Abstract 2439: Cross-linker evaluation in the design of antibody-SN-38 conjugates for cancer therapy

Antibody (MAb) conjugates of the potent topoisomerase I inhibitor, SN-38, prepared using the bifunctional derivative, ‘CL2-SN-38’ (maleimido-[x]-Phe-Lys-PABOCO-20-O-SN-38), were previously shown to produce significant and selective therapeutic efficacies of human tumor xenografts in nude mice. With a view to examining alternative linkers for possible improvement, we prepared a variant, namely maleimido-[x]-Lys-PABOCO-20-O-SN-38 (‘CL2A-SN-38’), and two other derivatives with cross-linker attachment at the 10-position of SN-38 instead of the 20-position: maleimido-[x]-Phe-Lys-N(Me)-(CH2)2-N(Me)-CO-10-O-SN-38 (‘CL2D-SN-38’) and maleimido-[x]-Phe-Lys-PABOCO-N(Me)-(CH2)2-N(Me)-CO-10-O-SN-38 (‘CL2E-SN-38’). In these, PAB is p-aminobenzyl moiety and ‘x’ contains a short PEG group to increase solubility. Conjugates of the humanized anti-CEACAM5 MAb, hMN-14 (labetuzumab), were prepared by reacting the maleimide-appended SN-38 derivatives with mildly reduced hMN-14. The SN-38/MAb molar substitutions were in the range of 5.9 to 6.3. In vitro studies showed a similarity in the stabilities, in mouse serum, of CL2 and CL2A-based linkers, but an order of magnitude greater stability for conjugates with linker at the 10-position; however, CL2D-SN-38 linker was inert to dipeptide cleavage by cathepsin B at the lysosomal pH of ∼ 5. The latter finding was also reflected in the loss of cytotoxicity due to hMN-14-CL2D-SN-38 conjugate in the LS174T colon cancer cell line in vitro. All conjugates were also evaluated in an exploratory preclinical therapy experiment in nude mice with aggressive s.c. LS174T human colon cancer xenografts. Conjugates were administered i.p. in groups of tumor-bearing animals (n = 10) at a schedule of 0.39 mg/kg of SN-38 equivalent twice weekly × 2 weeks. Untreated animals received saline. Mean tumor volumes (cm3) of animals in various treatment groups, on day 18, were: 0.073 ± 0.035 for the hMN-14-CL2-SN-38 positive control group; 0.055 ± 0.032 for the hMN-14-CL2A-SN-38 group; 0.518 ± 0.239 for the hMN-14-CL2D-SN-38 group; 0.490 ± 0.224 for the hMN-14-CL2E-SN-38 group; and 1.09 ± 0.88 for the untreated group. Median survival times (days) were 54.5, 55, 30, 25, and 26.5 for these groups, respectively. These results showed that a cleavable peptide was not necessary in the SN-38 conjugate design, as hMN-14-CL2A-SN-38 was equivalent to hMN-14-CL2-SN-38 in efficacy, and that hMN-14-CL2E-SN-38 was surprisingly much less efficacious than the CL2 and CL2A-base