A comparison of DFO and DFO* conjugated to trastuzumab-DM1 for complexing 89Zr - In vitro stability and in vivo microPET/CT imaging studies in NOD/SCID mice with HER2-positive SK-OV-3 human ovarian cancer xenografts.

Desferrioxamine (DFO) is conjugated to antibodies to chelate 89Zr for PET, but DFO forms a hexadentate complex with Zr4+ that exhibits instability contributing to bone uptake of 89Zr, while the cationic charge of the Zr4+-DFO complex may promote normal tissue uptake of the radioimmunoconjugates (RICs). DFO* is a novel chelator that forms a more stable octadentate and neutral complex with 89Zr. Our aim was to compare the in vitro stability of [89Zr]Zr-DFO*-human IgG (hIgG) and [89Zr]Zr-DFO-hIgG RICs, and the in vivo PET imaging properties of the antibody-drug conjugate (ADC), trastuzumab-DM1 (T-DM1), labeled with 89Zr by conjugation to DFO or DFO*. SCN-pPhe-DFO and SCN-pPhe-DFO* were reacted with hIgG at a 14.6-fold excess or with T-DM1 at a 4.1-fold or 10-fold excess, respectively, purified and labeled with 89Zr. The number of DFO* introduced was determined by measuring the absorbance at 245/252 nm and the protein concentration was measured at 280 nm. The stability of [89Zr]Zr-DFO*-hIgG was studied in vitro in human plasma, and by challenge with a 385-fold excess (0.1 mM) of DFO or EDTA. An inverse stability study was performed with [89Zr]Zr-DFO-hIgG challenged with 0.1 mM DFO*. The HER2 binding affinity of [89Zr]Zr-DFO*-T-DM1 was measured in a direct (saturation) binding assay using SK-BR-3 human breast cancer cells or SK-OV-3 human ovarian cancer cells. The biodistribution of [89Zr]Zr-DFO*-T-DM1 and [89Zr]Zr-DFO-T-DM1 were compared in non-tumor bearing Balb/c mice and in NOD/SCID mice with s.c. SK-OV-3 xenografts at 96 h post-intravenous injection (p.i.). MicroPET/CT images were obtained at 96 h p.i. of the RICs. hIgG and T-DM1 were conjugated to 4.5-5.3 and 3.1 chelators (DFO or DFO*), respectively, and labeled with 89Zr to a final radiochemical purity of 91-99%. [89Zr]Zr-DFO*-hIgG was stable in vitro in human plasma or to challenge with 0.1 mM EDTA, but incubation with 0.1 mM DFO caused 26.0 ± 2.1% loss of 89Zr after 5 days. In contrast, incubation of [89Zr]Zr-DFO-hIgG with 0.1 mM DFO* resulted in 77.0 ± 3.9% loss of 89Zr after 5 days. [89Zr]Zr-DFO*-T-DM1 retained high affinity binding to HER2 on SK-BR-3 and SK-OV-3 cells with a Kd = 2.2 ± 0.3 nM and 1.9 ± 0.3 nM, respectively, and Bmax = 3.4 ± 0.1 × 105 and 1.1 ± 0.04 × 105 receptors/cell, respectively. Biodistribution studies of [89Zr]Zr-DFO-T-DM1 and [89Zr]Zr-DFO*-T-DM1 in Balb/c and NOD/SCID mice revealed significantly lower uptake in bone, liver, kidneys, and spleen for [89Zr]Zr-DFO*-T-DM1 than [89Zr]Zr-DFO-T-DM1. Uptake of [89Zr]Zr-DFO*-T-DM1 and [89Zr]Zr-DFO-T-DM1 in SK-OV-3 tumors was moderate [5.0 ± 1.8% injected dose/g (%ID/g) and 6.3 ± 0.6%ID/g, respectively; P = 0.18]. Tumors were imaged with both RICs. We conclude that DFO* conjugated to T-DM1 provides more stable complexation of 89Zr and therefore, [89Zr]Zr-DFO*-T-DM1 would be more useful than [89Zr]Zr-DFO-T-DM1 to probe the delivery of T-DM1 to tumors by PET, which we previously found is correlated with response to treatment with T-DM1 in mouse tumor xenograft models. This study is the first to directly compare the PET imaging properties of [89Zr]Zr-DFO*-T-DM1 and [89Zr]Zr-DFO-T-DM1 in a HER2-overexpressing tumor xenograft mouse model. Our results indicate that [89Zr]Zr-DFO*-T-DM1 provides superior imaging properties due to the greater stability of the [89Zr]Zr-DFO* than [89Zr]Zr-DFO complex.