Evaluation of the Anti-Tumor Activity of the Humanized Monoclonal Antibody NEO-201 in Preclinical Models of Ovarian Cancer

Abstract

Purpose: Despite high initial response rates with cytoreductive surgery, conventional chemotherapy and the incorporation of biologic agents, ovarian cancer patients often relapse and die from their disease. New approaches are needed to improve patient outcomes. This study was designed to evaluate the antitumor activity of NEO-201 monoclonal antibody (mAb) in preclinical models of ovarian cancer where the NEO-201 target is highly expressed. Experimental Design: Functional analysis of NEO-201 against tumor cell lines was performed by antibody-dependent cellular cytotoxicity (ADCC) assays. Binding of NEO-201 to tumor tissues and cell lines were determined by immunohistochemistry (IHC) and flow cytometry, respectively. Further characterization of the antigen recognized by NEO-201 was performed by mass spectrometry. Ovarian cancer models were used to evaluate the anti-tumor activity of NEO-201 in vivo. NEO-201 at a concentration of 250 g/mouse was injected intraperitoneally (IP) on days 1, 4, and 8. Human PBMCs were injected IP simultaneously as effector cells. Results: Both IHC and flow cytometry revealed that NEO-201 binds prominently to the colon, pancreatic, and mucinous ovarian cancer tissues and cell lines. Immunoprecipitation of the antigen recognized by NEO-201 was performed in human ovarian, colon, and pancreatic cancer cell lines. From these screening, carcinoembryonic antigen-related cell adhesion molecule 5 (CEACAM5) and CEACAM6 were identified as the most likely targets of NEO-201. Our results confirmed that NEO-201 binds different types of cancers; the binding is highly selective for the tumor cells without cross reactivity with the surrounding healthy tissue. Functional analysis revealed that NEO-201 mediates ADCC killing against human ovarian and colorectal carcinoma cell lines in vitro. In addition, NEO-201 inhibited tumor growth in the presence of activated human PBMCs in orthotopic mouse models of both primary and metastatic ovarian cancer. Importantly, NEO-201 prolonged survival of tumor-bearing mice. Conclusions: These data suggested that NEO-201 has an antitumor activity against tumor cells expressing its antigen. Targeting an antigen expressed in tumors, but not in normal tissues, allows patient selection for optimal treatment. These findings strongly indicate that NEO-201 warrants clinical testing as both a novel therapeutic and diagnostic agent for treatment of ovarian carcinomas. A first in human clinical trial evaluating NEO-201 in adults with chemo-resistant solid tumors is ongoing at the NIH clinical Center.

Keywords: antibody-dependent cellular cytotoxicity; carcinoembryonic antigen-related cell adhesion molecule (CEACAM6); carcinoembryonic antigen-related cell adhesion molecule 5 (CEACAM5); monoclonal antibody; natural killer cell; tumor-associated antigen.

Figure 1

Expression profile of the antigen recognized by NEO-201 in patient tumor tissues and human cancer cell lines. (A,B) NEO-201 binding in colon (n = 21), lung (n = 24), breast (n = 19), and ovarian (n = 11) cancers, lymphoma (n = 11), glioblastoma multiforme (n = 7), and melanoma (n = 11) was measured by immunohistochemistry (IHC). Additionally, the degree of expression of the antigen recognized by NEO-201 in 627 tissues from more than 10 different ovarian cancer subtypes including serous, germ cell, clear cell, endometroid, mucinous, sarcoma, transitional, and adenocarcinoma NOS, was evaluated by tissue microarray analysis (TMA). Tissues were scored for positive vs. negative expression of the antigen recognized by NEO-201 and for percentage of positive tumor tissue. Those tissues with a complete staining of the membrane in more than 10% of the samples analyzed were given a 2+ score, while those with a complete staining of the membrane in <10% of the tissue analyzed were given a 1+. (C) Cell pellet from the ovarian cancer cell lines OV90, PEO1, PEO4, and colorectal cancer cell lines SW480 and LS174T were screened for NEO-201 binding by IHC. (D) NEO-201 binding on cancer cell line model was confirmed by FACS analysis. Cells were incubated with NEO-201 PE-conjugated antibody and then analyzed.

Figure 2

NEO-201 binds to CEACAM 5 and 6. (A) To identify the antigen recognized by NEO-201, 100 μg of proteins extracted from OV90, CFPAC1, and OVCAR8 was initially immunoprecipitated with 1 μg, 10 ng of NEO-201 and protein A beads. One microgram of NEO-201 showed the best results in terms of protein isolation and was used to perform intraperitoneal (IP) injection for the protein sample that was then analyzed by mass spectrometry. Peptide suggesting the glycosylated form of CEACAM-5 and−6 was identified as the most likely antigen recognized by NEO-201. (B) Flow cytometry analysis of the tumor cell lines ASPC-1, BxPC-3, CFPAC-1, and LS174T was performed to discriminate the native form of CEACAM5 and CEACAM6 from the NEO-201-reactive variant forms of CEACAM5 and CEACAM6. NEO-201 binding on cancer cell line model was assessed using NEO-201 Pacific Blue conjugated antibody. (C) NEO-201 binding to different forms of CEACAMs, other than 5 and 6, was measured by ELISA. (D,F) To confirm the binding of NEO-201 to CEACAM-5 and−6, CEACAM-negative cells HEK293T were transfected with CEACAM-5 and 6 siRNA. The two proteins, which express a different molecular weight, were equally bound by NEO-201. As proof of concept, CEACAMs and NEO-201-positive cells LS174T were knock-down for CEACAM 5/6 genes, which resulted in an about 25% partial loss of NEO-201 binding when a single gene was knock-down, and a further decreased to about the 50% loss of NEO-201 binding by the combined knockdown of both CEACAMs. (E) Western blot confirmed that NEO-201 binds an antigen with two different molecular weights in cell line models.

Figure 3

NEO-201 decreases cell viability by activating natural killer (NK)-mediated antibody-dependent cellular cytotoxicity (ADCC) in vitro. (A) NEO-201-positive OV90 and LS174T cells lines were selected to assess the effect of NEO-201 on NK-mediated ADCC. Cells were incubated with either highly active natural killer (haNK), NEO-201 at a dose of 1 μg/ml, or a combination of NEO-201 with haNK at different effector-to-target (E:T) ratios. ADCC activity was assessed by non-radioactive ADCC assay. Experiment and results from (A) were analyzed according to the Celigo program manufacture as described in the Material and Methods section. (B) In order to select the best NK human donor to use for in vivo experiments, the OV90 cell line was incubated with the NK isolated from the PBMCs from six different donors in the presence of 1 μg/ml of NEO-201. To evaluate specific lysis, NK derived from a selected human donor were activated with IL-2 and used for the experiment in two different effector cells (E):target cells (T) ratios: 50:1 or 100:1. Chromium released assay was used to evaluated specific lysis. Experiments depicted in this figure reflect the mean ± SD of three independent experiments. Ordinary one-way ANOVA was used for statistical analysis. *P < 0.05.

Figure 4

Antiproliferative response to NEO-201 and survival in model of ovarian cancer in vivo. (A) OV90 cells were injected in the mice ovary bursa and observed for their ability to develop tumors. Tumors were allowed to grow until they reached at least 100 mm³, about 8 weeks post cell injection. Tumor size was followed by weekly ultrasound. After randomization, each mouse received one cycle of treatment, which consists of two doses of drug on day 1 and day 7 of week 10. Curves represent tumor volume when treated ×3 with either control {phosphate-buffered saline (PBS)/IgG}, PBMCs (8 million cultured in 200 U/ml IL2 overnight), NEO-201 (250 μg), or their combinations. Kruskal–Wallis test was used to evaluate significance (p = 0.6001). (B) Representative ultrasound images of tumor-bearing mouse ovaries. Purple lines indicate cross sectional measurements from which volume was calculated. (C) OV90 cells were injected in the mice peritoneum to reproduce an orthotopic model of peritoneal carcinomatosis from ovarian cancer spreading. Tumors were allowed to grow for 2 weeks before being randomized in the four groups of treatment as described above. Curves represent survival. Significance was evaluated by Mantel–Cox. *P < 0.05, **P < 0.01, ***P < 0.001.