Alpha imaging confirmed efficient targeting of CD45-positive cells after astatine-211 (²¹¹At)-radioimmunotherapy for hematopoietic cell transplantation

Abstract

Alpha-radioimmunotherapy targeting CD45 may substitute for total body irradiation in hematopoietic cell transplantation (HCT) preparative regimens for lymphoma. Our goal was to optimize the anti-CD45 monoclonal antibody (MAb; CA12.10C12) protein dose for astatine-211 (²¹¹At)-radioimmunotherapy, extending the analysis to include intra-organ ²¹¹At activity distribution and α-imaging-based small-scale dosimetry, along with immunohistochemical staining.

Methods

Eight normal dogs were injected with either 0.75 (n=5) or 1.00 mg/kg (n=3) of ²¹¹At-B10-CA12.10C12 (11.5–27.6 MBq/kg). Two were euthanized and necropsied 19–22 hours post injection (p.i.), and six received autologous HCT three days after ²¹¹At-radioimmunotherapy, following lymph node and bone marrow biopsies at 2–4 and/or 19 hours p.i. Blood was sampled to study toxicity and clearance; CD45 targeting was evaluated by flow cytometry. ²¹¹At localization and small-scale dosimetry were assessed using two α-imaging systems: α-camera and iQID.

Results

Uptake of ²¹¹At was highest in spleen (0.31–0.61 %IA/g), lymph nodes (0.02–0.16 %IA/g), liver (0.11–0.12 %IA/g), and marrow (0.06–0.08 %IA/g). Lymphocytes in blood and marrow were efficiently targeted using either MAb dose. Lymph nodes remained unsaturated, but displayed targeted ²¹¹At localization in T lymphocyte-rich areas. Absorbed doses to blood, marrow, and lymph nodes were estimated at 3.1, 2.4, and 3.4 Gy/166 MBq, respectively. All transplanted dogs experienced transient hepatic toxicity. Liver enzyme levels were temporarily elevated in 5 of 6 dogs; 1 treated with 1.00 mg MAb/kg developed ascites and was euthanized 136 days after HCT.

Conclusion

²¹¹At-anti-CD45 radioimmunotherapy with 0.75 mg MAb/kg efficiently targeted blood and marrow without severe toxicity. Dosimetry calculations and observed radiation-induced effects indicated that sufficient ²¹¹At-B10-CA12.10C12 localization was achieved for efficient conditioning for HCT.

Introduction

Hematopoietic cell transplantation (HCT) is a curative treatment option for many malignant and non-malignant hematologic disorders. A key element is the HCT conditioning regimen, which prevents relapse after remission by reducing the underlying disease burden and facilitates allogeneic engraftment through host immunosuppression. In recent years, efforts have been made to increase the absorbed dose to target sites while reducing the radiation-induced adverse effects, through substitution of radiolabeled monoclonal antibodies (MAb) for standard pre-treatment total body γ-irradiation. The radioimmunotherapy regimens have generally utilized β-emitters such as iodine-131, which are suboptimal for killing hematopoietic cells due to their long particle range and low linear energy transfer. Conversely, short-ranged α-emitters deposit very high energies over a few cell diameters (40–100 μm), making α-radioimmunotherapy an appealing option for improved conditioning specificity and efficacy (1).

Our group has previously explored α-radioimmunotherapy preparative regimens for HCT in preclinical models, concluding that conditioning with either bismuth-213 (²¹³Bi)- or astatine-211 (²¹¹At)-anti-CD45 radioimmunoconjugates was efficacious, without significant non-hematopoietic toxicity (2–5). ²¹¹At was chosen for subsequent α-radioimmunotherapy studies because of its superior myelosuppression, lower cost, greater availability and clinically less challenging half-life (7.2 hours for ²¹¹At versus 46 minutes for ²¹³Bi).

CD45 is a cell surface tyrosine phosphatase that regulates T cell and B cell activation and maturation. It is broadly expressed (~200,000 copies/cell) on all hematopoietic cells except platelets and erythrocytes, regardless of malignant transformation (6). Our overall goal is to achieve efficient targeting and pharmacokinetics with minimal normal tissue toxicity using ²¹¹At-anti-CD45-radioimmunoconjugates to replace total body irradiation or high-dose chemotherapy conditioning in HCT preparative regimens for hematopoietic diseases. In this study we aimed to optimize the anti-CD45 MAb protein dose in a canine model, evaluating the results through blood counts, plasma chemistry, flow cytometry, immunohistochemical staining, α-imaging, and small-scale α-dosimetry.

Material and Methods

Antibodies and Antibody Conjugates

Anti-canine CD45 MAb CA12.10C12 (2,3,5,7) was conjugated with closo-decaborate(2-) to produce CA12.10C12-B10 for ²¹¹At-labeling, as previously described (8). Flow cytometry was performed as described by Sandmaier et al. (2). MAbs were produced and purified at the Biologics Production Facility at Fred Hutchinson Cancer Research Center (FHCRC; Seattle, WA).

Radioactivity

²¹¹At was produced using a Scanditronix MC-50 cyclotron at the University of Washington, Seattle, WA, and isolated through a “wet chemistry” method (8). Subsequent ²¹¹At-labeling of CA12.10C12-B10 was performed using the procedure of Chen et al. (5). All radioactive materials were handled according to approved protocols at FHCRC and the University of Washington.

Dogs

Dogs were raised at FHCRC and given standard immunizations. The experimental protocol was approved by the FHCRC Institutional Animal Care and Use Committee, and the study executed according to principles outlined in the Guide for the Care and Use of Laboratory Animals (Institute of Laboratory Animal Resources) (9).

Flow Cytometry

Surface bound anti-CD45 was detected through flow cytometry of cell suspensions prepared from samples of blood, lymph nodes, and bone marrow aspirates, as previously described (2,5). Antigen saturation was determined by comparing the mean fluorescence intensity (MFI) of cells incubated with either FITC-conjugated goat anti-mouse F(ab′)₂ (FMF) or FITC-conjugated anti-CD45 MAb (CD45F).

Biodistribution, Pharmacokinetics, and Toxicity

Two anti-CD45 MAb dose levels (0.75 mg/kg, n=5; 1.00 mg/kg, n=3) were evaluated with regards to in vivo distribution, pharmacokinetics, and normal organ toxicity. The starting level was based on previous canine studies, which showed that 0.50 mg/kg insufficiently saturated available CD45 antigens (2,3,5). Unlabeled CA12.10C12 (0.05 mg/kg) was injected 30–60 minutes before ²¹¹At-MAb infusion to prevent non-specific Fc receptor binding. Clearance of MAb and ²¹¹At was assessed using blood collected from 5 minutes before to 22 hours after radioimmunoconjugate injection, using an enzyme-linked immunosorbent assay (ELISA), as previously described (2), and by radioactivity measurements, respectively. Eight dogs were infused with CA12.10C12-B10 labeled with 14.6–36.7 MBq ²¹¹At/mg MAb (Table 1); two were euthanized and necropsied without HCT 19–22 hours post injection (p.i.). Harvested tissues were weighed and measured for radioactivity, and the results expressed as the percentage of injected radioactivity per gram (%IA/g) after corrections for background and decay. Six dogs received autologous HCT three days after the ²¹¹At-MAb infusion. Marrow was aspirated from humeri and femora, at least two weeks before radioimmunotherapy treatment, and then processed and stored as previously described (10,11). Biopsies of lymph nodes and bone marrow were taken at an early (2–4 hours) and/or a late (19–22 hours) time point after ²¹¹At-MAb infusion. Samples were split for flow analysis, α-imaging, immunohistochemistry, and radioactivity measurement.

TABLE 1.

Dogs Treated with ²¹¹At-Anti-CD45 Radioimmunotherapy

Dog ID	Age (mo)	Weight (kg)	MAb dose* (mg/kg)	211At activity (MBq/kg)	Specific 211At activity (MBq/mg MAb)	Autologous transplant	Sample time (h p.i.)	Transfusions (d post HCT)	Survival (wk)
H573	12	7.7	0.75	14.6	19.5	No	22	—	—
H585	9	9.4	0.75	27.6	36.7	No	19	—	—
H543	20	12.8	0.75	11.5	15.3	Yes	19	7, 11, 19	48†
H689	7	8.7	0.75	13.9	18.5	Yes	2, 19	7	32†
H522	22	11.8	0.75	13.9	18.6	Yes	2	8	46†
H629	8	9.5	1.00	14.6	14.6	Yes	2, 19	6, 8, 12	20‡
H638	10	12.3	1.00	18.4	18.4	Yes	4, 19	6, 10, 14, 16, 19, 22	32†
H632	10	8.4	1.00	18.7	18.7	Yes	2, 19	7, 9	32†

^*Including 0.05 mg/kg of unlabeled MAb

^†End of study

^‡Severe ascites

MAb = Monoclonal antibody

HCT = Hematopoietic cell transplantation

Toxicities were evaluated by measuring complete blood counts, blood urea nitrogen, creatinine, and liver enzymes; at baseline, daily for the first two months or until full hematopoietic recovery, and then weekly until the end of the study. Necropsies at euthanasia included gross examination and tissue harvest for pathological assessment of microscopic abnormalities.

Alpha Imaging

Two α-imaging systems were used for high-resolution ex vivo assessment of ²¹¹At localization and microdosimetry: α-camera (12) and ionizing-radiation Quantum Imaging Detector (iQID) (13). Cryosections (10–12 μm) of popliteal lymph nodes and bone marrow cores were placed on a scintillation film (EJ-440; Eljen Technology) and imaged as previously described (14) [Miller et al. Med Phys. 2015. (in press)]. Consecutive sections were stained with hematoxylin and eosin (H&E) for histological comparison with the imaged intra-organ ²¹¹At distribution.

Dosimetry

Absorbed doses to blood were estimated individually from the blood samples, by creating time-activity curves (%IA/g as a function of time) assuming 100 %IA in blood at injection (t=0). Blood volumes were derived from individual dog weights and a standardized total blood volume of 102.6 mL/kg (15). The cumulated ²¹¹At activity (Ã) was calculated after exponential curve fitting for each time-activity curve. Using an absorbed fraction of α-particles (ϕ_α) of 1 and including contributions exclusively from α-particles, the absorbed dose (D) in Gy was calculated using the equation

$$D=\frac{\stackrel{\sim }{A}}{m}{\mathit{\Delta }}_{\alpha }{\varphi }_{\alpha }$$

where m denotes tissue mass in kg, and Δ_α is the mean energy released per ²¹¹At decay (1.09 × 10⁻¹² J). Mean absorbed dose rates at biopsy were calculated and normalized to the individual injected activities (μGy/MBq-s) for bone marrow (core and aspirate) and lymph nodes using radioactivity measurements of the samples. Total absorbed doses (Gy/MBq) to marrow were estimated by interpolation between early and late time points using polynomial curve fits of the time-activity curves, assuming zero tissue radioactivity at t=0.

Small-scale dosimetry was performed for marrow and lymph nodes through quantification of the radioactivity concentration (Bq/g) at biopsy in the α-imaged samples. Individual section masses were estimated using the area and thickness of each imaged section, and a tissue density of 1 g/cm³. Small-scale mean absorbed dose rates were calculated as described above, and compared to those calculated for “macroscopic” (i.e. non-sectioned) biopsies using traditional whole-tissue dosimetry methods (16). Dose rate differences between various sub-compartments in α-imaged samples were also examined.

Image-based small-scale 3D-dosimetry was performed for lymph nodes from two dogs, using voxel dose-point kernels and α-camera imaging of serial sections (17). Briefly, series of 13 consecutive 12-μm sections were imaged, registered, and stacked to a 3D-matrix of 12×12×12-μm³ voxels. Voxel dose-point kernels for ²¹¹At were generated using Monte Carlo simulation and convolved with the radioactivity images using MATLAB (The MathWorks, Inc.), to estimate the absorbed dose rate in each voxel. Dose rate distribution maps were thereby achieved for the centermost section (7/13) of each series.

Histology and Immunohistochemistry

Formalin-fixed, paraffin-embedded lymph node samples were stained on a Leica Bond Rx system with antibodies against CD3 (3 μg/ml, Serotec MCA1477), MAC 387 (0.2 μg/ml, Dako M0747), and cleaved caspase-3 (0.3 μg/ml, Biocare Medical CP229B) after ethylenediaminetetraacetic acid (EDTA)-based antigen retrieval. The CD3 antibody was followed by a rabbit anti-rat secondary, all antibodies were then detected and visualized using Power Vision HRP polymers and Refine DAB (Leica Biosystems). Slides were scanned using an Aperio AT scanscope system (Leica) with 40× image capture magnification and analyzed using the ImageScope viewing software.

Results

²¹¹At-B10-CA12.10C12

Astatination of CA12.10C12-B10 was achieved with 74–91 % labeling yields. The radiochemical purity of the ²¹¹At-B10-CA12.10C12 was always >98 %, as determined by radio-HPLC or instant thin layer chromatography.

Biodistribution and Pharmacokinetics

The in vivo distributions in selected tissues from non-transplanted dogs (H573 and H585) are shown in Fig. 1A. The ²¹¹At accumulation was highest in lymphocyte-rich target tissues such as spleen (0.31–0.61 %IA/g), lymph nodes (0.02–0.16 %IA/g), and bone marrow core (0.06–0.08 %IA/g), and in the liver (0.11–0.12 %IA/g). Uptake levels in other tissues were low. The activity concentration in bone marrow core for all dogs, at one or two biopsy time points, revealed no distinct trend in terms of accumulation/release over time (paired t-test; p=0.61). Clearance of circulating ²¹¹At from blood is shown in Fig. 1B. The mean %IA/g did not vary significantly between the two MAb doses (unpaired t-test; p=0.62), whereas the mean anti-CD45 MAb concentration in plasma was higher for 1.00 mg/kg (p=0.01; Supplemental Fig. 1).

FIGURE 1.

Biodistribution and clearance of ²¹¹At-anti-CD45 MAb. (A) Radioactive content expressed as the percentage of injected activity per gram (%IA/g ± SD, n=3) in selected tissues from two dogs euthanized 19–22 hours p.i. (B) Retention of ²¹¹At in blood for all treated dogs, determined by radioactivity measurements of repeated blood samples (%IA/g ± SD, 0.75 mg/kg, n=5; 1.00 mg/kg, n=3).

Anti-CD45 Saturation

Flow cytometry data for lymphocytes in marrow and lymph nodes are summarized in Tables 2 and 3, showing MFI and the relative proportion (%) of FMF and CD45F staining, respectively. Bone marrow lymphocytes were rapidly and efficiently targeted by anti-CD45 MAb, as demonstrated by high intensity and percentage binding of FMF at the early time points. The corresponding CD45F numbers were low, indicating near CD45 saturation. In contrast, FMF staining of lymph node samples resulted in low MFI regardless of time point or MAb dose. This result, in combination with high CD45F intensity, confirmed the abundance of free antigen. Flow cytometry dot plots of FITC versus forward scatter indicated cell depletion in marrow at the later time point, but no similar effect was seen for lymph nodes (data not shown). The antigen binding differed significantly in terms of FMF and CD45F MFI in blood lymphocytes after ²¹¹At-MAb infusion, confirming the efficient saturation of CD45 sites on circulating lymphocytes (Supplemental Fig. 2).

TABLE 2.

Mean Fluorescence Intensity after Staining of CD45+ Cells (Gated on Lymphocytes)

Dog ID	Bone marrow core				Lymph node
	FMF		CD45F		FMF		CD45F
	2–4 h	19–22 h	2–4 h	19–22 h	2–4 h	19–22 h	2–4 h	19–22 h
H573*	—	7245	—	1292	—	1288	—	7500
H585*	—	1009	—	213	—	402	—	7027
H543*	—	17931	—	669	—	2358	—	5257
H689*	28172	10082	553	1841	2178	1011	12133	6020
H522*	10432	—	82	—	505	—	6935	—
H629†	17956	1000	109	65	758	2774	6041	6025
H638†	37176	32831	220	548	1780	1057	15113	13754
H632†	24298	38886	155	564	1517	4478	14112	17067

^*MAb dose 0.75 mg/kg

^†MAb dose 1.00 mg/kg

FMF = FITC-conjugated goat anti-mouse F(ab′)₂

CD45F = FITC-conjugated anti-CD45 MAb

TABLE 3.

Ratio (%) of Stained CD45+ Cells (Gated on Lymphocytes)

Dog ID	Bone marrow core				Lymph node
	FMF		CD45F		FMF		CD45F
	2–4 h	19–22 h	2–4 h	19–22 h	2–4 h	19–22 h	2–4 h	19–22 h
H573*	—	44.7	—	51.0	—	9.9	—	97.0
H585*	—	13.7	—	6.1	—	2.7	—	98.2
H543*	—	94.2	—	15.2	—	17.2	—	82.4
H689*	95.8	81.7	11.7	87.8	8.0	10.0	93.7	96.8
H522*	91.4	—	0.4	—	4.2	—	97.0	—
H629†	95.6	5.2	0.1	0.8	5.4	24.4	94.1	88.3
H638†	97.2	92.8	0.8	5.3	6.8	6.8	96.8	97.2
H632†	78.9	93.6	0.5	7.5	6.2	35.8	97.9	97.9

^*MAb dose 0.75 mg/kg

^†MAb dose 1.00 mg/kg

FMF = FITC-conjugated goat anti-mouse F(ab′)₂

CD45F = FITC-conjugated anti-CD45 MAb

Alpha Imaging

Images acquired using α-camera and iQID showed heterogeneous ²¹¹At distribution in lymph nodes at all time points, with predominant localization to T lymphocyte-rich regions in the paracortex, as determined from consecutive H&E-stained cryosections (Fig. 2, Supplemental Fig. 3). The lymph node medulla and follicles in the superficial cortex contained little radioactivity. Follicles in the superficial cortex presented as clearly defined dark, round areas in contrast to the high-intensity deep cortical units (Fig. 2A, C). Quantitative image analysis showed that the activity concentrations in high-activity lymph node regions could be a factor of 8 up to 65 times higher than those in lower-activity regions. Apart from the bone matrix, which exhibited no quantifiable activity uptake, the marrow samples displayed a more homogeneous uptake profile at both time points, with high levels of ²¹¹At activity in all regions containing hematopoietic cells (Fig. 2B, D).

FIGURE 2.

Distribution of ²¹¹At in cryosectioned tissue samples imaged using the iQID camera (A–B), and corresponding H&E-stained sections (C–D). Panels (A) and (C) show lymph node sections from dog H638, biopsied 19 hours p.i; (B) and (D) show sections of bone marrow core from dog H632, biopsied 2 hours p.i. Green and yellow lines indicate lymph node follicles and paracortex (including visibly apoptotic areas), respectively, identified in the H&E section. Overlaid with the α-image, the outlined regions coincide well with low- and high-activity areas. Horizontal scale bars indicate 1 mm, colorbars show ²¹¹At activity in μBq.

Histopathology and Immunohistochemistry

Visual examination of H&E slides and corresponding α-images revealed an apparent association between apoptotic cells and high-signal areas, in particular at the later time points. Radiation-induced damage and associated features (cellular debris and macrophage infiltration) were distinctly increased in lymph nodes biopsied at 19 hours p.i. Immunohistochemical staining of paraffin-embedded lymph nodes allowed more detailed analysis compared with frozen samples, because of superior morphology preservation; however, few samples were large enough to enable both methods. The most comprehensive data set was achieved for the H629 2-hour biopsy seen in Fig. 3. Note that the paraffin-embedded sections were cut from the same lymph node as the α-imaged cryosection in Fig. 2B and E, albeit not consecutively. The panels in Fig. 3 show representative regions including cortex, paracortex and medulla, demonstrating the general morphology (A), T cell localization (B), apoptosis markers (C), and macrophage localization (D). As expected, T cells were abundant in paracortex but typically scarce in the cortical follicles. B cell staining was not achieved, due to unsatisfactory antibody performance in paraffin-embedded canine tissues. Similarly, efforts to assess the distribution of CA12.10C12-B10 in untreated lymph nodes were unsuccessful. At 2 hours p.i., staining for cleaved caspase-3 indicated widespread apoptosis. Macrophages containing multifocal apoptotic cellular debris were present in medullary cords and interfollicular zones, as well as within occasional germinal centers.

FIGURE 3.

Immunohistochemical staining of a formalin-fixed paraffin-embedded lymph node from dog H629, biopsied 2 hours after ²¹¹At-B10-CA12.10C12 injection. (A) H&E-stained section, including representative areas of cortex, paracortex and medulla. Panels (B), (C), and (D) show staining for T cells (CD3), apoptosis (cleaved caspase-3), and macrophages (MAC 387), respectively, in corresponding areas of serially cut sections from the same lymph node. The scale bar represents 200 μm.

Dosimetry

The absorbed dose to blood ranged from 0.013 to 0.030 Gy/MBq, corresponding to total absorbed doses of 1.6–5.7 Gy (mean 3.1 Gy) for the mean injected ²¹¹At activity of 166 MBq (Table 4). Only minor intersubject variability was exhibited in absorbed dose rates to blood over time, analogous to the ²¹¹At blood clearance (Fig. 1B).

TABLE 4.

Mean Absorbed Dose to Blood after Injection of ²¹¹At-Anti-CD45 MAb

Dog ID	211At activity (MBq/kg)	Injected 211At activity (A) (MBq)	Absorbed dose to blood (D) (Gy)	D/A (Gy/MBq)
H573*	14.6	112	3.3	0.030
H585*	27.6	259	5.7	0.022
H543*	11.5	147	2.2	0.015
H689*	13.9	120	1.6	0.013
H522*	13.9	165	2.5	0.015
H629†	14.6	139	2.7	0.020
H638†	18.4	227	3.0	0.013
H632†	18.7	157	3.4	0.022

^*MAb dose 0.75 mg/kg

^†MAb dose 1.00 mg/kg MAb = Monoclonal antibody

Concordance was seen between mean absorbed dose rates for marrow derived from either whole-tissue dosimetry of macroscopic samples (core and aspirate) or α-imaging (core), ranging from 0.05–0.60 to 0.01–0.11 μGy/MBq-s at early and late time points, respectively (Supplemental Fig. 4A). Polynomial curve fitting of the macro-biopsy data resulted in mean absorbed doses to marrow of 2.4 Gy/166 MBq. For lymph nodes, the dose rates were 0.03–0.61 and 0.01–0.20 μGy/MBq-s early and late, respectively (Supplemental Fig. 4B). Interestingly, two 2-hour data points for H632 were markedly elevated, indicating rapid and high lymph node uptake of ²¹¹At. A polynomial fit using these two high-dose data points and the mean of all data points at 19–22 hours p.i. resulted in a mean absorbed dose to a high-activity lymph node of 3.4 Gy/166 MBq of ²¹¹At.

Fig. 4 compiles the lymph node data from the 3D small-scale dosimetry for dogs H638 and H689. Highly non-uniform dose rate distributions were observed in all three cases (Fig 4A, C, D), with small, “hot” volumes (80–180 mGy/h) dispersed in main volumes of lower dose rate levels (0–20 mGy/h). Comparison with H&E-stained sections for H638 (Fig. 4B) indicated co-localization of areas with high dose rate and lymphocyte-rich subregions within the lymph node.

FIGURE 4.

Dose rate images derived by 3D-dosimetry using α-camera imaging. (A) The dose rate distribution at biopsy for H638 4 hours p.i., with a corresponding H&E-stained section in (B). Panels (C) and (D) display the dose rate distribution for H689 2 and 19 hours p.i., respectively. Colorbars express the dose rate in mGy/h; the scale bar (top right) indicates 1 mm.

Toxicity

Aspartate aminotransferase levels (Supplemental Fig. 5A) remained primarily within the normal range, except for dog H632, which experienced a brief initial spike with a 2-fold increase in numbers above the upper limits of normal (ULN) 14–28 days after HCT, before returning to normal on day 42. H629 displayed a more gradual effect with a maximum on day 136 after HCT, at which point it was euthanized due to severe ascites that started developing on day 119. Alkaline phosphatase levels were initially elevated (1.1–2.6 × ULN) overall, but returned to normal between day 28 and 70 after HCT, except for dog H689 whose numbers remained just above the reference range (Supplemental Fig. 5B). Bilirubin and creatinine levels were unremarkable for all dogs (Supplemental Fig. 5C, D).

All six transplanted dogs experienced transient pancytopenia (Supplemental Fig. 5E–G). The median nadir levels for granulocytes, lymphocytes, and platelets were 9 cells/μL (range 0–24 cells/μL), 74 cells/μL (range 57–124 cells/μL), and 4,500 platelets/μL (range 3,000–6,000 platelets/μL), respectively, at a median time of 3 days (range 3–5 days), 1 day (range–1 to 2 days), and 7 days (range 6–8 days) after HCT, respectively. The first of three consecutive days after nadir with absolute neutrophil counts ≥ 500/μL occurred at a median of 9 days after HCT (range 8–20 days), and the first of seven consecutive days after nadir with platelet counts ≥ 20,000/μL occurred at a median of 21 days p.i. (range 18–24 days). A median of 3 platelet transfusions were administered per dog (range 1–6), with a transfusion threshold of 10,000/μL.

All subjects except H629 were euthanized in good condition after 32 to 48 weeks of follow-up. H543, H689, H522, and H632 displayed no macroscopic or histologic evidence of toxicity in the liver, kidney, or other organs. Pathological examination of H638 revealed signs of subacute bronchitis and increased iron stores in liver macrophages following six blood transfusions. The necropsy of H629 showed mild hepatic abnormalities, including features sometimes attributed to early sinusoidal obstructive syndrome. Its seemingly mild liver toxicity did not correspond to the level of ascites found at the time of euthanasia—a discrepancy that remains unexplained. No signs of renal toxicity were seen in any of the transplanted dogs.

Discussion

Radioimmunotherapy is a highly suitable treatment option for lymphoma because of the synergistic combination of high radiosensitivity of hematological malignancies and accessibility of circulating radioimmunoconjugates to targets in lymph nodes, marrow and spleen. Our group has previously demonstrated the safety and efficacy of CD45-targeted α-radioimmunotherapy in dogs as a potential substitute for external beam total body irradiation and high-dose systemic chemotherapy in HCT pretreatment regimens (2,3,5). The current study further assessed ²¹¹At-anti-CD45 radioimmunotherapy conditioning for use in upcoming therapy studies of spontaneously occurring canine lymphoma—a feasible and clinically relevant model, thanks to the many similarities with human lymphoma. Here, the administered MAb dose was optimized and the analysis extended to include intra-organ ²¹¹At activity distribution and small-scale dosimetry at the micrometer range, using two α-imaging techniques along with immunohistochemical staining.

The starting dose of 0.75 mg/kg proved satisfactory for targeting lymphocytes in blood and marrow but left a large majority of lymph node binding sites vacant. Increasing the dose to 1.00 mg/kg did not improve the ratios; the only significant effects were an expected elevation in circulating anti-CD45 MAb levels and a minor shift in FMF fluorescence intensity corresponding to increased targeting of blood lymphocytes. Interestingly, the binding of FITC-labeled anti-CD45 MAb also increased slightly; although significant, this effect was less pronounced. The results indicate that considerably higher amounts of anti-CD45 MAb likely would have been required to approach saturation in lymph nodes, causing increased non-target irradiation by circulating radioimmunoconjugates. Therefore, further dose escalation was not explored.

The two MAb doses generated no distinguishable differences in hematologic toxicity or hepatic/renal function, and no trends were seen with regards to the administered ²¹¹At activity level. No serious toxicity was seen after administration of the lower MAb dose (0.75 mg/kg). The main adverse event occurred in H629, which was euthanized due to severe ascites 20 weeks after receiving the higher MAb dose (1.00 mg/kg; 14.6 MBq ²¹¹At/kg) and HCT. The clinical presentation and histological features of its liver closely resembled those previously reported by our group for two non-transplanted dogs who were treated with ²¹¹At-anti-CD45 radioimmunotherapy using 0.50 mg MAb/kg (5). In that radiation dose escalation study, hepatic aberrations of clinical relevance were observed in dogs treated with more than 15.0 MBq/kg. Eight additional dogs received DLA-identical marrow transplants following conditioning with 7.5–23.1 MBq/kg of ²¹¹At-anti-CD45 radioimmunotherapy (0.50 mg MAb/kg), without any cases of evident liver failure. The mechanisms behind this tolerance discrepancy were not fully elucidated, but based on rat studies conducted by Harb et al. (18) it was hypothesized that radiation-induced hepatic damage may have been repaired by bone marrow-derived progenitor cells in transplanted dogs. In an earlier set of experiments, similar toxicity was seen in one dog after pretreatment with α-radioimmunotherapy (326 MBq/kg of ²¹³Bi-anti-CD45 MAb; 0.50 mg/kg) before DLA-identical transplantation (2).

As expected, the biodistribution showed higher accumulation of ²¹¹At in the liver than in other non-target organs, which is typical for radioimmunotherapy due to internalization of circulating immunoglobulins by hepatic Kupffer cells and endothelial cells (19). In addition, slightly increased %IA/g is expected for anti-CD45-radioimmunoconjugates due to presence of CD45-expressing cells of hematopoietic origin in the liver (20). Even so, CD45 is attractive for radioimmunotherapy because of its abundance on target cells, and non-internalizing and non-shedding features. Most importantly, anti-CD45 MAbs do not compete with any anti-CD20 pretreatment (e.g. rituximab) for binding to target antigens, which may limit the efficacy of anti-CD20 radioimmunotherapy (21).

The estimated mean absorbed dose to blood of 3.1 Gy per 166 MBq of ²¹¹At was comparable to standard external total body irradiation dose levels of 2–3 Gy (2). However, this level of absorbed dose increases markedly for ²¹¹At-radioimmunotherapy if accounting for the relative biological effectiveness (RBE) of α-emitters, generally assumed to be in the range of 3–5 (16). Corresponding, non-RBE-adjusted, absorbed doses to marrow and lymph nodes calculated using traditional whole-tissue dosimetry were comparable to the dose to blood, but it should be noted that the lymph node estimations were based on a limited number of time points. Nevertheless, the dose levels of 3 Gy and above are a strong indication of satisfactory targeting, which has not been demonstrated in this model before.

Alpha imaging revealed substantial dose rate variations within the lymph nodes, but congruence between “hot” areas and lymphocyte-rich regions sustained the perception of effective CD45 targeting. Furthermore, markedly higher dose rates (~factor 2) were seen in these high-activity areas at the early time points—the total absorbed dose to lymphocyte-rich target regions could thus be a factor of 2 higher than the 3.4 Gy/166 MBq that was calculated from the mean activities, should this be a characteristic time progression for each lymph node. Widespread indications of cell death observed in immunohistochemically stained sections supported the preliminary dose estimates, although more accurate dosimetry requires additional time points. Ongoing ²¹¹At-radioimmunotherapy studies in dogs will extend the dosimetric analysis further.

Concordance was also seen between flow cytometry data and in vivo activity profiles. The profuse ²¹¹At radioactivity registered in marrow matched the widespread targeting indicated by flow analysis; likewise, the heterogeneous activity distributions in lymph nodes corresponded well with the lower overall targeting discerned by the flow assays. CA12.10C12 supposedly targets all isoforms of CD45 expressed by T and B lymphocytes, yet our α-images consistently demonstrated absence of ²¹¹At in B lymphocyte-containing follicles at all studied time points. A possible explanation could be that T cell-associated isoforms (180- to 220-kDa) are more abundant or easier to access than B cell isoforms (205- to 220-kDa).

Conclusion

Efficient targeting of ²¹¹At-anti-CD45 radioimmunotherapy to blood and marrow was achieved using a MAb dose of 0.75 mg/kg, without severe toxicity. Although antigen saturation was not reached, dosimetry calculations and observed radiation-induced effects in immunohistochemically stained lymph node samples indicated that sufficient ²¹¹At localization was achieved for use in HCT pretreatment regimens at levels of injected activity that should be tolerable for normal tissues. Our results highlight the importance of small-scale dosimetry for understanding therapeutic outcomes after α-radioimmunotherapy. Ongoing studies will reveal if efficient ²¹¹At-B10-CA12.10C12 conditioning can be achieved for autologous HCT curative therapy of relapsed lymphoma in dogs.