Myelotoxicity and RBE of ²¹¹At-Conjugated Monoclonal Antibodies Compared with ^99mTc-Conjugated Monoclonal Antibodies and ⁶⁰Co Irradiation in Nude Mice

Radionuclides and External-Beam Irradiation

Two radionuclides, ²¹¹At and ^99mTc, were used for the internal irradiation. ²¹¹At was produced by the ²⁰⁹Bi(α,2n)²¹¹At reaction in a cyclotron (Scanditronix MC32 NI; PET Center, Rigshospitalet, Köpenhamn, Denmark) and isolated from the irradiated targets by a dry distillation procedure described by Lindegren et al. (7). ^99mTc was obtained from a Tc/Mo-generator (Mallinckrodt). The activity eluted in 1–2 mL was about 20–60 GBq. The injected activities of ²¹¹At (0.09–1.3 MBq) and ^99mTc (250–1,300 MBq) were measured using an ionization chamber (CRC-15R; Capintec). The external irradiation was accomplished using a ⁶⁰Co source (TEM Mobaltron therapy unit; Crawley, U.K.) with an activity of 18.87 TBq on March 19, 2003.

mAbs

Two mAbs were used, MX35 and 88BV59. MX35 is a murine IgG1-class mAb, developed and characterized at the Sloan-Kettering Institute, directed toward a cell-surface glycoprotein antigen of about 95 kDa on NIH:OVCAR-3 tumor cells (8). The antigen is expressed strongly and homogeneously on 90% of human epithelial ovarian cancers (9). Dr. Kenneth Lloyd (Memorial Sloan-Kettering Cancer Center, New York) provided MX35. 88BV59 is a human IgG1-class mAb that recognizes a polypeptide complex antigen of 35–43 kDa identified in human colorectal adenocarcinoma (10) and on ovarian tumor cells. The antigen specificity of the antibodies was of no importance in this study since the mice had no tumors and the binding to normal tissue was low.

Conjugation of Antibodies

²¹¹At was coupled to antibodies according to a protocol previously developed by Lindegren et al. (11). Briefly, ²¹¹At (15–30 MBq) in chloroform was added to a reaction vial and the solvent was evaporated to dryness. One nanomole of N-succinimidyl-3-(trimethylstannyl)benzoate and 0.5 nmol of N-iodosuccinimide in methanol/1% acetic acid were added to give a final volume of 12 μL. After a 15-min reaction, unreacted astatine was reduced using sodium metabisulfite and the organic fraction of the labeling solvent was evaporated under a gentle stream of nitrogen. The antibody was added in 0.5 mol/L borate buffer (pH 8.5) to the crude labeling mixture and conjugation was allowed to proceed for 30 min.

The ^99mTc conjugation was achieved using a modified SnCl conjugation method, in which ^99mTc-sodium pertechnetate is reduced from the oxidation state of +7 to +4 and the disulfide bonds on the protein are reduced to sulfhydryl groups to allow the binding of ^99mTc to the protein. For MX35, the phosphate-buffered saline (PBS) buffer was replaced by physiologic saline using a Sephadex G-25 PD-10 column blocked with PBS/1% bovine serum albumin (BSA) and rinsed with physiologic saline. SnCl, dissolved in reversed-osmosis water, was added to the antibody and the mixture was incubated for 45 min at 37°C, followed directly by the addition of ^99mTc-sodium pertechnetate and further incubation for 5 min at 37°C. Finally, diethylenetriamine pentaacetic acid (DTPA) was added, dissolved in physiologic saline, to bind unreacted ^99mTc. ^99mTc-DTPA was then separated from conjugated antibody on a Sephadex G-25 PD-10 column, blocked with PBS/1% BSA, and rinsed with physiologic saline. For 88BV59, the method began by dissolving SnCl in distilled water, which was added to the freeze-dried antibody. Thereafter, the procedure was the same as for MX35.

Animals

Nude female BALB/c nu/nu mice (age, 8- to 12-wk old; weight, 18.9–22.7 g) were used throughout the studies (Mollegaard & Bomholtgaard). The animals were kept at 22°C and 50–60% humidity with a light/dark cycle of 12 h. They were given standard food and water ad libitum. The animals were allowed to adapt for at least 1 wk after purchase. During the long-term and the 5-d toxicity studies, the mice were kept in separate lead-shielded cages for the first 24 h to avoid the effect of cross-irradiation between the animals from ^99mTc. With a separation of ∼30 cm between the cages and a lead shielding of 1–2 mm between them, the cross-irradiation was considered to be negligible. After 24 h, 10 animals were kept in each cage. The Ethics Committee of Göteborg University approved all experiments.

Study Groups

Figure 1 shows an experimental flow chart indicating the times at which the different experiments in studies A, B, and C were performed.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Experimental flow chart indicates times at which the different experiments in studies A, B, and C were performed.

Dosimetry and RBE Analysis

The mean energy deposited in various bone marrow sites of the mouse was simulated for the α-particles emitted from ²¹¹At and the electrons emitted from ^99mTc. The geometries and distribution of the different bone marrow sites was simulated according to a model by Muthuswamy et al. (12): vertebra, spheres with a diameter of 0.2 mm (21% of bone marrow); skull, spheres with a diameter of 0.17 mm (12%); limb bones, cylinders with diameters of 0.9 mm and lengths of 20 mm (20%); and ribs + clavicula + sternum + pelvis, slab with a thickness of 0.3 mm (47%).

In-house Monte Carlo software was used to simulate the energy deposited in each geometry. The random generator from Matlab (Mathworks) was used to randomly locate the nuclides and choose the directions of radiation emission. The energy deposited within the marrow sites was calculated by integrating over the track length within the marrow. The scaled point kernel of Berger (13) was used for electrons, and the stopping power values from ICRU Report 49 (14) were used for the α-particles. All integration was performed by numeric summation.

To be able to calculate the absorbed dose to the bone marrow, the ratio between the activity concentrations in bone and blood as well as bone marrow and blood was defined: Eq. 1 Eq. 2

The absorbed dose (D) to the bone marrow was then calculated by assuming a baseline cumulated activity concentration in bone marrow equal to the cumulated blood activity concentration multiplied with BBLR, assuming an absorbed fraction equal to 1. For the excess over baseline, the cumulated activity concentration in bone marrow was set to the cumulated blood activity concentration multiplied with BMBLR − BBLR. The absorbed fraction (φ) was in this case calculated using the Monte Carlo software for the different bone marrow geometries. Hence, the absorbed dose to the bone marrow, originating from α-particles and electrons, can be written as: Eq. 3 where C̃_blood is the cumulated blood activity concentration and Δ is the sum of the radiation energy emitted per nuclear transition.

The contribution of the photons to the absorbed dose to the bone marrow was included for ^99mTc and was calculated based on the whole-body retention curves. In this case, the cumulated whole-body activity concentration was determined and the absorbed fraction of the photons was calculated by approximating the shape of the mouse to a sphere with a radius of r = (3m/4πρ_m)^1/3, where m is the mass of the mouse and ρ_m is the mean tissue density. It was further assumed that both the activity and the bone marrow were homogeneously distributed throughout this sphere. The absorbed fraction was calculated as: Eq. 4 where μ_en is the energy absorption coefficient, equal to 0.027 and 0.70 cm⁻¹ for the dominant photon energies 140 keV and 18.3 keV, respectively, and r is the radius of the sphere (15).

For equitoxic effects, the in vivo RBE can thus be defined as a quotient between the absorbed dose arising from electrons and photons from ^99mTc and that due to α-particles from ²¹¹At: Eq. 5 where φ^211_At(α) and φ^99m_Tc(CE) is the absorbed fraction for α-particles from ²¹¹At and electrons from ^99mTc in the bone marrow, respectively. φ^99m_Tc(γ) is the absorbed fraction of the photons from ^99mTc in the whole animal. Δ^211_At(α) is the sum of the radiation energy emitted per nuclear transition for α-particles from ²¹¹At. Δ^99m_Tc(CE) and Δ^99m_Tc(γ) is the sum for the radiation emitted per nuclear transition for electrons and photons from ^99mTc, respectively. The integrals represent the cumulated activity concentrations in the blood and the whole animal for ²¹¹At and ^99mTc, respectively. The contribution to the absorbed dose to the bone marrow from photons and electrons originating from ²¹¹At was negligible.

Study A: Long-Term Study

The nadir of the WBC counts was reached at days 4–7 after injection for both the ²¹¹At- and ^99mTc-conjugated mAb 88BV59 (Fig. 2). The nadir of the WBC counts for the ⁶⁰Co-irradiated animals was reached at days 3–5 after irradiation (data not shown). Day 5 was chosen as a point of reference for measuring the myelotoxicity. The WBC counts seemed to be completely recovered 14 d after injection. No difference in recovery time was observed between ²¹¹At-, ^99mTc-88BV59, or ⁶⁰Co-irradiated animals. The variations in WBCs, RBCs, PLTs, and HGB (Fig. 2) during the observation time for the control groups were insignificant, and the mean values ± SEM were 5.7 ± 0.3 (×10⁹/L), 6.7 ± 0.4 (×10¹²/L), 681 ± 29 (×10⁹/L) and 130 ± 4.9 (g/L), respectively.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

WBC counts, RBC counts, PLT counts, and HGB as function of time after injection during the long-term study, when animals were injected intraperitoneally with ²¹¹At-conjugated (▴) and ^99mTc-conjugated (▪) mAb 88BV59 (n = 17 and 30, respectively), giving an absorbed dose to bone marrow of approximately 0.3 and 1.3 Gy, respectively. The nadir of WBC counts was reached at days 4–7. Results from control animals (unlabeled antibody 88BV59 [n = 20] and PBS [n = 30]) are also plotted (•). Since no significant (P > 0.05) difference between unlabeled antibody and PBS was observed, the 2 control groups were put together. Day 5 was chosen as the reference point when comparing myelotoxicity of different combinations of nuclide and antibody (Fig. 5). Values are means ± SEM.

Study B: Biodistribution

The organ-to-blood ratios for the various combinations of radioimmunocomplexes did not differ dramatically (Table 1). For ^99mTc-88BV59, an elevated organ-to-blood ratio of 1.9 ± 0.24 was found 6 h after injection in the kidneys. For ²¹¹At-88BV59, a ratio of 8.93 ± 1.56 was found in the thyroid 18 h after injection. When investigating the BMBLR at 3, 6, and 18 h after injection, we observed no significant variation with time for any of the radioimmunocomplexes. The difference between the various combinations of radionuclide and mAb was small. The mean value of the BMBLR (at 3, 6, and 18 h) for ²¹¹At-mAbs (MX35 and 88BV59) and ^99mTc-mAbs (MX35 and 88BV59) was 0.58 ± 0.06 and 0.56 ± 0.06, respectively (Table 1). The weight of the bone marrow samples varied between 4 and 10 mg.

Study C: Five-Day Toxicity and RBE

The whole-body retention curves decline exponentially with a final biologic half-life of ∼18 h for both the ²¹¹At- and ^99mTc-mAbs (MX35 and 88BV59) (Fig. 3). The curves in Figure 3 also include animals injected intravenously, as no variation was seen on the whole-body retention for these animals compared with those injected intraperitoneally.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Whole-body retention for intraperitoneally (i.p.) or intravenously (i.v.) injected ²¹¹At-MX35 (▴, n = 14 [i.p.] and n = 4 [i.v.]), ²¹¹At-88BV59 (▵, n = 16 [i.p.] and n =4 [i.v.]), ^99mTc-MX35 (▪, n = 16 [i.p.] and n = 4 [i.v.]), and ^99mTc-88BV59 (□, n = 10 [i.p.]) at 0, 1, 3, 6, 12, and 18 h after injection. Whole-body retention curves were used to estimate cumulated activity (i.e., AUC) and, hence, the photon contribution to absorbed dose from ^99mTc for each individual animal. Values are means ± SEM.

For the intraperitoneally injected radioimmunocomplexes, the maximum activity concentration in the blood varied between 14 %IA/g and 21 %IA/g 3 h after injection (Fig. 4). For the intravenous injections, a value of 42–56 %IA/g was observed after 1 h.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Blood activity kinetics for intraperitoneally injected ²¹¹At-MX35 (▪, n = 14), ²¹¹At-88BV59 (▴, n = 16), ^99mTc-MX35 (▪, n = 16), and ^99mTc-88BV59 (–, n = 10) as well as for intravenously injected ²¹¹At-MX35 (□, n = 4), ²¹¹At-88BV59 (×, n = 4), and ^99mTc-MX35 (Δ, n = 4) at 1, 3, 6, 12, and 18 h after injection. Individual curves for each animal were used to estimate cumulated activity (i.e., AUC) and, hence, the contribution from α-particles or electrons, originating from decays of ²¹¹At and ^99mTc, respectively, to absorbed dose to bone marrow for each individual animal. Values are means ± SEM.

The absorbed fractions (φ) for α-particles and electrons in the bone marrow were calculated to 0.88 and 0.75, respectively. The absorbed fraction of the photons from ^99mTc was calculated to 0.033 for 140 keV and 0.52 for 18.3 keV.

For the ²¹¹At-mAbs (MX35 and 88BV59), the absorbed doses to the bone marrow varied between 0.01 and 0.60 Gy, causing a suppression of the WBC counts of 1%–90%. For the ^99mTc-mAbs, the absorbed doses to the bone marrow varied between 0.39 and 1.92 Gy, causing a suppression of the WBC counts of 23%–89%. Two animals that received 2 Gy to the bone marrow from irradiation of ^99mTc did not survive until day 5. In Figure 5, individual suppression in WBC counts is given as a function of the individually calculated absorbed doses to the bone marrow. The data points for 2 groups (n = 10 each) with no individual absorbed doses are shown as mean WBC count suppression and mean absorbed dose.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Relative suppression of WBC counts as function of absorbed dose to bone marrow in nude mice when injected with ²¹¹At-conjugated (▴, n = 38) and ^99mTc-conjugated (▪, n = 20) mAbs or when irradiated externally with a ⁶⁰Co source (□, n = 19). The y-axis is logarithmic, where 1 represents no suppression and 0.1 represents 90% suppression of WBC counts. Solid lines represent exponential fits to ²¹¹At and ^99mTc data. Dashed lines flanking each full line represent 95% confidence interval. In vivo RBE is defined as quotient of absorbed dose from ^99mTc and that from ²¹¹At for equal toxic effect—that is, for equal degrees of suppression of WBC counts. The 2 data points with error bars originate from preliminary experiments with ^99mTc-88BV59, where cumulated blood activity concentrations and suppression of WBC counts were investigated in separate animals (▪, n = 10).

For the externally irradiated animals, the absorbed dose to the bone marrow was 1.4, 1.9, and 2.4 Gy, causing a suppression of the WBC counts of 47%–90% (Fig. 5). Exponential fits were made to the ²¹¹At, ^99mTc, and external irradiation data and are shown in Figure 5 (together with the 95% confidence intervals).

The in vivo RBE for myelotoxicity for the suppression of WBC counts for the α-particles from ²¹¹At, calculated from the equations describing the exponential fits in Figure 5, was determined to be 3.4 ± 0.6 in relation to the electrons from ^99mTc and 5.0 ± 0.9 in relation to external irradiation using ⁶⁰Co.