Radioimmunotherapy with α-Particle–Emitting ²¹³Bi-C-Functionalized trans-Cyclohexyl-Diethylenetriaminepentaacetic Acid-Humanized 3S193 Is Enhanced by Combination with Paclitaxel Chemotherapy

A broad number of tumor-associated antigens have been identified as targets for monoclonal antibody (mAb) therapy of cancer, including the Lewis Y (Le^y) antigen, a carbohydrate type 2 blood group-related antigen overexpressed as either glycolipid or cell surface glycoproteins on tumor cell surfaces. Expression of Le^y has been described in 60% to 90% of tumors and metastases of epithelial origin, including breast, colon, prostate, non–small cell lung, ovarian, and gastric cancers, and therefore, Le^y is an attractive target for mAb therapy (1). The humanized anti-Le^y mAb humanized 3S193 (hu3S193) is currently under investigation in phase I/II trials in patients with Le^y-positive tumors. The specificity of hu3S193 for Le^y makes this mAb an attractive candidate for immunotherapy as well as the targeted delivery of cytotoxics, such as radioisotopes, to tumor cells (2). Accordingly, the biodistribution and antitumor efficacy of radiolabeled hu3S193 have been investigated in preclinical tumor xenograft animal studies of radioimmunotherapy (3, 4).

To date, the majority of human radioimmunotherapy studies targeting solid tumors have been done using β-particle–emitting radioisotopes, such as iodine-131 and yttrium-90, and clinical trials of these constructs have been primarily done in patients with bulky disease (5). However, studies have increasingly found that the greatest prospect for successful clinical applications of radioimmunotherapy of solid tumors is in the treatment of limited tumor burden, such as disseminated and micrometastatic disease (6, 7). However, the radiochemical properties of β-emitter radioisotopes currently used in radioimmunotherapy can be disadvantageous for such applications. The long path lengths (2-10 mm) of β-emitters mean that only a fraction of the emitted energy is deposited in small-volume disease, with the remainder passing to normal tissues, increasing the risk of nonspecific cytotoxicity (8). The short range of α-particle emitters (50-80 μm) is more suited to the treatment of small-volume disease, as the high-energy (5-8 MeV) emissions are deposited directly over 2 to 4 cell diameters, resulting in a high absorbed dose and linear energy transfer (9). The high linear energy transfer of α-emitters in part contributes to their higher tumor relative biological effectiveness of approximately 3 to 4 compared with an equivalent dose of β-emitter (10–12).

Although more than 100 α-emitting radioisotopes are known to exist, various factors, such as half-life, labeling chemistry, and general availability, have greatly limited the number of feasible options for α-particle radioimmunotherapy of cancer (13). Subsequently, only bismuth-212 (t_1/2 = 60.6 min), bismuth-213 (²¹³Bi; t_1/2 = 45.6 min), astatine-211 (t_1/2 = 7.2 h), and actinium-225 (t_1/2 = 10 days) have been extensively explored in preclinical radioimmunotherapy (9). To date, the most advanced of these is ²¹³Bi, which has been evaluated in patients using mAb HuM195, which is reactive to the CD33 antigen overexpressed on human leukemia cells, and also with mAb 9.2.27, which targets metastatic melanoma (14, 15). As ²¹³Bi is produced from decay of actinium-225, a ²¹³Bi generator has been developed, which produces up to 25 mCi of pure, chemically reactive ²¹³Bi that can be readily conjugated to antibodies through radiometal chelates, such as C-functionalized trans-cyclohexyl-diethylenetriaminepentaacetic acid (CHX-A"-DTPA; refs. 8, 16, 17). However, because of the limited 45.6-min half-life of ²¹³Bi, efficacy is likely only to be observed when targeting readily accessible micrometastases and hematologic malignancies (18–20).

Radioimmunotherapy using ²¹³Bi and other α-emitters has been explored in several tumor types and has shown efficacy in diverse applications, such as pretargeted and fractionated locoregional radioimmunotherapy (21, 22). The purpose of the current study was to characterize the in vitro and in vivo activity of ²¹³Bi-CHX-A"-DTPA-hu3S193 to determine its suitability for the treatment of micrometastatic Le^y-positive cancers. In addition, the possible enhancement of α-particle radioimmunotherapy by combination with paclitaxel chemotherapy was assessed. A solid rationale of combining targeted radiation with paclitaxel now extends to the clinic, and enhancement of α-emitter radioimmunotherapy cytotoxicity by paclitaxel through cell cycle arrest and increased DNA double-strand breaks has been observed in vitro (23, 24). Finally, the effect of tumor size and morphology on antitumor efficacy was assessed, as these factors will likely have critical importance to the clinical efficacy of α-emitter radioimmunotherapy (6).

Cell lines and antibodies. MCF-7, a Le^y-positive breast adenocarcinoma cell line, has been described previously (3). Cell viability, as determined by trypan blue exclusion, exceeded 90% in all experiments. hu3S193, a complementary determining region–grafted IgG1 antibody specific for the Le^y antigen (1), and isotype control huA33 (25) were produced by the Biological Production Facility, Ludwig Institute for Cancer Research (Melbourne, Victoria, Australia).

Chelation and radiolabeling.²¹³Bi was eluted from an actinium-225 generator obtained from the Institute of Power Engineering and Physics (Obninsk, Russia). Radiolabeling of hu3S193 with ²¹³Bi was achieved using the bifunctional metal ion chelate CHX-A"-DTPA (26).

Immunoreactivity and affinity. Determination of the radiolabeled hu3S193 immunoreactivity was done according to the “Lindmo” cell binding assay using Le^y-positive MCF-7 cells as described previously (3). The affinity constant (K_a) and the number of antigen binding sites per cell were determined by Scatchard analysis (3). The immunoreactivity of the control huA33 IgG was determined by single-point binding assay, where 10 × 10⁶ A33 antigen-positive SW1222 cells were incubated with 20 ng radiolabeled huA33 for 30 min at room temperature. Cells were washed thrice, and immunoreactivity was expressed as a percentage of activity in duplicate radioconjugate standards.

Intracellular trafficking of hu3S193. hu3S193 was labeled with Alexa Fluor 546 using the protein labeling kit (Molecular Probes) according to the manufacturer's instructions. MCF-7 cells (3.0 × 10⁵) in 1 mL culture medium were seeded onto 10-mm glass coverslips (ProSciTech) in 24-well plates and incubated overnight. Cells were washed thrice in 0.25% human serum albumin before incubation with 1:75 dilution of hu3S193-Alexa Fluor 546 in a humidified chamber at 4°C for 30 min. Cells were again washed and incubated in 1 mL of 37°C culture medium in 24-well plates for 0, 5, 10, 15, and 30 min for endosomal trafficking or 0, 15, 30, 60, and 90 min to assess hu3S193 localization to lysosomes. Following incubation, cells were human serum albumin washed, fixed for 30 min in 4% (w/v) paraformaldehyde, and permeabilized with 0.1% Triton X-100 detergent for 1 min (Sigma Chemical). After permeabilization, cells were incubated with either 1:200 mouse anti-human early endosomal antigen-1 (EEA-1) or 1:50 mouse anti-human lysosomal-associated membrane protein-1 (LAMP-1), respectively, for 20 min at room temperature (BD Biosciences). Mouse antibodies were then probed with Cy2 fluorophore-conjugated (green fluorescence) donkey anti-mouse antibody (Jackson ImmunoResearch Laboratories) for 20 min at room temperature. Coverslips were finally washed once in PBS and once in distilled water and mounted on glass slides (SuperFrost Plus, Menzel Glaser) using Fluoromount G mounting medium (Southern Biotechnology).

The intracellular localization of hu3S193 and the location of endosomes and lysosomes were imaged using a Bio-Rad MRC-1024 confocal microscope attached to an inverted microscope (Nikon Eclipse TE300) equipped with a 60× oil immersion lens with a coverslip correction collar (Nikon CFI Plan Apo 60×, 1.2 numerical aperture). Colocalization between hu3S193 and either of the compartments is shown by the presence of yellow fluorescence on merging of the separate red (hu3S193) and green (EEA-1 or LAMP-1) fluorescence images.

[³H]thymidine proliferation assay. Measurement of [³H]thymidine incorporation into DNA of proliferating cells was used to determine the cytotoxicity of ²¹³Bi-radiolabeled hu3S193 on the proliferation of MCF-7 breast carcinoma cells. Cells at a density of 5 × 10³ per well were plated out in microwell 96-well plates (Nalgene Nunclon). Immediately following the addition of cells into the wells, ²¹³Bi-hu3S193 ranging in concentration from 25 to 200 μCi/mL was added to wells in medium to give respective final concentrations of 2.5, 5, 10, and 20 μCi/well in the treatment groups. Corresponding control wells were treated with medium alone and unlabeled hu3S193 was equal in concentration (∼5 μg/well) to radiolabeled hu3S193. To ensure even distribution of the radioconjugate, cells were mixed by gentle pipetting every 10 min for 2 h while incubated at 37°C. Cells were further incubated 24 or 72 h before the addition of 1 μCi/well [³H]thymidine (Perkin-Elmer Life and Analytical Sciences), with cells then incubated overnight (∼12 h) to allow [³H]thymidine incorporation. Cells were washed in ice-cold PBS, fixed in 10% (w/v) trichloroacetic acid, and washed again in PBS before being lysed in 100 μL 0.1 mol/L NaOH. A 50 μL sample of the lysis solution was then mixed with 5 mL of scintillation fluid (Ultima, Gold, Perkin-Elmer Life and Analytical Sciences), and activity was measured on an automated scintillation counter (LS6500, Beckman Instruments).

Apoptosis cell death ELISA. Apoptosis resulting from treatment with radiolabeled hu3S193 was assessed using the Cell Death Detection ELISA^PLUS (Roche Diagnostics), a quantitative sandwich ELISA, according to the manufacturer's protocol. MCF-7 cells were prepared as for the [³H]thymidine proliferation assay and treated with radiolabeled hu3S193 for either 24 or 72 h.

Immunoblotting analysis of DNA damage. MCF-7 cells (5 × 10⁵) were incubated with 50 and 100 μCi/well ²¹³Bi-hu3S193 in six-well culture plates (Nalgene Nunclon) for 4 h at 37°C. Cells were then washed in ice-cold PBS, whole-cell lysates were prepared, and proteins were separated by SDS-PAGE as described previously (27). Following transfer to polyvinylidene difluoride Immobilon-P transfer membrane (Millipore Corp.), proteins were detected by immunoblotting with 1:200 anti-γ-H2AX antibody (Calbiochem), which detects phosphorylated histone 2AX, a marker of double-stranded DNA breakages (28). Membranes were then incubated with horseradish peroxidase–conjugated anti-rabbit IgG (Chemicon International), and specific protein bands were visualized using enhanced chemiluminescence (ECL reagent, GE Healthcare Biosciences).

Animal models and xenograft establishment. MCF-7 xenografts were established in female BALB/c nude (nu/nu) mice of 3 to 4 weeks of age (Animal Resource Centre, Perth, Western Australia, Australia) as described previously (3). The growth of MCF-7 tumors in vivo was assessed over time, with individual tumors removed 2, 3, 5, 7, 9, 12, and 21 days after injection of cells. The morphology, Le^y expression, vascularization, and encapsulation of the tumor cells were assessed histologically. Tumor volume was regularly measured using digital calipers and the following formula: tumor volume (mm³) = (length × width²) / 2 of the tumor (3). Estrogen pellets implanted in mice to support the estrogen-dependent MCF-7 tumors were reimplanted on day 85 of the study to allow assessment of tumor growth to study end at day 160. All animal studies were approved by the Animal Ethics Committee of the Austin Hospital (Heidelberg, Victoria, Australia).

²¹³Bi-hu3S193 in minimal residual disease. To assess the efficacy of ²¹³Bi-hu3S193 α-emitter radioimmunotherapy in minimal residual disease, mice were injected with ²¹³Bi-hu3S193 or controls 2 days after injection of MCF-7 cells. Mice (n = 5 per group) were injected by tail vein injection with single doses of 15, 30, or 60 μCi ²¹³Bi-hu3S193 (9.2 μg protein) or control ²¹³Bi-huA33 in 0.1 mL saline. Control groups received saline vehicle or 9.2 μg unlabeled hu3S193 (to match protein dose of mAbs across all therapy groups) in equivalent volumes to the radiolabeled antibodies. The day of mAb injection was designated day 0 of the study.

²¹³Bi-hu3S193 in combination with paclitaxel. Mice (n = 5 per group) with 2-day-old xenografts were injected by tail vein injection with single doses of 15, 30, or 60 μCi ²¹³Bi-hu3S193 (9.2 μg protein). Twenty-four hours following mAb injection, mice were injected i.p. with paclitaxel (Anzatax, Mayne Pharma) at a subtherapeutic 300 μg dose (3 mg/mL) in an aqueous solution containing 100 μL paclitaxel diluted in 200 μL saline for injection (0.9% saline, Pharmacia & Upjohn). This dose has previously been shown not to have significant efficacy when given alone and did not show toxicity in BALB/c nude mice (29, 30).

²¹³Bi-hu3S193 in established xenografts. The therapeutic efficacy of hu3S193 was assessed using 7-day established MCF-7 xenografts. Mice (n = 5 per group; tumor volume = 125.4 ± 51.5 mm³) were injected with a single dose of 30, 60, or 120 μCi ²¹³Bi-hu3S193 (20 μg protein) by tail vein injection 7 days after establishment of the xenografts. Separate control groups received saline vehicle or 20 μg unlabeled hu3S193 in equivalent volumes to the radiolabeled antibodies. A nonspecific control arm using ²¹³Bi-huA33 was not included in this established xenograft model, as specificity of ²¹³Bi-hu3S193 had previously been established in the earlier minimal residual disease models.

Statistical analysis. Statistical analysis of therapy studies was carried out using Prism (version 4.0; GraphPad Software, Inc.). Unpaired two-tailed t tests were done on mean (±SD) tumor volumes at the termination of the PBS control group (day 72). Significance was set at the 95% level, with results declared statistically significant if P < 0.05. Survival curves of mice in the therapy models were also constructed, and log-rank tests were done to assess differences in survival, again using Prism. Treatment of in vivo xenografts resulting in tumor volumes ≤100 mm³ on the day of saline control groups termination was defined as a partial response, where complete regression of tumor xenografts was defined as complete responses.

Immunoreactivity of radiolabeled hu3S193.²¹³Bi-hu3S193 and control ²¹³Bi-huA33 were prepared at high specific activities (6-6.5 mCi/mg) for in vitro and in vivo characterization. Assessment of radiochemical purity by instant TLC consistently showed >99% ²¹³Bi bound to hu3S193 and huA33. Immunoreactivity of ²¹³Bi-hu3S193 as determined by the Lindmo assay was 85%, with an apparent affinity of 1.4 × 10⁷ mol/L⁻¹ as determined by Scatchard analysis. Immunoreactivity of ²¹³Bi-huA33 to A33 antigen-positive SW122 cells as determined by single-point binding assay was 88.3%.

Intracellular trafficking of hu3S193. The intracellular trafficking of hu3S193 through endosomes was investigated using a specific marker of the endosomal pathway, EEA-1, by confocal microscopy (Fig. 1A). Endosomes are visualized as green punctate vesicles interspersed around the cell periphery. hu3S193 was rapidly internalized into MCF-7 cells within 5 min (data not shown). Maximal trafficking of hu3S193 to endosomes was observed following 15-min incubation, with colocalization clearly evident between hu3S193 and EEA-1 at this time (Fig. 1A, white arrows).

The fate of hu3S193 after trafficking through early endosomes was assessed using antibody to LAMP-1, as macromolecules such as antibodies are trafficked to lysosomes for degradation (Fig. 1B). Lysosomes were visualized as large intracellular bodies by staining with anti-LAMP-1 antibody and Cy2-conjugated secondary antibody. Evident localization of hu3S193 to lysosomes was first observed after 30-min incubation, with clear colocalization with LAMP-1 apparent (Fig. 1B, white arrows). Colocalization between hu3S193 and LAMP-1 remained evident up to 120-min incubation (data not shown).

Cytotoxicity of ²¹³Bi-hu3S193 in vitro.²¹³Bi-hu3S193 showed potent and specific effects on the growth of MCF-7 cell in vitro. As shown in Fig. 2, the cytotoxicity of ²¹³Bi-hu3S193 was dose dependent, with decreased cell viability apparent with increasing doses of ²¹³Bi-hu3S193. All doses caused potent and significant (P < 0.0001) antiproliferative effect compared with controls, with the highest dose of 20 μCi causing near complete cytotoxicity, with only 1.96% viable cells at 24-h incubation (Fig. 2A). The calculated dose causing 50% cytotoxicity (EC₅₀) for ²¹³Bi-hu3S193 was 1.0 μCi/well.

Specificity of the ²¹³Bi-hu3S193–mediated cytotoxicity was assessed using isotype control ²¹³Bi-huA33 mAb, which does not bind MCF-7 cells. Minimal cytotoxicity was observed using ²¹³Bi-huA33 even at the maximal 20 μCi/well dose, with no significant difference in cell viability relative to untreated control (P = 0.62). Some apparent cytotoxicity is apparent following 72-h incubation, although it is again not significantly different from untreated controls (P = 0.25) and significantly less (P < 0.0001) than that observed with Le^y-specific ²¹³Bi-hu3S193 (Fig. 2B). The EC₅₀ for ²¹³Bi-huA33 was ∼200 μCi.

Induction of apoptosis and DNA damage by ²¹³Bi-hu3S193. A clear and specific induction of apoptosis following treatment with ²¹³Bi-hu3S193 and appropriate controls was apparent following both 24- and 72-h incubation (Fig. 3A and B). No dose-specific effect was observed, with levels of apoptosis relatively equivalent at the 10 and 20 μCi doses. Levels of apoptosis increased over time, with 10 μCi ²¹³Bi-hu3S193 observed to cause a 3.4-fold increase in apoptosis relative to untreated controls following 72-h incubation, whereas only a 1.8-fold increase was apparent at 24 h. ²¹³Bi-huA33 was not observed to cause any significant induction of apoptosis (Fig. 3A and B).

The ability of ²¹³Bi-hu3S193 to cause double-stranded DNA breaks was assessed by Western blotting for γ-H2AX. As shown in Fig. 3C, clear and specific expression of γ-H2AX is present in cells exposed to ²¹³Bi-hu3S193 at 50 and 100 μCi relative to control cells treated with either medium alone or 20 μg/mL hu3S193.

Efficacy of ²¹³Bi-hu3S193 in MCF-7 minimal residual disease model. The results of ²¹³Bi-hu3S193 single-agent treatment in the minimal residual disease model are shown in Fig. 4A. Xenografts in mice injected with saline or 9.2 μg unlabeled hu3S193 grew relatively quickly, with mice euthanized on days 72 and 79 of the study, respectively, due to tumor burden reaching ethical limits. ²¹³Bi-hu3S193 caused significant retardation of xenograft growth relative to saline controls at all three doses investigated (P < 0.001 at all doses) and was evidently superior to control ²¹³Bi-huA33 at equivalent doses (Fig. 4B). Accordingly, mice that received ²¹³Bi-huA33 at doses of 15, 30, and 60 μCi were euthanized on days 79 (15 and 30 μCi) and 83 (60 μCi). In contrast, tumors treated with 15 μCi ²¹³Bi-hu3S193 showed retarded tumor growth, and mice were euthanized on day 128 of the study. This 49-day difference in survival represents a significant (P = 0.0021) survival advantage over the equivalent dose of ²¹³Bi-huA33. The higher doses of 30 and 60 μCi ²¹³Bi-hu3S193 showed even more marked antitumor effect relative to huA33 controls (P < 0.0001 and 0.0001, day 72) and prolongation of mice survival to completion of the study at day 153, indicating a dose-dependent effect of the ²¹³Bi-hu3S193 treatment (Fig. 4C and D). No significant weight loss or other toxicities were identified that were directly attributable to radioimmunotherapy treatments. One mouse from the 60 and 30 μCi ²¹³Bi-hu3S193 treatment groups died on days 21 and 42, respectively, of the study, which decreased the survival advantage over isotype controls (P = 0.037 and 0.035), although no cause was apparent on postmortem examination and no other adverse events were observed.

Combined ²¹³Bi-hu3S193 radioimmunotherapy and paclitaxel chemotherapy in MCF-7 minimal residual disease model. The efficacy of combined ²¹³Bi-hu3S193 radioimmunotherapy and 300 μg paclitaxel chemotherapy on the growth of newly established MCF-7 xenografts is presented in Fig. 5. Single-agent paclitaxel was observed to cause moderate, yet significant, antitumor effect relative to saline controls (P = 0.0023) at day 72, and mice were euthanized on day 83. Treatment with 15 μCi ²¹³Bi-hu3S193 combined with paclitaxel caused significant suppression of tumor growth relative to saline controls (P < 0.0001), although no significant improvement (P = 0.13) over radioimmunotherapy alone was apparent. However, mice receiving 15 μCi ²¹³Bi-hu3S193 and paclitaxel showed a significant increase in survival (P = 0.0021) over ²¹³Bi-hu3S193 and survived to study completion (Fig. 5D). Treatment with 30 and 60 μCi ²¹³Bi-hu3S193 and paclitaxel also resulted in very significant suppression of tumor xenograft growth compared with saline controls (P < 0.0001 and 0.0001) and survival to study completion. At these doses, a significant improvement in efficacy was observed with combined radioimmunotherapy and paclitaxel (P = 0.0052 and 0.0312) relative to radioimmunotherapy alone on day 72. Combined radioimmunotherapy and paclitaxel produced definable tumor responses at all doses, with a partial response observed in 40%, 20%, and 60% of mice receiving 15, 30, and 60 μCi ²¹³Bi-hu3S193 and paclitaxel, respectively. Complete and durable resolution of tumors was observed in 20% of mice receiving 15 μCi ²¹³Bi-hu3S193 and in 40% of mice receiving 30 and 60 μCi doses.

Efficacy of hu3S193 in established MCF-7 xenografts. The antitumor efficacy of ²¹³Bi-hu3S193 in established 7-day-old xenografts was assessed in addition to the minimal residual disease model (Fig. 6A). Unlabeled low-dose hu3S193 (20 μg) did not show any antitumor effect relative to saline controls (P = 0.8847), and both groups were euthanized 63 days after study commencement due to tumor burden. Significant antitumor efficacy (P < 0.0001) following treatment with ²¹³Bi-hu3S193 was observed at each of the respective dose levels (30, 60, and 120 μCi). A dose-dependent effect was evident, with tumor volumes in mice that received 120 μCi ²¹³Bi-hu3S193 significantly smaller than those that received the 30 μCi dose (P = 0.0045, day 63). Some possible toxicity was observed following 120 μCi ²¹³Bi-hu3S193, as a mouse died on days 14 and 35, respectively (Fig. 6B). An average weight loss of 10% original weight was observed in the remaining mice that received the 120 μCi dose of ²¹³Bi-hu3S193, although this resolved within 3 weeks after injection and these animals survived until study completion. However, because of the two early deaths following 120 μCi ²¹³Bi-hu3S193, this treatment did not improve survival relative to saline controls (P = 0.296). Conversely, significantly improved survival compared with controls (P = 0.0034) was observed following 30 and 60 μCi ²¹³Bi-hu3S193, as toxicity was not evident at these doses.

Radioimmunotherapy with short-path-length (50-80 μm) and high-energy (5-8 MeV) α-particle–emitting radioisotopes is ideally suited to the treatment of small tumors, residual disease, and micrometastatic disease, where promising efficacy has been observed in several preclinical and clinical studies (10, 31). To date, preclinical hu3S193 radioimmunotherapy has been assessed exclusively preclinically using β-emitting radioisotopes in models of established disease, where significant efficacy was observed (4, 30). However, such isotopes are unsuitable for single-cell killing and for treatment of micrometastatic disease (9, 32). Studies by McDevitt et al. (19) have suggested that α-emitter radioconjugates prepared at high specific activities have greater efficacy than respective radioconjugates at lower specific activities. In the current study, hu3S193 was prepared at high specific activities averaging 6.5 mCi/mg (240.5 MBq/g) for both in vitro analyses and in vivo studies of animal models of breast cancer, which were comparable with other reports (16, 33).

Internalization of radiolabeled antibodies is not required for cytotoxicity, as the radiation emitted can result in lethal irradiation over the effective path length of the radioisotope (11, 34). However, internalizing mAbs are advantageous in radioimmunotherapy, as radiation emitted from internalized radioisotopes has an increased probability of traversing the cell and hitting the nucleus relative to surface-bound radioisotope (35). Accordingly, the internalization and intracellular trafficking of hu3S193 were examined. Rapid internalization of hu3S193 was observed, with evident intracellular localization occurring within 5 min. Such rapid internalization is especially important for efficient delivery of radiation dose when using short-lived radioisotopes, such as ²¹³Bi, in α-radioimmunotherapy (19). Evident localization of hu3S193 to MCF-7 cell lysosomes was observed 30 min after internalization as apparent by colocalization of hu3S193 with LAMP-1. Such lysosomal localization has been previously observed with another anti-Le^y mAb, BR96 (36). It is also well established that radiometal-DTPA-lysyl peptide complexes formed from lysosomal catabolism of mAbs conjugated with radiometals, such as ²¹³Bi, through DTPA chelates accumulate in lysosomes and cannot escape, further ensuring maximal cellular irradiation (37). Together, the rapid internalization and subsequent intracellular accumulation of hu3S193 suggest it is ideally suited to efficient delivery of ²¹³Bi radiation to tumor cells.

The efficacy of ²¹³Bi-hu3S193 for single-cell killing was assessed using the [³H]thymidine incorporation assay, where potent and specific killing of MCF-7 cells was apparent following treatment with ²¹³Bi-hu3S193. A dose-dependent effect was observed, and an EC₅₀ of 1.0 μCi was determined. The potency of ²¹³Bi-hu3S193 in killing MCF-7 cells is comparable with the ²¹³Bi-radiolabeled construct α-PAI2, where a 37% survival value (D₀) of 2.3 μCi was observed following treatment (38). Various other constructs ²¹³Bi-labeled constructs have been explored in radioimmunotherapy targeting melanoma, colorectal, prostate, and pancreatic cancers, with low survival values generally comparable with the EC₅₀ of ²¹³Bi-hu3S193 observed in the current study (39–42). Specificity of the ²¹³Bi-hu3S193 was apparent by the failure of nonspecific ²¹³Bi-huA33 to produce significant toxicity, resulting in a much higher EC₅₀.

It is thought that high linear energy transfer radiation from α-emitter radiation results in irreparable double-stranded DNA breaks and severe chromosomal damage leading to the rapid and irreversible induction of apoptosis and cell death (8, 12, 43). In the current study, significant (P < 0.0001) induction of apoptosis as assessed by DNA fragmentation was observed after treatment with ²¹³Bi-hu3S193 and was most apparent following 72-h incubation. These results agree with several other studies assessing α-particle radioimmunotherapy in lymphoma and prostate cancers, where treatment was shown to induce cellular apoptosis (12, 44). Phosphorylation of histone H2AX (γ-H2AX) is known to occur in response to double-stranded DNA breaks and was assessed via immunoblotting (28). Robust expression of γ-H2AX was only observed in MCF-7 cells treated with 50 or 100 μCi ²¹³Bi-hu3S193, indicating significant and specific induction of double-stranded DNA breakages following ²¹³Bi-hu3S193 treatment. This is consistent with numerous other studies where foci of γ-H2AX expression were observed following α-emitter irradiation (45).

Efficacy of ²¹³Bi-hu3S193 was initially assessed in vivo in a dose titration study, where significant retardation of tumor xenograft growth relative to control mice injected with either saline, unlabeled hu3S193, or isotype control mAb was observed following single doses of 15, 30, or 60 μCi ²¹³Bi-hu3S193. Although hu3S193 does have potent immune effector functions, the lack of antitumor efficacy following unlabeled mAb treatment in this study was not unexpected and agrees with our previous studies (1, 30). Single low doses of antibody were used, and it is also possible that hu3S193 is not able to effectively activate murine complement in the BALB/c nude mouse. The efficacy observed following systemic delivery of ²¹³Bi-hu3S193 in this study compares favorably with others, where only doses approaching 100 μCi were observed to mediate considerable antitumor efficacy (19, 46, 47).

The chemotherapeutic paclitaxel has been identified in several studies to have radiosensitizing properties and has been combined with both external beam radiotherapy and radioimmunotherapy in several preclinical and clinical studies (24, 29, 48). A study by Supiot et al. (23) has examined sensitisation to α-radioimmunotherapy using paclitaxel and doxorubicin, where paclitaxel was observed to enhance the efficacy of α-radioimmunotherapy of myeloma cell lines through mediating cell cycle arrest and increasing DNA strand breaks, providing a solid rationale for the investigation of combining α-particle radioimmunotherapy with chemotherapies. Combined treatment of ²¹³Bi-hu3S193 and paclitaxel resulted in enhanced efficacy at the 30 and 60 μCi dose relative to radioimmunotherapy alone (P = 0.0052 and 0.0312) at day 72 and enhanced survival at the 15 μCi dose (P = 0.0027). Enhancement of α-particle radioimmunotherapy in vivo by combination with paclitaxel has not yet been previously reported and is a promising avenue available to ensure maximal efficacy of radioimmunotherapy without reducing therapeutic margins.

Radioimmunotherapy with α-emitting radioisotopes is expected to be most effective in the treatment of minimal or micrometastatic disease, as their short path lengths result in more efficient delivery of radiation than that which can be achieved using β-emitter radioisotopes, such as yttrium-90 (13). Conversely, as tumor size increases, the efficacy of α-radioimmunotherapy decreases, as increasingly limited penetration of mAb, combined with the limited range of α-particles, results in a lower fraction of irradiated tumor (6, 8). In the current study, tumors treated with 30 or 60 μCi ²¹³Bi-hu3S193 2 days after implantation were significantly (P = 0.0062 and 0.0025) smaller at study completion than tumors treated after 7 days of growth, therefore indicating that tumor size and morphology are important determinants of α-emitter radioimmunotherapy. This inverse relationship between tumor control and tumor volume has been observed following local injection of ²¹³Bi-labeled mAbs in models of melanoma and prostate cancer (42, 49). A recent study by Elgqvist et al. examining ²¹¹At radiolabeled MX35 in i.p. OVCAR-3 xenografts further shows the importance of tumor size, with reduced tumor-free fractions, lowered mean absorbed dose, and increased percentage of tumor receiving zero dose all correlating with increased tumor volume (6, 32).

In summary, ²¹³Bi-hu3S193 has shown potent and specific activity in vitro and an ability to cause MCF-7 cell death through the induction of DNA damage and apoptosis. Treatment of a minimal residual disease model with ²¹³Bi-hu3S193 resulted in significant reduction relative to nonspecific controls, and efficacy of radioimmunotherapy was enhanced by combination with paclitaxel chemotherapy. Studies in established MCF-7 xenografts suggest that greatest efficacy of ²¹³Bi-hu3S193 will be observed in the treatment of cell clusters of early micrometastatic disease. Together, these findings support the further investigation of hu3S193 radioimmunotherapy with α-particle–emitting radioisotopes.

We thank Rushika Perera and Stephen Cody for their assistance in confocal microscopy techniques.