Direct radiolabeling of proteins can result in the loss of targeting activity, requires highly customized procedures, and yields heterogeneous products. Here we describe a novel imaging complex comprised of a standardized (99m)Tc-radiolabeled adapter protein noncovalently bound to a "Docking tag" fused to a "Targeting protein". The assembly of this complex is based on interactions between human 109-amino acid (HuS) and 15-amino acid (Hu-tag) fragments of ribonuclease I, which serve as an "Adapter protein" and a Docking tag, respectively.
Methods: HuS modified with hydrazinonicotinamide (HYNIC) was radiolabeled using (99m)Tc-tricine to a specific activity of 3.4-7.4 MBq/microg. Protein complexes were then formed by mixing (99m)Tc-HuS with equimolar amounts of either Hu-tagged VEGF(121) (Hu-VEGF [vascular endothelial growth factor]) or Hu-tagged anti-VEGFR-2 single-chain antibody (Hu-P4G7) and incubating on ice for 15 min. 4T1 luc/gfp luciferase-expressing murine mammary adenocarcinoma cells (1 x 10(4)) were implanted subcutaneously or injected intravenously into BALB/c mice. Bioluminescent imaging (BLI) was performed 10 d later. Immediately after BLI visualization of tumor, 18.5-37 MBq of tracer (5-10 microg of protein) were injected via tail vein. One hour later planar or SPECT images were obtained, followed by killing the mice.
Results: There was significantly (P = 0.0128) increased uptake of (99m)Tc-HuS/Hu-VEGF (n = 10) within subcutaneous tumor as compared with (99m)Tc-HuS/Hu-P4G7 (n = 5) at biodistribution assay (2.68 +/- 0.75 vs. 1.8 +/- 0.21; tumor-to-subcutaneous tissue [ratio of specific activities], respectively), despite similar molecular weights. The focal (99m)Tc-HuS/Hu-VEGF uptake seen on planar images (3.44 +/- 1.16 [tumor to soft-tissue background]) corresponded directly to the locations of tumor observed by BLI. Region of interest analyses of SPECT images revealed a significant increase of (99m)Tc-HuS/Hu-VEGF (n = 5) within the lungs with BLI-detectable pulmonary tumor nodules as compared with controls (n = 4) (right: 4.47 +/- 2.07 vs. 1.79 +/- 0.56; left: 3.66 +/- 1.65 vs. 1.62 +/- 0.45, tumor lung [counts/pixel]/normal lung [counts/pixel], respectively).
Conclusion: (99m)Tc-HuS/Hu-VEGF complex is stable for at least 1 h in vivo and can be effectively used to image mouse tumor neovasculature in lesions as small as several millimeters in soft tissue. We expect that a similar approach can be adapted for in vivo delivery of other targeting proteins of interest without affecting their bioactivity.