Abstract: A circulating isoform of EGFR has been identified as an emerging serum biomarker in cancer patients. Serum concentrations of EGFR have been evaluated as screening, diagnostic, prognostic, and theragnostic markers for diverse human malignancies, including breast, ovarian, endometrial, colorectal, and lung cancer. Yet full-length EGFR is rarely shed from normal or malignant cells. Here we demonstrate that an alternately spliced EGFR isoform, termed ‘sEGFR,’ gives rise to the circulating sEGFR protein through proteolytic cleavage of a cell surface precursor. This cleavage event is inhibited by fibronectin and stimulated by both an anti-Δ5/α1 integrin antibody and the metalloprotease activator 4-aminophenylmercuric acetate. Preliminary results identify ADAM17/TACE as at least one mediator of sEGFR shedding. Shed sEGFR binds to EGF with high affinity. Two FDA-approved therapeutic anti-EGFR antibodies, cetuximab (Erbitux) and panitumumab (Vectibix), inhibit shedding of sEGFR, implicating sEGFR as a potential alternative target for these antibodies, consistent with previous studies indicating that baseline serum sEGFR concentrations may predict responsiveness to anti-EGFR-targeted therapy. These observations bear striking similarity to aspects of trastuzumab regulation of HER2, shedding, and, by analogy, may have implications for the design of future clinical trials using EGFR-directed therapeutics, including cetuximab and panitumimab. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 3624. doi:1538-7445.AM2012-3624

Abstract: Trastuzumab is the first Food and Drug Administration (FDA)‐approved therapeutic targeting a HER‐family receptor tyrosine kinase (HER2/ErbB2/neu). Although trastuzumab is effective in the treatment of HER2‐positive breast cancer, a substantial proportion of patients will not respond to trastuzumab‐based regimens (primary resistance), and those who do respond will often lose clinical benefits (i.e., secondary resistance). Although multiple mechanisms underlying the development of secondary trastuzumab resistance have been identified, few studies have specifically examined the basis of primary trastuzumab resistance. Here, we review these studies, which together demonstrate that trastuzumab induces phenotypic changes in tumor cells, even when they are not growth inhibited by trastuzumab, including changes in gene expression. These changes have important clinical implications, including the sensitization of malignant cells to other therapeutic drugs. In light of these observations, we propose that the conventional definition of resistance as it pertains to trastuzumab and, perhaps, to other targeted therapeutics, may require revision. The results of these studies will be useful in informing the direction of future basic and clinical research focused on overcoming primary trastuzumab resistance.

Abstract: In July, 2009 the United States’ Food and Drug Agency (FDA) approved the use of a new biologically-targeted drug called ‘cetuximab’ (Erbitux) for the treatment of patients with advanced colorectal cancer (CRC). Cetuximab is a chimeric antibody directed against the extracellular domain of the epidermal growth factor receptor (EGFR). Paradoxically, EGFR expression, as it is currently measured by commercially available assays, has not been useful for predicting who will respond to cetuximab. We propose that one problem with these assays is their inability to measure EGFR and soluble EGFR (sEGFR) isoforms independently. We previously have documented the expression of an sEGFR isoform in human tissues (1), and have generated antibodies directed against a 31-mer peptide (residues 604-634) located within the unique 78 amino acid carboxy-terminal sequence of sEGFR (2). Interestingly, this anti-sEGFR antibody recognizes a second protein of ∼140-kDa (p140), which we have identified as human α5-integrin. Biochemical and mass spectrometry analyses show that a short region in the calf domain of α5-integrin shares sequence identity with the 31-mer sEGFR peptide. We have used this 31-mer antibody with an antibody directed against the intracellular domain of full-length EGFR to develop a novel EGFR vs. sEGFR/α5-integrin expression assay. This assay demonstrates that sEGFR/α5-integrin is co-expressed with EGFR in normal as well as malignant tissues, including in a number of carcinoma-derived cell lines. We also observe that high expression of EGFR and sEGFR/α5-integrin in a large cohort of ovarian cancer patients defines a strong adverse prognostic determinant of survival. These results implicate EGFR, sEGFR, and integrins as cancer biomarkers and demonstrate the critical importance of independently assessing both full-length and soluble EGFR isoforms in prognostic EGFR expression assays. *–JAW and ATB contributed equally to this study (1). Reiter et al. (2001) Genomics. 71(1):1-20 (2). Christensen et al. (2002) Hybrid Hybridomics. 21(3):183-9 Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 101st Annual Meeting of the American Association for Cancer Research; 2010 Apr 17-21; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2010;70(8 Suppl):Abstract nr 4649.

Abstract: The epidermal growth factor receptor (EGFR) is a clinically validated target for the treatment of diverse cancers, as demonstrated by the FDA's approval of five drugs directed against this receptor (cetuximab, panitumumab, erlotinib, gefitinib, and lapatinib), which are used for the treatment of metastatic breast cancer (MBC), colorectal cancer (CRC), non-small cell lung cancer, head and neck cancer, and pancreatic cancer. Yet the clinical utility of measurements of EGFR expression has been limited. Assessment of EGFR expression in tumors is insufficient to stratify cancer patients for therapy, even with EGFR-targeted drugs. Here we propose many of the limitations associated with measurement of EGFR expression result from failure to independently assess all four isoforms of EGFR. In addition to full-length EGFR, three naturally occurring EGFR isoforms are synthesized from alternative EGFR transcripts that arise from either read-through of an exon-intron boundary (1.8 kb EGFR), or by incorporation of one of two alternate exons, i.e., 15A (2.4 kb EGFR) or 15B (3.0 kb EGFR). The best studied of these alternate isoforms is the product of the 3.0 kb EGFR transcript, which we have designated sEGFR. sEGFR is a 90/110-kDa protein consisting of most of the EGFR extracellular domain followed by a unique carboxy-terminal cysteine-rich 78-amino acid sequence. While the biosynthesis of sEGFR is predicted to follow a classic secretory fate, sEGFR is unexpectedly a cell surface protein. How sEGFR associates with the cell surface is unclear, but recent studies demonstrate that sEGFR is proteolytically released from the cell surface through a regulated process (1). This process is important since it gives rise to serum sEGFR, as demonstrated by mass spectrometric analysis (2). Quantification of both serum and tumor sEGFR has been shown by our laboratory and others to have diagnostic and prognostic utility in cancer patients, as outlined below. Ovarian cancer: EGFR is a controversial tumor marker in epithelial ovarian cancer (EOC). In ∼40 published studies, only half report that EGFR is a prognostic marker (3). IHC studies using antibodies specific for sEGFR vs. full-length EGFR (Yale cohort: n=196) demonstrate that sEGFR and EGFR are expressed independently in both normal and malignant ovarian tissues. Moreover, while neither sEGFR nor EGFR are independent prognostic markers, combined tumor sEGFR and EGFR expression is a powerful negative predictor of overall survival (4). In addition, four laboratories have demonstrated the diagnostic utility of serum sEGFR in EOC (2, 5-7). Serum sEGFR is characteristically lower in women with EOC, even early stage EOC, compared to healthy controls of similar age and menopausal-status. Serum sEGFR concentrations remain stable or rise in EOC patients immediately following surgery (8). The fact that low serum sEGFR concentrations are negatively correlated with tumor burden in EOC, whereas high tumor sEGFR expression is negatively correlated with survival suggests that the regulation of sEGFR release into the bloodstream is intricately associated with the progression of this disease. Breast cancer: A similar reciprocal pattern between serum sEGFR and tumor EGFR expression has been reported in MBC patients: high tumor EGFR expression and low baseline serum sEGFR are both correlated with shorter overall survival (9-12). Paradoxically, however, serum sEGFR decreases in response to chemotherapy (12). Baseline serum sEGFR also has been evaluated as a theragnostic marker for MBC patients treated with aromatase inhibitors and SERMs (9, 13, 14), chemotherapy (10, 12), and trastuzumab (15), however, these studies report mixed results, perhaps because of differences in trial design or methods of sEGFR assessment. Strikingly, however, increased concentrations of serum sEGFR can predict the development of breast cancer 17 months prior to clinical diagnosis of this disease (16). Endometrial cancer: We have completed the first analysis of serum sEGFR in patients with endometrial cancer (17). In this GOG (229C) study baseline serum EGFR expression (n=23) was the only significant prognosticator of overall patient survival in patients treated with gefitinib. Lung cancer: Baseline serum sEGFR concentrations have not been associated with improved survival in patients treated surgically (18, 19), or in patients treated with either chemo- or radiation therapy (20). High baseline sEGFR, however, has been correlated with response to gefitinib treatment (21). Serum sEGFR also is a leading candidate diagnostic marker in lung cancer patients, where age and gender affect the ability of this biomarker to discern lung cancer patients from healthy controls (22). In preliminary studies both human and mouse circulating sEGFR levels can be used to detect lung cancer (22, and S. Hanash, personal communication); the future use of this biomarker in animal models of cancer may, therefore, provide an excellent opportunity for exploring the limits of early detection using this new serum biomarker. Colorectal cancer: Two EGFR-directed antibodies, cetuximab and panitumumab, are FDA approved for the treatment of patients with CRC. Recent studies have shown that tumor expression of downstream effectors such as PTEN, BRAF, and KRAS are negative predictors of the efficacy of these drugs, whereas EGFR expression does not predict responsiveness to either of these antibodies (23). In contrast, high baseline serum sEGFR is strongly associated with best objective response in patients (n=42) treated with gefinitinb/FOLFOX6 for 10 cycles followed by gefitinib maintenance (24), and also with improved survival in patients (n=118) with third-line irinotecan and cetuximab treatment (25). While the basis for this association is not yet understood, we recently have demonstrated that both cetuximab and panitumumab bind to sEGFR with high affinity (26). We estimate that peak cetuximab concentrations are likely in only 28-fold molar excess relative to circulating sEGFR concentrations in treated patients. Since baseline serum sEGFR concentrations are characteristically decreased in many cancer patients (27), some patients would be predicted to receive much greater effective doses of cetuximab than others; to date, however this effect has not been considered in tailoring treatment for lung cancer patients. Baseline serum sEGFR also has been reported to have diagnostic utility in CRC; Abdel-Aziz and coworkers have shown that sEGFR can differentiate between CRC patients (n=48) and normal subjects (n=20) with a sensitivity of 71% at 100% specificity (28), and Spindler and colleagues have shown a significant difference between pretreatment sEGFR concentrations in controls (n=126) vs. patients (n=118) (25). Conclusions: Quantification of EGFR expression in human tumors has proven problematic, in part, because of the coincident expression of multiple EGFR isoforms in human tissues, a phenomenon that has been overlooked in the development of diagnostic assays. The independent assessment of tumor sEGFR vs. EGFR in a large EOC cohort as reported here, is clinically informative, and provides the first proof of principle of the benefit of independently measuring individual EGFR isoforms. Tissue sEGFR is a precursor of serum sEGFR, and serum sEGFR has been demonstrated to have independent diagnostic, prognostic, and theragnostic utility in diverse malignancies. In addition, serum sEGFR appears to be an unanticipated “first target” for at least two EGFR-directed therapeutic antibodies. While study of variant EGFR isoforms is in its infancy, these early studies demonstrate a previously unrecognized level of complexity in EGFR-based diagnostics that may account for current paradoxes associated with the clinical utility of EGFR-based diagnostic assays, similar to the ‘HER2 Conundrum’ described by Allison (29). References: 1. J. A. Wilken, M. Perez-Torres, E. M. Cora, A. T. Baron, N. J. Maihle, Submitted, (2010). 2. A. T. Baron et al., Cancer Epidemiol Biomarkers Prev 12, 103 (2003). 3. J. M. Lafky, J. A. Wilken, A. T. Baron, N. J. Maihle, Biochim Biophys Acta 1785, 232 (2008). 4. A. T. Baron et al., Submitted, (2010). 5. G. P. Bertenshaw et al., Cancer Epidemiol Biomarkers Prev 17, 2872 (2008). 6. R. G. Moore et al., Gynecol Oncol 108, 402 (2008). 7. B. Nolen et al., Gynecol Oncol 112, 47 (2009). 8. A. T. Baron et al., Cancer Epidemiol Biomarkers Prev 8, 129 (1999). 9. C. Souder et al., Cancer 107, 2337 (2006). 10. M. T. Sandri et al., Cancer 110, 509 (2007). 11. J. Marx III et al., Proc Am Soc Clin Oncol 21, 1743 (2002). 12. V. Muller et al., Anticancer Res 26, 1479 (2006). 13. J. M. Lafky et al., Cancer Res 65, 3059 (2005). 14. J. M. Marx et al., Proc Am Soc Clin Oncol 21, (2002). 15. G. Hudelist et al., Eur J Cancer 42, 186 (2006). 16. S. J. Pitteri et al., AACR Annual Meeting, Abstract #4815 (2010). 17. K. K. Leslie et al., AACR Annual Meeting, Abstract #LB-134 (2009). 18. H. Sasaki et al., Int J Clin Oncol 8, 79 (2003). 19. Y. Lemos-Gonzalez, F. J. Rodriguez-Berrocal, O. J. Cordero, C. Gomez, M. Paez de la Cadena, Br J Cancer 96, 1569 (2007). 20. W. Jacot, J. L. Pujol, J. M. Boher, P. J. Lamy, Br J Cancer 91, 430 (2004). 21. V. Gregorc et al., Clin Cancer Res 10, 6006 (2004). 22. A. T. Baron et al., Submitted, (2010). 23. Z. Saridaki, V. Georgoulias, J. Souglakos, World J Gastroenterol 16, 1177 (2010). 24. M. G. Zampino et al., Cancer Chemother Pharmacol 63, 139 (2008). 25. K. G. Spindler, D. Olsen, I. Brandslung, A. Jakobsen, J Clin Oncol 27, e22096 (2009). 26. J. A. Wilken, A. T. Baron, N. J. Maihle, Submitted, (2010). 27. A. T. Baron, J. A. Wilken, D. E. Haggstrom, S. T. Goodrich, N. J. Maihle, IDrugs 12, 302 (2009). 28. M. M. Abdel-Aziz, M. Lotfy, I. M. El-Kady, M. Abozaid, Cancer Detect Prev 32, 329 (2009). 29. M. Allison, Nat Biotechnol 28, 117 (2010). Citation Information: Cancer Prev Res 2010;3(12 Suppl):CN09-03.

Abstract: Although trastuzumab (Herceptin) is an important advance in the treatment of breast cancer, a significant proportion of patients do not respond to trastuzumab either alone or in combination with chemotherapy. In this study, we observe that epidermal growth factor receptor (EGFR) and HER3 expression is substantially increased after long-term trastuzumab exposure of HER2-positive breast carcinoma–derived cell lines that show primary resistance to trastuzumab. Furthermore, long-term trastuzumab exposure of trastuzumab-resistant cell lines induces de novo sensitivity to the EGFR-targeted agents gefitinib or cetuximab in two of three cell lines accompanied by increased EGFR expression. Together, these results indicate that primary trastuzumab resistance is not synonymous with lack of responsiveness to trastuzumab and, importantly, suggest that trastuzumab priming may sensitize trastuzumab-resistant tumors to other HER family-directed therapeutics. [Cancer Res 2009;69(6):2191–4]