IVD Technology Magazine
IVDT Article Index

Originally published January, 1998

Performance characteristics of a novel immunoassay for cancer detection

That T. Ngo, Michael C. Cress, and Ronald J. Moore

Clinical traits characteristic of certain cancers provide clues for the development of new generations of tumor markers.

Proteases have been implicated in a number of malignant conditions, and researchers have observed increased secretion of proteases into the interstitial fluid around growing tumors. These proteases inevitably act on proteins, including those in the coagulation cascade leading to the formation of fibrin. Furthermore, fibrin is very frequently observed at the invading periphery of malignant neoplasms.¹ Malignant cells also characteristically possess high levels of plasminogen activator, which should induce local fibrinolysis.²

In view of the concurrent increase in the formation of fibrin and in the secretion of proteases in malignant conditions, it is logical to conclude that the measurement of serum fibrinogen degradation product (FDP) levels may represent a useful measure of malignancy. This article describes the results of early studies to explore the viability of an immunoassay, called Oncochek, for the detection of FDPs as indicators of the presence of various cancers.

Background

Tumor cells release proteases into interstitial fluid at a higher rate than normal cells.^3,4 At least four lines of evidence support the concept that this increased protease activity contributes directly to the invasiveness of tumor cells and to the destruction of the adjacent host tissue.^5—7

First, in the case of breast cancer metastases, four classes of proteases appear to be involved in disease progression.⁸ These include cysteine proteases (cathepsins B and L), aspartyl proteases (cathepsin D), collagenases (metalloproteases), and serine proteases (urokinase and plasminogen). Increased expression of the collagenases has been correlated with increased invasiveness of some tumor cells.

Second, down-regulation of these enzymes by genetic means reduces both the invasiveness and metastases of the tumor.⁹ Third, addition of tissue metalloproteinase inhibitors to tumor cells blocks cell invasion in vitro. Fourth, the administration of either natural or synthetic metalloproteinase inhibitors has been shown to prevent metastasis in a simple lung colonization model.⁹

Protease release by tumor cells can also result in the proteolysis of plasma proteins. Theoretically, the extent of proteolytic degradation of these proteins can be correlated with the activity of the tumor cells and used indirectly to evaluate their tumor burden or degree of malignancy.

Principle of the Technology

Taking advantage of the foregoing information, the authors have devised a method for detecting proteolytic degradation products of plasma proteins with minimal interference from the parent protein (the protease substrate) and have begun to explore the utility of the method as a cancer detection assay. Specifically, the method measures unique epitopes that are manifested secondary to proteolytic degradation of fibrinogen. These epitopes are either sterically or immunochemically unreactive in the native fibrinogen molecule.

Assay specificity is achieved by the use of two different antibodies in a two-site, solid-phase enzymometric assay. The more highly specific antibody, which is immobilized to the solid phase, consists of a murine monoclonal to a glycine-histidine-arginine-proline-leucine-aspartate-lysine-cysteine (GHRPLDKC) octapeptide. The first seven amino acids of this peptide represent an internal sequence within the ß-chain of fibrinogen, which is near the amino terminus and is exposed after initial plasminolysis (residues 15—21).¹⁰ After capture of the proteolytic degradation products of fibrinogen by the immobilized monoclonal antibody, the immune complex is detected by using a highly specific conjugate consisting of polyclonal antifibrinogen labeled with horse-radish peroxidase.

Assay Procedure. Calibrators, controls, and samples in a diluent buffer are added in 100-µl amounts to wells of an antibody-coated microwell plate. After a one-hour incubation at room temperature and a wash step, 100-µl aliquots of conjugate are added to each well. The plate is again incubated for 30 minutes at room temperature, washed, and incubated with 100 µl TMB substrate. After 15 minutes at 25°C the reaction is stopped by adding 100 µl 0.1 N HCl into each well. The absorbance at 450 nm is proportional to the level of FDP in the assay.

Analytical Performance

Calibration Curve. Calibrators for the assay were prepared by controlled plasminolysis of fibrinogen to produce FDPs, which were formed in a time-dependent fashion (data not shown). Intact fibrinogen that was not subject to prior treatment with plasmin was unreactive in the current assay. Plasmin-digested fibrinogen exhibited a curvilinear dose response over the concentration range of 0—1250 ng/ml fibrinogen equivalents (see Figure 1). Sera from cancer patients, which typically contained elevated FDP levels, exhibited dilutional parallelism and linearity to the FDP calibration curve over a dilution range from 5- to 80-fold (see Figure 2).

Figure 1. Standard curve for measurement of fibrinogen degradation products (FDPs) by Oncochek.

Figure 2. Dilution of a high-titer patient sample.

Specificity. The Oncochek assay selectively measures FDPs. Neither fibrinogen fragment D (FD), fibrinogen fragment E (FE), nor native fibrinogen (FG) register any significant overt cross-reactivity with the Oncochek assay (Table I). Absorbance values at 450 nm reflect the relative immunoreactivities in this assay system.

Figure 3(A) illustrates results indicating that FD affects FDP measurements in the Oncochek assay in a pattern consistent with noncompetitive inhibition or covert cross-reactivity.¹¹ This inhibition pattern is consistent with the mechanism that FD binds to the solid phase of capture antibody, thus reducing the antibody sites available for binding FDPs. The double reciprocal plots of FE and FG inhibition studies are consistent with the absence of interaction between MAb and FE and FG (see Figures 3(B) and 3(C)). They are also consistent with the results presented in Table I, which shows the lack of response by FE and FG in the Oncochek assay.

Figure 3. Selectivity of the Oncochek assay: (A) noncompetitive inhibition by fibrinogen fragment D; (B) double reciprocal plot for fibrinogen; and (C) double reciprocal plot for fibrinogen fragment E.

Clinical Performance. Results of the Oncochek assay indicate that FDP levels in the sera of patients with various types of cancer are significantly elevated in comparison to normals. For example, FDP levels in the sera of normal control subjects were compared with those in the sera of patients with five types of cancers. Each group consisted of 50 patients and included breast, colon, lung, ovarian, and prostate cancers. The data presented in Figure 4 were subjected to a receiver-operating-characteristics (ROC) analysis to assess the relationship between the sensitivity and specificity of the assay at various threshold concentrations of FDP. By ROC analysis, using an upper limit of normal corresponding to 96% specificity, sensitivities of 84, 82, 82, 34, and 60% were achieved for breast, colon, lung, ovarian, and prostate cancers, respectively (see Table II). If an elevation in the value of either the Oncochek assay or the organ-specific marker (or both) was used as a prediction of the presence of cancer, sensitivities approximating 90% or great-er were achieved for breast, colon, and lung cancers.

Table II. Clinical sensitivity of the Oncochek immunoassay compared to that of various organ-specific markers.

Results shown in Table II and Figure 4 suggest that the Oncochek immunoassay can detect multiple cancers with a high degree of specificity and clinical sensitivity. When it is used with established organ-specific markers, improved clinical sensitivity may be achieved for breast, colon, and lung cancers.

Figure 4. FDP levels in the sera of normal control subjects and patients with five types of cancer (n = 50).

Discussion

When used in conjunction with the recognized organ-specific tumor marker for breast, colon, and lung cancers, the unique epitope detected by the Oncochek immunoassay system appears to offer increased clinical sensitivity. The results for ovarian (CA 125) and prostate (PSA) cancers are less dramatic. Prospective clinical studies are needed to elucidate the interrelationships between Oncochek, the established tumor markers, and disease progression.

Biochemical specificity of the assay, which is central to its clinical utility, was ensured by use of a monoclonal antibody against a GHRPLDKC octapeptide conjugated to bovine serum albumin. The first seven amino acids of the octapeptide corresponds to amino acids 15 to 21 of the ß-chain of human fibrinogen.¹² This antibody recognizes fragment D, the proteolytic product of fibrinogen plasminolysis, but does not recognize either fragment E or intact native fibrinogen. Thus, in the current assay format, the immobilized monoclonal antibody will capture both fragment D and FDPs, but only FDPs are detected in the assay by the polyclonal antifibrinogen that has been labeled with horseradish peroxidase.

Conclusion

The novel immunoassay for cancer detection described in this article offers three major performance advantages. First, it enables immunochemical measurements of proteolytic degradation products in the presence of, and without interference by, the parent protein molecules (the substrate). Second, the assay detects multiple cancers with a high degree of specificity and sensitivity. Third, when used together with established organ-specific markers, the overall clinical performance may be improved. The authors are currently expanding their initial studies to include a larger patient population as well as other types of cancers and benign disease states such as inflammation and infection.

References

1. Hiramoto R, Berneck J, Jurandowski J, et al., "Fibrin in Human Tumors," Cancer Res, 20:592—593, 1960.

2. Ossowski L, Quigley JP, Kellerman GM, et al., "Fibrinolysis Associated with Oncogenic Transformation," J Exp Med, 138:1056—1064, 1973.

3. Sylven B, "Lysosomal Enzyme Activity in the Interstitial Fluid of Solid Mouse Tumour Transplants," Eur J Cancer, 4:463—474, 1968.

4. Sylven B, "Cellular Detachment by Purified Lysosomal Cathepsin B," Eur J Cancer, 4:559— 562, 1968.

5. Poole AR, Tiltman KJ, Recklies AD, et al., "Differences in Secretion of the Proteinase Cathepsin B at the Edges of Human Breast Carcinomas and Fibroadenomas," Nature, 273:545—547, 1978.

6. Keppler D, Abrahamson M, and Sordat B, "Secretion of Cathepsin B and Tumor Invasion," Biochem Soc Trans, 22:43—49, 1994.

7. Pietras RJ, Szego CM, Roberts JA, et al., "Lysosomal Cathepsin B—Like Activity: Mobilization in Prereplicative and Neoplastic Epithelial Cells," J Histochem Cytochem, 29:440—450, 1981.

8. Dickson RB, Shi YE, and Johnson MD, "A Novel Matrix-Degrading Protease in Hormone-Dependent Breast Cancer," Biochem Soc Trans, 22:49—52, 1994.

9. Goldberg GI, and Eisen AZ, "Extracellular Matrix Metalloproteinases in Tumor Invasion and Metastasis," in Regulatory Mechanisms in Breast Cancer, Lippman ME, and Dickson RB (eds), Boston, Kluwer Academic Publishers, pp 421— 440, 1990.

10. Chung DC, Que BG, Rixon MW, et al., "Characterization of Complementary Deoxyribonucleic Acid and Genomic Deoxyribonucleic Acid for the ß Chain of Human Fibrinogen," Biochemistry, 22:3244—3250, 1983.

11. Suelter CH, A Practical Guide to Enzymology, New York, Wiley, p 248, 1985.

12. Haverkate F, and Timan G, "Protective Effect of Calcium in the Plasmin Degradation of Fibrinogen and Fibrin Fragments D," Thromb Res, 10:803—812, 1977.

That T. Ngo is president and CEO, Michael C. Cress is research scientist, and Ronald J. Moore is vice president of operations at AMDL, Inc. (Tustin, CA). The authors wish to thank Robert L. Oak and Thara Rhamankutty for their technical assistance in preparing this article.