Abstract 52: Mechanistic and benchmarking studies of ADCT-502, a pyrrolobenzodiazepine (PBD) dimer-containing antibody-drug conjugate (ADC) targeting HER2-expressing solid tumors

ADCT-502 is an ADC composed of an engineered version of humanized IgG1 trastuzumab, directed against human HER2, site-specifically conjugated to the highly cytotoxic PBD-based linker-drug tesirine (drug-antibody ratio of 1.7). In vitro, ADCT-502 has highly potent and targeted cytotoxicity against various solid cancer cell lines. In vivo, ADCT-502 demonstrates strong and durable antitumor activity in mouse xenografts with various levels of HER2, but is inactive in a HER2-negative xenograft. ADCT-502 is stable, well tolerated and has a favorable PK profile both in rat and cynomolgus monkey. The current study aimed to define further the mechanism of action of ADCT-502 and to benchmark its activity in xenograft models against ado-trastuzumab emtansine (T-DM1), the ADC currently approved for the treatment of HER2+ metastatic breast cancers. ADCT-502 bound and internalized efficiently in JIMT-1 cells (HER2+) and co-localized with lysosomes within 2 hours. PBD dimers bind in the DNA minor groove and exert cytotoxicity via the formation of DNA interstrand cross-links. Following a 2-hour exposure to ADCT-502, DNA interstrand cross-linking peaked between 12 and 24 hours, after which cross-links persisted at least 36 hours. In contrast, cross-link formation by an equimolar concentration of warhead alone, peaked immediately following drug exposure and a non-targeted ADC did not produce DNA crosslinks in these cells. Moreover, ADCT-502 showed indirect bystander killing activity in HER2-negative MDA-MB-468 cells incubated with conditioned medium from ADCT-502-treated HER2+ SK-BR-3 cells. In vivo, antitumor activity of ADCT-502 was compared to T-DM1 in both cell line- and patient-derived-xenograft (PDX) models. For example, in a HER2 1+, FISH- breast cancer PDX, ADCT-502 showed dose-dependent antitumor activity resulting in 1/8 and 8/8 TFS after a single dose at 0.1 and 0.2 mg/kg, respectively. Conversely, a single dose of T-DM1 at 30 mg/kg showed only marginal activity compared to the control. Similarly, in a HER2 1+, FISH- esophageal cancer PDX, while a single dose of ADCT-502 at 0.44 mg/kg resulted in strong and durable antitumor activity, single doses of T-DM1 at either 10 or 30 mg/kg showed no activity compared to the control. These data confirm that the mechanism of cell killing of ADCT-502 is via target-specific internalization and subsequent cross-linking of DNA. They also show superior in vivo antitumor activity of ADCT-502 compared to T-DM1 in various tumor xenografts, including those with low HER2 levels. Taken together, these results support the development of ADCT-502 not only in patients that have become resistant/refractory to T-DM1, but also in patients whose tumors express low levels of HER2, and are not eligible for treatment with T-DM1.