²¹³Bi (α-Emitter)–Antibody Targeting of Breast Cancer Metastases in the neu-N Transgenic Mouse Model

Abstract

Treatment failure in breast cancer is largely the failure to control metastatic dissemination. In this study, we investigated the efficacy of an antibody against the rat variant of HER-2/neu, labeled with the α-particle emitter ²¹³Bi to treat widespread metastases in a rat/neu transgenic mouse model of metastatic mammary carcinoma. The model manifests wide-spread dissemination of tumor cells leading to osteolytic bone lesions and liver metastases, common sites of clinical metastases. The maximum tolerated dose was 120 μCi of ²¹³Bi-7.16.4. The kinetics of marrow suppression and subsequent recovery were determined. Three days after left cardiac ventricular injection of 10⁵ rat HER-2/neu–expressing syngeneic tumor cells, neu-N mice were treated with (a) 120 μCi ²¹³Bi-7.16.4, (b) 90 μCi ²¹³Bi-7.16.4, (c) 120 μCi ²¹³Bi-Rituximab (unreactive control), and (d) unlabeled 7.16.4. Treatment with 120 μCi ²¹³Bi-7.16.4 increased median survival time to 41 days compared with 28 days for the untreated controls (P < 0.0001); corresponding median survival times for groups b, c, and d were 36 (P < 0.001), 31 (P < 0.01), and 33 (P = 0.05) days, respectively. Median survival relative to controls was not significantly improved in mice injected with 10-fold less cells or with multiple courses of treatment. We concluded that α-emitter ²¹³Bi-labeled monoclonal antibody targeting the HER-2/neu antigen was effective in treating early-stage HER-2/neu–expressing micrometastases. Analysis of the results suggests that further gains in efficacy may require higher specific activity constructs or target antigens that are more highly expressed on tumor cells.

Introduction

Metastasis to distant organs is the major cause of mortality in breast cancer patients. The 5-year relative survival rate is 26.1% compared with 97.9% for patients with localized tumors at diagnosis (1). To eradicate microscopic disseminated disease, systemic adjuvant therapies are necessary after surgery and radiation therapy of the primary tumor. Multicenter clinical trials with new chemotherapy regimens (2), aromatase inhibitor–based endocrine therapy (3, 4), and locoregional radiation therapy (5) have shown improved recurrence-free survival in advanced breast cancer patients. HER-2/neu is a cell surface tyrosine kinase associated with aggressive tumor behavior and poor prognosis and is overexpressed on ~20% of breast cancers (6). Targeting tumor antigen HER-2 using monoclonal antibody (mAb) Trastuzumab has also shown significant clinical benefit. Patients receiving Trastuzumab in combination with standard chemotherapy had a 52% decrease in recurrence compared with patients in the chemotherapy-alone group (7, 8). Trastuzumab, as a single agent, has an objective response of only 35% even in patients with 2+ and 3+ HER-2–positive breast cancer, as assessed by immunohistochemistry (9). One of the possible mechanisms for Trastuzumab resistance includes deficiency of the PTEN protein (10).

Radioimmunotherapy using the α-emitter ²¹³Bi delivers a cytotoxic radiation dose to tumors independent of the underlying signaling pathways. Compared with more commonly used β-emitter ⁹⁰Y and ¹³¹I, α-particles travel a very short distance (~80 μm) and deposit highly focused energy along their path compared with β-particles (average linear energy transfer of 100 keV/μm versus 0.2 keV/μm; ref. 11). As a result, α-particles can efficiently kill single cells and micrometastases with limited toxicity to surrounding normal tissues. Furthermore, the high prevalence of DNA double-strand breaks caused by α-radiation reduce the possibility of repair of sublethal damage, thereby making targeted α-particle therapy less susceptible to the majority of tumor resistance mechanisms. The short half-life of ²¹³Bi is well suited to targeting hematologic malignancies and prevascularized micrometastases. Thus far, efficacy of ²¹³Bi killing has been shown against PSMA-expressing prostate tumor spheroids in vitro and intramuscular tumors in vivo (12, 13). In three mouse models of intraperitoneal metastases of colon, pancreatic, and stomach cancer, ²¹³Bi-labeled antibodies have been able to improve survival rates in these mice (14, 15). Efficacy of ²¹³Bi-labeled antibodies to target lung metastases and melanoma have also been shown (16, 17). Clinical trials have shown safety, feasibility, and imaging of ²¹³Bi-labeled anti-CD33 antibody localization in targeting myeloid leukemia (18, 19).

Preclinical studies of antibody-mediated cytotoxic agents have largely been performed in xenograft models. In such models, the target antigen is exclusively expressed on the tumor. This is generally not the case in human studies. The neu-N mouse model (20) that expresses rat HER-2/neu on various normal organs, as well as the tumor cells, was therefore used here. We have previously shown that left cardiac ventricular (LCV) injection of syngeneic tumor cells in this model leads to wide-spread metastatic dissemination, including osteolytic bone metastases and also liver metastases (21).

In this study, we showed the efficacy of ²¹³Bi-labeled anti-rat-HER-2/neu mAb, 7.16.4, in the treatment of wide-spread breast cancer micrometastases in rat HER-2/neu transgenic mice. We hypothesized that ²¹³Bi-labeled whole antibody would be able to sterilize early micrometastases, easily accessible from the vasculature, whereas its toxicity to cross-reactive normal organs would be limited due to slow antibody localization. Maximal tolerated dose (MTD) was determined. The therapeutic efficacy of multiple treatment courses was also examined.

Materials and Methods

Mice, cell lines, and mAbs

neu-N transgenic mice, at ages 6 to 8 wk, that overexpress rat HER-2/neu under the mouse mammary tumor virus promoter were maintained and obtained from Harlan. All experiments involving the use of mice were conducted with the approval of the Animal Care and Use Committee of The Johns Hopkins University School of Medicine. The rat HER-2/neu–expressing mouse mammary tumor cell line NT2.5 was established from spontaneous mammary tumors in female neu-N mice (22). An NT4 cell line that does not express rat HER-2/neu was also derived similarly. The NT lines are grown in RPMI media containing 20% fetal bovine serum, 0.5% penicillin/streptomycin (Invitrogen), 1% L-glutamine, 1% nonessential amino acids, 1% sodium pyruvate, 0.02% gentamicin, and 0.2% insulin (Sigma) and maintained at 37°C in 5% CO₂. The hybridoma cell line for 7.16.4 was kindly provided by Dr. M. Greene (University of Pennsylvania). 7.16.4 collected from ascites of athymic mice was purified by a HiTrap protein G column (GE Healthcare Biosciences) using the Biologic LP purification system (Bio-Rad) and dialyzed into PBS using Centricon YM-10 filter units (Millipore). Rituximab (IDEC Pharmaceuticals Corp.), an anti-human CD20 mAb, and HuIgG, a control human IgG mAb, were used as negative controls.

Antibody conjugation, ²¹³Bi production, mAb radiolabeling, and purification

7.16.4 was conjugated to N-[2-amino-3-(p-isothiocyanato-phenyl)propyl]-trans-cyclohexane-1,2-diamine-N,N′,N′,N″,N″-pentaacetic acid (SCN-CHX-A″-DTPA), as described before (23, 24), and purified with Centricon YM-10 filter units (Millipore). The average number of chelates per antibody was typically between 1.0 and 2.0, as determined by the yttrium arsenazo spectrophotometric method (25). ²²⁵Ac was provided by the Institute for Transuranium Elements and purchased from Isonics Corp. The α-emitter ²¹³Bi was eluted from an ²²⁵Ac/²¹³Bi generator built in-house after a published procedure (26). 7.16.4 conjugated to the chelate were incubated with ${\text{BiI}}_4^{-}/{\text{BiI}}_5^{2-}$ (at 10 mCi/mg) for 8 min in a reaction buffer (pH 4.5) containing 3 mol/L ammonium acetate (Fisher Scientific) and 150 mg/mL L-ascorbic acid (Sigma) preheated to 37°C. The ²¹³Bi-labeled 7.16.4 was quenched with 10 μL of 100 mmol/L EDTA and purified by size exclusion PD-10 column or MicroSpin G-25 column (GE Healthcare BioSciences). The reaction efficiency and purity of the radioimmunoconjugates were determined with instant TLC using silica gel impregnated paper (Gelman Science, Inc.). Immunoreactivity of ²¹³Bi-7.16.4 was evaluated by incubating ~5 ng of ²¹³Bi-7.16.4 with excessive antigen binding sites (1 × 10⁷ NT2.5 cells) twice on ice for 30 min each time. Antibody immunoreactivity was calculated as the percentage of ²¹³Bi-7.16.4 bound to the cells.

Specific kill of ²¹³Bi-7.16.4 against single cells and tumor spheroids in vitro

Specific cell kill by ²¹³Bi-7.16.4 was determined by colony formation assay. Serially diluted ²¹³Bi-7.16.4 or ²¹³Bi-HuIgG ( from 0.08 to 40 μCi/mL) was incubated with 5 × 10⁴ NT2.5 cells in 200 μL medium for 24 h. Three different specific activities (0.7, 2.8, and 5.6 mCi/mg) of ²¹³Bi-7.16.4 and one specific activity (3.1 mCi/mg) of ²¹³Bi-HuIgG were used. ²¹³Bi-7.16.4 (2.8 mCi/mg) was also incubated with NT4 cells. After incubation, cells were harvested and plated in 25 cm² tissue culture flasks. Approximately, 250 total cells were plated per flask after dilution. The colonies formed after ~2 to 3 wk of culture were counted, and fraction of cells survived versus no treatment controls were calculated. Tumor spheroids were cultured using liquid-overlay method, as described previously (12). NT2.5 tumor spheroids (24 spheroids per group, ~200 μm in diameter) were treated with 5 μCi/mL (7.5 mCi/mg), 20 μCi/mL (0.9 or 7.5 mCi/mg), and 40 μCi/mL (7.5 mCi/mg) of ²¹³Bi-7.16.4. NT2.5 spheroids were also treated with ²¹³Bi-HuIgG at 20 and 40 μCi/mL (6.2 mCi/mg) as controls. Spheroid size was measured twice per week under an inverted microscope fitted with an ocular scale (Axiophot 2; Carl Zeiss Ltd.). The major and minor diameters, d_max and d_min, were used to calculate spheroid volume by $V=\pi \times d_{max}\times d_{\text{min}}^2/6$(12).

Biodistribution of 7.16.4 in neu-N mice

neu-N mice were inoculated s.c. in the mammary fat pad or through LCV injection with 1.0 × 10⁵ NT2.5 cells suspended in 100 μL PBS. Tumors were palpable and reached an average diameter of 5 mm 14 d after s.c. inoculation. At 3 wk after LCV injection, mice began to exhibit hind-limb paralysis and onset of abdominal swelling due to accumulation of ascites fluid (21). Mice (six per group for subcutaneous tumors and three per group for metastatic tumors) were given i.v. injection of 20 μCi ¹¹¹In–labeled 7.16.4 (~2.0 mCi/mg). At 1, 6, 12, 24, and 72 h post injection, the mice were euthanized and radioactivity in tumor (or metastatic tumor from the spleen), blood, heart, lung, liver, spleen, kidneys, stomach, small intestines, muscle, and femur were determined in a γ-scintillation counter. Results were corrected for physical decay and presented as percentage of injected activity per gram (%ID/g).

Determination of MTD

MTD was defined as the highest dose that allows 100% survival of the mice with no significant body weight loss (>15%). Reversible immune suppression was usually observed and was acceptable. Three injections were given to mice (six per group) on three consecutive days with total doses of 600, 300, 225, 150, 120, and 90 μCi. Mice were weighed twice per week for 120 d. All mice were euthanized, and major organs were collected and examined for histopathology. Immune suppression was monitored by counting CD3+ T cells using TruCount tubes (Becton Dickinson Biosciences). PE-Cy5–conjugated anti-mouse CD3, FITC-conjugated CD4, and PE-conjugated CD8 mAb were purchased from BD Pharmigen. Briefly, 10 μL blood were collected from the mouse tail and incubated with anti-CD3, CD4, CD8 mAbs for 30 min followed by 15 min incubation with erythrocytes lysis buffer. The CD3+ T cells were then counted on a FACSCalibur (Becton Dickinson).

The murine model of breast cancer metastases

The murine model for rat HER-2/neu–expressing breast cancer metastases has been described previously (21). Briefly, neu-N mice, ages 6 to 8 wk, were injected with 1.0 × 10⁵ NT2.5 cells suspended in 100 μL cold PBS via the left cardiac ventricle after anesthesia with a ketamine (90 mg/kg) and xylazine (10 mg/kg) mixture. Establishment and progression of metastases in multiple organs, including bones, liver, spleen, etc., was confirmed by histopathology. Necropsy was performed on every mouse to confirm successful LCV injection; mice that were not successfully injected could be identified by tumor confined to the chest wall. Unsuccessfully injected mice were excluded from the analysis.

Efficacy of ²¹³Bi-labeled antirat HER-2/neu 7.16.4

To evaluate therapeutic efficacy of ²¹³Bi-7.16.4 in treating metastases, 3 d after tumor cell inoculation by LCV injection, neu-N mice were given three consecutive daily i.v. injections of (a) ²¹³Bi-7.16.4 with a total dose of 120 μCi, n = 28; (b) ²¹³Bi-7.16.4 with a total dose of 90 μCi, n = 9; (c) ²¹³Bi-Rituximab with a total dose of 120 μCi, n = 10; (d) unlabeled 7.16.4 4 mg/kg, n = 8 or 0.4 mg/kg, n = 9; (e) no treatment control, n = 18. Mice were observed and weighed thrice per week and were euthanized if significant body weight loss (>15%) or hind limb paralysis appeared. For comparison, neu-N mice inoculated with subcutaneous rat HER-2/neu tumors were also treated with (a) 120 μCi ²¹³Bi-7.16.4, 3 d after tumor cell inoculation, n = 6; (b) 120 μCi ²¹³Bi-7.16.4, 7 d after tumor cell inoculation, n = 6; (c) unlabeled 7.16.4 0.4 mg/kg, n = 6; (d) 120 μCi ²¹³Bi-Rituximab, 3 d after tumor cell inoculation, n = 6; (e) no treatment control, n = 9. To investigate whether smaller tumor burden would improve the efficacy, neu-N mice were inoculated with 1 × 10⁴ NT2.5 cells through intracardiac ventricular injection and treated with 120 μCi ²¹³Bi-7.16.4, n = 14. Treatment of the metastatic tumors using two courses of 120 μCi ²¹³Bi-7.16.4 was also studied. A 10-d interval between the two courses was chosen based upon the kinetics of marrow suppression, and recovery after treatment with 120 μCi ²¹³Bi-7.16.4. Neu-N mice inoculated with NT2.5 cells were treated with 120 μCi ²¹³Bi-7.16.4 3 d after inoculation and treated again 10 d after the first treatment with (a) 120 μCi ²¹³Bi-7.16.4, n = 10; (b) 90 μCi ²¹³Bi-7.16.4, n = 8; (c) 4 mg/kg unlabeled 7.16.4, n = 8. The survival time for each group was plotted as a Kaplan-Meier survival curve.

Histopathology and in vivo metastatic cell kill

neu-N mice were inoculated with 1.0 × 10⁵ NT2.5 cells by LCV injection and treated with 120 μCi ²¹³Bi-7.16.4 3 wk postinoculation. Mice were sacrificed on the following day of the last therapeutic dose. Major organs with most frequent metastatic sites, including femurs, spine, liver, and spleen, were collected for histopathology analysis, and the TUNEL assay for apoptosis was performed using In situ Cell Death Detection kit, Fluorescein (Roche) according to the supplier’s instructions. The apoptotic cells were observed with a Carl Zeiss Axiovert 200 fluorescence microscope (Zeiss).

Dosimetry

Biodistribution data using ¹¹¹In-7.16.4 were converted to ²¹³Bi-7.16.4 biodistribution, assuming radioisotope properties do not affect antibody localization and clearance. An exponential expression was fitted to the time activity data, and the cumulative activity in each organ was calculated by analytic integration of the fitted expression. Absorbed doses were calculated as described in Sgouros et al. (19). The mean energy emitted per nuclear transition for α particles, Δ_α, and β electrons, Δ_e, is 1.33 × 10⁻¹² and 1.05 × 10⁻¹³ Gy kg/Bq s, respectively. Absorbed dose is thus calculated as D = Ã × Δ_α/M + Ã × Δ_e/M, wherein D is absorbed dose, Ã is the total number of disintegrations in an organ or tumor, M is the weight of the organ or tumor. Absorbed dose from ²¹³Bi on the surface of a single cell was calculated using Monte Carlo microdosimetry. Tracks of randomly emitted α-particles by the sources were followed assuming that α-particles transverse in straight line. Energy deposited along the track when the α-particle crosses a target cell nucleus is calculated by $z=\frac{1}{m}{\int }_{t1}^{t2}S(x)\text{d}x$, with continuous slowing down approximation. z is the specific energy, Gy; S(x) is the stopping power in liquid water, (MeV·cm⁻¹); t1 and t2 are the residual ranges with which the α-particle enters and leaves the cell nucleus, cm; and m is the mass of the target nucleus assuming unity density (g).

Statistical analysis

The statistical significance of differences between two groups was analyzed with the t test and Kaplan-Meier survival analysis using MedCalc (MedCalc Software). Differences with P values of <0.05 were considered statistically significant.

Results

Specific killing of ²¹³Bi-7.16.4 against single cells and tumor spheroids in vitro

The labeling efficiency of ²¹³Bi on 7.16.4-Chx-A″-DTPA was typically between 85% and 90%, and all ²¹³Bi-7.16.4 used in this study reached a purity of ~98% after size exclusion purification. The immunoreactivity of ²¹³Bi-7.16.4 was between 80% and 85%. ²¹³Bi-7.16.4 thus prepared was effective against both rat HER-2/neu–expressing NT2.5 single cells and tumor spheroids, and the efficacy was both dose and specific activity dependent (Fig. 1). In single cells, ²¹³Bi-7.16.4 was able to completely kill 5 × 10⁴ NT2.5 cells with a concentration of 10 μCi/mL and specific activity of 5.6 mCi/mg. At a specific activity of 2.8 mCi/mg, it was still possible to eradicate 100% of NT2.5 cells at a concentration of 20 μCi/mL. Further decreasing the specific activity to 0.7 mCi/mg made it impossible to eradicate all NT2.5 cells even with a radioactivity concentration of 40 μCi/mL (~70% killing). Lower activity concentration at the same specific activity also led to reduced efficacy. At a specific activity of 5.6 mCi/mg, only half of the NT2.5 cells was killed with a concentration of ~1.0 μCi/mL. Nonspecific ²¹³Bi-HuIgG was not able to kill NT2.5 cells at specific activity of 3.1 mCi/mg, neither could ²¹³Bi-7.16.4 kill rat HER-2/neu–negative NT4 cells. Analogous results were observed when using ²¹³Bi-7.16.4 to eradicate NT2.5 tumor spheroids. The specific, radiolabeled antibody, at a specific activity of 7.5 mCi/mg and concentrations of 5, 20, and 40 μCi/mL, was able to completely inhibit tumor spheroid growth, whereas at 20 μCi/mL (specific activity of 0.9 mCi/mg), it was able to significantly slow the growth of NT2.5 tumor spheroids (Fig. 1B). Nonspecific HuIgG, with concentration of 20 and 40 μCi/mL (specific activity of 6.2 mCi/mg), was able to slightly slow the growth of tumor spheroids, but not as much as the specific antibody.

Figure 1.

Specific killing of ²¹³Bi-labeled 7.16.4 in NT2.5 single cells (A) and multicellular tumor spheroids (B). A, specific activity of ²¹³Bi-7.16.4 was varied from 0.7, 2.8, to 5.6 mCi/mg. ²¹³Bi-labeled nonspecific human IgG (3.1 mCi/mg) and rat HER-2/neu–negative NT4 cells were used as controls. B, activity concentration of ²¹³Bi-7.16.4 was varied from 5, 20, to 40 μCi/mL at 7.5 mCi/mg and 20 μCi/mL at 0.9 mCi/mg. ²¹³Bi-labeled nonspecific human IgG with activity concentrations of 20 and 40 μCi/mL (6.2 mCi/mg) were used as controls.

Biodistribution of 7.16.4 in neu-N mice bearing subcutaneous NT2.5 tumors and metastases

The biodistribution of ¹¹¹In-7.16.4 in neu-N mice bearing a subcutaneous tumor and metastases after LCV injection is shown in Fig. 2A and B. In neu-N mice bearing subcutaneous tumors, ¹¹¹In-7.16.4 was cleared from blood with an effective half-life of ~26.3 hours. ¹¹¹In-7.16.4 continued to accumulate in the tumors during the 72-hour period, with %ID/g increasing up to ~38%. Liver, spleen, and kidney are the normal organs that reached the highest %ID/g. Because ²¹³Bi has a half-life of 45.6 minutes, the biodistribution of 7.16.4 in the initial 6 hours is critical in determining the dose that tumor and each normal organ will receive. Although subcutaneous tumors reached close to 38% ID/g of ¹¹¹In-7.16.4, the uptake at 0.5 and 2.0 hours was only 3.7% and 11.0% ID/g, respectively, much less than that of blood (39.3% and 23.4% ID/g) and liver (12.6% and 11.5% ID/g). However, in the mice bearing metastases 3 weeks after left ventricular injection, much faster accumulation of ¹¹¹In-7.16.4 in the metastatic tumors was observed with %ID/g of 16.0% and 21.5% at 0.5 and 2 hours, respectively, after tracer injection. Both liver and spleen have higher %ID/g (17.3% and 17.1% at 0.5 hour, respectively) due to the presence of extensive metastatic tumors in these organs. At 3 weeks after tumor inoculation, some of the mice started to develop signs of hind limb paralysis accompanied by enlarged bladder, which caused slow clearance of antibody from the body and subsequently higher normal organ activity in these mice.

Figure 2.

Biodistribution of ¹¹¹In-labeled 7.16.4 in rat HER-2/neu transgenic neu-N mice bearing subcutaneous NT2.5 tumors (A; six mice per group were sacrificed at 0.5, 2, 6, 24, and 72 h after antibody injection) and metastatic tumors (B; three mice per group were sacrificed at 0.5, 2, and 6 h after antibody injection). C, immune suppression of ²¹³Bi-7.16.4 at or below MTD of 120 μCi. Histopathology staining of normal marrow (left) and marrow of mice treated with ²¹³Bi-7.16.4 (right). D, suppression and recovery kinetics of CD3+ T lymphocyte after ²¹³Bi-7.16.4 treatments.

MTD

The MTD of ²¹³Bi-7.16.4 was found to be 120 μCi. At dose of 600 μCi, all mice died within 1 week, whereas mice from the 300, 225, and 150 μCi group died between day 10 and day 14 after ²¹³Bi-7.16.4 injection. Subsequent histopathology analysis showed gastrointestinal tract hemorrhage in the higher dose group (600 and 300 μCi), and all groups of mice showed dose-dependent marrow depletion and opportunistic infection after ²¹³Bi-7.16.4 injections (Fig. 2C). Histopathology of all surviving mice found no evidence of long-term toxicity. Monitoring of the CD3+ T-lymphocyte population after ²¹³Bi-7.16.4 injections revealed a rapid recovery of the CD3+ T-cell population from marrow depletion for all doses under MTD (Fig. 2D). At MTD of 120 μCi, 9 days after ²¹³Bi-7.16.4 injections, CD3+ T lymphocyte populations rebounded to ~64% of the original level. For lower doses, by day 5 after ²¹³Bi-7.16.4 injection, CD3+ lymphocytes had recovered to ~100%, 91%, and 71% of the original levels for 30, 60, and 90 μCi, respectively.

Radioimmunotherapy of breast cancer metastases using ²¹³Bi-7.16.4

Kaplan-Meier survival curves showed a dose-dependent improvement in survival of neu-N mice bearing breast cancer metastases treated with ²¹³Bi-7.16.4 3 days after LCV injection (Fig. 3A). Median survival time improved from 28 days for control group to 41 days for 120 μCi ²¹³Bi-7.16.4 treatment group (P < 0.0001) and to 36 days for 90 μCi ²¹³Bi-7.16.4 treatment group (P < 0.001). ²¹³Bi-labeled nonspecific mAb Rituximab (120 μCi) improved median survival to 31 days (P < 0.01) probably due to nonspecific irradiation of metastases from activity in the circulation. Treatment with unlabeled anti-rat HER-2/neu 7.16.4 slightly improved survival of the neu-N mice to 33 days for both 0.4 mg/kg (P = 0.017) and 4 mg/kg (P = 0.047) treatment groups compared with 29.5 days in the untreated group (Fig. 3B). Because the amount of 7.16.4 injected in the group of 120 μCi ²¹³Bi-7.16.4 is approximately equal to 0.4 mg/kg, the improvement in survival time of this group at 33 to 42 days (P = 0.0001) is primarily attributed to tumor cell killing from α-radiation. In the subcutaneous tumor model, 120 μCi ²¹³Bi-7.16.4 was able to slow the growth of NT2.5 tumor (Fig. 3C). At 18 days after tumor inoculation, the tumors in the treated group had an average volume of 465.6 mm³ compared with 1350.5 mm³ (P < 0.02) and 1076.2 mm³ (P < 0.05) in the control and 0.4 mg/kg unlabeled antibody groups, respectively. Tumors treated with 120 μCi ²¹³Bi-7.16.4 7 days after inoculation and tumors treated with 120 μCi ²¹³Bi-Rituximab also showed a trend toward slower growth with an average tumor volume at 900.1 mm³ and 889.9 mm³ at 18 and 20 days after tumor inoculation, although both of the results are not statistically significant compared with untreated control and unlabeled 7.16.4 treatment.

Figure 3.

Therapeutic efficacy of ²¹³Bi-7.16.4 to treat rat HER-2/neu micrometastases induced with LCV injection. Kaplan-Meier survival curve of neu-N transgenic mice treated at 3 d after LCV inoculation of 1 × 10⁵ NT2.5 cells with 120 μCi (n = 28), 90 μCi (n = 9) ²¹³Bi-labeled 7.16.4 and 120 μCi ²¹³Bi-Rituximab (n = 10; A) and unlabeled 7.16.4 at 0.4 mg/kg (n = 9) and 4 mg/kg (n = 8; B). C, growth of subcutaneous tumors after treatment with 120 μCi ²¹³Bi-7.16.4 at 3 or 7 d after tumor inoculation (n = 6), and unlabeled 7.16.4 at 0.4 mg/kg, 3 d after tumor inoculation (n = 6).

To examine the sensitivity of the therapeutic response to initial metastatic burden, neu-N mice were inoculated with 10-fold less NT2.5 cells (1 × 10⁴) and the therapeutic response was evaluated. The median survival improved to 41 days compared with the control group (P = 0.0002; Fig. 4A). However, the increase in survival relative to controls was not significantly different from that obtained in mice injected with 10⁵ cells (P = 0.12). The effect of repeated dosing was also examined. After a first course treatment of 120 μCi ²¹³Bi-7.16.4, all subsequent treatments with 120 μCi, 90 μCi ²¹³Bi-7.16.4, or 4 mg/kg unlabeled 7.16.4 failed to improve survival compared with the single treatment group (Fig. 4B).

Figure 4.

A, Kaplan-Meier survival curve of neu-N mice treated with 120 μCi μCi ²¹³Bi-7.16.4 (n = 14). Inoculation of metastases was induced by left cardiac injection of 1 × 10⁴ NT2.5 cells. B, Kaplan-Meier survival curve of neu-N mice with two courses of treatments after LCV inoculation of NT2.5 cells. Neu-N mice were initially treated with 120 μCi ²¹³Bi-7.16.4, and 10 d later they were treated again with either 120 μCi (n = 10), 90 μCi (n = 8) ²¹³Bi-7.16.4, or 4 mg/kg unlabeled 7.16.4 (n = 8).

Histopathology of in vivo cell kill

H&E staining of micrometastases treated with 120 μCi ²¹³Bi-7.16.4 3 weeks after LCV injection is shown in Fig. 5A. Compared with micrometastases in untreated control mice (Fig. 5B), the density of tumor cells was greatly reduced. For larger tumors, there seemed to be tumor cells surviving the cell killing at the center of the metastatic tumor (arrow). The distance from the edge of the tumor to theses cells is between 100 and 200 μm, consistent with dosimetric estimation of the maximum tumor size that a single course of treatment with ²¹³Bi-antibody can kill. TUNEL staining showed a higher fraction of apoptotic cells in micrometastases treated with ²¹³Bi-7.16.4 (Fig. 5C) compared with untreated control (Fig. 5D).

Figure 5.

A, histopathologic H&E staining of rat HER-2/neu–expressing micrometastases after treatment with 120 μCi ²¹³Bi-7.16.4. Scale bar, 50 μm. B, untreated control. Metastatic tumor (T) present in the vertebra (V) above an intervertebral disc (IVD) is invading and expanding, locally compressing the spinal cord (SC). Scale bar, 200 μm. TUNEL staining (corresponding to the areas in the square) of micrometastases treated with 120 μCi ²¹³Bi-7.16.4 (C) and untreated control (D). In the untreated control, only cells on the edge of the spinal cord are brightly stained.

Dosimetry

Estimated absorbed doses from 120 μCi ²¹³Bi-7.16.4 to tumor, each organ in the subcutaneous tumor model, and the dose to the metastatic tumor are shown in Table 1. In the subcutaneous model, the blood dose is 6.7 Gy at the MTD of 120 μCi ²¹³Bi-7.16.4, with 6.2 Gy coming from the α-radiation. The s.c. implanted tumor only received 2.6 Gy primarily because most of the ²¹³Bi decays have occurred before substantial accumulation of the antibody in the subcutaneous tumor. In the metastatic model, the absorbed dose to metastatic tumors was 4.8 Gy, consistent with the faster access and binding of 7.16.4.

Table 1.

Estimated tumor and normal organ absorbed doses per 120 μCi ²¹³Bi-7.16.4 in rat HER-2/neu–expressing subcutaneous and metastatic tumor model

	α Dose (Gy)	β Dose (Gy)
Blood	6.20	0.49
Heart	1.10	0.09
Lung	1.45	0.11
Liver	2.24	0.18
Spleen	1.97	0.16
Kidney	1.68	0.13
Stomach	0.38	0.03
Left femur	0.47	0.04
Right femur	0.48	0.04
Subcutaneous tumor	2.44	0.19
Micrometastasis*	4.43	0.35

^*Dose to micrometastases was calculated using the biodistribution data shown in Fig. 2B.

Discussion

Using a transgenic model that expresses the rat variant of HER-2/neu, we find that ²¹³Bi-labeled anti–HER-2/neu mAb is effective in prolonging the survival of mice with widespread mammary carcinoma metastases. Preclinical and clinical studies have shown the promise of α-particle radioimmunotherapy in non–Hodgkin’s lymphoma (27), leukemia (28, 29), disseminated peritoneal (14, 30), and lung diseases (16). In breast cancer, the most frequent sites of metastatic progression are the bone, lung, and liver; these are found to harbor metastases in over 50% of breast cancer patients; more widely disseminated disease is commonly observed in ~45% of the patients (31).

The improved efficacy of ²¹³Bi-labeled antibody relative to the unlabeled antibody suggests that the radiation doses delivered by ²¹³Bi provide immune-mediated and molecular signaling pathway independent cell killing. Using Monte Carlo microdosimetry, we calculated that, on average, 12 ²¹³Bi decays on the surface of NT2.5 cells (cell and nucleus radius of 7 and 4 μm) are needed for a single α-particle traversal of the cell nucleus. At specific activities of 5.6, 2.8, and 0.7 mCi/mg used in the in vitro cell killing studies, 32.6, 16.3, and 4.1 ²¹³Bi decays per cell, respectively, are estimated, assuming that the 1.6 × 10⁵ rat HER-2/neu sites per cell are occupied. The daily injected activity of 40 μCi ²¹³Bi-7.16.4 was at a specific activity of between 4 and 6 mCi/mg, yielding a maximum of between 23.3 and 34.9 ²¹³Bi per cell if all sites are immediately saturated. If, based on the biodistribution results, we assume a delay of one Bi-213 half-life, 11.6 to 17.5 Bi-213 decays per target cell are obtained by assuming that equivalent binding of antigen sites was achieved in vivo. The calculated absorbed dose to the metastatic tumors was 4.78 Gy. Assuming this is applicable to the 1 × 10⁵ cells initially inoculated, it can be estimated that ~12 Gy is needed (thrice more in specific activity) to achieve 100% tumor control probability using a typical D₀ value of 0.7 Gy (32). Depending upon the K_D, a high concentration of antibody is needed to establish the concentration gradient for optimal binding to antigen sites. The higher the tumor burden, the higher the antibody concentration required. It is therefore essential to design a ²¹³Bi delivery vehicle that is able to achieve both higher specific activity and better penetration with higher antigen affinity simultaneously.

Targeting an antigen with higher cell-surface expression can also overcome the limitation of specific activity. We have previously shown that ²¹³Bi-Trastuzumab showed better killing of tumor spheroids expressing high and medium levels of HER-2 compared with those with low level of expression (33). The expression of 1.6 × 10⁵ rat HER-2/neu is sufficient to be targeted by ²¹³Bi-7.16.4, although the inability of the radioimmunoconjugate to completely eradicate the NT2.5 cells suggest that rat HER-2/neu expression levels are not adequate. The interesting finding that treatment of a smaller tumor burden with 1 × 10⁴ cells does not significantly improve mouse survival seems to also attribute the lack of cure to lower specific activity or antigen expression rather than excessive tumor burden. A metastatic model with higher level of HER-2/neu expression is desirable to examine such a hypothesis.

The finding that sequential courses of treatments using ²¹³Bi-7.16.4 failed to further improve survival of these mice might be each, or a combination, of the following possibilities. First, tumor cells surviving the first course grow to a tumor burden over the 10-day marrow recovery period that is too high for the second treatment to be effective given the limitations on specific activity. As noted earlier, the 10-day period was dictated by marrow recovery kinetics; initiating the second treatment at 5 days after the first course caused severe toxicity in these mice (data not shown). Optimal timing of the second treatment, therefore, is determined by the relative repopulation rate of marrow versus tumor after the first course of treatment. Including stem cell transfer or low-dose interleukin 2 in between the two courses might help to accelerate immune reconstruction. Using a carrier that clears more quickly from the circulation to deliver the α-emitter in the second course might also be helpful in reducing marrow dose and thereby allowing an earlier second administration. This would have to be balanced against possible second organ toxicity, in particular, for low molecular weight carriers the kidneys would be at risk. The second possible explanation for the absence of efficacy in the two-course treatment relative to the single course may be a selection process whereby the tumor cells surviving the first course express lower levels of rat HER-2/neu. Because antigen expression level is the selective pressure for both courses of treatment, these cells will almost certainly survive the second course after 10 days of expansion. If this is indeed the case, a different kind of cell kill mechanism, such as chemotherapy, might be necessary to completely eradicate the remaining tumor cells.

In conclusion, we have showed that ²¹³Bi-labeled mAb can be effective in arresting widespread dissemination of breast cancer metastases. The metastatic model used closely resembles the clinical progression of breast cancer metastases and provides a background expression of the target antigen not typically modeled in xenograft studies.