Stabilization of an 211At-Labeled Antibody with Sodium Ascorbate.

211At, an α-particle emitter, has recently attracted attention for radioimmunotherapy of intractable cancers. However, our sodium dodecyl sulfate polyacrylamide gel electrophoresis and flow cytometry analyses revealed that 211At-labeled immunoconjugates are easily disrupted. Luminol assay revealed that reactive oxygen species generated from radiolysis of water caused the disruption of 211At-labeled immunoconjugates. To retain their functions, we explored methods to protect 211At-immunoconjugates from oxidation and enhance their stability. Among several other reducing agents, sodium ascorbate most safely and successfully protected 211At-labeled trastuzumab from oxidative stress and retained the stability of the 211At-labeled antibody and its cytotoxicity against antigen-expressing cells for several days.

Introduction

Radioimmunotherapy (RIT) is defined as targeted radionuclide therapy using radiolabeled antibodies. RIT has expanded the applications of radiotherapy from focusing on local tumors to targeting scattered tumors, such as distant metastases and disseminated lesions. As for β-particles, two kinds of radiopharmaceuticals, which target the CD20 molecule on the surface of lymphoma cells, 90Y-labeled rituximab and 131I-labeled rituximab, have already shown clinical benefits against CD20-positive non-Hodgkin B-cell lymphoma.[1] Compared with β-particles, α-particles have more potent linear energy transfer (LET) and a shorter path range. Owing to their high-energy emission within a short path length, α-particles can selectively eliminate target cells with minimal radiation damage to the surrounding normal tissues when delivered selectively to tumor tissues. These properties render α-particles an attractive tool for treating intractable tumors.[2−5]211At is an α-emitter with a short half-life (7.2 h) and does not yield cytotoxic daughter isotopes during its decay; the first branch (58.2%) decays through electron capture to 211Po (half-life: 516 ms), which decays through α-particle emission to 207Bi (half-life: 31.55 y). The second branch (41.8%) directly decays through α-particle emission to 207Bi. 207Bi results in stable 207Pb via its metastable states after the electron capture. To harness the short path of α-particles and potent LET, 211At must be precisely delivered to the target. For delivering 211At to the desired regions, 211At-labeled small molecules, including uridine analogues,[6] benzylguanidine (a norepinephrine analogue),[7] biotin analogues,[8] a phenylalanine derivative,[9] and bisphosphonate complexes, have been previously designed.[10] Furthermore, 211At-labeled antibodies were reportedly tested to deliver highly cytotoxic 211At to the target in preliminary investigations and preclinical situations.[11−18] The 211At-labeled anti-Tenascin mAb 81C6 was administered locally to 18 patients with recurrent malignant brain tumors,[11] and the 211At-labeled MX35 F(ab′)2, targeting the sodium-dependent phosphate transport protein 2B, was intraperitoneally administered to nine patients with ovarian cancer.[12] In addition to the promising results obtained by these studies, the procedure for the production of 211At-labeled antibodies under the current Good Manufacturing Practices (cGMP) toward clinical application was recently reported.[19] To maximize the functions of 211At-labeled antibodies, the quality of the conjugates must be validated. Therefore, in the present study, we evaluated the quality of an 211At-conjugated antibody. Several reports on the disruption of radioimmunoconjugates by reactive oxygen species (ROS), determined by functional analyses, have been published.[20−24] In this paper, we clearly show the disruption of the 211At-conjugated antibody by ROS generated from water radiolysis through various approaches including sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE). The ROS concentration was measured using a luminol assay system, and the degradation was suppressed by quenching ROS by addition of sodium ascorbate (SA), a safe (or non-cytotoxic) reducing agent.

Results and Discussion

Evaluation of the Effects of SA Concentration on the Stability of 211At-Labeled Trastuzumab by SDS–PAGE, Autoradiography, and Flow Cytometry Assay

211At-labeled trastuzumab was prepared according to the previously described procedures.[13,25,26] The effects of different concentrations of SA on the stability of 211At-labeled trastuzumab were evaluated using SDS–PAGE, and autoradiography was performed on the day of 211At labeling and on the following day (Figure ). Even on the day of 211At labeling, 211At-labeled trastuzumab was slightly disrupted in the absence of SA (Figure a,b). On the following day, 211At-labeled trastuzumab was completely disrupted in the presence of less than 6 × 10–4 mg/mL SA (Figure c,d). These results indicate that the astatinated antibodies were disrupted in a time-dependent manner. In contrast, 211At-labeled trastuzumab in the presence of more than 6 × 10–2 mg/mL SA was still stable on the following day, thus indicating the concentration-dependent protective effects of SA.[27]

Figure 1

Sodium dodecyl sulfate–polyacrylamide gel (SDS–PAGE) and autoradiography for 211At-labeled trastuzumab. SA concentrations (mg/mL) are indicated. SDS–PAGE and autoradiography were performed to determine the effects of SA on the stability of the immunoconjugate. (a) Polyacrylamide gel on day 0 and (b) autoradiograph of 211At-labeled trastuzumab on day 0 and (c) polyacrylamide gel after 1 d and (d) 211At-labeled trastuzumab after 1 d.

The binding activity (binding activity = MI – MINC) of 211At-labeled trastuzumab to high (SK-BR-3) and low (MCF-7) human epidermal growth factor receptor 2 (HER2)-expressing cell lines was investigated using flow cytometry, where MI is defined as median intensities of samples incubated with trastuzumab, Sn-trastuzumab, or 211At-trastuzumab and MINC is median intensities of negative control, which is the sample incubated with only the secondary antibody.

Binding activities of trastuzumab, N-[2-(maleimido)ethyl]-3-(trimethylstannyl)benzamide-conjugated trastuzumab (Sn-trastuzumab), and astatinated trastuzumab (211At-trastuzumab) in phosphate-buffered saline (PBS) for SK-BR-3 cells were 2.01 × 105, 1.98 × 105, and 3.91 × 103, respectively (Figure ). Binding activities of 211At-trastuzumab in 6 × 10–4, 6 × 10–2, and 6 mg/mL SA to SK-BR-3 cells were 1.29 × 104, 1.33 × 105, and 1.59 × 105, respectively. In the presence of more than 6 × 10–2 mg/mL SA, the binding activity of 211At-trastuzumab was maintained. The binding affinities of trastuzumab and functionalized trastuzumab to MCF-7 cells were weak because of their low HER2 expression levels.

Figure 2

Flow cytometry analysis of the binding activity of 211At-labeled trastuzumab to human epidermal growth factor receptor 2 (HER2)-expressing cells. 211At-labeled trastuzumab with or without SA at the indicated concentrations facilitated binding with breast cancer cell lines having different HER2 expression levels. SK-BR-3: high HER2 expression; MCF-7: low HER2 expression. Sn-trastuzumab is N-[2-(maleimido)ethyl]-3-(trimethylstannyl)benzamide-conjugated trastuzumab. Negative control is a sample incubated with only the secondary antibody. The flow cytometry analysis was performed 6 d after 211At labeling.

In Vitro Cytotoxicity Evaluation of 211At-Labeled Trastuzumab

The cytotoxicity of 211At-labeled trastuzumab was evaluated using the WST-8 assay. The cytotoxic effects of 211At-labeled trastuzumab on cancer cells depended on HER2 expression levels on the cell surface. Astatinated trastuzumab killed SK-BR-3 cells more efficiently than free 211At. However, the differences in the cytocidal effects on MCF-7 cells between free 211At and the immunoconjugates were minor. Based on the protective effects on the binding activities of 211At-trastuzumab, SA contributed to the cytocidal effects of the immunoconjugates in a dose-dependent manner (Figure ). 211At-trastuzumab in 6 × 10–2 and 6 mg/mL SA exerted greater cytocidal effects on SK-BR-3 cells than 211At-trastuzumab in PBS and 6 × 10–4 mg/mL SA.

Figure 3

Cytotoxic effects of 211At-labeled trastuzumab in breast cancer cell lines. The cytotoxic effects of 211At-labeled trastuzumab with and without SA on breast cancer cell lines with different expression levels of human epidermal growth factor receptor 2 (HER2) were determined using the WST-8 cell count assay. SK-BR-3: high HER2 expression; MCF-7: low HER2 expression. N = 4. Data are presented as mean ± standard deviation (SD) values.

Detection of ROS Using Luminol Assay

We speculated that antibody damage was caused by ROS generated through the interaction of water molecules and α-particles emitted from 211At. ROS can be detected using chemiluminescent luminol assay. The luminol assay can measure the global levels of ROS, such as H2O2, O2–, and OH•, with high sensitivity under physiological conditions and can be sensitized by the addition of horseradish peroxidase.[28]

Thus, we performed the luminol assay to quantify the ROS in 211At and 211At-conjugated antibody solutions (adjusted to approximately the same radioactivity) with different SA concentrations (Figure ). On addition of SA at a low concentration (6 × 10–4 mg/mL), the same level of luminol reactivity was detected as in the SA-free sample in both 211At and 211At-labeled trastuzumab. Although low levels of ROS were detected in 6 × 10–2 mg/mL SA 211At solution, the levels were below the detection limit in 211At-labeled antibody solution. The reason for this difference is unclear at this moment. The intensity of chemiluminescence in the samples with 6 mg/mL SA was lower than the detection limit. These results suggest that SA greatly contributes to the reduction in the amount of ROS.

Figure 4

Detection of ROS using luminol assay. (a) SA was added at different concentrations to 211At in PBS. The lower (boxed) graph is an expansion of the 6 mg/mL SA addition protocol. (b) SA was added at different concentrations to 211At-labeled trastuzumab in PBS. The lower (boxed) graphs are expansions of 6 × 10–2 and 6 mg/mL SA addition protocols. RLU = relative luminescence units. Data are presented as mean ± SD values.

Scope of Reducing Agents

The scope of reducing agents, in addition to SA, potentially applicable for reducing ROS levels was investigated. We used agents with other mechanisms of reduction, namely, l-cysteine, sodium hydrosulfite, and maltose. The reducing ability of l-cysteine and sodium hydrosulfite is because of the redox potential of the sulfur atom, whereas that of maltose is inherent in its hemiacetal structure. The concentrations of these reducing agents were set to 6 × 10–2 mg/mL based on the SA threshold. We performed a luminol assay to assess the ROS-quenching abilities of various reducing agents in solutions of free 211At and 211At-labeled trastuzumab (Figure ). The order of the reducing abilities of the agents in the 211At-labeled trastuzumab solution was SA > l-cysteine ≥ sodium hydrosulfite > maltose (Figure b). We found that 6 × 10–2 mg/mL l-cysteine did not completely quench ROS in the 211At-labeled antibody solution, although it efficiently quenched ROS in the solution of free 211At (Figure a). At present, the reason for the difference in ROS concentration in the solutions of free 211At and 211At-labeled antibody in the presence of 6 × 10–2 mg/mL l-cysteine is unclear.

Figure 5

Quenching potential of different reducing agents for ROS in 211At or 211At-labeled trastuzumab solutions, as assessed using the luminol assay. (a) Reducing agents (6 × 10–2 mg/mL) were added to 211At in PBS. (b) Reducing agents (6 × 10–2 mg/mL) were added to 211At-labeled trastuzumab in PBS.

SDS–PAGE revealed that maltose or sodium hydrosulfite could not adequately protect the astatinated antibodies (Figure ). However, SA and l-cysteine protected the immunoconjugates from oxidative stress. These results are in agreement with the ROS concentration measured by luminol assay.

Figure 6

(a) Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS–PAGE) of Sn- and 211At-labeled trastuzumab (IgG) in the presence of various reducing agents. Concentration of each reducing agent is 6 × 10–2 mg/mL. SDS–PAGE analysis was performed 4 d after 211At labeling. (b) Flow cytometry analysis of the binding activity of 211At-labeled trastuzumab to different breast cancer cell lines in the presence of various reducing agents. Cell lines with high (SK-BR-3) and low (MCF-7) human epidermal growth factor receptor 2 (HER2) expression levels were treated with 211At-labeled trastuzumab in the presence of various reducing agents. Flow cytometry analysis was performed 6 d after 211At labeling. SA: sodium ascorbate; Cys: l-cysteine; SHS: sodium hydrosulfite; Mal: maltose. Concentrations of reducing agents are 6 × 10–2 mg/mL. RLU = relative luminescence units.

Accordingly, 211At-labeled trastuzumab solution containing l-cysteine or SA displayed high binding activity to SK-BR-3 and MCF-7 cells (Figure ), depending on the surface HER2 expression levels, and these results were comparable to those obtained for the naked and linker-attached trastuzumab before astatination. However, compared with the 211At-labeled trastuzumab protected with SA or l-cysteine, the astatinated antibodies with maltose or sodium hydrosulfite did not retain their binding activity. The binding activities of trastuzumab and Sn-trastuzumab to SK-BR-3 cells were 1.50 × 105 and 1.42 × 105, respectively. Binding of 211At-trastuzumab in PBS, 211At-trastuzumab in 6 × 10–2 mg/mL sodium hydrosulfite, and 211At-trastuzumab in 6 × 10–2 mg/mL maltose to SK-BR-3 cells became weak, with activities of 1.84 × 103, 7.77 × 102, and 7.57 × 103, respectively. 211At-trastuzumab in 6 × 10–2 mg/mL SA and 211At-trastuzumab in 6 × 10–2 mg/mL l-cysteine retained binding activities of 6.61 × 104 and 7.77 × 104, respectively. In MCF-7 cells, the binding activities of trastuzumab, Sn-trastuzumab, 211At-trastuzumab in PBS, 211At-trastuzumab in 6 × 10–2 mg/mL SA, 211At-trastuzumab in 6 × 10–2 mg/mL l-cysteine, 211At-trastuzumab in 6 × 10–2 mg/mL sodium hydrosulfite, and 211At-trastuzumab in 6 × 10–2 mg/mL maltose were 8.11 × 103, 7.87 × 103, 2.96 × 102, 6.78 × 103, 7.28 × 103, 1.33 × 102, and 1.08 × 103, respectively. Regarding cytotoxicity, 211At-labeled trastuzumab with l-cysteine had the same potency as 211At-labeled trastuzumab with SA, and the astatinated antibodies with these protectants exerted greater cytotoxic effects than the immunoconjugate without the protectant (Figure ). However, the astatinated antibodies with maltose or sodium hydrosulfite displayed antitumor activities that were less than or similar to those of the immunoconjugate without the protectant.

Figure 7

Cytotoxic effects of 211At-labeled trastuzumab in breast cancer cell lines. The cytotoxic effects of 211At-labeled trastuzumab in the presence of reducing agents,SA and l-cysteine, and on breast cancer cell lines with different expression levels of human epidermal growth factor receptor 2 (HER2) were determined using the WST-8 assay. SK-BR-3: high HER2 expression; MCF-7: low HER2 expression. SA: sodium ascorbate; Cys: l-cysteine; SHS: sodium hydrosulfite; Mal: maltose. N = 4. Data are presented as mean ± SD values.

Conclusions

In this study, we longitudinally investigated the quality of conjugates after labeling of antibodies with 211At, a promising α-emitter applicable for targeted alpha therapy. Our results indicate that the radioimmunoconjugates were severely degraded within 1 day of labeling with 211At. Although these devastating effects of α-particle emitters on macromolecules, such as proteins, have been reported previously, the mechanism underlying the destruction of macromolecules including antibodies in vivo has not been clarified as of now. Here, we particularly focused on the high LET of 211At. High-LET particles can strongly induce radiolysis of exposed materials. When the water is irradiated, numerous types of radicals, primarily ROS in water solutions, are produced. In this study, using luminol assay, SDS–PAGE, flow cytometry, and cytotoxicity assays, we clearly show that 211At-labeled trastuzumab was degraded by ROS generated from the radiolysis of water.

The mechanism underlying antibody denaturation upon 211At labeling provides insights into the protection against damage caused by radioactive conjugated antibodies.

Certain reducing agents or radial scavengers can suppress ROS upon 211At conjugation.

Our assays were performed 4–6 days after 211At conjugation, with and without the reducing agents. 211At-conjugated trastuzumab was stable in the presence of SA for several days. SDS–PAGE, flow cytometry, and cytotoxicity assays revealed that the minimum concentration of SA for protection of 211At-labeled trastuzumab is 6 × 10–2 mg/mL. Protected 211At-labeled trastuzumab maintained its binding activity and potent antitumor effects on antigen-expressing cells. Although numerous 211At-labeled antibodies have been reported so far, the stability of 211At-labeled immunoglobulin G has remained largely unclear, except in studies using large quantities of 211At-conjugated samples.[19] Overall, our results clearly indicate that protection from oxidative stress is required for 211At immunoconjugation.

The stability and biodistribution of 211At-labeled antibodies are important considerations for α-particle conjugation; hence, the structure of the conjugated antibody needs to be maintained, as observed upon the addition of SA. In the presence of 6 × 10–2 mg/mL SA, the 211At-labeled antibody was stable for several days.

The selection of reducing agents is important in RIT using 211At-labeled antibodies. In this case, reducing agents should have an efficient ROS-quenching ability and less toxicity. The results of our experiments indicate that SA and l-cysteine are good candidates for use as reducing agents. Because SA is frequently added as a stabilizing agent in clinically approved formulations, SA addition is highly practicable. The toxicity of reducing agents is also important when clinical application is considered. LD50 (rat) of SA is 11,900 mg/kg, whereas that of l-cysteine is 1890 mg/kg.[29] Considering that the LD50 of SA is extremely high, SA is currently the most effective and safest candidate as a ROS scavenger in the view of toxicity as well.