Therapeutic efficacy of α-radioimmunotherapy with different activity levels of the ²¹³Bi-labeled monoclonal antibody MX35 in an ovarian cancer model

The ²¹³Bi was produced by a ²²⁵Ac/²¹³Bi generator as described previously [14, 15] and was eluted from the generator according to the standard protocol provided by Institute for Transuranium Elements (ITU) in Karlsruhe, Germany. In short, 600 μL of a 0.1 M HCl/0.1 M NaI solution was pumped through the generator column to elute the ²¹³Bi. Subsequently, the ²¹³Bi solution was pH-adjusted by addition of 4 M sodium acetate and a 20% L-ascorbic acid solution (the ascorbic acid additionally protects the antibody conjugate from radiolysis) [16].

MX35, a murine IgG₁ mAb, was used in the experiments. The MX35 mAb was developed at the Sloan-Kettering Cancer Center (New York, NY, USA) [17, 18] and was produced from hybridoma cells obtained from the Ludwig Institute for Cancer Research, Zürich, Switzerland. It recognizes the transport protein NaPi2b, which is homogeneously expressed in approximately 90% of human epithelial ovarian cancer cells and to only a small degree in normal tissues [19], making it suitable for epithelial ovarian cancer therapy.

The ovarian carcinoma cell line OVCAR-3 (American Type Culture Collection, Rocksville, MD, USA), was used in all in vitro and in vivo experiments. The cells were cultured in T-75 culture flasks in RPMI 1640 medium supplemented with 10% fetal calf serum, 1% L-glutamine, and 1% penicillin-streptomycin and were grown in a humidified atmosphere of 95% air/5% CO₂ at 37 °C.

Female immunodeficient BALB/c (nu/nu) mice (Charles River Laboratories International, Freiburg, Germany), 5–6 weeks of age, were used in the in vivo experiments. The housing of the animals was as previously described [13]. The animal study was performed according to the guidelines from the ethical committee and the legislations for animal research in Sweden and with approval from the Committee on the Ethics of Animal Experiments of the University of Gothenburg, Sweden.

The antibody-chelator conjugation was performed essentially as described previously [13], at room temperature (RT), overnight, in 0.2 M carbonate buffer, pH ~8.5. The chelator [(R)-2-amino-3-(4-isothiocyanatophenyl)propyl]-trans-(S,S)-cyclohexane-1,2-diamine-pentaacetic acid (CHX-A′′-DTPA) in dimethyl sulfoxide (0.018 M) was added to 2–4 mg/mL of mAb in 15 × molar excess. Subsequently, the buffer was changed to 0.1 M citrate, pH 5.5, using a NAP-5 column (GE Healthcare, Buckinghamshire, UK). For long-term storage, the buffer of the antibody conjugate was changed to phosphate buffered saline (PBS). The equipment was essentially kept metal-free, and the buffers included were made metal-free by incubation with Chelex 100 resin (Bio-Rad Laboratories, Hercules, CA, USA) as described previously [13]. The number of available chelators attached after labeling was estimated to be two per antibody according to evaluation by an arsenazo III spectrophotometric assay [20].

The ²¹³Bi-labeling of the MX35 conjugate was performed and evaluated as previously described, [13] but with some modifications. A total of 0.1 mg of the antibody conjugate was added to the pH-adjusted ²¹³Bi eluate, and the reaction was continued for 5 min. After quenching with 10 μl of 1.5 mg/mL 2-[Bis[2-bis(carboxymethyl)amino] ethyl]amino]acetic acid (DTPA), the labeled product was purified using a NAP-10 column (GE Healthcare) and eluted with PBS.

Prior to the in vivo experiments, the immunoreactivity of the radiolabeled MX35 was evaluated in vitro using suspensions of live OVCAR-3 cells as previously described [13]. Briefly, 10 ng of the labeled antibodies were added to duplicates of 0.5 mL of serially diluted cell suspensions (maximum 5 × 10⁶ cells/mL). The samples were incubated for 3 h at RT with vigorous agitation, centrifuged for 5 min, and washed twice with 1 mL of PBS. Following the last wash and centrifugation, the cell pellets were measured in a γ counter (Wallac, Finland) and compared with reference solutions of radiolabeled MX35. The fraction of bound activity over the total applied activity was calculated, as well as the immunoreactive fraction according to the Lindmo assay [21].

Therapeutic efficacy was investigated for two different activity levels of ²¹³Bi-labeled MX35. Sixty mice were inoculated i.p. with 1 × 10⁷ OVCAR-3 cells. Two weeks after cell inoculation, 40 mice were injected i.p. with ²¹³Bi-MX35: 20 were given 9 MBq/mL, equal to 30 μg of ²¹³Bi-MX35 in 1 mL of PBS, and 20 were given 3 MBq/mL, equal to 10 μg of ²¹³Bi-MX35 in 1 mL of PBS. A control group of mice received unlabeled MX35, 30 μg, in 1 mL PBS.

According to the ethical license, the mice were weighed prior to treatment and then every 7 to 10 days during treatment to monitor for any ascites discomfort. To monitor for acute myelotoxicity, white blood cell (WBC) counts and platelet counts in the blood from the tail vein of 5 mice from each group were analyzed at 6 and 14 days after treatment. The samples were analyzed in a microcell counter (Sysmex F-820; Toa Medical Electronics Co., Ltd., Kobe, Japan). The animals were sacrificed 8 weeks after therapy, and the abdominal cavity was opened to investigate any presence of ascites, microscopic, and/or macroscopic tumors. Microscopic tumor was defined as tumor sizes ranging from single cells to clusters of million cells not detected by the naked eye at autopsy, whereas macroscopic tumors are visible and detectable at autopsy, often more than 1 mm with or without ascites. To investigate occurrence of microscopic tumors, tissues from the abdominal wall, mesentery, and spleen were taken from all animals for paraffin sectioning and hematoxylin and eosin (H&E) staining. Additional samples were taken from suspected lesions at other locations. Between 10 and 30 sections at 50 μm distance were processed for microscopy. To support the H&E-stained slides in microscopic tumor identification, immunohistochemistry (IHC) for the MX35 antigen was also performed in representative cases. Both the animal dissection and the histological analysis were blinded, i.e., performed without any knowledge of the treatment received.

Analyses were performed to evaluate possible statistical significance between the two treated groups and the control group. Calculations were performed for the two-sided alternative hypothesis using Fisher’s exact test.

In the mouse model, the dose-limiting organ could possibly be the bone marrow. Previously presented biodistribution data for i.p. injected ²¹³Bi-MX35 were used for bone marrow dosimetry [12, 13]. The time-integrated activity per unit mass, Ã/m (Bq s kg⁻¹) was calculated from the area under the activity concentration curve for blood. The absorbed dose D was calculated as:where Δ is the mean energy of the α particles per nuclear transformation for ²¹³Bi (1.33 × 10⁻¹² J), and Φ is the absorbed fraction of the α particles (assumed to be 1). A previously obtained bone marrow-to-blood ratio of 0.58 (for ²¹¹At-labeled MX35) [22] was then applied to estimate the resulting absorbed dose to the bone marrow of i.p. injected ²¹³Bi-MX35.

The absorbed dose to the peritoneum was calculated as 50% of the equilibrium dose of the injected fluid. The antibody concentration in the i.p. fluid was assumed to be equal to the antibody concentration in the injected solution during the time of the ²¹³Bi decay.

Biokinetic modeling and microtumor dosimetry was performed as previously described by Palm et al. [12]. The modeling was adjusted for mouse kinetics, for the specific activity achieved in the study and the activity concentrations used, i.e., 3 and 9 MBq/mL. The number of antigen per cell was set to 700 000. Microdosimetry was performed for single cells and clusters with radii 9, 30, and 50 μm.

The radiochemical yield after the ²¹³Bi-labeling was 53 ± 12% (mean value ± standard deviation; without correction for decay), and the radiochemical purity was 91 ± 6.0%. The cell-binding assay showed that the fraction of bound activity after 3 h was 0.83 for the highest concentration of cells in the cell suspensions. The immunoreactive fraction of the radiolabeled mAbs in vitro, i.e., the calculated fraction of conjugated antibodies able to bind to the cells in a situation of infinite antigen access, was plotted and calculated to be 0.91. The specific activity at the time of injection was on average 45.6 GBq/μmol, which corresponds approximately a radionuclide:antibody ratio of 1:3000.

The tumor-free fractions (TFFs) of the groups treated with 3 or 9 MBq/mL of ²¹³Bi-labeled MX35 were 0.55 and 0.78, respectively, (Table 1). No animals in the groups receiving ²¹³Bi-MX35 developed ascites. However, four animals in the control group receiving cold MX35 had obvious ascites production. The control group also had the largest fraction of animals with macroscopic tumor incidence. The overall TFF of the control group was 0.15 (Table 1). The samples taken for microscopy evaluation from the peritoneal cavity showed a wide range of tumor progression, from complete absence of visual tumor cells to large tumors. Examples of visualizations of macroscopic and microscopic tumors by H&E staining and IHC are shown in Fig. 1.

The statistical analysis performed using Fisher’s exact test showed that the difference in TFF between the mouse group receiving 3 MBq/mL of ²¹³Bi-MX35 and the control group and between the group receiving 9 MBq/mL of ²¹³Bi-MX35, and the control group was statistically significant (p = 0.02) and (p = 0.0002) respectively. The difference between the group receiving 3 MBq of ²¹³Bi-MX35 and the group receiving 9 MBq/mL of ²¹³Bi-MX35 was, however, not statistically significant (p = 0.18).

In the group treated with 3 MBq/mL of ²¹³Bi-MX35, the WBC counts were on average 4.3 × 10⁹/L 6 days after treatment, as shown in Fig. 2. In the 9 MBq/mL group, the WBC counts were slightly lower with an average of 3.9 × 10⁹/L. However, the WBC counts of the control animals were 3.4 × 10⁹/L, i.e., lower than in both of the treated groups. Fourteen days after treatment, the WBC counts had increased by 22% on average in all animals. The increase in WBC counts was lowest in the control group (8%). Thus, no bone marrow toxicity could be observed from the WBC levels of the animals in any of the treated groups.

Neither could we demonstrate significant hematological toxicity from the platelet counts after any of the treatments, as shown in Fig. 3. The group receiving 9 MBq/mL of ²¹³Bi-MX35 had the lowest platelet counts 6 days after treatment (mean value = 777 × 10⁹/L). However, the variation between the animals was large, and the mean value for the group receiving 9 MBq/mL was similar to the value of the control group 14 days after treatment.

When using previous biodistribution data [13] to calculate the absorbed dose in the mice receiving 3 MBq/mL of ²¹³Bi-MX35, the absorbed dose to blood was approximately 1.3 Gy, resulting in an estimated absorbed dose to the bone marrow of 0.8 Gy. Accordingly, the mice administered 9 MBq/mL of ²¹³Bi-MX35 received an absorbed dose to blood of approximately 3.9 Gy, yielding an estimated dose to the bone marrow of 2.3 Gy. The absorbed dose to the kidneys was estimated to be 0.45 and 1.35 Gy for the 3 and 9 MBq/mL level, respectively. The injected activity concentrations of ²¹³Bi-MX35 at 3 and 9 MBq/mL would result in absorbed doses to a human peritoneum of 7.9 and 23.6 Gy, respectively, according to the dosimetric calculations. The calculated dose to microtumors from both non-specific and specific, i.e., cell-bound, irradiation are shown in Table 2.