couple more eortc posters
223Trastuzumab-DM1: mechanisms of action and mechanisms of resistance
G. LiC.T. Fields, K.L. Parsons, J. Guo, G.D. Lewis PhillipsGenentech Inc., Research Oncology South San Francisco USA; Genentech Inc., Translational Oncology South San Francisco USA
Overexpression of the HER2 receptor tyrosine kinase (RTK) occurs in approximately 20–25% of human breast cancer. Patients with elevated HER2 levels show more rapid disease progression and worse survival than breast cancer patients with HER2-negative disease. Trastuzumab (Herceptin®) is a humanized antibody developed specifically to treat HER2-positive breast cancer and is used in both the adjuvant and metastatic setting. Although trastuzumab is efficacious for many patients, a number of patients will have disease progression through trastuzumab therapy. We have recently developed an antibody–drug conjugate comprised of trastuzumab covalently linked through a stable linker to the microtubule inhibitory agent DM1, as an alternative treatment for patients whose cancers overexpress HER2. Trastuzumab-MCC-DM1 (T-DM1) is currently undergoing clinical testing in metastatic HER2-positive breast cancer patients. We performed cell culture studies to compare the potency of T‐DM1 to chemotherapeutic agents and to determine T‐DM1 mechanisms of action (MOA). Viability and clonogenic assays on HER2-amplified breast cancer cells show that T‐DM1 is more potent than taxane or vinca alkaloid agents. Treatment with T‐DM1 results in G2/M cell cycle arrest, as measured in cell cycle experiments or by Western analysis of the G2/M markers cyclin B1 and phospho-histone H3. Subsequently, G2/M-arrested cells undergo apoptosis as indicated by PARP cleavage and loss of XIAP expression. Like trastuzumab, T‐DM1 suppresses signaling through the PI3 kinase pathway. However, T‐DM1 inhibits PI3 kinase signaling in trastuzumab-insensitive cells; this activity was shown to be mediated by the DM1 component of T‐DM1. Anti-proliferative activity of T‐DM1 was also compared to T‐DM1 components (trastuzumab, DM1, SMCC linker) and identified catabolites (Lys-MCC-DM1 and MCC-DM1). Only T‐DM1 and DM1 potently inhibit growth of HER2-overexpressing breast cancer cells. Mechanisms of resistance to T‐DM1 were investigated in 3 cell lines developed to have acquired T‐DM1 resistance. One T‐DM1-resistant line shows upregulation of multi-drug resistance (MDR) transporters. Increased expression of EGFR and other RTKs, as well as several erbB ligands was also observed. Mutations in beta 1‐tubulin were not detected. These studies demonstrate the potent anti-tumor activity of T‐DM1 compared to conventional chemotherapeutic agents, show multiple MOA for T‐DM1, and shed light on potential mechanisms of resistance.
235 Negatively-charged sulfonate group in linker improves potency of antibody–maytansinoid conjugates against multidrug-resistant cancer cells
Y. KovtunG. Jones, C. Audette, M. Mayo, B. Leece, R. Zhao, L. Clancy, X. Sun, R. Chari, R. SinghImmunoGen Inc., Cell Biology Waltham MA USA; ImmunoGen Inc., Pharmacology Waltham MA USA; ImmunoGen Inc., Chemistry Waltham MA USA; ImmunoGen Inc., Biochemistry Waltham MA USA
Antibody–maytansinoid conjugates (AMCs) targeting cancer cell-surface antigens are in clinical trials against several cancers. The AMCs in the clinic employ disulfide- or thioether-bonds to link cytotoxic maytansinoid molecules to the antibody. In pre-clinical models, disulfide-linked AMCs demonstrate superior in vivo activity compared to thioether-linked conjugates for several reasons including bystander killing. Cancer cells can be multi-drug resistant due to overexpression of drug efflux transporters such as MDR1. A goal of this study was to enhance the potency of disulfide-linked conjugates against multidrug-resistant cancer cells. Since MDR1 favors neutral substrates, we hypothesized that the incorporation of a negatively-charged sulfonate group in the disulfide linker would improve retention of the polar metabolite inside the cell and enhance conjugate potency to MDR1-expressing cells. We compared the cytotoxic potencies of disulfide-linked anti-EpCAM AMCs with a neutral linker (SPDB) and a negatively-charged linker (sulfo-SPDB) in several EpCAM-expressing cells with different levels of MDR1. The conjugates had similar cytotoxicities toward MDR1-negative cells, but the sulfo-SPDB conjugate was 10‐ to 30‐fold more potent than the SPDB conjugate toward MDR1-positive COLO 205MDR and HCT‐15 cells. An MDR1 inhibitor enhanced the cytotoxic potency of the SPDB conjugate to a level similar to that of the sulfo-SPDB conjugate. The sulfo-SPDB linker allowed preparation of AMCs with high numbers of maytansinoid molecules per antibody (6–8), which showed increased cytotoxic potency over conjugates with the usual level of payload (3–4) against the MDR1 cells. Importantly, the sulfo-SPDB conjugate demonstrated greater activity than the SPDB conjugate against multidrug-resistant HCT-15 tumors in SCID mice. The advantage of the sulfo-SPDB linker was further demonstrated for anti-CanAg antibody conjugates against COLO 205MDR cells, where the sulfo-SPDB linked conjugate was more active than the SPDB conjugate in arresting cells in the G2‐M phase. Similar bystander killing activities were observed for sulfo-SPDB and SPDB conjugates: both conjugates eradicated mixed-cell populations of antigen-positive and antigen-negative cells in culture. Thus, we have developed a new linker that improves the potency of the disulfide-linked conjugate to multidrug resistant cancer cells, while preserving the conjugate activity towards non-resistant cells.
236 Antibody–maytansinoid conjugates targeting folate receptor 1 for cancer therapy
O. AbV.S. Goldmacher, L.M. Bartle, D. Tavares, C.N. Carrigan, S. Xu, M. Okamoto, H. Johnson, K.R. Whiteman, T. ChittendenImmunoGen Inc., Cell Biology Waltham MA USA; ImmunoGen Inc., Antibody engineering Waltham MA USA; ImmunoGen Inc., Translational research Waltham MA USA; ImmunoGen Inc., Biochemistry Waltham MA USA; Im
Background: Folate receptor 1 (FOLR1) is highly expressed in ovarian cancers and several other epithelial malignancies. We wished to examine if conjugates of anti-FOLR1 antibodies with the highly cytotoxic maytansinoid derivative, DM4, would be effective in antigen-selective elimination of FOLR1-expressing cancer cell lines in vitro and in eradication of FOLR1-expressing xenograft tumors in mice.
Materials and Methods: A panel of anti-FOLR1 monoclonal antibodies was humanized using ImmunoGen's resurfacing technology. Affinities of these antibodies were examined on FOLR1-expressing cells by flow cytometry. Antibodies were conjugated to DM4 using the disulfide-containing SPDB linker by previously described methods. On the average, these conjugates contained 3.5 to 4 DM4 molecules per antibody. The cytotoxic activity of the conjugates in vitro and anti-tumor activity in vivo were analyzed on the cultured FOLR1-positive KB cell line and on KB-derived xenograft tumors in immunodeficient mice, respectively. Data for a representative conjugate huFR107–SPDB–DM4 are reported here. Immunohistochemistry (IHC) was performed on formalin fixed paraffin embedded ovarian carcinoma arrays with monoclonal antibody BN3.2 (Leica).
Results: The humanized FOLR1 antibody, huFR107, bound to FOLR1-expressing cells with a KD of 0.1 nM, and the huFR107–DM4 conjugate retained similar high affinity binding. The huFR107–SPDB–DM4 conjugate was potent in killing KB cells in vitro (IC50 of 70 pM). This activity was antigen-selective, since the cytotoxicity of huFR107–SPDB–DM4 for KB cells was at least 300‐fold lower in the presence of an excess of huFR107 and also for FOLR1-negative cells. HuFR107–SPDB–DM4 was highly active in eradicating subcutaneous KB xenografts in mice. A single intravenous injection of the conjugate at 5 mg/kg completely eradicated the tumors, while tumor growth in mice treated with a non-targeting huAb–SPDB–DM4 conjugate was similar to that of PBS-treated control mice. IHC evaluation revealed that the expression of FOLR1 in KB xenografts was comparable to that found on 57% of ovarian clinical tumors (N = 67).
Conclusions: Anti-FOLR1–DM4 conjugates were found to exhibit specific and highly potent activity against FOLR1-expressing cancer cells, both in vitro and in vivo. Our results suggest that antibody-maytansinoid conjugates targeting FOLR1 constitute a promising approach for the treatment of FOLR1-expressing tumors.
237 Mechanistic pharmacokinetic/pharmacodynamic (PK/PD) modeling of xenograft tumor response of Trastuzumab‐DM1-antibody drug conjugates
R. WadaH.K. Erickson, G. Lewis Phillips, C.A. Provenzano, D.D. Leipold, J. Pinkas, E. Mai, M. Gupta, H. Johnson, J. TibbittsQuantitative Solutions Menlo Park USA; ImmunoGen Inc. Waltham USA; Genentech Inc., Research Oncology South San Francisco USA; Genentech Inc, Pharmacokinetics and Pharmacodynamics South San Francisco U
Trastuzumab-DM1 (T-DM1) is an antibody–drug conjugate (ADC) in development for the treatment of HER2+ metastatic breast cancer. Nonclinical studies of T‐DM1, which employs a non-reducible thioether linker, showed slightly greater efficacy in mouse models of HER2+ breast cancer than T‐SPP-DM1, which employs a reducible disulfide linker. Previous studies of the two conjugates also found differences in the pharmacokinetics, uptake of conjugate into tumors, and the active products of intracellular catabolism of the conjugates, but similarities in the conjugate in-vitro potency, in vitro conjugate catabolism kinetics, and tumor catabolite concentrations. The objective of this study was to use PK/PD modeling to explore the differences in plasma and tumor conjugate and catabolite concentrations between these two ADCs and to better understand the mechanisms for their relative efficacy. A mechanistic PK/PD model was assembled which allowed prediction of tumor conjugate and catabolite concentrations from plasma PK data. The tumor catabolite concentrations, the presumed active agent, were then used to predict tumor response. This model is consistent with the proposed mechanism of action of ADCs.
The PK/PD model fit the data well, based on visual inspection and evaluation of estimate error. Tumor response was well-predicted from catabolite concentrations. Consistent with similarities seen in in vitro efficacy and tumor response with T‐DM1 and T‐SPP-DM1, the tumor PD based on catabolite concentrations were similar between T‐DM1 and T‐SPP-DM1 with a time to cell death of approximately 1–1.5 days, a maximal tumor kill rate of approximately 0.3 day−1, and a 50% tumor kill concentration of approximately 180 pmol catabolite/gram of tumor. Tumor catabolism half-life was 1 day, similar to in vitro data. Differences in the catabolite efflux rate were found to explain the inconsistency between tumor conjugate concentrations and catabolite concentrations noted for these ADCs.
This study describes the use of a mechanistic model of ADC PK/PD, incorporating knowledge from tumor uptake and catabolism studies to improve the understanding of the pharmacologic behavior of these molecules.