Targeted alpha immunotherapy of CD20-positive B-cell lymphoma model: dosimetry estimate of 225Ac-DOTA-rituximab using 64Cu-DOTA-rituximab

The aim of this study was to evaluate the radiation dosimetry of alpha-emitter 225Ac-DOTA-rituximab using Monte Carlo simulation of 64Cu-DOTA-rituximab. CD20 expression was evaluated in lymphoma cell lines (Jurkat and Raji). DOTA-rituximab was conjugated and then chelated by 64Cu. Tumor xenograft models were established in BALB/c-nu mice. Animal PET/CT imaging was obtained after tail vein injection with and without a pre-dose of 2 mg of cold rituximab. Specific binding of tumors was evaluated by an organ distribution assay and autoradiography. CD20 expression in tumor tissues was evaluated by immunohistochemistry. The residence time was calculated using 64Cu-DOTA-rituximab PET/CT acquisition data using OLINDA/EXM software. 225Ac-DOTA-rituximab tumor dosimetry was performed using Monte Carlo simulation with 64Cu-DOTA-rituximab PET/CT images. Specific binding of Raji cells (CD20 positive) was 90 times that of Jurkat cells (CD20 negative) (p < 0.0001). Immunoreactivity was more than 75%. PET/CT imaging with 64Cu-DOTA-rituximab was specifically observed in tumors. The radioactivity of the tumor was much higher than that of other organs, and tumor uptake was related to CD20 expression. The predicted human dose for the administration of 64Cu-DOTA-rituximab was measured as the effective dose (1.07E-02 mSv/MBq). In the tumor region, equivalent doses of 225Ac-DOTA-rituximab (14 SvRBE5/MBq) were much higher (74-fold) than those of 64Cu-DOTA-rituximab (0.19 SvRBE5/MBq) (p < 0.01). Tumor dosimetry of 225Ac-DOTA-rituximab can be estimated via the Monte Carlo simulation of 64Cu-DOTA-rituximab. 225Ac-DOTA-rituximab can be employed for lymphoma as targeted alpha therapy.

Recently, targeted alpha therapy (TAT) was attempted and exhibited excellent results in certain diseases. TAT has been found to overcome resistance to beta-emitters because the linear energy transfer (LET) of beta-emitters is low (0.1-1.0 keV/μm), while that of alpha-emitters (50-230 keV/μm) is high [13]. In this regard, it is a good strategy to apply rituximab RIT with alpha-emitter radionuclides such as Actinum-225 ( 225 Ac).
Ac-225 decays into 221 Fr, and finally to the more stable 209 Bi, yielding the four alpha-emitters [14]. Some studies using the PSMA inhibitor, PSMA-617, demonstrated that 225 Ac-PSMA-617 contributes to the dramatic benefit in metastatic castration-resistant prostate cancer patients [15,16]. Other studies utilizing antibodies also showed the effectiveness and safety of TAT [17][18][19]. Accordingly, 225 Ac is a potent candidate for the radionuclide of RIT because its half-life of 225 Ac is 9.92 days [20] and rituximab has been known as an antibody with a long half-life, which is 2-7 days [21].
However, no study has evaluated 225 Ac-DOTArituximab in CD20-positive human B-cell lymphoma. In this study, we estimated the feasibility of RIT with 225 Ac-DOTA-rituximab in a CD20-positive B-cell lymphoma model through Monte Carlo simulation using 64 Cu-DOTA-rituximab. To perform this, we established a CD20-positive human lymphoma xenograft model to evaluate targeted imaging for non-invasive in vivo monitoring of lymphoma using 64 Cu-DOTA-rituximab PET and evaluated the correlation between CD20 expression and specific binding of 64 Cu-DOTA-rituximab in tumor tissue. For further studies, we estimated dosimetry using biodistribution results and quantitative analysis of 64 Cu-DOTArituximab PET imaging.

Experimental
Antibody and cell lines CD20-targeted rituximab (Mabthera) was purchased from Roche. Human lymphoma cell lines (Jurkat and Raji) were used. Both cells were obtained from American Type Culture Collection (ATCC) and maintained in RPMI 1640 with 10% fetal bovine serum (Gibco) containing 1% antibiotics (penicillin G, 100 unitsmL, and streptomycin 10 µgml; Gibco). Cells were incubated at 37 °C in a 5% CO 2 atmosphere.

Western blot
Protein concentrations were determined using a BCA protein assay kit (Thermo Scientific). The membranes were blocked for 1 h at room temperature and incubated with either anti-CD20 (#sc-58985) or β-actin (#A5441, Sigma-Aldrich) primary antibodies overnight at 4 °C. An enhanced chemical luminescence reagent (Roche) was used, and luminescent signals were measured with a ChemiDoc imaging system (Bio-Rad).

64
CuCl2 was produced at the Korea Institute of Radiological and Medical Sciences (Seoul, Korea) by 50 meV cyclotron irradiation [22]. The ratio of 64 CuCl 2 activity per DOTArituximab was determined to be 2 MBq per 1 mg. DOTArituximab conjugate was incubated with dried 64 CuCl2 in 1 mM sodium acetate buffer at 40 °C for 30 min. The radiolabelling yield was evaluated by instant thin-layer chromatography (iTLC) without additional purification. The stability of 64 Cu-DOTA-rituximab was analyzed through iTLC after incubation in human and mouse serum, and phosphate-buffered saline (PBS) at 37 °C for various times (1,2,6,24,48, and 60 h).

Cell binding assays and immunoreactivity
Cell binding with 64 Cu-DOTA-rituximab was performed using Jurkat and Raji cells. Both, Jurkat and Raji cell setups (5 × 10 5 cells/500 µL in tube) were incubated for 3 h in triplicate. Nonspecific binding (competitive inhibition) was performed by adding 11 µM of cold rituximab. After incubation, the cells were rinsed twice with cold PBS containing 1% bovine serum albumin (BSA). Cell bound radioactivity (count per minute) was evaluated using a γ-counter (Wizard 1480, Perkin-Elmer). Specific binding (%) was calculated using total binding and non-specific binding data. To evaluate immunoreactivity [23], 3.1 nM of 64 Cu-DOTA-rituximab was added to Raji cells diluted from 1 × 10 7 cells/tube to 0.016 × 10 7 cells/tube in 500 µL serum-free medium. The incubation time, washing, and calculation methods were the same as above. The immunoreactive fraction was determined by performing a linear regression analysis of the double inverse plot of total/bound activity versus normalized cell concentration. The immunoreactive fraction was then obtained from the inverse of the intercept on the plot. Data analysis was performed using GraphPad Prism software.

Animal experiments
Six-week-old female BALB/c-nude mice were obtained from NARA Bio, Inc. All animal experiments were approved by the Institutional Animal Care and Use Committee of KIRAMS (2018-0061). For establishment of lymphoma xenograft mouse models (n = 5; for PET/CT imaging, n = 3; for biodistribution), Raji cells (5 × 10 7 / in 200 µL PBS) were transplanted subcutaneously into the right thigh. Smallanimal PET imaging and biodistribution studies were performed when tumor sizes were > 0.5 cm in diameter.

Small-animal SPECT imaging
225 AC-DOTA-rituximab mouse image was performed by the Inveon multimodality scanner combines SPECT and CT components in a common gantry. Images were obtained at 2, 24, 48 and 72 h after intravenous injection (i.v.) of 0.37 MBq of 225 AC-DOTA-rituximab per mouse with 83 keV and 440 keV. CT data acquisitions were performed using a voltage of 80 kVp and then SPECT data acquisition were performed using 1-MME-3.0 pinhole collimator which has one 3 mm size pinhole. The acquired data were reconstructed by MAP3D algorithm (iterations 16 and subset 6) with PSF (Point spread function) mode and sensitivity mode. Image analysis was performed by AMIDE (A Medical Image Data Examiner).

Biodistribution study
Biodistribution studies were performed to evaluate the uptake of 64 Cu-DOTA-rituximab in tumor-bearing mice or normal mice. All mice were intravenously injected with 1.85 MBq of 64 Cu-DOTA-rituximab. Tumor-bearing mice (n = 3) were killed 48 h after i.v. injection. For dosimetry, normal mice (n = 4/group) were killed at 1, 2, 6, 24, 48, and 72 h after i.v. injection. Various organs containing tumors and blood samples were weighed, and the radioactivity was measured. The γ-counter data were represented by the percentage of injected activity per gram of tissue (%IA/g).

Autoradiography and Immunofluorescence
After the γ-counting of tumors from the biodistribution study, tumor tissues were frozen using optimal cutting temperature (OCT) compound at − 80 °C. A cryostat microtome (CM1800, Leica Instruments) was used for frozen sections of tumors tissue (15 μm depth in non-coating slide). The frozen sections were exposed on a film for 7 days in a deep freezer, and the film was scanned with BAS-5000 (Fujifilm). The image intensity of photostimulated luminescence was analyzed using Fujifilm Multi Gauge software, version 3.0 (Fujifilm).
For immunostaining, a cryostat microtome was used for frozen sections of the tumors (7 μm depth in coating slide). Briefly, the slides were rinsed using PBS for 10 min and fixed in 4% paraformaldehyde for 10 min. After being washed twice, the slides were incubated in Triton X-100 for 10 min to permeate the tumors tissue. Normal goat serum (1%) was used for non-specific binding and incubated with anti-CD20 (#sc-58985) primary antibody overnight at 4 °C. To visualize specific binding of the antibody, fluorescence-labeled secondary antimouse antibody (Bethyl lab) was added to the slides for 1 h. Immunofluorescence images were obtained using IN Cell Analyzer 2200 (GE Healthcare). Renal uptake of non-equilibrium progeny 213 Bi within 6.33 h and its contribution to kidney radiation dose was added by 213 Bi SPECT image with 440 keV energy window.

S value and absorbed dose calculation
The specific tumor and organ S value was acquired in Monte Carlo simulation from CT density information. CT density and PET radioactivity information were used as input data of Monte Carlo simulation. The S value of the tumor and organs was calculated using simulated organ specific dose map. The S value equation is as follows: where y i is the number of energy, E i is the energy per radiation, and m is the mass of the target region.
The absorbed dose ( D r T ) was calculated using the following formula: where A 0 is the initial injected radioactivity and ∼ A r j is the radiotracer residence time of a source organ r i . S(r T ← r s ) is the dose deposited in the target r T per unit of cumulated activity in source r s . The absorbed dose of each organ and tumors was calculated for each organ, including contributions from the self-dose with the cross-dose from other segmented regions [24].

Statistical analysis
All statistical analyses were performed using GraphPad Prism software. All data are evaluated as means ± standard deviation (SD) and are representative of at least two separate biological experiments performed in triplicate. Statistical significance between groups was compared using one-way ANOVA and unpaired Student's t test. P < 0.05 was considered statistically significant.

Specific CD20 targeting of 64 Cu-DOTA-rituximab
The mass difference between DOTA-rituximab and rituximab was about 1093 (m/z), and consequently we confirmed that the number of DOTA chelates per rituximab was approximately 2.8 ± 0.21 (Supplementary Table 1). The radiolabeling efficiency was over 95% ( Supplementary  Fig. 1a), and the stability of 64 Cu-DOTA-rituximab was over 90% in human and mouse serum, and in PBS (Supplementary Fig. 1b). CD20 expression was observed only in Raji cells (Fig. 1a). Specific binding of 64 Cu-DOTA-rituximab in Raji cells was 90-fold higher than that in Jurkat cells (p < 0.0001). Between non-specific binding and total binding in Raji cells, the binding effect of 64 Cu-DOTA-rituximab was found to be significantly higher in total binding in Raji cells, by as much as 95-fold (p < 0.01) (Fig. 1b). Immunoreactivity was observed at 76%, which was similar to that observed in a study by Natarajan et al. [25] (Fig. 1c and d).

Cu-DOTA-rituximab PET/CT imaging in lymphoma models
To evaluate the in vivo PET imaging of lymphoma using 64 Cu-DOTA-rituximab, we established CD20-positive lymphoma xenograft models using Raji cells. Confirmation of the formation of the lymphoma model was done using 18 F-FDG. Localized 18 F-FDG uptake was observed at the xenograft site of the lymphoma model ( Supplementary Fig. 2). In the 64 Cu-DOTA-rituximab PET study, focal 18 F-FDG uptake was found in CD20 positive tumors at 24 h and 48 h (not shown at 2 h), but there was less uptake in the cold rituximab pre-injection group (Fig. 2a). These uptake differences were remarkable at 48 h maximum intensity projection image from 64 Cu-DOTA-rituximab PET/CT (Fig. 2b). In biodistribution results at 48 h, the distribution of 64 Cu-DOTA-rituximab in tumor groups revealed similar results, compared to the normal mice group without xenograft (Fig. 2c). Autoradiography of the extracted tumor indicated a high correlation between autoradiography and immunohistochemistry with CD20 positive areas of the tumor (Fig. 2d). Consequently, we demonstrated that 64 Cu-DOTA-rituximab specifically targeted CD20 positive tumors.

Discussion
The current study aimed to evaluate radiation dosimetry for 225 Ac-DOTA-rituximab in CD20-positive B-cell lymphoma xenograft models. We demonstrated that 64 Cu-DOTA-rituximab specifically targeted CD20-positive lymphoma xenograft in terms of cell binding assay, biodistribution, autoradiography, and tissue staining as well as in vivo PET/CT imaging. In addition, we calculated the equivalent dose for 225 Ac-DOTA-rituximab TAT using Monte Carlo simulation of 64 Cu-DOTA-rituximab PET imaging.
The results of the present study correspond with those of a previous study, in which the biodistribution of Cu-DOTA-rituximab binding. a Western blot analysis of CD20. β-actin was used as an internal control. CD20 expression was observed in Raji cells. b Cell binding assay of 64 Cu-DOTA-rituximab. Specific binding (SB) of Raji cells was 90-fold higher than that of Jurkat cells (p < 0.0001). c, d Immunoreactivity assay (Lindmo assay) using Raji cells. d Was analyzed using GraphPad software from (c). Immunoreactivity was 76% 227 Th-DOTA-rituximab was similar to that of 64 Cu-DOTArituximab [27][28][29]. However, the equivalent dose of 227 Th-DOTA-rituximab (50 Sv RBE5 /MBq) [27] was lower than those of 64 Cu-DOTA-rituximab (0.19 Sv RBE5 /MBq), and 225 Ac -DOTA-rituximab (14 Sv RBE5 /MBq) because they are different radionuclides. Furthermore, these values were calculated using different methodologies. We maintain that this study's equivalent dose is more accurate than the radiation dose of 227 Th-DOTA-rituximab because the absorbed dose was estimated using the area under the curve of the time activity curve, [27] while this study analyzed the equivalent dose using the tumor S value of the Monte Carlo simulation. 64 Cu-DOTA-rituximab was previously reported by Natarajan et al. as a radiopharmaceutical agent for diagnostic PET imaging [25]. This result was first presented from synthesis to dosimetry for human patients using a transgenic mouse model expressing huCD20 in the kidney. Biodistribution study with (n = 3) and without (n = 4) lymphoma at 48 h. Accumulation in the tumor was the second highest. d Autoradiography ( 64 Cu-DOTA-rituximab) and immunostaining (CD20) images showed a correlation However, mouse xenograft models bearing human lymphoma cells were not studied using 64 Cu-DOTA-rituximab. Therefore, to obtain a tumor-targeting effect with organ distribution pattern of 64 Cu-DOTA-rituximab, we developed a CD20-positive lymphoma xenograft model. In our study, 64 Cu-DOTA-rituximab was also specifically bound to only CD20 positive lymphoma cells (Fig. 1). Furthermore, specific uptake in tumor was observed through PET/ CT imaging, and the distribution pattern of 64 Cu-DOTArituximab in tumor models was evaluated (Fig. 2). Additionally, we confirmed biodistribution and residence time in normal mice ( Supplementary Fig. 3). In a normal biodistribution study, the half-life of 64 Cu-DOTA-rituximab in mouse blood was estimated to be approximately 6-21 h [30]. Consequently, these results suggest that 64 Cu-DOTArituximab specifically targets CD20-positive human B-cell lymphoma.
Generally, the pharmacokinetics of therapeutic radiopharmaceuticals are analyzed using diagnostic radionuclides. For example, some studies prefer to use 124 I-mIBG PET/ CT images for 131 I-mIBG-targeted radionuclide therapy as a means for better treatment planning [31]. In addition, the radiation dosimetry of 213 Bi-PSMA-617 was estimated using 68 Ga-PSMA-617 imaging in patients for clinical application of alpha radionuclide therapy [32]. As a positron emitter, 64 Cu is also an attractive diagnostic radionuclide for PET imaging [33]. Using 64 Cu, Woo et al. calculated the absorbed dose of 67 Cu-DOTA/NOTA-trastuzumab to evaluate the therapeutic effect of HER2-positive breast cancer [26]. Based on this method, we estimated the radiation dosimetry of 225 Ac-DOTA-rituximab using 64 Cu-DOTA-rituximab PET/ CT imaging from CD20-positive human B-cell lymphoma models (Fig. 3). The OLINDA/EXM software is widely used for the analysis of normal organ dosimetry. However, this software has been used for human digitalised phantoms, and its limitation is that it cannot calculate abnormal regions such as the tumor region. For this reason, we performed a Monte Carlo simulation to calculate the absorbed dose in the tumor regions. Our results demonstrated that 225 Ac-DOTArituximab showed a significantly larger equivalent dose than 64 Cu-DOTA-rituximab (Fig. 3c). Consequently, a low dose of 225 Ac-DOTA-rituximab can be applied to provide targeted therapy by releasing high LET alpha-emitters.
High LET alpha particles are promising candidates for killing target cells. Among them, 225 Ac plays an important role in TAT and can be used efficiently and safely [34]. Some studies have shown that high energy within a short range induces DNA damage at least 100-fold higher than beta particles [35], and DNA double-strand breakage (DSB) has therapeutic effects [36]. With regard to radiation dose, another study revealed that the radiation dose (Gy/Bq) of 225 Ac was 900 times that of lutetium-177 ( 177 Lu) and 14 times that of radium-223 ( 223 Ra) through Monte Carlo simulation [37]. This result could explain why equivalent doses of 225 Ac-DOTA-rituximab in major organs were higher than those of 64 Cu-DOTA-rituximab and tumor ( Fig. 3a and b). Similarly, TAT using 225 Ac-DOTATOC triggers much higher γH2AX-foci formation from DSB than 177 Lu-DOTATOC [37]. Therefore, we can infer that 25 Ac-DOTA-rituximab could be a more effective treatment option than 131 I-rituximab for NHL patients. We have performed a clinical trial of RIT for NHL patients using 131 I-rituximab for 16 years [8][9][10][11][12]. RIT demonstrated excellent results, but some refractory patients showed resistance to RIT. We expect that it is necessary to treat these refractory patients using more powerful RIT agents. We calculated the probable radiation dose to the tumor and found that 225 Ac-DOTA-rituximab revealed a higher radiation dose to the tumor than 131 I-rituximab.
Our study has several limitations. First, we analyzed only the 64 Cu-DOTA-rituximab results in terms of biological experiments. To evaluate the therapeutic effect of 225 Ac-DOTA-rituximab more precisely, we need further studies using 225 Ac-DOTA-rituximab radiopharmaceutical using an in vivo tumor model. In particular, the evaluation of time activity curve for late reacting organs such as kidneys should have been considered on the therapeutic radiopharmaceutical dosimetry. Long time reacting radioactive of non-equilibrium 213Bi in kidney should be considered and the absorbed dose of 213Bi was 1.48 times as much as that of 225Ac. Radioactive of non-equilibrium 213 Bi within 6 h was 0.81%, long time reacting activation in kindney should be considered 4.8 times [38]. Furthermore, clinical trials are warranted to verify the adverse effects and therapeutic effects of 225 Ac-DOTA-rituximab. Fortunately, some studies demonstrated that biodistribution patterns using trastuzumab were significantly related between 64 Cu-DOTA and 225 Ac -DOTA groups in murine models [39,40]. Watabe et al. [41] also observed that 64 Cu/ 225 Ac-FAPI-04 in murine tumor model had a similar trend according to physiological accumulation. These results could be supported our hypothesis that appropriateness of 64 Cu as a surrogate marker of 225 Ac. Second, while it has been considered that the lymphoma transgenic mice model will be more clinically relevant than the lymphoma xenograft model, we still preferred to utilize the CD20-positive B-cell lymphoma xenograft model due to the difficulty in evaluating the radiation dosimetry using the small tumor from transgenic mice models. Hopefully, we succeeded in labeling the 225 Ac-DOTA-rituximab group (data not shown). In addition, comparison of the therapeutic effect of 131 I-rituximab and 225 Ac-DOTA-rituximab is ultimately required to confirm the cytotoxicity.

Conclusion
Tumor dosimetry of 225 Ac-DOTA-rituximab can be estimated via the Monte Carlo simulation of 64 Cu-DOTArituximab. 225 Ac-DOTA-rituximab can be employed for lymphoma as targeted alpha therapy.