Inhibition of tumor metastasis: functional immune modulation of the CUB domain containing protein 1

Abstract

Despite significant progress and notable successes in tumor therapy, malignant disease remains an extremely difficult problem in today's health care setting. There is, however, an increasing application of new therapies targeting proteins specifically upregulated on tumor cells. These innovative therapeutic approaches are aimed at molecules that contribute to malignant development and progression but spare normal tissues. The CUB domain containing protein 1 (CDCP1) is such a tumor-associated protein and, thus, a potential candidate for targeted cancer immunotherapy. Herein, we describe the generation of function-blocking human antibodies against CDCP1 that were obtained from human scFv phage display libraries using subtractive panning protocols on CDCP1 expressing cancer cells and immunopurified CDCP1 protein. One of the isolated anti-CDCP1 antibodies, namely, C20Fc, efficiently blocked experimental metastasis of human carcinoma cells, including HeLa cells stably transfected with CDCP1 and prostate carcinoma cells PC-hi/diss naturally expressing CDCP1, in both chick embryo and mouse model systems. The C20Fc antibody also reduced colony formation of CDCP1 expressing cells in a soft agar assay for anchorage-independent cell growth. Specific targeting of CDCP1 by C20Fc mediated the delivery of a toxin-conjugated antibody complex, thus, providing evidence for antibody internalization and specific killing of CDCP1-positive tumor cells. Our findings indicate a functional role for CDCP1 in human cancer and underscore the therapeutic potential of function-blocking anti-CDCP1 antibodies targeting both primary and metastatic carcinoma cells.

Figure 1

Selected scFv antibodies on phage by whole cell ELISA. Positive scFv antibody expressing phage clones were selected by whole cell ELISA with CDCP1 expressing HeLa cells (HeLa-CDCP1) and Neo expressing HeLa cells (HeLa-neo). Data are the mean of triplicate wells. Error bars indicate the SD.

Figure 2

Alignment of anti-CDCP1 antibodies. Multiple sequence alignment was performed with AlignX (Invitrogen). Identical residues of three antibodies are shown in red, of two antibodies are shown in blue. Different residues are shown in black and weakly similar in green. CDR regions of antibodies, linker and the Fc hinge are indicated above the sequences.

Figure 3

Binding of scFv-Fc antibodies to CDCP1 positive cells. scFv-Fc antibodies recognize cancer cells based on their expression of CDCP1. Several independent experiments on each of these cell types gave similar results. Shown are flow cytometric analyses with binding of scFv-Fc antibodies to cells detected as fluorescence signal of fluorescein labeled anti-human IgG Fc goat 2^nd antibody. 41-2 mouse monoclonal antibody was used as a positive control of anti-CDCP1 binding with using a fluorescein labeled anti-mouse IgG goat 2^nd antibody.

Figure 4

Experimental metastasis of CDCP1 positive cells in chick embryos (A) and mice (B) is sensitive to anti-CDCP1 C20Fc. (A) HeLa-CDCP1 cells (left panel) or PC-hi/diss (right panel) were injected i.v. into chick embryos (5×10⁴ cells per embryo) along with 25 μg of control A4Fc or anti-CDCP1 scFv-Fc C14Fc, C20Fc and C25Fc. On day 5 following inoculations, the levels of CAM colonization were determined by Alu-qPCR. *, P<0.05 in unpaired t-test with Welch's correction; **, P<0.05 in two-tailed unpaired t-test. (B) HeLa-CDCP1 cells (left panel, 1 × 10⁶ cells/mouse) or PC-hi/diss (right panel, 5×10⁵ cells/mouse) were injected i.v. into immunodeficient NOD-SCID mice (4–6 mice per group) along with 100 μg of control A4Fc or anti-CDCP1 C20Fc. Additional injections of corresponding antibodies were performed i.p. on days 1 and 2 after cell injections. Levels of lung colonization by tumor cells were determined 4 weeks (HeLa-CDCP1 cells) and 8 weeks (PC-hi/diss cells) after cell inoculations. Data are means ± SEM of numbers of human cells per 10⁶ host cells determined by qPCR. ***, P<0.05 in a two-tailed unpaired t-test.

Figure 5

Antibody effect for anchorage independent colony formation of cancer cells in soft agar. Colony formation of CDCP1 expressing cell lines, PC-3LC and HeLa-CDCP1 or CDCP1 negative cell line, HeLa-Neo, was studied in 0.3 % LMA containing complete media for 2 weeks. The number of colonies was counted in eight random fields from each plate using a microscope. Data shown are the mean ± S.D for the number of colonies in each field. Significant difference is indicated by asterisks (*) when P value was less than 0.05 (t-test).

Figure 6

CDCP1 antibody-mediated internalization and tumor cell killing. Internalization was studied using anti-CDCP1 antibodies with an appropriate saporin secondary conjugate for indirect immunotoxin experiments. A PBS vehicle control without antibody served as blank. Pre-incubated 40 ng primary antibodies and 40 ng/well of goat anti-mouse saporin (mAb-ZAP) or anti-human secondary saporin (Hum-ZAP) conjugates were added to each well. Percent cell viability compared to nontreated control was determined for primary antibodies with appropriate secondary saporin conjugate. Experiments were done in triplicate over three plates, and a representative graph for the study is shown.