Pretargeted radioimmunotherapy using genetically engineered antibody-streptavidin fusion proteins for treatment of non-hodgkin lymphoma

Abstract

Purpose: Pretargeted radioimmunotherapy (PRIT) using streptavidin (SAv)-biotin technology can deliver higher therapeutic doses of radioactivity to tumors than conventional RIT. However, "endogenous" biotin can interfere with the effectiveness of this approach by blocking binding of radiolabeled biotin to SAv. We engineered a series of SAv FPs that downmodulate the affinity of SAv for biotin, while retaining high avidity for divalent DOTA-bis-biotin to circumvent this problem.

Experimental design: The single-chain variable region gene of the murine 1F5 anti-CD20 antibody was fused to the wild-type (WT) SAv gene and to mutant SAv genes, Y43A-SAv and S45A-SAv. FPs were expressed, purified, and compared in studies using athymic mice bearing Ramos lymphoma xenografts.

Results: Biodistribution studies showed delivery of more radioactivity to tumors of mice pretargeted with mutant SAv FPs followed by (111)In-DOTA-bis-biotin [6.2 ± 1.7% of the injected dose per gram (%ID/gm) of tumor 24 hours after Y43A-SAv FP and 5.6 ± 2.2%ID/g with S45A-SAv FP] than in mice on normal diets pretargeted with WT-SAv FP (2.5 ± 1.6%ID/g; P = 0.01). These superior biodistributions translated into superior antitumor efficacy in mice treated with mutant FPs and (90)Y-DOTA-bis-biotin [tumor volumes after 11 days: 237 ± 66 mm(3) with Y43A-SAv, 543 ± 320 mm(3) with S45A-SAv, 1129 ± 322 mm(3) with WT-SAv, and 1435 ± 212 mm(3) with control FP (P < 0.0001)].

Conclusions: Genetically engineered mutant-SAv FPs and bis-biotin reagents provide an attractive alternative to current SAv-biotin PRIT methods in settings where endogenous biotin levels are high.

Fig 1

Schematization of the IF5 scFvSAv gene and its derivative tetravalent FP. A) The scFvSAv fusion gene is comprised of the scFv gene of the murine IgG1 anti-CD20 1F5 Ab fused to the full length SAv gene; B) a diagrammatic representation of the tetravalent FP that spontaneously forms after secretion into the periplasmic space of E. coli; C) Representative HPLC tracing of a purified FP (S45A-SA); SDS-PAGE analyses of unpurified and iminobiotin-purified 1F5(scFv)₄SAv WT FP (D), Y43A-SAv FP (E), and S45A-SAv FP (F), respectively. The mass of the intact FPs (174 kDa) and the monomers (44 kDa) are indicated on each SDS-PAGE gels (4-12% Tris-glycine; (Invitrogen, Carlsbad, CA) stained with Coomassie blue. Lane M, migration of pre-stained molecular weight standards (SeeBlue plus; Invitrogen, Carlsbad, CA). Lane 1, loaded E. coli crude lysate, containing 1F5(scFv)₄SAv WT FP. Lanes 2 and 3, flow-through and wash fractions, respectively, from an iminobiotin purification column. Lanes 4 and 5, iminobiotin-purified FP without boiling (lane 4) or denatured by boiling for 5 minutes to resolve the tetramers into monomers (lane 5).

Fig 2

Flow cytometry showing equivalent binding of the 1F5-WT-SAv, 1F5-Y43A-SAv, and 1F5-S45A-SAv FPs to the surface of CD20-expressing Ramos cells (Fig. 2B). For comparison, binding of a non-binding negative control FP, CC49-WT-SAv (Fig 2B), a non-binding, isotype matched, negative control Ab, HB8181 (Fig. 2A) and of the parental 1F5 anti-CD20 Ab (positive control, Fig. 2A) are also shown.

Fig 3

Cell binding assay comparing the binding of ⁹⁰Y-DOTA-biotin and ⁹⁰Y-DOTA-bisbiotin to Ramos cells previously saturated with 1F5-WT-SAv FP, 1F5-Y43A-SAv FP, 1F5-S45A-SAv FP, or CC49-SA FP (negative control) or to untreated cells incubated with PBS. Ramos cells (1.0 × 10⁶ per sample) were incubated with 25 μl of 20ug/ml 1F5-Y43A-SAv, 1F5-S45A-SAv, 1F5-WT-SAv, or CC49-WT-SAv FPs in a 96 round bottom ELISA plate at 4 °C for 60 minutes. Cells were then washed twice in PBS. Next, 100 ng/ml of 90Y-DOTA–biotin or ⁹⁰Y-DOTA–bis-biotin were added to the cells and incubated at 4°C for 60 minutes. Cells were washed twice with PBS and then bound radioactivity was measured in a gamma counter.

Fig 4

Blood Clearance of 125I-1F5-S45A-SA FP using either a monovalent biotin CA (NAGB,◆) or a synthetic bis-biotin CA (BBTG-CA,▴). CAs (5.8 nmol) were injected 20 hours after administration of FP, then mice were bled serially and blood radioactivity measured in a gamma counter. Control mice received 2.8 nmol of FPs without any CA (●). Results are expressed as the percent of the injected dose of radioactivity present per gram of blood (%ID/g) on a logarithmic scale.

Fig 5

Biodistributions of 111In-labeled biotin compounds in tumors and normal organs of athymic mice bearing Ramos xenografts pretargeted with 1F5-S45A-SAv, 1F5-Y43ASAv, or 1F5-WT-SAv FP. Mice were fed either normal or biotin-deficient diets for 5-6 days before injection of 2.8 nmol of FPs, as indicated. Twenty-four hours later 5.8 nmol of either monovalent NAGB CA or divalent bis-biotin CA was injected. After another 4 hours, mice were injected with 1.2 nmol of 111In-labeled DOTA-biotin or 111In-labeled bisbiotin. Results are shown for mice sacrificed either 24 hours (A) or 48 hours (B) after administration of the radiolabeled biotin ligands.

Fig 6

Growth of Ramos xenografts in athymic mice pretargeted with CC49-SAv FP (nonbinding negative control, ●), 1F5-S45A-SAv FP (▴), 1F5-Y43A-SA FP (∎), 1F5-WT-SAv (◆) or 1F5-WT-SAv with a biotin deficient diet (⦿) followed by the indicated CA and 90Y-DOTA-bisbiotin or 90Y-DOTA-biotin. Mice were fed a normal, biotin-replete diet unless otherwise specified.