CUNICAL CHEMISTRY, Vol. 40, No. 12, 1994. Immunological Recognition and Clinical Significance of Nicked Human Chorionic. Gonadotropin in Testicular ...
CUN. CHEM.40/1 2, 2306-2312 (1994)
#{149} Endocrinology
and Metabolism
Immunological Recognition and Clinical Significance of Nicked Human Chorionic Gonadotropin in Testicular Cancer Rudolf Hoermann,”5 Peter Berger,2 Gerald Laurence A. Cole,4 and Klaus Mann’
Spoettl,3
Florian
GiHesberger,3
Andrew
Kardana,4
We studied the physical properties, immunological recognition, and clinical significance of nicked human chorionic gonadotropin (hCGn) in testicular cancer. Upon sodium dodecyl sulfate-polyacrylamide gel electrophoresis under reducing conditions, the f3-subunitof hCGn (hCG/3n) dissociated into two peptides, and, by reversed-phase chromatography, hCGn was found to be less hydrophobic than hCG. Immunologically, hCGn lacked two epitopes specific for holo-hCG (ci, c2), whereas at least five hCGf3 epitopes ($l-5) were preserved, and, as a result, recognition of hCGn by different hCG assays varied widely. In 309 sera and 88 urine samples from patients with seminomatous or nonseminomatous testicular cancer, hcG-only, hCG + hCGn, and hCG + hCGn + hCGI3 + hCGn + hCG core-fragment assays gave parallel results. hCGn was more abundant in urine than in serum samples. In conclusion, hCGn lacks two conformationally dependent epitopes of hcG, causing a change in hydrophobicityand explaining itsfailure to react in certain holohCG assays. Recognition of hcGn, however, does not seem to be crucial in the routine use of serum hCG as a tumor marker in patients with testicular cancer.
thought to exhibit similar immunological reactivities and to be indistinguishable from pregnancy hCG by standard hCG immunoassays (9-16). More recent studies have identified nicks or missing peptide linkages in a proportion of the hCG molecules, particularly between residues 45 and 46, and 47 and 48 of the (3-subunit (17-21). The type and extent of nicking has been reported to vary considerably in samples from both pregnant women and patients with gestational trophoblastic disease, and nicking may interfere with the immunological recognition of hCG by some monoclonal antibodies (m.Abs) (21-24). To obtain more information on the occurrence, immunological behavior, and clinical significance of nicked hCG (hCGn), we sought to define immunological properties of hCGn with a panel of mAbs, to examine hCGn recognition by different immunometric assays, and, most importantly, to assess the clinical performance of an hCG IRMA sensitive to hCGn, compared with an hCG IRMA with poor recognition of hCGn, in serum and urine samples from patients with testicular cancer and gestational trophoblast disease.
IndexIngTerms: tumor markers/immunoassays/chromatography,
MaterIals and Methods hCG Standard Preparations
reversed-phase/electrophoresls, Human
chorionic
polyacrylamide gel
gonadotropm
(hCG) is a useful and well-established marker of trophoblastic tumors (1_7).6 It is secreted by the trophoblastic cells of the placenta during pregnancy and in the human testis (1, 8). Similarly, tumor cells of trophoblastic origin are able to produce and secrete hCG in large quantities in hydatidiform mole, choriocarcinoma, or testicular cancer. Although hCG derived from trophoblastic tumors may differ from its pregnancy counterpart with respect to its molecular size, carbohydrate structure, and biological properties, tumor-derived hCG has generally been ‘Department of Endocrinology, University of Essen, Hufelandstr. 55, D-45122 Essen, Germany. for Biomedical Aging Research of the Austrian Academy of Sciences, A-6020 Innsbruck, Austria. ‘Medical Department II, Klinikum Grosshadern, University of Munich, D-81377 Munich, Germany. 4Department of Obstetrics and Gynecology, Yale University, New Haven, CT 06510. ‘Author for correspondence. Fax 49-201-723-5976. 6Nonstandard abbreviations: hCG, human choriomc gonadotropin; hCGn, nicked human chorionic gonadotropin; hCGa, a-subunit of hCG; hCG, a-subunit of hCG; hCGn, p-subunit of hCGn; hCGcf, hCG core fragment; IRP, International Reference Preparation; mAb, monoclonal antibody; SDS, sodium dodecyl sulfate; and PAGE, polyacrylamide gel electrophoresis. Received August 8, 1994; accepted September 22, 1994. 2306
CUNICAL CHEMISTRY, Vol. 40, No. 12, 1994
Highly a-subunit
purified preparations of intact hCG (CR123), of hCG (hCGa, CR123a), and (3-subunit of
hCG (hCG/3,CR123/3) were supplied by the Hormone Distribution Program of the NI-ADDK, Bethesda, MD. The 1st International Reference Preparation for hCG (IRP 75/537) and the 1st IRP for hCG/3 (75/551) were obtained from the National Biological Standards Board, Hertfordshire, UK. The hCGn preparation (C5) was proviously prepared and intensively characterized (21-24). It is 100% nicked between positions (347 and (348 (21). Monoclonal Antibodies mAbs 11/7, 3/6, 12/17, and 2/6 were a generous gift from K. Siddle, Cambridge University, UK (25,26), and the mAb FB12 was kindly donated by J. M. Bidart, Institut Gustave-Roussy, Villejuif, France. The panel of our mAbs was previously described in detail (16, 2729). Specificities of the mAbs are shown in Table 1.
Serum and UrineSamples We collected 292 serum samples from 95 patients with nonseminomatous testicular cancer [stages I to IV according to World Health Organization classification (30)], 17 samples from 14 patients with hCG-positive seminoma, 34 samples from 14 women with choriocarcinoma (n = 10) or hydatidiform moles (n = 4), and, for
Table 1. SpecIfIcity Code
Epltope
hCG
hCGn
hCG/3
of mAbs. fre.
hCGpn
hCG/3c1
INN-hCG-2
131
+
+
+
+
+
-
+
-
a
+
INN-hCG-22
132
+
+
+
+
INN-bLH-1
(33
+
+
+
+
INN-hCG-24 INN-hCG-58 3/6 12/17 FBi2
(34
+
+
+
+
+
-
(+)
(35
+
+
+
+
+
-
+
13 13
+
+
+
+
+
-
+
+
+
+
+
-
+
+
+
+
-
-
+
+
+
2/6
INN-hCG-10
/39
(36or$7 ci
INN-hCG-40 INN-hCG-45 INN-hCG-26
11/7
INN-FSH-158
+
-
c2 c3 c4
+
-
clorc2
+
a5
+
+
+
+
+
-
+
+
hFSH
hLH
(+)L
-
(+) -
(+) (+) + (+)
-
+
+
-
ahLH, human luteinizinghormone; hFSH, human follIcle-stimulatinghormone. a, epitopeonthea.subunit; (3,hCG$epitope;C, conformational holo-hCGepitope. b
(+), weakly reactive.
comparison, 24 samples from postpartum women. All samples were referred to our laboratory for routine determination of hCG during the last 5 years and stored frozen at -40#{176}C until use. In addition, 88 urine samples from 71 patients with nonseminomatous testicular cancer were analyzed. The procedures were performed in accordance with the guidelines of our local ethics committee and the Helsinki Declaration.
Purification of Urinary hCG hCG was purified from urine samples (seven patients with nonseminomatous testicular cancer, one choriocarcinoma, four pregnant women) by a three-step method involving DEAE anion-exchange chromatography (DEAE-Trisacryl M; Serva, Freiburg, Germany), hydrophobic interaction chromatography (Phenyl-Sepharose; Pharmacia, Freiburg, Germany), and gel filtration (Su-
perdex 75; Pharmacia).
The preparations
were devoid of
appreciable free hCG/3 contamination, as judged from their elution pattern on the Superdex 75 column and their immunological reactivity in the free hCGJ3 assay.
Gel Chromatographyand Reversed-Phase Chromatography To demonstrate
purity
of the substances,
we subjected
hCG and hCGn to gel chromatography on a column of Superdex 200 (1.6 x 60 cm; Pharmacia), which was preequilibrated and eluted (60 mlIh, 2-mL fractions) with a solution of 0.01 moJJL Tris-HC1 and 0.15 molfL NaCl, pH 7.5. Hydrophobicity of hCG and hCGn was assessed by reversed-phase chromatography on a Vydac C4 protein column (0.46 x 25 cm; Sigma, Deisenhofen, Germany). Prior to injection, the samples were incubated for 1 h at 37#{176}C in 1 milL trifluoroacetic acid, dissociating the dimer into its subunits. The column was eluted isocratically at 0.5 mI.lmin for 5 mm followed by a 100 mLIL to 500 mL/L acetonitrile gradient for 85 mm and a 500 mLIL to 1000 mLIL gradient for 10 mm. Fractions collected (0.5 mL each) were analyzed by varions immunoassays.
Sodium Dodecyl Suffate-Polyacrylamide Gel Electrophoresis(SDS-PAGE) and lmmunoblothng SDS-PAGE of hCG samples was conducted under reducing (10 g/L dithiothreitol and acetylation with iodoacetamide) and nonreducing conditions according to the method described by Goerg et al. (31): sample buffer 0.0625 mol/L Tris-HC1, 10 g/L SDS, 0.1 g/L bromphenolblue, gradient gel 230 x 120 x 0.48 mm, T (total acrylaniide) 10-20%, C (bisacrylainide) 3%, 600 V, 50 mA, 30 W, 10#{176}C, 2 h. After electrophoresis, proteins were transferred to nitrocellulose filters by using the semidry blotting technique of Kyhse-Andersen (32) and bands were visualized with our mAbs by the ECL chemiluminescence system (Amersham, Braunschweig, Germany) (33).
Immunoassays IRMAs were constructed and three of them were used to specifically measure in serum and urine samples either intact hCG, hCG + hCGn, or hCG + hCGn + hCGp + hCG(3n + hCGj3cf. The assay measuring hCG + hCGn was newly developed; the other two assays have been described previously (26,33-35). For standardization of the assays the 1st IRP for hCG was used. mAb-coated tubes were incubated with 100 pL of sample and 100 iL of phosphate-buffered saline containing 10 g/L bovine serum albumin for 2 h at room temperature, washed twice with 10 mmolJL Tris-HC1 buffer (pH 8.0) containing 0.05 g/L Tween 20 (1 mL), incubated with ‘25I-labeled second antibody (-100 000 counts/min; 2-3 ng in 100 j.iL) for 2 h, washed again, and finally counted. Intraassay CVs of the assays were from 5% to 10% at different concentrations, and the interassay CVs ranged from 9% to 16%. Reference values for nonpregnant patients were defined by the 95th percentile of 200 healthy subjects of both sexes including 46 postmenopausal women (36). Statistics Statistical analysis of the data was performed culation of r and the use of the Mann-Whitney
by calU-test.
CLINICALCHEMISTRY, Vol.40,No.12,1994
2307
Results PhysicalPropertiesof Purified hCGn The purified preparations of hCGn (C5, 100% nicked hCG) and hCG (CR123) showed minor differences in their behavior on a Superdex 200 gel chromatography column, each eluting as a single hCG + hCGn + hCG(3 + hCG/3n + hCGf3cf immunoreactive peak at nearly identical positions (data not shown). Contamination of the two preparations with either hCGI3 or hCG/3n was shown to be minimal by both the pattern of elution and the immunological determination of free hCG/3 + hCGpn + hCGf3cf in the fractions. Vydac C4 chromatography, on the other hand, revealed marked discrepancies in the hydrophobic properties between the (3-subunit of hCGn and hCG(3 released from hCG, with hCG/3n being less hydrophobic, eluting far ahead of the position of hCG(3 (Fig. 1). A minute amount of hCG/3nlike material was also released from hCG (Fig. 1). When subjecting hCGn and hCG to SDS-PAGE under nonreducing conditions, migration of hCGn and hCG was similar, whereas under reducing conditions hCG/3n, unlike hCG/3, dissociated into two peptides, each approximately half the size of intact hCGI3.
Immunological Propertiesof Purified hCGn To define the immunological epitopes present on hCGn, we subjected it to SDS-PAGE and immunostaining with a panel of mAbs recognizing distinct epitopes hcGn
hCGos
u)
10
a
8
ii
6 4
on the surface of the intact hCG molecule. hCGn was stained by two mAbs (INN-hCG-45, INN-hCG-26) recognizing conformationally dependent epitopes specific for the intact dimeric hCG molecule (c3 and c4), but not by three other antibodies directed against determinants ci and c2 (INN-hCG-i0, INN-hCG-40, and 11/7). However, hCGn was recognized by all seven anti-hCG/3 mAbs used, which recognize at least five epitopes on the /3-subunit of hCG ((31-/35) in its free or combined form (Fig. 2). To permit a quantitative comparison of the immunereactivities of hCG and hCGn, we constructed 10 different two-site assays. In these assays, the immunological activity of hCGn varied widely (Table 2). Three assays, either for intact hCG, hCG + hCGn, or hCG + hCGn + hCG(3 + hCG/3n + hCG/3cf, were selected for analysis of patients’ samples. Table 3 summarizes the characteristics of these assays.
of Serum and UrinaryhCG by Reversed-Phase Chromatography Analysis
Because reversed-phase chromatography had revealed significant differences in the hydrophobicity between hCG/3n and hCG(3, serum and urine samples were screened by this method for the presence of hCG/3n-like material. In 17 sera from testicular cancer, 2 sera from choriocarcinoma, and 7 pregnancy sera that were tested, hCG/3 immunoreactivity invariably eluted in a single peak at a position typical for hCGf3. In the purified urinary hCG preparations obtained from seven patients with testicular cancer, one patient with choriocarcinoma, and four pregnant women, however, hCGf3n-like material dissociated from the a-subunit was detected in significant quantities, accounting for 5% to nearly 100% of the total hCGf3 + hCG/3n + hCG/3cf immunoactivity in the individual samples (Fig. 3). The chromatographic data were in good agreement with SDS-PAGE under reducing conditions, in that the less hydrophobic hCG/3n was also smaller-because it dissociated-than the
0 35
40
45
50
55
60
65
15
hCG 10
I?
hCGn 5
mAb
z
z
z
z
z
z
z
z
z -J
0 30
35
40
45
50
55
60
65
Minutes Fig.1. Reversed-phase chromatography of hCG (NIH CR123) (a) and hCGn (C5)(b) on Vydac C4 column. Under theconditions used, the hCGformsdissociatedintotheirconstitutive aand (3-subunits;immunoreactivityof the eluted fractions was analyzed by hCG$ + hCG(3n + hCGpcf assay. The data reveal that hCG(3nisless hydrophobicthan hCG(3.Whereas most of the hCG$ releasedfromthe hCG standard eluted in a position typCaI of Intact hCG(3,smallerportions eluted in the positions of oxidizedhCGP and hCG(3n. 2308
CLINICAL CHEMISTRY, Vol. 40,No. 12,1994
Epitope cl,2cl
c2 c3 c4 l
2
3
4
D5
Fig.2.SOS-PAGE (nonreducing conditions) and immunoblothng of hCGn (C5)and hCG (NIH CR123) with a panel of mAbs. The mAbsrecognizeeitherconformatlonal(c) epltopeson dimerichCG (cic4) or epitopes present on the hCG/3 subunit in both its combined and free
form ((31-(35), as specified in Table 1 Apparently. hCGn lacks two ofthe holo-hCG-speciflcepitopes(ciand c2),but noneofthe hCGPepitopestested.
Table 2. ImmunologIcal actMty of hCGn (C5) In various two-site Immunometric assays. Assay design
Antibody combination
Immunoactivity,
j35a
3/6-1117 3/6-12/17 3/6-INN-FSH-158
44 618 744
ssa
3/6-anti-hCG$ (goat)
669
ci or c2-J3 c1-
1 117-ariti-hCGp (goat) INN-hCG-1O--antf-hCG (goat)
10 3
c2-
INN-hCG-40--anti-hCGfl (goat)
3
c3-$5 c4-
INN-hCG-45--anti-hCG INN-hCG-26-anhi-hCGp
INN-hCG-2-anti-hCG
(goat) (goat)
(goat)
lU/L’
Ii 15 10 6
0
31
36
36
37
36
41
43
46
47
40
51
63
66
67
603 1082
810
Standarduzed with the 1st lAP 75/537. $.hCG(3epitope; c, contormatlonal holo-hCG epitope;FSH, follicle-stimulating hormone.
60
CHOI
Ii 16
10
6
more hydrophobic hCG/3 (Fig. 3). Remarkably, hCGf3n was frequently found to be at the same time abundantly present in urine and totally absent in the serum of the same patient.
0
31
33
36
36 30
rHT1
36
hCG Measurements in Serum and Urine from Patients with Cancer and Gestational Trophoblast Disease
Testicular
Keeping sera for 72 h at room temperature did not result in an increase in the ratio of hCG + hCGn to hCG, nor did it cause a change in the reversed-phase chromatography profile (data not shown). hCG only, hCG + hCGn, and hCG + hCGn + hCG(3 + hCG(3n + hCG/3cf immunoactivities, measured in 292 serum samples collected from patients with nonseminomatous testicular cancer before and during chemotherapy, covered a wide range of hCG concentrations, from
