Web Exclusive

Clinical advancement of EpCAM-specific antibodies and EpCAM’s newly discovered role in cancer development are putting the target back in center stage.

In order to be active, antibody-based cancer therapies require a target antigen firmly bound to the surface of cancer cells. By binding to the surface target, the antibody can deliver a deadly signal to the cancer cell. In an ideal treatment scenario, a target antigen is abundantly present and accessible on every cancer cell and is absent, shielded, or much less abundant, on normal cells. This situation provides the basis for a therapeutic window in which a defined amount of the antibody-based therapeutic effectively hits cancer cells, but spares normal cells.

Table 1. Criteria for selection of surface target antigens for antibody-based therapeutic approaches to treat cancer. High frequency and high level of target antigen expression on individual cells of a tumor High frequency of antigen expression within the selected patient population No reduction or loss of antigen with tumor progression and metastasis Specific tumor cell surface binding Absence of substantial antigen shedding No substantial internalization of target antigen upon antibody binding Either absence, low-level expression or inaccessibility of target antigen on cells of normal tissue No reduction or loss of antigen expression by concomitant conventional therapies A negative correlation of target expression with survival parameters of patients A functional role of the antigen for tumor biology such that antigen loss is of considerable disadvantage for tumor cells Expression on cancer stem cells (Source: Micromet)

Surface antigens for antibody-based therapies need several more properties to qualify as targets, a selection of which is listed in Table 1. Certain antibody-based therapeutics, such as immunotoxins, rely on internalization of the target-bound payload, while other approaches need the target antigen to stay on the cell surface in order to allow sustained engagement of cytotoxic cells of the patient’s immune system. Shedding of the target antigen is not desirable as the soluble antigen will absorb the antibody in the blood stream, requiring compensation of the lost antibody by higher dosing. What is desirable is for the target antigen to act in an important functional role for the phenotype of cancer cells, because this will reduce the chance of selecting escape variants lacking the target antigen. Even if target-negative cancer cells remain after treatment, these may be severely handicapped without the surface antigen’s function in support of survival, proliferation, invasiveness, anti-apoptosis, or neovascularisation of cancer cells.

Surface antigens currently targeted by marketed antibody therapeutics teach us a great deal about the ideal properties of a target antigen.

For the treatment of solid tumors, these are:

• Her-2/neu (Herceptin, Roche, Genentech)
• EGFR (Erbitux, BMS, ImClone, Merck Serono; Vectibix, Amgen)

For the treatment of blood-borne tumors, these are:

• CD20 (Rituxan/Mabthera, Roche, Genentech; Bexxar, GSK)
• CD20 (Bayer Schering, Cell Therapeutics)
• CD52 (CAMPATH, Bayer Schering)
• CD33 (Mylotarg; Wyeth)

Important properties of Her-2/neu and EGFR are that both are signalling molecules for which cancer cells show a high dependence, and that cancer cells contain a much higher copy number of molecules compared to normal cells, a fact which contributes to a therapeutic window.

A large number of target antigens have been identified and investigated during the past 30 years for potential use in antibody-based therapies of solid tumors. It is, however, surprising that, until today, only two cell surface antigens, Her-2/neu and EGFR, are being targeted by marketed products. On one hand, this may be due to certain shortcomings of most potential targets. On the other hand, the antibodies’ mode of action selected for a particular surface target antigen may not have been the most appropriate one in the cases where clinical trials failed due to lack of efficacy. With the recent positive opinion of the European Medicines Agency (EMEA) to grant market authorization to the antibody catumaxomab (Removab; Fresenius Biotech, Gräfelfing, Germany), a third target is likely to be added to this short list. With insulin-like growth factor type 1 receptor, a fourth surface antigen is being successfully used for treatment of solid tumors by several antibodies, which are in various stages of clinical development.

Catumaxomab, which will be used to treat cancer patients with malignant ascites,¹ is specific for a well-established, tumor-associated surface target antigen called epithelial cell adhesion molecule (EpCAM). EpCAM is currently being targeted by several more antibody-based therapeutic approaches, which are in different stages of clinical development (Table 2). EpCAM was first described 30 years ago as 17-1A antigen, eliciting a dominant antibody response in mice to human colon carcinoma tissue.² Owing to its high immunogenicity in mice and frequent expression by various human cancers, EpCAM was later independently described by many other laboratories as a tumor-associated antigen. This led to more than a dozen synonyms for EpCAM, and more than 500 publications. The first monoclonal antibody ever administered for cancer therapy to humans was directed at EpCAM and is called CO17-1A.³ It was later named edrecolomab, or Panorex (Glaxo Wellcome, Centocor), and was the first to show signs of clinical activity in cancer patients as an adjuvant treatment after surgical resection of primary colorectal tumors when compared to best supportive care.⁴ While these data led to market approval in Germany, subsequent larger studies showed an inferiority of Panorex vis-à-vis an established chemotherapy,⁵ leading to withdrawal of its market authorization.

Table 2. EpCAM-directed, antibody-based approaches for cancer treatment in clinical development. (Source: Micromet)
Therapeutic Compound (Alternative Name)	Class	Development Status	Indication	Company
Catumaxomab (Removab®)	Bispecific, trifunctional antibody (hybrid of mouse IgG2a and rat IgG2b)	Phase 3 Positive CHMP (EMEA) opinion granted February 2009	Ovarian cancer Gastric cancer Lung cancer	Trion Pharma/ Fresenius Biotech
Proxinium® Vicinium® (VB4-845)	Immunotoxin (single-chain antibody fused to pseudomonas exotoxin)	Phase 2/3 Phase I/II	Head & neck cancer Bladder cancer	Viventia
Adecatumumab (MT201)	Human monoclonal IgG1	Phase 2 Phase 1	Breast, prostate and colorectal cancer Prostate, metastatic breast cancer (+ docetaxel)	Micromet Inc./ Merck Serono
MT110	BiTE® antibody	Phase 1	Gastrointestinal and lung cancer	Micromet Inc.
ING-1	Human monoclonal IgG1	Phase 1	Various adenocarcinomas	Xoma Inc.
EMD 273 066 (huKS-IL2)	Antibody/cytokine fusion (humanized mAb KS1/4 with human IL-2)	Phase 1	Hormone-refractory prostate cancer	Merck Serono

Only very recently, new biological functions of EpCAM were discovered that are similar to those that initially qualified Her-2/neu and EGFR as target antigens for antibody treatment of breast and colorectal cancer patients, respectively. A study in the January 2009 issue of Nature Cell Biology⁶ showed that when EpCAM was expressed as a complete protein, or in the form of its small intracellular portion in non-malignant cells, these cells could form tumors in mice. Moreover, EpCAM was found to be a signalling protein, very much like Her-2/neu and EGFR, which can promote cancer cell proliferation. EpCAM can activate an intracellular pathway leading to up-regulation of proto-oncogene c-myc and to cyclins in the cell nucleus.

But how can a membrane protein like EpCAM reach the cell nucleus? The new study revealed that when EpCAM molecules bind to each other on the cell surface, two kinds of proteases in the cell membrane are induced to clip off a short intracellular piece of EpCAM, called EpICD.⁶ The small EpICD peptide then combines in the cytoplasm with adaptor protein FHL-2 and beta-catenin, moves as part of a large complex into the cell nucleus, and once there, engages a transcription factor called LEF/TCF (Fig. 1). Beta-catenin and LEF are likewise signalling components of the so called wnt pathway, which is implicated in growth and survival of cancer and pluripotent stem cells. In human colorectal cancer tissue, fluorescent antibodies specific for EpICD light up the cell nucleus, while in normal epithelial tissues they light up the membrane-bound form, showing that only cancer cells contain the activated, nuclear form of EpCAM.

click to enlarge Figure 1. Nuclear signalling by EpCAM. In normal cells (left), EpCAM is intact and EpCAM’s intracellular domain (EpICD) found associated with the plasma membrane. In cancer cells (right), EpICD is found in the cell nucleus in a large complex with transcription factor TCF/LEF, -catenin and adaptor protein FHL2. One signal inducing the release of EpICD from membrane-bound EpCAM by two proteases (TACE and presenilin-2) is homophilic association of two neighbouring EpCAM molecules. For details, see reference no. 6. (Source: Micromet)

These novel findings explain a number of features of EpCAM that, in the past, were difficult to reconcile with a mere function of EpCAM as a cell adhesion molecule. First, EpCAM is found on most human adenocarcinoma, including cancers of colorectal, breast, lung, gastric, bladder, prostate, ovarian, and pancreatic origin. For instance, in colorectal cancer, more than 98% of patients show an intense and frequent expression of EpCAM on cancer cells in the primary tumor.⁷ Second, EpCAM is not lost from cancer cells when they de-differentiate and progress to the metastatic stage. Third, in some cancers, such as breast, ovarian, and certain squamous cell carcinomas, EpCAM expression is either de novo or highly up-regulated compared to normal epithelial tissues. Fourth, when EpCAM expression is knocked down in cancer cells by anti-sense or siRNA, cells cease to proliferate, move and invasively grow in soft agar. Conversely, ectopic expression of EpCAM in quiescent cells confers these properties, and leads to their serum growth factor-independent growth.⁸ Fifth, EpCAM has now been added to the list of cancer stem cell markers.⁹ Cancer stem cells are thought to constantly repopulate tumors and to be responsible for chemoresistance and tumor relapse. EpCAM expression has been found on cancer stem cells derived from breast, colon, prostate, liver, and pancreas tumors. Sixth, EpCAM antibodies are being used for isolation of circulating tumor cells, which have a significant prognostic potential for disease progression.¹⁰ All these features are well explained by the newly discovered signalling and oncogenic function of EpCAM, and reinforce its value as a target antigen for treatment of cancer.

Like Her-2/neu and EGFR targets, EpCAM is also expressed on normal tissues. For instance, a variety of epithelial cells in the gastrointestinal tract, bile ducts, and pancreas express EpCAM. This raised initial concerns with regard to a therapeutic window. In this context, it is important to note that two high-affinity monoclonal antibodies to EpCAM caused acute pancreatitis in patients.^11,12 On the other hand, edrecolomab, and certain other EpCAM-directed antibodies (see Table 2), did not show this particular adverse event in clinical trials. It appears that the binding affinity and epitope recognition need to be carefully selected for antibodies directed against this target antigen. It has been proposed that EpCAM on cancer cells is more accessible to antibodies than the EpCAM on normal epithelial tissue.^11,12 In the latter situation, EpCAM may be shielded in complexes with other membrane proteins, and sequestered within tight junctions of properly structured epithelia. However, on the surface of cancer cells, complexes of EpCAM with protein partners may be altered and certain epitopes of EpCAM no longer buried in epithelial structures. This differential accessibility of EpCAM is finding strong support from in vivo targeting studies with an antibody in mice that express transgenic human EpCAM on normal and tumor tissue.¹³

Apart from catumaxomab, which is a bispecific, trifunctional antibody binding in addition to EpCAM to CD3 on T-cells, and to antigen-presenting cells, there is also an EpCAM/CD3-bispecific antibody called MT110 is in clinical development.¹⁴ MT110 belongs to the class of BiTE antibodies, which can engage cytotoxic T cells for highly effective lysis of cancer cells. One representative, the CD19/CD3-bispecific antibody blinatumomab, has shown partial and complete tumor regression in lymphoma patients.¹⁵ MT110 is in a dose-escalating Phase 1 trial with gastrointestinal and lung cancer patients. Adecatumumab (MT201), a human IgG1 monoclonal antibody targeting EpCAM, has been investigated in various Phase 1 and 2 clinical trials in prostate and breast cancer patients. The developer of MT201, Micromet, is about to initiate a Phase 2 trial in colorectal cancer patients. Its safety profile is distinct from that of high-affinity anti-EpCAM antibodies supporting the existence of a therapeutic window.¹⁶ In retrospective analyses, breast cancer patients expressing high levels of EpCAM on their primary tumor have shown a significant increase in time to tumor progression, and an inhibition of the appearance of new lesions when treated with the higher of two doses of adecatumumab. This was surprising because, in other studies, EpCAM expression on tumors was shown to have a negative prognostic impact on the survival of breast cancer patients. Adecatumumab works in these patients also in combination with docetaxel, where objective tumor responses were seen in 43 percent of patients expressing high levels of EpCAM on their primary tumors, while none were observed in patients with low EpCAM expression.¹⁷

After many failed attempts to use EpCAM as target antigen, there are currently a number of approaches under development that may finally exploit this well-established, tumor-associated antigen in therapies for solid tumors.

References

1. A. Burges et al., Clin. Cancer Res. 13, 3899 (2007).
2. M. Herlyn et al., Proc. Natl. Acad. Sci. USA 76 : 1438 (1979).
3. D. Herlyn et al., Hybridoma 5 Suppl 1: S51 (1986).
4. G. Riethmüller, J. Clin. Oncol. 16: 1788 (1998).
5. C.J. Punt et al., Lancet 360: 671 (2002).
6. D. Maetzel et al., Nat. Cell Biol. 11: 162 (2009).
7. P. Went et al., Br. J. Cancer 94: 128 (2006).
8. M. Münz et al., Oncogene 23: 5748 (2004).
9. J.E. Visvader and G.J. Lindeman, Nat. Rev. Cancer 8: 755 (2008).
10. D.F. Hayes et al., Clin. Cancer Res. 12 : 4218 (2006).
11. P.A. Baeuerle and O. Gires, Br. J. Cancer 96: 417 (2007).
12. P.A. Baeuerle and G. Riethmüller, in: Tumor-associated antigens: identification, characterization and clinical application (O. Gires, B. Seliger, eds.) Wiley-VHC, Weinheim, pp. 179-200 (2009).
13. P.M. McLaughlin et al., Cancer Res. 61: 4105 (2001).
14. K. Brischwein et al., Mol. Immunol. 43: 1129 (2006).
15. R. Bargou et al., Science 321: 974 (2008).
16. C. Dittrich et al., Annual Meeting of AACR-NCI-EORTC, abstract no. A71 (2007).
17. M. Schuler et al., Annual Meeting of ESMO, abstract no. 485p (2008).

About the Authors
Patrick Baeuerle holds a PhD in biology from the Munich University, and did post-doctoral research at the MIT. In 1998, he joined Micromet from Tularik Inc., where he headed drug discovery. He was Germany's most frequently cited biomedical scientist of the past decade, and 38th worldwide.

Dominik Rüttinger, MD, PhD, is a board-certified surgical oncologist trained in tumor immunology at the Earle A. Chiles Research Institute in Portland, Oregon. He holds a faculty position at the University of Munich and serves as Senior Medical Director for Micromet. He is also an advisor to GTAC UK, UK’s Gene Therapy Advisory Committee.