Dosimetric evaluation and radioimmunotherapy of anti-tumour multivalent Fab' fragments

We have been investigating the use of cross-linked divalent (DFM) and trivalent (TFM) versions of the anti-carcinoembryonic antigen (CEA) monoclonal antibody A5B7 as possible alternatives to the parent forms (IgG and F(ab')(2)) which have been used previously in clinical radioimmunotherapy (RIT) studies in colorectal carcinoma, Comparative biodistribution studies of similar sized DFM and F(ab), and TFM and IgG, radiolabeled with both (131)I and (90)Y have been described previously using the human colorectal tumour LS174T nude mouse xenograft model (Casey et al (1996) Br J Cancer 74: 1397-1405). in this study quantitative estimates of radiation distribution and RIT in the xenograft model provided more insight into selecting the most suitable combination for future RIT. Radiation doses were significantly higher in all tissues when antibodies were labelled with (90)Y. Major contributing organs were the kidneys, liver and spleen, The extremely high absorbed dose to the kidneys on injection of (90)Y-labelled DFM and F(ab')(2), as a result of accumulation of the radiometal would result in extremely high toxicity. These combinations are clearly unsuitable for RIT, Cumulative dose of (90)Y-TFM to the kidney was 3 times lower than the divalent forms but still twice as high as for (90)Y-IgG. TFM clears faster from the blood than IgG, producing higher tumour to blood ratios. Therefore when considering only the tumour to blood ratios or the total absorbed dose, the data suggests that TFM would be the most suitable candidate. However, when corrected for equitoxic blood levels, doses to normal tissues for TFM were approximately twice the level of IgG, producing a two-fold increase in the overall tumour to normal tissue ratio. In addition RIT revealed that for a similar level of toxicity and half the administered activity, (90)Y-IgG produced a greater therapeutic response. This suggests that the most promising A5B7 antibody form with the radionuclide (90)Y may be IgG. Dosimetry analysis revealed that the tumour to normal tissue ratios were greater for all (131)I-labelled antibodies. This suggests that (131)I may be a more suitable radionuclide for RIT, in terms of lower toxicity to normal tissues. The highest tumour to blood dose and tumour to normal tissue ratio at equitoxic brood levels was (131)I-labelled DFM, suggesting that (131)I-DFM may be best combination of antibody and radionuclide for A5B7. The dosimetry estimates were in agreement with RIT results in that twice the activity of (131)I-DFM must be administered to produce a similar therapeutic effect as (131)I-TFM. The toxicity in this therapy experiment was minimal and further experiments at higher doses are required to observe if there would be any advantage of a higher initial dose rate for (131)I-DFM. (C) 1999 Cancer Research Campaign.