%ID/g observed previously with NR-LU-10-SA antibody (10). Given the fast clearance of the DOTA-biotin through the kidneys, the ratios of tumor:blood and tumor:bone marrow are more favorable than when 90Y B3 is given without pretargeting. The average tumor:blood area under the curve ratio in our pretargeting studies was 9.3, which compares favorably with the 2.4 ratio seen with 90Y directly conjugated to B3.
Significant decreases of WBC and lymphocyte counts were observed 1–2 weeks after therapy. The platelet counts in the treated and control groups failed to show any significant changes. These results are in line with those reported previously with NR-LU-10-SA, where rapid targeting was observed and where dose-limiting toxicity was not observed at doses of 29.6 MBq (800 μCi). The limited toxicity to the marrow was most likely attributable to the rapid clearing of the unbound 90Y-labeled DOTA-biotin through the kidney, such that with the long physical half-life of 90Y this would not have given a large dose to the whole body, bone marrow, or kidney. Evidence of toxicity was observed when pretargeted 90Y-labeled DOTA-biotin was used with HAT-SA where the mice died or were sacrificed because of tumor ulceration or significant body weight loss within 41 days. Additional studies will be required to determine the cause of toxicity in this control group.
Whereas preclinical animal studies are useful and necessary to help optimize delivery strategies and show proof of principle, use in humans remains the final goal. Given the favorable targeting seen previously in vivo with 111In-CHX-B3, the findings in this pretargeting study encourage us to pursue this system in humans (12).
In summary, B3-SA using a three-step pretargeting approach allows rapid and specific delivery of radionuclides to tumor xenografts. High tumor uptake with fast clearance from the vasculature and low nontumor uptake resulted in high tumor doses and good radioimmunotherapeutic responses. These findings support the use of this reagent for a clinical trial in patients.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The abbreviations used are: MAb, monoclonal antibody; %ID/g, injected dose per gram; SA, streptavidin; sCA, synthetic clearing agent; DOTA, 1,4,7,10-tetraazacyclododecane-N,N′,N″,N‴-tetraacetic acid; PSL, photoluminescence stimulated unit; HAT-SA, humanized anti-Tac monoclonal antibody-streptavidin; CBC, complete blood count.
We thank Debby Drumm-Guisseppe for expert technical assistance in animal experiments and assays.