Preclinical characterization of novel LIV1 antibody drug conjugates

Background: LIV1 is a member of the zinc transporter family. With limited normal tissue expression, LIV1 was found to be overexpressed with high prevalence in breast (93%), prostate (72%) and lung (10%) cancers, and considered as an attractive cell surface target for developing ADC therapeutics. We have generated 48D6, a proprietary novel humanized anti-LIV1 mAb with high affinity, specificity, internalization ability, unique epitope and improved pharmacokinetics (PK) profile in mice. Then we developed 48D6 based ADCs using glycotransferase mediated site-specific conjugation with Topo I inhibitor (ADC-2) or MMAE (ADC-3), and characterized their anti-tumor activities in preclinical PDX/CDX models and exploratory toxicity in mice. Methods: PK properties were studied with single dose (i.v.) of 3 or 10 mg/kg of both naked antibodies and ADCs in Balb/c mice. To evaluate anti-tumor activities, nude mice were subcutaneously implanted with LIV1-expressing PDX tumor blocks. When tumors reached ~200mm3, mice were treated (i.v.) with isotype control-ADC or LIV1 ADCs. To explore the combination efficacy of LIV1 ADCs and PD-1 antibody, we established LIV1/PDL1 co-expressing NCI-H460-LIV1 CDX model in human PBMC reconstituted mice. For exploratory tox study, mice were i.v. injected with LIV1 ADC or 48D6 at 10, 30, 60 mg/kg every week for 3 times, then recovered for 5 weeks. Results: The T1/2 of of 48D6 antibody is 13.8~15.6 days relative to 1.7~3.3 days for Ladiratuzumab analog in mice. After conjugation into ADCs, the T1/2 of 48D6 ADC-2 ranged 10.4~11.6 days while that of SGN-LIV1A analog in mice is 3.7~3.9 days. Topo I inhibitor payload based ADC-2 displayed strong anti-tumor activities in LIV1 expressing NSCLC and breast PDX models. However, for LIV1 expressing prostate PDX models, two doses of ADC-2 did not inhibit tumor growth significantly. Thus MMAE-based ADC-3 replaced ADC-2 from the 3rd dose. Interestingly ADC-3 inhibited the growth of the prostate tumor significantly. In a LIV1 high expressing prostate PDX, the tumor growth was suppressed by ADC-3 for over 70 days after the dosing was stopped on Day 28. In the LIV1/PDL1 co-expressing NCI-H460-LIV1 CDX model, combination group inhibited tumor growth significantly better than single agent group. In exploratory tox study, ADC-2 were well tolerated following repeated administrations in mice at all doses tested. Slight lesions were observed in 60 mg/kg group during the treatment period and fully recovered at the end of the recovery period. Based on these results, the maximum tolerated dose (MTD) of ADC-2 in mice was determined at 60 mg/kg. Conclusions: LIV1 targeting ADC-2 and ADC-3 exhibited strong anti-tumor activities as monotherapy in PDX models and the anti-tumor activity is further enhanced by checkpoint inhibitor. ADC-2 displayed excellent tolerability profile in mice. These results support further investigation of our LIV1 ADCs in LIV1 positive solid tumors.