The number of tumor markers that could be utilized in some experimental or human situation is extremely large, but to date only a relatively small number have been used in screening, detec- tion, and diagnosis, prognosis, and follow-up of human neoplastic disease. Table 17.13 provides a listing of a variety of markers that are utilized for any one of these parameters in specific hu- man neoplasms. One of the first generally accepted markers was originally recognized by Gold and Freedman (1965) as a glycoprotein, which they termed carcinoembryonic antigen (CEA). They found that it was produced in large quantities in the large bowel and initially was felt to be relatively specific for that lesion (cf. Moertel et al., 1993). Even some patients with colonic pol- yps had elevated CEA levels (Doos et al., 1975). It is now apparent that a large number of neo- plasms and even normal tissues produce varying amounts of CEA. Furthermore, CEA is a family of proteins that are normally found at high levels in fetal gut and appear to be involved in inter- cellular adhesion (Benchimol et al., 1989) as well as signal transduction (Dráber and Skubitz,1998). Another glycoprotein is CA 125, which has been used as a marker of ovarian neoplasms, but is even more generally found in normal tissues as well as other neoplasms (cf. Jacobs and Bast, 1989).
Notice from the table that a number of markers are still in the experimental stage, and actually only those so indicated are approved by the U.S. Food and Drug Administration for use in the United States. In mesenchymal neoplasms as well as lymphomas and leukemias, it is ap- parent that the cytogenetic markers are primarily for diagnostic purposes. Similarly, but not shown in the table, is the use of cytoskeletal proteins as tissue markers for a variety of different types of neoplasms that aid the pathologist in specific diagnoses (Virtanen et al., 1984; Table16.4). If one analyzes the tumor markers associated with any specific type of neoplasm, a wide variety of different markers, both in tissue and body fluid, may be seen. Table 17.14 presents a short list of markers found in two different types of lung cancer at two different stages (Chapter10). However, this is only a small part of the total number of markers that have been described in human bronchogenic carcinoma (e.g., Coombes et al., 1976).
As discussed in Chapter 10 (Table 10.7), a variety of tissue markers in the form of changes in expression and mutation in proto-oncogenes and tumor suppressor genes have been utilized to make predictions about the natural history of neoplastic growth in specific individuals—i.e., the prognosis of the disease in such persons. Within the last few years, it has also been possible to detect mutations of tumor suppressor genes and proto-oncogenes in plasma DNA of cancer pa- tients (e.g., Silva et al., 1999; Kopreski et al., 2000). In addition, antibodies to the p53 tumor suppressor protein have been found in human cancer patients with a specificity of 96% but sen- sitivity of 30% (Soussi, 2000). Whether such markers will be a useful adjunct to prognosis, de- tection, or therapy still remains to be seen.
One of the most frequently utilized tumor markers at the present time is the prostate- specific antigen (PSA). This protein is a glycoprotein with a molecular mass of 33 to 34 kDa existing in at least five isomers. It is a serine protease capable of catalyzing the hydrolysis of a variety of proteins. PSA is not synthesized exclusively by the prostate, since both normal tissues (breast and endometrium) as well as nonprostatic neoplasms (breast, and salivary glands) also secrete this material into the serum (cf. Duffy, 1996). In men, the normal serum PSA values increase with age (DeAntoni, 1997). PSA concentrations in black men are significantly higher than those in whites (Morgan et al., 1996), leading to a differential discrimination in prostate cancer diagnosis between these two races based on PSA concentrations. Recently, it has been noted that PSA in the serum is predominantly bound to α-1-antichymotrypsin (PSA-ACT) with a smaller percentage bound to a macroglobulin. Recent studies have suggested that measurement
CEA, carcinoembryonic antigen; CA-125, cancer antigen 125; SCC Ag, squamous cell carcinoma antigen; TPA, tissue polypeptide antigen; AGP, alpha-1-acid glycoprotein; NSE, neuron-specific enolase; CK BB, BB isoenzyme of creatine kinase; ChrA, chromogranin A; CA 19-9, cancer antigen 19-9. Adapted from Strauss and Skarin, 1994.
of PSA not bound to protein may increase the sensitivity of the test (Potter et al., 1999). While the sensitivity of the assay, especially if free PSA is taken into account, may reach 100%, the specificity of the assay ranges from 20% to 95% in a variety of series (Potter et al., 1999). In Figure 17.17 may be noted serum PSA levels in healthy controls as well as a variety of condi- tions in which one might expect to see an increase in the marker. While patients with prostate cancer exhibit a very high median level, the overlap with patients having the benign condition, benign prostatic hyperplasia (BPH)—as well as with other genitourinary cancers or diseases in patients who had undergone prostatectomy—is still exhibiting some overlap. It is possible that the analysis of free PSA may tighten these numbers so there will be much less overlap, increas- ing the sensitivity of the assay. Despite these shortcomings, this tumor marker is being utilized for screening and monitoring of men over 50 years of age together with a measurement of other clinical parameters.
Thus, it is apparent that tumor markers have some usefulness in certain neoplasms, especially in the diagnosis of mesenchymal neoplasms and the usefulness of PSA in moni- toring the presence of prostatic carcinoma. Like so much in cancer research, the potential for
Figure 17.17 Serum PSA levels in healthy male controls, patients with prostate cancer, patients with benign prostatic hyperplasia (BPH), patients with other genitourinary cancers or diseases, and patients who had undergone prostatectomy. The individual dots each represent a single patient and the horizontal line show the mean PSA levels for each group. (Adapted from Barak et al., 1989, with permission of the authors and publisher.)
the usefulness of these markers, many of which contribute to the host-tumor relationship, is as yet unfulfilled.