A N-acetylgalactosamino Dendron-Clearing Agent for High-Therapeutic Index DOTA-Hapten Pretargeted Radioimmunotherapy

Abstract

Clearing agents (CAs) can rapidly remove non-localized targeting biomolecules from circulation for hepatic catabolism, thereby enhancing the therapeutic index (TI), especially for blood (marrow), of the subsequently administered radioisotope in any multi-step pretargeting strategy. Herein we describe the synthesis and in vivo evaluation of a fully synthetic glycodendrimer-based CA for DOTA-based pretargeted radioimmunotherapy (DOTA-PRIT). The novel dendron-CA consists of a non-radioactive yttrium-DOTA-Bn molecule attached via a linker to a glycodendron displaying sixteen terminal α-thio-N-acetylgalactosamine (α-SGalNAc) units (CCA α-16-DOTA-Y³⁺; molecular weight: 9059 Da). Pretargeting [¹⁷⁷Lu]LuDOTA-Bn with CCA α-16-DOTA-Y³⁺ to GPA33-expressing SW1222 human colorectal xenografts was highly effective, leading to absorbed doses of [¹⁷⁷Lu]LuDOTA-Bn for blood, tumor, liver, spleen, and kidneys of 11.7, 468, 9.97, 5.49, and 13.3 cGy/MBq, respectively. Tumor-to-normal tissues absorbed-dose ratios (i.e., TIs) ranged from 40 (e.g., for blood and kidney) to about 550 for stomach.

Graphical Abstract

Introduction

The therapeutic indices (TIs; tumor-to-normal tissue-absorbed dose ratios) of radioimmunotherapy (RIT) should be maximized for safe and effective treatment of solid tumors.¹ RIT with radiolabeled-IgG antibodies suffers from low TIs due to the unfavorable pharmacokinetics of the IgG carrier—slow localization in the targeted tumor and rapid, high uptake in the radiosensitive bone marrow and other reticuloendothelial tissues—and is thus often ineffective at maximum tolerated activities.¹ We developed a S-2-(4-aminobenzyl)-1,4,7,10-tetraazacyclododecane tetraacetic acid (DOTA) chelate-based pretargeting strategy (DOTA-PRIT) for theranostic imaging and treatment of solid tumors with DOTA-radiohaptens.^2-5 DOTA-PRIT consists of separate, temporally spaced injections of three reagents: (1) a tetravalent bispecific IgG-single chain (IgG-scFv; 210 kDa) antibody (BsAb) with high affinity for (a) a tumor antigen (the IgG portion) and (b) the radiometal complex of S-2-(4-aminobenzyl)-1,4,7,10-tetraazacyclododecane tetraacetic acid chelate ([M]-DOTA-Bn; the “C825” scFv portion)⁶; (2) a clearing agent (CA) to rapidly reduce circulating BsAb after sufficient time is given for BsAb to accumulate at antigen-positive tumor⁷; and (3) a radiolabeled DOTA-hapten such as the DOTA-Bn complex with the theranostic (i.e., gamma- and beta-emitting) isotope lutetium-177 ([¹⁷⁷Lu]LuDOTA-Bn).⁸ To target the radioactivity to tumor, the low-molecular weight circulating DOTA-hapten rapidly enters the tumor parenchyma and binds to the intra-tumoral BsAb and is otherwise rapidly cleared via renal excretion.⁸ DOTA-PRIT has been applied to tumor antigens encompassing a wide variety of solid and hematological tumor types, targeting tumor antigens that include GD2⁵, GPA33^2,3, HER2⁴, CD20⁹, CD38¹⁰, and CEA⁶.

For high-TI DOTA-PRIT, the CA step is interposed between injection of BsAb and radiolabeled DOTA-hapten.^{2-6, 9, 10} The currently utilized CA is a 500-kD dextran-DOTA hapten conjugate (dextran-CA) designed to bind to the anti-DOTA(M)-scFv domains of the BsAb in circulation via a DOTA(Y) moiety displayed on the dextran scaffold and remove it from the blood via recognition and catabolism by the reticuloendothelial system (RES).⁷ By virtue of its large size, undesirable binding of CA to tumor-associated BsAb is minimal because of its poor intra-tumoral extravasation.¹¹ The use of dextran-CA for DOTA-PRIT and other pretargeting approaches (e.g., based on bioorthogonal inverse electron demand Diels–Alder reaction¹²) is facilitated by ease of synthesis and relatively low cost of starting materials as well as a favorable safety profile of dextran-conjugates in animals.

Although highly effective, the use of a dextran-CA has drawbacks. As a naturally occurring glucose polymer, the dextran scaffold is inherently polydisperse, thus presenting challenges related to reproducible batch-to-batch fabrication as well as in vivo use. Also, enzymatic degradation by RES dextran-1,6-glucosidase could potentially lead to the introduction of hapten fragments into the circulation, which will compete with the radiohapten itself for binding by tumor-BsAb.¹³ Similar issues applied to albumin-based CAs^14,15 during clinical streptavidin-biotin PRIT, prompting the development of a dendrimer-based CA (dendron-CA) consisting of a biotin attached via a linker to a glycodendron displaying 16 terminal α-thio-N-acetylgalactosamine (α-SGalNAc) units for active hepatocyte targeting.¹⁶ The α-SGalNAc are ligands for the hepatocyte Ashwell-Morell receptor, a lectin involved in the recognition, binding, and clearance of asialoglycoproteins.¹⁷ This biotin-dendron-CA was shown to be effective and well-tolerated in patients (e.g., with no side effects during phase I/II studies of PRIT of non-Hodgkin’s lymphoma¹⁸).

Herein, we sought to adapt the biotin-dendron-CA for DOTA-PRIT. We synthesized a novel DOTA-dendron-CA consisting of 16 terminal α-SGalNAc residues presented using a dendrimer core in place of the dextran and verified its effectiveness in vivo using model DOTA-PRIT.

Results and Discussion

The dendron-CA (Figure 1; see SI for complete synthesis details) was prepared by mixing an advanced fourth-generation dendrimer intermediate, bearing 16 terminal α-SGalNAc residues and a free amine at the stem¹⁹, and commercially available S-2-(4-isothiocyanatobenzyl)-1,4,7,10-tetraazacyclododecane tetraacetic acid (p-SCN-Bn-DOTA) under basic conditions to provide the corresponding DOTA-thiourea-dendrimer.

Figure 1.

N-acetylgalactosamino dendron-clearing agent designed for use with DOTA-PRIT, CCA-16-DOTA-Y³⁺ dendron-CA.

For recognition and binding by the anti-DOTA antibody C825 (equilibrium dissociation constant (K_D) = 15.4 ± 2.0 pM for Y-DOTA-Bn⁶, the DOTA-thiourea-dendrimer was loaded with yttrium using Y²⁰ chloride hexahydrate under slightly acidic conditions. The final removal of all sugar acetate groups (48 of them) was performed under degassed conditions using 0.2 N NaOH in methanol. The dendron-CA CCA α-16-DOTA-Y³⁺ final product was purified by HPLC and fully characterized by liquid chromatography mass spectrometry to ensure product integrity and purity.

To verify the DOTA-dendron-CA’s effectiveness in vivo as well as to determine the kinetics of removal of DOTA-PRIT-BsAb from circulation, we initially conducted experiments in healthy (tumor-free) athymic nude mice using model radioiodinated-anti-GPA33/anti-DOTA huA33-C825 BsAb (¹³¹I-BsAb) as a tracer³. Initially, groups of animals (n = 3/group) were injected with 250 μg (1.19 nmol) of ¹³¹I-BsAb (19-22 μCi), a BsAb dose previously optimized for DOTA-PRIT.³ Twenty-four hours later, mice were injected with dendron-CA (25 μg, 2.76 nmol) or vehicle control. Serial tail-vein blood collection was performed at various times up to 4 hours post-injection of CA (or 28 hours post-injection of ¹³¹I-BsAb), and the ¹³¹I-activity concentration was determined by weighing each sample and radioassay in an ¹³¹I-calibrated gamma counter. As shown in below in Figure 2, there was initially no significant difference in blood ¹³¹I-activity at baseline (i.e., at 24 hours post-injection of the ¹³¹I-BsAb immediately prior to the injection of the CA) between groups, but the CA was very effective in decreasing circulating ¹³¹I-BsAb, as the average blood activity concentration significantly dropped by 64% at 5 minutes post-injection from 6.7 %ID/g at baseline to 2.4 %ID/g.

Figure 2.

Effect of dendron-CA CCA-16-DOTA-Y³⁺ on circulating ¹³¹I-BsAb in vivo. Athymic normal (tumor-free) mice were intravenously injected a t = 0 with ¹³¹I-BsAb (19-21 μCi; 250 μg, 1.19 nmol), followed with either vehicle (saline) or dendron-CA (25 μg; 2.76 nmol) at t = 24 hours. Serial blood sampling was conducted at various time points from t = 1-28 h post-injection of ¹³¹I-BsAb. Data is presented as mean ± SD. For some points, the error bars would be shorter than the height of the symbol. ***P <0.001 compared with vehicle.

At 1 hour post-injection of CA, the average blood activity was 1.3 %ID/g (−81% change from baseline) or 5.8 %ID/g (−13% change from baseline) in those mice given CA or vehicle, respectively. Typically, a 1- to 4-hour time interval between CA and hapten is used for DOTA-PRIT; the average blood activity at 4 hours was essentially unchanged from that at 1 h post-injection of CA (1.2 %ID/g, (−82% change from baseline) or 5.1 %ID/g (−13% change from baseline)) in those mice given CA or vehicle, respectively.

Next, we conducted CA dose-escalation studies using a model DOTA-PRIT system (targeting sub-cutaneous human colorectal cancer xenograft SW1222 in athymic nude mice with anti-GPA33-DOTA-PRIT) to evaluate the relationship between administered dendron-CA dose and subsequent uptake of [¹⁷⁷Lu]LuDOTA-Bn in tumor and normal tissues. Briefly, groups of SW1222 tumor-bearing mice (n = 3-4/group) were injected with BsAb (250 μg, 1.19 nmol; t = −28 hours) followed by varying amounts of CA (0-25 μg; 0-2.76 nmol; t = −4 hours). An additional group of tumored animals were given dextran-CA (62.5 μg; 0.125 nmol; 7.625 nmol (Y)DOTA) in place of the dendron-CA for comparison. Four hours after CA administration, the mice were injected with [¹⁷⁷Lu]LuDOTA-Bn (150-167 μCi; 30-33 pmol). Notably, we used a 24-hour time interval between BsAb and CA; we chose this time interval (e.g., to 48 hours) due to the reported limited serum stability of the C825 scFv; according to Orcutt, et al.²¹ the binding activity of the C825 scFv decreases to about 60% after 7 days in mouse serum at 37 degrees C. We intend to conduct further studies to clarify optimal timing, with the goal of achieving the highest possible therapeutic index, while maintaining sufficient tumor localization to achieve high tumor radiation-absorbed dose.

Based on previously published work, non-tumor-associated BsAbs-[¹⁷⁷Lu]LuDOTA-Bn is metabolized in the liver and spleen, as evidenced by previous ex vivo serial biodistribution studies documenting relatively slow clearance of ¹⁷⁷Lu-activity from those tissues in mice administered DOTA-PRIT plus [¹⁷⁷Lu]LuDOTA-Bn.³ As shown in Table 1, ex vivo biodistribution studies in groups of SW1222 tumor-bearing mice administered DOTA-PRIT plus [¹⁷⁷Lu]LuDOTA-Bn showed CA-dose dependent tumor-to-blood uptake (i.e., activity concentration) ratios of [¹⁷⁷Lu]LuDOTA-Bn at 24 hours post-injection of ¹⁷⁷Lu-activity; e.g., average tumor-to-blood ratios were 2.9, 26, and 59 for vehicle (0 μg), 15 μg, or 20 μg of dendron-CA, respectively.

Table 1.

Dose escalation of dendron-CA during anti-GPA33 DOTA-PRIT + [¹⁷⁷Lu]LuDOTA-Bn. ¹⁷⁷Lu-activity concentrations and corresponding tumor-to-blood ratios at 24 hours post-injection of [¹⁷⁷Lu]LuDOTA-Bn are shown. Please see SI Table 1 for additional data. Data is presented as mean ± SEM.

Dose of CA	n	Blood (%ID/g)	Tumor (%ID/g)	Tumor/Blood
0 μg vehicle	4	11.9 ± 0.36	34.7 ± 4.04	2.9 ± 0.4
2.5 μg/0.276 nmol	4	11.1 ± 1.15	32.5 ± 1.76	2.9 ± 0.3
5 μg/0.552 nmol	4	9.47 ± 0.28***	41.2 ± 3.55	4.4 ± 0.4
10 μg/1.10 nmol	4	6.42 ± 1.25**	31.5 ± 6.34	4.9 ± 1.4
15 μg/1.66 nmol	4	1.27 ± 0.21***	33.3 ± 4.42	26.3 ± 5.5
20 μg/2.21 nmol	4	0.46 ± 0.13***	27.2 ± 5.17	59.2 ± 20.0
25 μg/2.76 nmol	3	0.40 ± 0.18***	30.8 ± 5.07	76.4 ± 36.2
Dextran-CA 62.5 μg/0.125 nmol dextran/7.625 nmol (Y)DOTA	4	0.45 ± 0.09***	34.6 ± 4.50	77.3 ± 19.2

^***P <0.001 compared with vehicle

^**P <0.01 compared with vehicle

^*P <0.05 compared with vehicle

A 25-μg dose of dendron-CA led an average tumor-to-blood uptake ratio of 76, almost identical to dextran-CA (average tumor-to-blood ratio of 77; see table S1). Furthermore, a 25-μg dose of dendron-CA resulted in low tissue uptake in other normal tissues, including liver and spleen. Notably, dosing from 5-25 μg with dendron-CA led to average tumor [177Lu]LuDOTA-Bn uptakes of 27-41 %ID/g at 24 hours post-injection 177Lu-activity, with no statistical differences between vehicle (i.e., 0 μg of CA) and dextran-CA controls (each ~35 %ID/g). Furthermore, it was apparent that the dendron-CA did not have a negative impact on subsequent tumor uptake of [177Lu]LuDOTA-Bn at doses ≤ 25 μg, as there were no statistical differences between vehicle and all studied doses of dendron-CA (2.5 – 25 μg). In data not shown, we found dendron-CA doses of 40-μg/mouse (n = 3), reduced uptake in tumor an average of 13.8 ± 1.16 %ID/g (significant P = 0.015 compared with 25-μg/mouse) with 0.28 ± 0.03 %ID/g and 0.67 ± 0.09 %ID/g for blood and kidney, respectively, at 24 hours post-injection 177Lu-activity (not significant P = 0.267 and P = 0.466, respectively, compared with 25-μg/mouse).

Lastly, using an optimized dendron-CA dose, serial biodistribution studies were performed to estimate absorbed doses and TIs during DOTA-PRIT with [¹⁷⁷Lu]LuDOTA-Bn. In order to calculate absorbed doses for PRIT plus [¹⁷⁷Lu]LuDOTA-Bn including dendron-CA, groups of nude mice bearing GPA33-expressing SW1222 xenografts (n = 5/group) was injected i.v. with the BsAb huA33-C825 (250 μg; 1.19 nmol) 28 h prior and i.v. with dendron-clearing agent CCA-16-DOTA-Y³⁺ (25 μg; 2.76 nmol) 4 h prior to administration of [¹⁷⁷Lu]LuDOTA-Bn (20 pmol, 3.7 MBq [100 μCi]). These mice were sacrificed at 1, 4, 24, or 48 h p.i. of [¹⁷⁷Lu]LuDOTA-Bn for biodistribution assay; dosimetry calculations were performed using the resulting organ time-activity data (Supplementary Information). The estimated absorbed doses (cGy/MBq) to tumor, blood, liver, and kidney for single-cycle PRIT plus dendron-CA were 468, 11.7 (TI: 40), 9.97 (TI: 47), and 13.3 (TI: 35), respectively (Table 2). In addition, absorbed doses were calculated for select tissues for DOTA-PRIT without dendron-CA by assuming that the respective areas under the curves (AUCs or residence times) and therefore the absorbed doses scaled as the 24-h %ID/g values (i.e., there was no difference in the ¹⁷⁷Lu-activity kinetics with and without CA); the calculated estimated absorbed doses for tumor, blood, liver, and kidney were 593, 126 (TI: 4.7), 77 (TI: 7.7), and 39 (TI: 15), respectively.

Table 2.

Absorbed doses for pretargeting of [¹⁷⁷Lu]LuDOTA-Bn anti-GPA33-DOTA-PRIT with dendron-clearing agent CCA-16-DOTA-Y³⁺ in nude mice carrying s.c. GPA33(+) SW1222 tumors. The therapeutic index (TI) is defined as estimated tumor/normal tissues absorbed dose ratio.

Tissue	Absorbed dose (cGy/MBq)	Therapeutic index
Blood	11.7	40
Tumor	468	---
Heart	2.66	176
Lung	10.7	44
Liver	9.97	47
Spleen	5.49	85
Stomach	0.86	545
Small Intestine	1.16	404
Large Intestine	1.87	250
Kidneys	13.3	35
Muscle	3.73	126
Bone	3.68	127

Overall, the TIs ranged from about 40 (e.g., for blood and kidney) to about 550 for stomach. Previously with dextran-CA, we demonstrated effective treatment of SW1222 tumor-bearing mice (e.g., 111.0 MBq delivered in a fractionated dose strategy yielded a complete-response rate of 100%, including survival beyond 140 days post-tumor inoculation in two of nine mice) with estimated absorbed doses of 7304, 100, and 588 cGy to tumor, blood, and kidney and no demonstrable toxicity.³ Therefore, we anticipate that effective and safe colorectal cancer therapy in xenografts in mice is feasible with DOTA-PRIT plus dendron-CA, since a tumor-absorbed dose of ~73 Gy can be achieved with an administered activity of 15.6 MBq (with 183 cGy to blood (marrow) and 207 cGy to kidney, which are below benchmark maximum tolerated doses (MTDs) based on human normal-tissue radiation dose tolerance estimates derived from clinical observations of 250 cGy and 2,000 cGy for bone marrow and kidney, respectively.²² Furthermore, we estimate that the maximum tolerated pretargeted [¹⁷⁷Lu] LuDOTA-Bn activity is 21 MBq, with the bone marrow as the dose-limiting organ. At this activity, the estimated absorbed dose delivered to tumor would be 9,828 cGy (98 Gy), with 246 cGy to blood (marrow) and 279 cGy to kidney.

Overall, these data suggest that this novel CCA-16-DOTA-Y³⁺ dendron-CA can be used in place of dextran-CA for enhanced blood clearance of a ¹³¹I-BsAb as well as for high-TI DOTA-PRIT.