Journal of Surgical Oncology 7:77-80 (1975)
Simultaneous Determination of Plasma and Urinary Carcinoembryonic Antigen levels .......................................................................................... .......................................................................................... C . P. KARAKOUSIS, M.D., E. D. HOLYOKE, M.D., and T. M. CHU, Ph.D. The results of simultaneous plasma and urinary CEA levels in 14 patients are described. The aim was t o determine whether urinary CEA derived from plasma or arose independently in the urinary tract. The method of reductio ad absurdum is followed. It is assumed that CEA in the urine is derived from plasma, and CEA clearance values for each patient are calculated. These varied widely from zero to a theoretical infinity. On this basis and from previous experimental evidence, it is concluded that urine is independent of plasma CEA. The finding, however, in some individuals of “anti-CEA”-like factor(s) distorting urinary CEA determinations makes for some reservations about the safety of the conclusion. It is, however, certain that with the presently available technology, present urinary and plasma CEA values cannot be correlated.
.......................................................................................... .......................................................................................... Studies of the chemical composition of CEA have revealed that it is a glycoprotein (Krupey, et al., 1968; Terry et al., 1972) located in the glycocalyx of tumor cells (Gold, 1973). The majority of clinical studies of CEA have been performed measuring its levels in plasma. There have been, however, some studies in patients with bladder tumor which have shown higher urinary values than are found in urine from normal individuals. It has been speculated that the urinary CEA in these patients was derived from the surface of these tumors and was thus carried in the urine (Hall et al., 1972). Urinary CEA values may then provide a more sensitive test than plasma values for the detection of urothelial carcinoma. CEA has also been detected in the urine of patients with active urinary tract infection but usually not in the presence simply of bacteriuria. Since CEA appears also in the urine of patients with urogenital tumors; the question arises as to its origin. The present work was aimed at studying the correlation between plasma and urine CEA values in patients with malignancy but with normal urinary tracts and renal function. I f such correlation existed the possible benefit would be a more “sensitive” test, particularly where the plasma values are too low to be detected with present techniques, since most substances excreted through the kidneys appear in a higher concentration in urine than in plasma. Departments of Clinical Laboratories and Surgery, Roswell Park Memorial Institute, Buffalo, N. Y.
77
@Alan R. Liss, Inc., 150 Fifth Avenue, New York, N.Y. 10011
78
Karakousis, Holyoke, a n d Chu
MATERIALS AND METHODS Urine was collected for 24 hr in 14 patients, most of them with metastatic colonic carcinoma. Plasma CEA was determined either immediately prior to the urine collection or at the time of completion (Table I). In two patients the tests were repeated with urine collected in two 12-hr periods with a plasma CEA determination for each period (Table 11). No patient had a clinical urinary tract infection at the time of urine collection. Plasma CEA was measured by a radioimmunoassay previously described (Chu and Reynoso, 1972). Urine specimens were treated in the same manner as described by Hall et al. (1972) before radioirnmunoassay.
RESULTS The CEA clearance was estimated for each patient based on the general formula C = (U/P)V, where U is the amount of CEA in ng per ml of urine, V is the volume (in ml) of urine excreted per min, and P is the amount of CEA in ng per ml of plasma (Best and Taylor, 1966).
TABLE I. No.
Plasma CEA and Clearance Values in 14 Patients
CEA plasma CEA urine (ng/ml) (ng/ml)
24 hr urine vol. (ml)
CEA clearance (mlimin)
I
0.3
16.3
1,380
2
40.4
24.2
1 380
0.6
3
84.8
0
1,840
0
4 5
1.o 7.8
42.3 63.6
1,250 390
36.7 2.1
6
7.5
13.4
1,188
1.6
I
0
3.3
2.860
-
8
22.9
25.8
820
0.6
9
3.4
c)
10
7.9
98.2
760
6.5
11
3.8
58.1
400
4.2
12
0.1
86.0
720
13
2.6
5.7
1,820
2.7
14
10.1
60.6
700
2.9
2,370
23
0
434
Diagnosis BUN Creal __ 7 1.2 Adeno Ca sigmoid with liver metastases 36 Adeno Ca rectum with advanced node metastases 13 Adeno Ca of pelvis with 1 paraortic node metastases 6 0.8 Adeno Ca of sigmoid 22 Recurrent adeno Ca of rectum Adeno Ca of sigmoid with 25 metastases 9 0.7 Adeno Ca of the sigmoid, invasive Advanced Ca of rectum with 14 liver metastases 6 Undifferentiated Ca of cervix, Stage 1B 15 Ca of R. breast with osseous metastases 21 0.9 Adeno Ca of gall bladder with liver metastases 42 1.2 Adeno Ca of rectum in a villous adenoma 34 1.1 Sq. cell Ca of Cx, Stage IIA 12 Mod. diff. sq. Ca of Cx, Stage IIIB
Plasma and Urinary CEA Values
79
Under normal conditions the glomerular filtration rate remains fairly stable, and clearance of a substance is characteristic of the way the kidney handles it. The clearance values for CEA were found to vary widely from 0 t o a theoretical infinity (Table I). In the two patients (Table 11) where two sequential 12-hr urine specimens (both day and night) were examined, there was considerable variation in each patient in the clearance, although their corresponding plasma values remained stable. No consistency was noted in diurnal variation.
DISCUSSION It has been reported that the molecular weight of CEA should be about 200,000 daltons (Terry, et al., 1972). This would suggest that it should not be filterable in the glomeruli (pores approximately in 100 A in diameter), since the smallest plasma protein, albumin, which normally appears in the glomerula filtrate either not at all or at a very low concentration (less than 0.2-0.3% of its concentration in plasma) (Best and Taylor, 1966), has a molecular weight of about 69,000. However, two possibilities are to be considered in the case of CEA: (1) the estimation of the CEA molecular weight has been done using extracts of hepatic metastasis and may not reflect the molecular weight of CEA as it circulates in plasma, and (2) degradation products of CEA which possess CEA activity may have molecular weights below 69,000 and consequently be filterable through the kidney. In a study of the metabolism of human colonic CEA in xenogeneic animals, the majority of intravenously injected radioactive CEA was removed from the circulation in 30 min (50%in the first 5 min, 20% in the next 25 min, and thereafter very slowly) and most radioactivity deposited in the liver (Shuster et al., 1973). The excreted radioactive materials in the urine after 10 min and up to 1 hr consisted of polypeptides with low molecular weight, while the radioactive components collected during the first 10 min of the experiment contained a small amount of immunoreactive material, as demonstrated by binding to goat CEA antiserum. However, at the end of 1 hr urinary excretion accounted for not more than 0.5% of the injected radioactive material. The disappearance rate of radioactive CEA from the plasma was shown to be not a saturation phenomenon and pointed to biological heterogeneity of the CEA molecules (Shuster et al., 1973). The clearance values of CEA in our study being at such variance from each other, with high plasma values corresponding to nondetectable urinary levels and vice versa, Sequential Plasma CEA and Clearance Values in Two Patients
TABLE 11.
CEA (W/ml) Urine vol(24 hr) Plasma Urine (ml)
No. ~~
~~~
Creat
Diagnosis
~~
15
22.9
25.8
16
23.2 23.2 3.4
105 3.5 0
2.9 2.7
CEA clearance (mlimin) BUN
10.0 3.8
820 800 (12 hr) 230 (12 hr) 2,370
720 (12 hr) 810 (12 hr)
0.6
14
-
4.9 0.04 0
14 14 6
-
3.4 1.5
6 6
Advanced Ca of rectum with liver metastases
-
-
Undifferentiated Ca of C, Stage 1 B
80
Karakousis, Holyoke, and Chu
suggest that urinary CEA did not come from the plasma CEA and therefme waD formed in the urinary tract. However, there are two Implicit as8umptlenfi made in this conclusion: (1) Kidneys of all patients handle CEA In the o me way, This i s probably a valid assumption since all normal kidneys function thniiarly. (2) CEA which is measured by radioimmunoassay is not only antigenically the same, but alno ha$ approximately the same molecular weight in the plasma and the urine for all patients, Thk may not be true, as has been indicated before (Gold et al., 1973), and could perhaps explain the lack of plasma correlation seen here. The urine specimens from normal healthy individual8 were a h ured as CQntFQlfar this study. An “anti-CEA”-like factor was found in some of there contra1 urines w d in some instances a large portion of CEA-IlZ5was precipitated by dialyzed urine samples, Tt is questionable whether this phenomenon is related only to the airaonyl p h s p h a t e gel or if it would also appear with the double antibody technique. Further atudies will be TIquired to determine if this anti-CEA substance can be demonstrated in m a n patients and if there might possibly be a means of making a correlation and determining a true urinary CEA. Correlation then with plasma CEA values might shed more light on the question of whether CEA is partially or wholly derived from plasma or is formed hdep6gdently in the urinary tract. However, since anti-CEA-like factors have not been detected in tho plasma with the present radioimmunoassay, the plasma CEA values should be valid when compared with urine CEA levels. The data in Tables I and I1 support the view that the urinary ‘‘€EX’ is a urinary constituent not derived from the plasma, and certainly there ir no correlation between urine and plasma CEA values, at least with the techniques employed in thip work, The two patients who had 12-hr urine collections demonstrated a wide varlrttion in the urinary “CEA” despite a stable plasma CEA. This also suggests that the presence af a major portion, at least, of what is measured as urinary “CEA” i s subject to wide variations probably due t o local factors, rather than being characteristic of the urine of theoe patients or related to plasma levels. ACKNOWLEDGMENTS We wish t o acknowledge the help of Dr. 0. Yung and Dr. L. Blurnenron for ths statistical analysis of plasma and urine CEA values and Miss J. Ogledzinski for preparation of this manuscript. Supported in part by U.S.P.H.S. Grants CA-f 5263,CA.I 3 126, CA-15437, and 1-CB-33858. REFERENCES Best, C.H., and Taylor, N.B. (1966). The Physiological Basis of Medical Practice. Willlomi m d Wilkins CP., Boston, pp. 1671-1672. Chu, T. M., and Reynoso, G. (1972). Evaluation of a new radioimmunoassay method for owofno= embryonic antigen in plasma with use of zirconyl phosphate gel. Clin. Chem. 1$:81$-932, Gold, P. (1971). Antigenic reversion in human cancer. Ann. Rev. Med. 22:85-94. Gold, P., Wilson, T., Romero, R., Shuster, J., and Freedman, S. 0. (1973). Immunology and oalonle cancer. Dis. Colon. and Rectum 16:358-365. Hall, R. R., Lawrence, R. J. D., Darcy, D., Stevens, V., James, R., Roberts, S., and Neville, M, A, (1972). Carcinoembroyonic antigen in the urine of patients with urothelial carcinoma. Brit, J. Med. 3:609-611. Krupey, J., Gold, P., and Freedman, S. 0. (1968). Physicochemical studies of the carcineembrynnio antigens of the human digestive system. J. Exp. Med. 128:387-397. Shuster, J., Silverman, M., and Gold, P. (1973). Metabolism of human carcinoembryonic antigen in xenogeneic animals. Cancer Res. 33:65-68. Terry, D. W., Henkart, A. P., Coligan, E. J., and Todd, C. W. (1972). Structural studies of the major glycoprotein in preparations with carcinoembryonic antigen activity. J. Exp. Med. 136:200-204.
Simultaneous Determination of Plasma and Urinary Carcinoembryonic Antigen levels .......................................................................................... .......................................................................................... C . P. KARAKOUSIS, M.D., E. D. HOLYOKE, M.D., and T. M. CHU, Ph.D. The results of simultaneous plasma and urinary CEA levels in 14 patients are described. The aim was t o determine whether urinary CEA derived from plasma or arose independently in the urinary tract. The method of reductio ad absurdum is followed. It is assumed that CEA in the urine is derived from plasma, and CEA clearance values for each patient are calculated. These varied widely from zero to a theoretical infinity. On this basis and from previous experimental evidence, it is concluded that urine is independent of plasma CEA. The finding, however, in some individuals of “anti-CEA”-like factor(s) distorting urinary CEA determinations makes for some reservations about the safety of the conclusion. It is, however, certain that with the presently available technology, present urinary and plasma CEA values cannot be correlated.
.......................................................................................... .......................................................................................... Studies of the chemical composition of CEA have revealed that it is a glycoprotein (Krupey, et al., 1968; Terry et al., 1972) located in the glycocalyx of tumor cells (Gold, 1973). The majority of clinical studies of CEA have been performed measuring its levels in plasma. There have been, however, some studies in patients with bladder tumor which have shown higher urinary values than are found in urine from normal individuals. It has been speculated that the urinary CEA in these patients was derived from the surface of these tumors and was thus carried in the urine (Hall et al., 1972). Urinary CEA values may then provide a more sensitive test than plasma values for the detection of urothelial carcinoma. CEA has also been detected in the urine of patients with active urinary tract infection but usually not in the presence simply of bacteriuria. Since CEA appears also in the urine of patients with urogenital tumors; the question arises as to its origin. The present work was aimed at studying the correlation between plasma and urine CEA values in patients with malignancy but with normal urinary tracts and renal function. I f such correlation existed the possible benefit would be a more “sensitive” test, particularly where the plasma values are too low to be detected with present techniques, since most substances excreted through the kidneys appear in a higher concentration in urine than in plasma. Departments of Clinical Laboratories and Surgery, Roswell Park Memorial Institute, Buffalo, N. Y.
77
@Alan R. Liss, Inc., 150 Fifth Avenue, New York, N.Y. 10011
78
Karakousis, Holyoke, a n d Chu
MATERIALS AND METHODS Urine was collected for 24 hr in 14 patients, most of them with metastatic colonic carcinoma. Plasma CEA was determined either immediately prior to the urine collection or at the time of completion (Table I). In two patients the tests were repeated with urine collected in two 12-hr periods with a plasma CEA determination for each period (Table 11). No patient had a clinical urinary tract infection at the time of urine collection. Plasma CEA was measured by a radioimmunoassay previously described (Chu and Reynoso, 1972). Urine specimens were treated in the same manner as described by Hall et al. (1972) before radioirnmunoassay.
RESULTS The CEA clearance was estimated for each patient based on the general formula C = (U/P)V, where U is the amount of CEA in ng per ml of urine, V is the volume (in ml) of urine excreted per min, and P is the amount of CEA in ng per ml of plasma (Best and Taylor, 1966).
TABLE I. No.
Plasma CEA and Clearance Values in 14 Patients
CEA plasma CEA urine (ng/ml) (ng/ml)
24 hr urine vol. (ml)
CEA clearance (mlimin)
I
0.3
16.3
1,380
2
40.4
24.2
1 380
0.6
3
84.8
0
1,840
0
4 5
1.o 7.8
42.3 63.6
1,250 390
36.7 2.1
6
7.5
13.4
1,188
1.6
I
0
3.3
2.860
-
8
22.9
25.8
820
0.6
9
3.4
c)
10
7.9
98.2
760
6.5
11
3.8
58.1
400
4.2
12
0.1
86.0
720
13
2.6
5.7
1,820
2.7
14
10.1
60.6
700
2.9
2,370
23
0
434
Diagnosis BUN Creal __ 7 1.2 Adeno Ca sigmoid with liver metastases 36 Adeno Ca rectum with advanced node metastases 13 Adeno Ca of pelvis with 1 paraortic node metastases 6 0.8 Adeno Ca of sigmoid 22 Recurrent adeno Ca of rectum Adeno Ca of sigmoid with 25 metastases 9 0.7 Adeno Ca of the sigmoid, invasive Advanced Ca of rectum with 14 liver metastases 6 Undifferentiated Ca of cervix, Stage 1B 15 Ca of R. breast with osseous metastases 21 0.9 Adeno Ca of gall bladder with liver metastases 42 1.2 Adeno Ca of rectum in a villous adenoma 34 1.1 Sq. cell Ca of Cx, Stage IIA 12 Mod. diff. sq. Ca of Cx, Stage IIIB
Plasma and Urinary CEA Values
79
Under normal conditions the glomerular filtration rate remains fairly stable, and clearance of a substance is characteristic of the way the kidney handles it. The clearance values for CEA were found to vary widely from 0 t o a theoretical infinity (Table I). In the two patients (Table 11) where two sequential 12-hr urine specimens (both day and night) were examined, there was considerable variation in each patient in the clearance, although their corresponding plasma values remained stable. No consistency was noted in diurnal variation.
DISCUSSION It has been reported that the molecular weight of CEA should be about 200,000 daltons (Terry, et al., 1972). This would suggest that it should not be filterable in the glomeruli (pores approximately in 100 A in diameter), since the smallest plasma protein, albumin, which normally appears in the glomerula filtrate either not at all or at a very low concentration (less than 0.2-0.3% of its concentration in plasma) (Best and Taylor, 1966), has a molecular weight of about 69,000. However, two possibilities are to be considered in the case of CEA: (1) the estimation of the CEA molecular weight has been done using extracts of hepatic metastasis and may not reflect the molecular weight of CEA as it circulates in plasma, and (2) degradation products of CEA which possess CEA activity may have molecular weights below 69,000 and consequently be filterable through the kidney. In a study of the metabolism of human colonic CEA in xenogeneic animals, the majority of intravenously injected radioactive CEA was removed from the circulation in 30 min (50%in the first 5 min, 20% in the next 25 min, and thereafter very slowly) and most radioactivity deposited in the liver (Shuster et al., 1973). The excreted radioactive materials in the urine after 10 min and up to 1 hr consisted of polypeptides with low molecular weight, while the radioactive components collected during the first 10 min of the experiment contained a small amount of immunoreactive material, as demonstrated by binding to goat CEA antiserum. However, at the end of 1 hr urinary excretion accounted for not more than 0.5% of the injected radioactive material. The disappearance rate of radioactive CEA from the plasma was shown to be not a saturation phenomenon and pointed to biological heterogeneity of the CEA molecules (Shuster et al., 1973). The clearance values of CEA in our study being at such variance from each other, with high plasma values corresponding to nondetectable urinary levels and vice versa, Sequential Plasma CEA and Clearance Values in Two Patients
TABLE 11.
CEA (W/ml) Urine vol(24 hr) Plasma Urine (ml)
No. ~~
~~~
Creat
Diagnosis
~~
15
22.9
25.8
16
23.2 23.2 3.4
105 3.5 0
2.9 2.7
CEA clearance (mlimin) BUN
10.0 3.8
820 800 (12 hr) 230 (12 hr) 2,370
720 (12 hr) 810 (12 hr)
0.6
14
-
4.9 0.04 0
14 14 6
-
3.4 1.5
6 6
Advanced Ca of rectum with liver metastases
-
-
Undifferentiated Ca of C, Stage 1 B
80
Karakousis, Holyoke, and Chu
suggest that urinary CEA did not come from the plasma CEA and therefme waD formed in the urinary tract. However, there are two Implicit as8umptlenfi made in this conclusion: (1) Kidneys of all patients handle CEA In the o me way, This i s probably a valid assumption since all normal kidneys function thniiarly. (2) CEA which is measured by radioimmunoassay is not only antigenically the same, but alno ha$ approximately the same molecular weight in the plasma and the urine for all patients, Thk may not be true, as has been indicated before (Gold et al., 1973), and could perhaps explain the lack of plasma correlation seen here. The urine specimens from normal healthy individual8 were a h ured as CQntFQlfar this study. An “anti-CEA”-like factor was found in some of there contra1 urines w d in some instances a large portion of CEA-IlZ5was precipitated by dialyzed urine samples, Tt is questionable whether this phenomenon is related only to the airaonyl p h s p h a t e gel or if it would also appear with the double antibody technique. Further atudies will be TIquired to determine if this anti-CEA substance can be demonstrated in m a n patients and if there might possibly be a means of making a correlation and determining a true urinary CEA. Correlation then with plasma CEA values might shed more light on the question of whether CEA is partially or wholly derived from plasma or is formed hdep6gdently in the urinary tract. However, since anti-CEA-like factors have not been detected in tho plasma with the present radioimmunoassay, the plasma CEA values should be valid when compared with urine CEA levels. The data in Tables I and I1 support the view that the urinary ‘‘€EX’ is a urinary constituent not derived from the plasma, and certainly there ir no correlation between urine and plasma CEA values, at least with the techniques employed in thip work, The two patients who had 12-hr urine collections demonstrated a wide varlrttion in the urinary “CEA” despite a stable plasma CEA. This also suggests that the presence af a major portion, at least, of what is measured as urinary “CEA” i s subject to wide variations probably due t o local factors, rather than being characteristic of the urine of theoe patients or related to plasma levels. ACKNOWLEDGMENTS We wish t o acknowledge the help of Dr. 0. Yung and Dr. L. Blurnenron for ths statistical analysis of plasma and urine CEA values and Miss J. Ogledzinski for preparation of this manuscript. Supported in part by U.S.P.H.S. Grants CA-f 5263,CA.I 3 126, CA-15437, and 1-CB-33858. REFERENCES Best, C.H., and Taylor, N.B. (1966). The Physiological Basis of Medical Practice. Willlomi m d Wilkins CP., Boston, pp. 1671-1672. Chu, T. M., and Reynoso, G. (1972). Evaluation of a new radioimmunoassay method for owofno= embryonic antigen in plasma with use of zirconyl phosphate gel. Clin. Chem. 1$:81$-932, Gold, P. (1971). Antigenic reversion in human cancer. Ann. Rev. Med. 22:85-94. Gold, P., Wilson, T., Romero, R., Shuster, J., and Freedman, S. 0. (1973). Immunology and oalonle cancer. Dis. Colon. and Rectum 16:358-365. Hall, R. R., Lawrence, R. J. D., Darcy, D., Stevens, V., James, R., Roberts, S., and Neville, M, A, (1972). Carcinoembroyonic antigen in the urine of patients with urothelial carcinoma. Brit, J. Med. 3:609-611. Krupey, J., Gold, P., and Freedman, S. 0. (1968). Physicochemical studies of the carcineembrynnio antigens of the human digestive system. J. Exp. Med. 128:387-397. Shuster, J., Silverman, M., and Gold, P. (1973). Metabolism of human carcinoembryonic antigen in xenogeneic animals. Cancer Res. 33:65-68. Terry, D. W., Henkart, A. P., Coligan, E. J., and Todd, C. W. (1972). Structural studies of the major glycoprotein in preparations with carcinoembryonic antigen activity. J. Exp. Med. 136:200-204.
