elevation of free ßsubunit of human choriogonadotropin ...hcg (9, 10). in nonpregnant women and...
TRANSCRIPT
[CANCER RESEARCH 52, 4628-4633. September 1. 1992]
Elevation of Free ßSubunit of Human Choriogonadotropin and Core ßFragment
of Human Choriogonadotropin in the Serum and Urine of Patients withMalignant Pancreatic and Biliary Disease1
Henrik Alfthan,2 Caj Haglund, Peter Roberts, and Ulf-HâkanStenman
Department of Obstetrics and Gynecolog); Haartmaninkatu 2, SF-00290 Helsinki [H. A., U-H. S.J, and Department of Surgery, Kasarmikatu 11-13 [C. H., P. R.J,Helsinki University Central Hospital, SF-OOI30 Helsinki, Finland
ABSTRACTHuman Choriogonadotropin (hCG), its free /} subunit (£fh('(>), and the
core ,(h( (. fragment (c/3hCG) were measured by highly sensitive time-
resolved immunofluorometric assays in the serum and urine of 29 patients with pancreatic cancer, 7 patients with biliary cancer, and 45patients with benign pancreatic or biliary diseases. The results werecompared with those of an age- and sex-matched reference population of
nonpregnant women and men. Of the various forms of hCG assayed inserum, /?hCG showed the best diagnostic accuracy, and c/3hCG was thebest marker in urine. Elevated serum concentrations of /ShCG wereobserved in 72% of the patients with pancreatic cancer, in 6 of 7 patientswith biliary cancer, and in 9% of those with benign disorders. The serumconcentrations of c/3hCG were elevated in 45%, 57%, and 2%, respectively, and those in urine in 55%, 71%, and 11%, respectively. The molarconcentrations of c/JhCG in serum were mostly lower than those of,1M (,. Thus i;h< (i secreted into serum appears to be the main source of
c/ShCG in urine. Provided that they are measured by sufficiently sensitive and specific assays, ßhCGin serum and c/ShCG in urine appear tobe useful markers for pancreatic and biliary cancer.
INTRODUCTION
hCG3 is a glycoprotein hormone consisting of two dissimilarsubunits designated a and ß.In serum and urine, hCG immu-noreactivity occurs in many different molecular forms. Duringpregnancy intact hCG is the main form in serum (1). The proportion of ßhCGis 1-8% (2-6), and that of the so-called corefragment of /ShCG comprises only about 0.03% of total hCGimmunoreactivity (7, 8). In urine c/3hCG is the major form ofhCG (9, 10). In nonpregnant women and in men the concentrations of hCG in serum (11) and hCG, 0hCG, and c/3hCG inurine increase with age. The serum concentrations of 0hCG aresimilar in women and men and they change very little with age,and in women they are 10-20% of those of hCG.4
hCG is an established marker for trophoblastic diseases, inwhich the level of 0hCG in serum may also be elevated (12). Introphoblastic cancer the ratio of/3hCG to hCG is higher than inbenign molar disease or pregnancy (13-15). Elevated levels ofhCG and ßhCGare frequently observed in nonseminomatoustesticular cancer (16), and in 1-2% of the patients, 0hCG is the
Received 12/31/91; accepted 6/23/92.The costs of publication of this article were defrayed in part by the payment of
page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1Supported by grants from the Finnish Cancer Foundation, the Finnish SocialSecurity Institution, the Sigrid Juselius Foundation, and the Magnus EhrnroothFoundation. Parts of the results have been presented as an abstract at the FifthWorld Congress on Gestational Trophoblastic Diseases, London. October 3-5,1990.
2 To whom requests for reprints should be addressed, at the Clinical Laboratory.Department of Obstetrics and Gynecology, Helsinki University Central Hospital,Haartmaninkatu 2, SF-00290 Helsinki, Finland
3 The abbreviations used are: hCG, human Choriogonadotropin; (ihCG, free dsubunit of hCG; ctfhCG, core dhCG fragment; IFMA, time-resolved immunofluorometric assay; ROC, receiver operating characteristic.
4 H. Alfthan. C. Haglund. J. Dabek, and U-H. Stenman. Concentrations ofhCG, rfhCG and oihCG in serum and urine of nonpregnant women and men. Clin.Chem.. in press, 1991.
only form of hCG found in serum (17). Thus, hCG and ßhCGare important serum markers for these cancers.
Elevated serum levels of hCG immunoreactivity are occasionally seen in patients with nontrophoblastic neoplasms like cancer of the stomach (23%), liver (17%), lung (19-33%), andurinary bladder (30%) (18-20), as well as adenocarcinoma (33-50%) (18) and endocrine carcinoma of the pancreas (69%) (21,22). Elevated levels of serum hCG have also been observed in10-18% of patients with nontrophoblastic gynecological cancer, but the hCG level in serum does not correlate with thecourse of the disease, especially when treatment has includedovariectomy. Therefore hCG in serum has not been considereda useful marker for these cancers (23, 24).
When sensitive and specific methods are used to measurehCG, concentrations in the serum offertile women and men arebelow 2 IU/liter, but they increase after the menopause to 2-9ID/liter (11). This may explain some earlier observations ofelevated serum levels in cancer patients. When postmenopausallevels were considered Ozturk et al. (4) did not find elevatedserum levels (>1.5 Mg/liter) of intact hCG in any of 178 patientswith various types of cancer.
In contrast to hCG in serum, hCG immunoreactivity in urineappears to be a useful test for cancer. Papapetrou et al. (25)observed elevated levels in 44% of patients with various nontrophoblastic malignancies, whereas the hCG levels in serumwere elevated in only 17%. The main immunoreactivity in urinewas shown to be a fragment of hCG (26). Later studies haveconfirmed that cßhCGis often elevated in the urine of patientswith gynecological cancer (27, 28), and the levels correlate withthe progress of the disease (29). Some patients with highlyelevated levels of cßhCGin urine have ßhCGin serum (25, 30),suggesting that cßhCGin urine is derived from 0hCG.
To evaluate the usefulness of various hCG assays for thediagnosis of pancreatic cancer and to clarify the origin ofcßhCG in the urine of cancer patients, we have determinedserum and urine levels of hCG, /3hCG, and cßhCGby highlysensitive assays. To evaluate the cancer specificity of hCG immunoreactivity we also studied patients with benign disease.
MATERIALS AND METHODS
Patient Samples. Parallel samples of serum and urine were obtainedfrom patients with ductal adenocarcinoma of the pancreas (n = 29),pancreatitis (n = 25), biliary cancer (n = 7), and benign biliary obstruction (n = 20) (Table 1). Four patients with pancreatic cancer had re-sectable tumors. All samples were stored at —¿�20°Cuntil analyzed.
Assays. hCG, ßhCG,and cßhCGwere quantitated in serum andurine by IFMAs as previously described (5, 7, 11). The sensitivity of theassays for hCG, ßhCG,and c/3hCG was 0.8, 0.45, and 0.44 pmol/liter,respectively. Cross-reaction of hCG and ßhCGin the assay for cßhCGwas 1.2% and 37%, respectively. Cross-reaction of hCG and cßhCGinthe assay for ßhCGwas 0.15% and <0.1%, respectively. In the hCGassay, the cross-reactions of ßhCGand c/3hCG were 0.6% and <0.1 %,
4628
Research. on February 9, 2021. © 1992 American Association for Cancercancerres.aacrjournals.org Downloaded from
tihCG AND CORE tfhCG IN MALIGNANT PANCREATIC AND BILIARY DISEASE
Table 1 Age distribution of the patients studied
PancreaticcancerBiliary
cancerBenign
diseasesAcutepancreatitisChronicpancreatitisBenign
biliary obstructionn29745141120Age
(years)Median737856364273Range45-9130-8118-9118-6624-7151-91
Table 2 Frequency of elevated values ofhCG, ßhCG,and cßhCGin serum ofpatients with malignant and benign pancreatic and hepatobiliary diseases
Cutoff levels based on the 97.5 percentile of benign diseases were used forc/JhCG. For the other forms of hCG the cutoff levels were based on the 97.5percentile of a reference group matched for age and sex.
respectively. The results for cßhCGin serum were corrected for cross-reaction with /3hCG.
Creatinine in urine was quantitated by the modified Jaffémethod(31) with a Kone Progress analyzer (Kone Instruments, Espoo, Finland), using reagents supplied by the manufacturer.
Gel Chromatography. One hundred-/«'samples of serum (n = 3) and1-ml samples of urine (n = 4) were fractionated by gel chromatographyon a 1.6 x 40 cm column of Sephadex G-75 (Pharmacia Fine Chemicals, Uppsala, Sweden). Elution was performed with 50 mmol/liter Tris(pH 7.7), 0.15 mol/liter NaCl, and 0.05 g/liter NaN3. The flow rate was15 ml/h, and 1-ml fractions were collected into tubes prefilled with 100M!assay buffer with a 10-fold protein concentration. hCG, /3hCG, andc/3hCG were quantitated in the fractions by IFMAs using a samplevolume of 200 M!.The elution profiles of the samples were comparedwith those of purified preparations of urinary hCG, |3hCG (5), andc/3hCG (32).
Statistical Analysis. The correlation between measurable concentrations of different markers was analyzed by the Spearman test. ROCcurve analysis was performed as described by Metz (33).
Reference Intervals. The reference intervals for hCG, /3hCG, andc/3hCG in serum and urine have been established in nonpregnantwomen and men.4 The upper reference limit of hCG in the serum of
women below and above 50 years of age is 8.6 and 15.5 pmol/liter,respectively. In men the corresponding figures are 2.1 and 6.1 pmol/liter, respectively. For ßhCGin the serum of women and men thefigures are 1.6 and 2.0 pmol/liter, and 1.9 and 2.1 pmol/liter, respectively. The upper reference limit of cßhCGin serum could not bereliably estimated because levels exceeding the detection limit occurredin less than 3% of the men and nonpregnant women studied. Thereforea provisional limit was calculated on the basis of the 97.5 percentile ofpatients with benign diseases. This gave a value of 1.1 pmol/liter.
In urine the upper reference limit of hCG in women below and above50 years of age is 0.82 and 1.34, and in men 0.15 and 0.82 pmol/mmolcreatinine, respectively. For ßhCGthe corresponding values are 0.24and 0.64, and 0.11 and 0.35 pmol/mmol creatinine, respectively, andfor dShCG 0.51 and 1.25, and 0.75 and 0.83 pmol/mmol creatinine,respectively.
PancreaticcancerBiliary
cancerBenign
diseasesAcute pancreatitisChronic pancreatitisBenign biliary
obstructionn29745
141120hCG4(14%)1
(14%)3
(7%)0 (0%)0 (0%)3(15%)/ihCG21
(72%)6
(86%)4
(9%)I (7%)0 (0%)3(15%)ctfhCG13(45%)4
(57%)1
(2%)1 (7%)0 (0%)0 (0%)hCG
+ÕJhCG12(41%)5(71%)3
(7%)0 (0%)0 (0%)3 (7%)
Conversion Factors. International units of hCG (third InternationalStandard) were converted to molar units by using an M, of 36,700 anda specific activity of 9,286 lU/mg (34). For ßhCGan M, of 22,000 wasused (35), and for c/3hCG an M, of 10,000 (32).
RESULTS
hCG, 0hCG, and c/3hCG ¡nSerum. Elevated concentrationsof ßhCGin serum were observed in 72% of the patients withpancreatic cancer and in 9% of those with benign diseases (Fig.1; Table 2). In pancreatic cancer the median concentration was2.4-fold and the highest level 120-fold the cutoff level. In patients with benign diseases the median level was 39% of thecutoff level and the maximum elevation 2.2-fold. In biliary cancer ßhCGwas elevated in 6 of 7 patients (86%) (Fig. 1). hCG inserum was elevated in 4 (14%) patients with pancreatic cancer,and the highest level was 1.8-fold the cutoff level. cßhCGwaselevated in 45% of the patients with pancreatic cancer and in 4of the 7 (57%) patients with biliary cancer. The highest valuewas 86-fold the cutoff level. The concentrations of ßhCGwerehigher than those of cßhCGin all but three samples. In thesesamples the elevation was small, 1.1-2.3-fold, and the concentration of ßhCGwas normal. Of the six patients with normalserum levels of ßhCGtwo had a small resectable tumor.
hCG, /ShCG, and c/8hCG in Urine. Elevated concentrationsof cßhCGin urine were observed in 55% of the patients withpancreatic cancer and in 11% of those with benign disease. Thecorresponding figures for urinary ßhCGwere 28% and 13%,respectively (Fig. 2; Table 3). The highest elevation of urinaryqShCG in pancreatic cancer was 58-fold, and in benign diseaseit was 1.7-fold the cutoff level. In biliary cancer urinary /3hCG
1000
S 100
iI
10 •¿� ,
0.1 •¿�-
S-hCGBenign
diseaseCancer•.--
3~°_£__*.E*AiI<50
>50 <50 >50 >50 <50>50Women
Men Women MenS-ßhCGBenign
diseaseo
y¿öä-i^^_A9c.â^B
86<50>50 <50>50Women
MenCancer•:
-__».«•
'm9>50
<50 >50
Women MenS-cßhCGBenign
disease Cancer•A
•¿�A-A"
^-^r§
A^ÛUAO 9PPÃŽ55 44* •¿�If<50
»0 <50 >50 >50 <50 >50Women Men Women Men
1000
100 i=
10 o
ieoü
0.1
Fig. l. Concentrations of hCG, ffhCG and cßhCGin serum of patients with acute pancreatitis (O), chronic pancreatitis (Q), benign biliary obstruction (A), pancreaticcancer (•),and biliary cancer (A). Horizontal bars, cutoff levels based on the 97.5 percentile of a reference group matched for age and sex. , detection limit ofthe assay.
4629
Research. on February 9, 2021. © 1992 American Association for Cancercancerres.aacrjournals.org Downloaded from
/JhCG AND CORE ßhCGIN MALIGNANT PANCREATIC AND BILIARY DISEASE
U-hCG U-ßhCG U-cßhCG
s s§öu og Iü E
10•¿�1
•¿�3.1
-
0.01 -Benign
disease¿£>
aQAnA<60
>50 <50Õ50Women
MenCancer•i
72:i550
<50>50Women
MenBenign
diseaseA-r
8103
A^pO•î
-Sfr<50>50 <50>50Women
MenCancer*
A*•
•¿�.•«4'.
«?50
<50iXWomen
MenBenign
disease£%§
^OOA<50
£50 <50«0Women
MenCancer
•¿�AAe"
A"»0
<50>50Women
Men
•¿�10
100
ü 03
§I1 ni nj
¿¡03
8 I01 S i
0.01
Fig. 2. Concentrations of hCG, 0hCG, and cffhCG in urine of patients with malignant and benign diseases. For details see Fig. 1.
was elevated in 4 of 7 (57%) and c/3hCG was elevated in 5 of 7patients (71%) (Fig. 2). The hCG levels in urine were normal inall samples.
Follow-up of Patients with Pancreatic Cancer. In a patientwith a resectable pancreatic tumor the concentrations of/3hCGin serum decreased after surgery and started to rise 2 monthsbefore a recurrence was clinically observed (Fig. 3). In anothertwo patients with normal preoperative serum levels of ßhCGnoelevation was seen in connection with a recurrence (not shown).
Receiver Operating Characteristics. ROC curve analysis wasused to evaluate the diagnostic value of various forms of hCGand to determine the optimal cutoff level for differentiationbetween pancreatic cancer and benign pancreatic and biliarydisease. This analysis showed that ßhCGin serum was a bettermarker than c0hCG in urine (Fig. 4). The other forms of hCGin serum and urine had much lower sensitivity and specificity.The cutoff value giving 90% specificity in benign disease was1.7 pmol/liter for serum 0hCG and 1.02 pmol/mmol creatininefor urinary cßhCG.Using these cutoff levels the frequency ofelevated levels of 0hCG in the serum of patients with pancreaticcancer was 72%, and that of urinary c/3hCG was 59%.
Correlations between the Concentrations of Different Formsof hCG. A strong positive correlation was seen between serum/3hCG and urinary cßhCGin cancer patients (r = 0.703, P <0.001). Similar correlations were observed between urinaryßhCGand cßhCG(r = 0.665, P < 0.001) and between serumand urinary 0hCG (r = 0.617, P = 0.001). The correlationbetween serum hCG and urinary cßhCGwas not significant(r = 0.101, P = 0.654). In benign disease the levels of urinaryc/3hCG correlated with the serum (r = 0.674, P = 0.002) andurinary levels of hCG (r = 0.575 P = 0.025) and ßhCG(r =0.482, P = 0.006) but not with those of serum ^hCG (r = 0.231,P =0.190).
Table 3 Frequency of elevated values ofhCG, fihCG, and cßhCGin urine ofpatients with malignant and benign pancreatic and hepatobiliaiy diseases
The cutoff levels are based on the 97.5 percentile of a reference group matchedfor age and sex.
1000l
PancreaticcancerBiliary
cancerBenign
diseasesAcute pancreatitisChronic pancreatitisBenign biliary
obstructionn29745
141120hCG1
(3%)0
(0%)1
(2%)0 (0%)1 (9%)0 (0%)ÖhCG8(28%)4(57%)6(13%)
2(14%)1 (9%)3(15%)cßhCG16(55%)5(71%)5(11%)
0 (0%)1 (9%)4(20%)hCG
+ ßhCG+qîhCG11(38%)3(43%)2
(4%)1 (7%)0 (0%)1 (5%)
0.10 100 200 300 400 500
Time (days)Fig. 3. Follow-up of a patient with pancreatic cancer. The levels decreased after
surgical removal of the tumor and started to increase 2 months before a recurrencewas observed clinically.
Gel Chromatography of Serum and Urine. ,3hCG and cßhCGin two serum and four urine samples were characterized by gelchromatography. In serum both /3hCG and c/3hCG eluted at thesame position as the corresponding purified substances. Meanrecovery after gel chromatography was 138% (range, 99-169%). In two of four urine samples part of the cßhCGimmu-noreactivity (18-27%) eluted between /3hCG and cßhCG,indicating an intermediate molecular size (Fig. 5). This componentreacted weakly in the 0hCG assay. The cßhCGimmunoreactiv-ity in serum eluted at the expected position (Fig. 6). The concentrations of cßhCGin these samples were 10, 27 and 38pmol/liter, respectively, corresponding to 15%, 37%, and 17%of the ßhCGlevel. In these patients the concentration of cßhCGin urine was 88-632 pmol/liter, and c/3hCG accounted for 82-96% of total hCG immunoreactivity.
DISCUSSION
This study shows that patients with pancreatic cancer oftenhave elevated concentrations of ßhCGin serum and cßhCGinurine, and their concentrations are strongly correlated. In contrast, the concentrations of hCG in serum and urine are normalor only marginally elevated. cßhCGimmunoreactivity also occurs in the serum of many cancer patients, but the levels arelower than those of ßhCG.It has been shown that c/ShCG is themain form of hCG found in urine after i.v. injection of purifiedßhCG,and cßhCGis thought to be formed by the proteolyticbreakdown of ßhCGin the kidneys (10, 36). Furthermore,
4630
Research. on February 9, 2021. © 1992 American Association for Cancercancerres.aacrjournals.org Downloaded from
ÃihCGAND CORE fihCG IN MALIGNANT PANCREATIC AND BILIARY DISEASE
1.0
§0.8 •¿�
IFig. 4. ROC curve analysis of different ^ 0.6forms of hCG in serum and urine as markers >for pancreatic cancer. The reference group ~consisted of patients with pancreatitis (n = 25) o 04and biliary obstruction (n = 20). A. curves for Q.hCG, /JhCG. and crfhCG in serum; B, curves 9jfor urine. »-I- 0.2
0.2 0.4 0.6 0.8 1.0False positive fraction
0 0.2 0.4 0.6 0.8 1.0False positive fraction
ßhCG cßhCG
O
Oü.eça
O20 70 8030 40 50 60
Fraction numberFig. 5. Gel chromatography on Sephadex G-75 of urine from a patient with
biliary cancer. fihCG and cfihCG in the fractions were quantitated by IFMA.
ßhCG cßhCG
re
8oü
20 30 40 50 60Fraction number
70 80
Fig. 6. Gel chromatography on Sephadex G-75 of serum from the patient inFig. 5. fihCG (O) and c/3hCG (•)¡nthe fractions »erequantitated by IFMA.
ßhCGhas been shown to be the major form of hCG immunore-
activity in tumor extracts and to be secreted by various tumorcells ¡nvitro (37). These and the present results suggest thatc/3hCG in the urine of cancer patients is mainly derived fromßhCGsecreted by the tumor and that part of it appears to beformed before excretion into urine.
Characterization of cßhCGimmunoreactivity in serum by gelchromatography indicated that it behaved as qShCG purifiedfrom urine. This is in agreement with our finding for cßhCGin
pregnancy serum (7). The cßhCGin pregnancy and chorioniccancer serum has been claimed to occur in a complex, which hasto be dissociated and separated by gel chromatography before itcan be detected (38). Our attempts to reproduce these findingshave not been successful.5
In two of four urine samples analyzed by gel chromatography, part of the cßhCGimmunoreactivity eluted with a molecular size between those of cjShCG and /3hCG. However, thelevels were low, which prevented further characterization. Thiscomponent could be an intermediate degradation product ofiShCG. Partially degraded forms of hCG have been demonstrated in the serum and urine of patients with chorionic cancer(39).
The median ratio of cßhCGto hCG in serum was 0.3. Because of this and the 37% cross-reaction of 0hCG in the c/3hCGassay the results for ßhCGin serum are less reliable than thosein urine. However, the gel Chromatographie experiments confirmed the results of the direct assays. In urine the cßhCGconcentrations were about 3-fold those of/3hCG. Therefore, thecross-reaction had only a minor effect (5%) on the urine values.Cross-reaction of hCG in the ßhCGassay (0.15%) was insignificant for interpretation of the results because the hCG concentrations were relatively low (<45 pmol/liter). The hCG assayprobably does not detect ßhCGat all, since the tracer antibodyin this assay is directed toward the a subunit of hCG. Theapparent cross-reaction of 0.6% is likely to be due to contamination of the |3hCG preparation with hCG.
Because elevated levels of ßhCGand c/3hCG may occur inbenign disease, we also compared the clinical performance ofthe various assays by ROC curve analysis using patients withbenign diseases as a reference group. At cutoff levels giving 90%specificity the sensitivity of serum /ShCG was 72% and that ofurinary cßhCGwas 59%. When the cutoff was lowered to give75% specificity the sensitivity of both markers was 79%. Thissuggests that the sensitivity and specificity of serum 0hCG andurinary c/3hCG are comparable with those of CA 19-9 andcarcinoembryonic antigen, which are established markers forpancreatic and biliary cancer. In a similar study, these markerswere shown to have sensitivities of 54% and 78% at cutoff levelsgiving specificities of 76% and 78%, respectively (40).
In earlier studies showing elevated levels of hCG immunoreactivity in the serum of cancer patients, radioimmunoassaysmeasuring both hCG and jShCG were used. On the basis of theshape of the dilution curves Braunstein et al. (18) drew the
5 H. Alfthan and U-H. Stenman, unpublished results.
4631
Research. on February 9, 2021. © 1992 American Association for Cancercancerres.aacrjournals.org Downloaded from
(JhCG AND CORE rfhCG IN MALIGNANT PANCREATIC AND BILIARY DISEASE
conclusion that the immunoreactivity consisted of hCG. In ourstudy, only four serum samples (14%) and one urine sample(3%) had elevated hCG levels, and the levels were only 2-3-foldthe cutoff level and only 50% higher than the highest levels (9IU/liter) occasionally observed in postmenopausal women (11).This finding is in agreement with the results of Ozturk et al. (4).For comparison, the highest concentrations of/3hCG were 120-fold and that of cßhCG86-fold the cutoff level. Therefore,/JhCG appears to be the major form of hCG causing elevatedhCG immunoreactivity in the serum of patients with pancreaticand biliary adenocarcinoma, but intact hCG may be expressedby some tumors. Alternatively, 0hCG produced by tumorscould recombine with pituitary a subunit in serum. This possibility seems unlikely because recombination of a and ßsubunitsusually requires much higher concentrations than observed inthe present study (41).
In patients with endocrine pancreatic cancer, ßhCGhas beenfound to be the main form of hCG immunoreactivity in serum,but «subunit may also be produced (21, 22). Expression of/ShCG is common in various tumors and cell lines (37), butßhCGin serum has not been found thus far to be a useful tumormarker (4). This has apparently been due to lack of sufficientlysensitive and specific assays. A high sensitivity is essential because of the low levels of ßhCGin normal serum. The methodused in the present study is 8-70-fold more sensitive than earlier ones (4, 15, 42). /3hCG in serum was characterized by gelchromatography, which confirmed that the molecular size ofimmunoreactive material corresponded to that of purifiedßhCG.The mean recovery of ßhCGafter chromatography was138%, suggesting that nonspecific interference by serum andurine components (43) caused a slight underestimation in theIFMAof/3hCG.
We have also analyzed whether total hCG-like immunoreactivity, i.e., the sum of hCG, ßhCG,and c/3hCG in serum, couldbe used, rather than ßhCGin serum or c/ShCG in urine. Thisroughly corresponds to what is measured by competitive immu-noassays, e.g., radioimmunoassay. The sensitivity of this assaymethod was 41% as compared with 72% for |3hCG. The concentration of |ohCG in postmenopausal serum is only 10-20%of the total hCG immunoreactivity, and in contrast to the levelsof hCG those of ßhCGincrease only slightly with age. Becauseof the large range in hCG levels even a 3-5-fold increase inßhCGwill not necessarily increase total hCG immunoreactivityabove the cutoff level. Therefore, ßhCGis superior to total hCGimmunoreactivity as a serum marker for pancreatic cancer,whereas cßhCGis the best urine marker. In this study /3hCG inserum was a better marker than cßhCGin urine, apparentlybecause the latter can be derived not only from 0hCG but alsofrom hCG.
The elevation of |3hCG in serum and cßhCGin urine inbenign diseases is a new finding. It may indicate that expressionofßhCG is not limited to neoplastic and trophoblastic cells butmay also occur in nonmalignant cells in connection with inflammatory reactions or tissue regeneration. This would actually not be surprising, considering the fact that other tumormarkers, e.g., CA 19-9, CA-50, and carcinoembryonic antigen,also show a similar elevation in benign diseases (40, 44). However, the elevation of /ShCG in serum and cßhCGin urine isinfrequent (9-11%) and moderate (2-fold) in comparison with9-15-fold elevations and 16-35% elevated levels of CA 19-9,CA-50, and carcinoembryonic antigen in patients with benignpancreatic and biliary diseases (45, 46). Thus the elevation in
benign disease does not invalidate the use of/3hCG and cßhCGas tumor markers.
ACKNOWLEDGMENTS
Drs. Robert E. Canfield, John O'Connor, Steven Birken, and Alex
ander Krichevsky of the College of Physicians and Surgeons of Columbia University (New York, NY) are gratefully acknowledged for theirkind gift of c/ShCG and monoclonal antibodies. The expert technicalassistance of Taina Grönholm. Maarit Leinimaa, and Anja Maki isgratefully acknowledged.
REFERENCES
1. Braunstein, G. D., Rasor. J.. Adler, D., Danzer, H., and Wade, M. E. Serumhuman chorionic gonadotropi levels throughout normal pregnancy. Am. J.Obstet. Gynecol., 75: 678-681, 1976.
2. Cole, L. A., Kroll, T. G., Ruddon, R. W., and Hussa, R. O. Differentialoccurrence of free ßand free n subunits of human chorionic gonadotropin(hCG) in pregnancy sera. J. Clin. Endocrinol. Metab., 58: 1200-1202, 1984.
3. Hay. D. L. Discordant and variable production of human chorionic gonadotropin and its free n- and rf-subunits in early pregnancy. J. Clin. Endocrinol.Metab.,«/: 1195-1200. 1985.
4. Ozturk, M.. Bellet. D.. Manil, L.. Hennen, G., Frydman, R., and Wands. J.Physiological studies of human chorionic gonadotropin (hCG). ahCG andrfhCG as measured by specific monoclonal immunoradiometric assays. Endocrinology, 120: 549-558. 1987.
5. Alfthan. H., Schröder.J., Fräser.R., Koskimies, A.. Halila, H., and Stenman.U-H. Choriogonadotropin and its ff subunit separated by hydrophobic-inter-action chromatography and quantified in serum during pregnancy by time-resolved immunofluorometric assays. Clin. Chem.. 34: 1758-1762, 1988.
6. Thomas. C. M. G., Reijnders, F. J. L.. Segers, M. F. G., Doesburg, W. H.,and Rolland, R. Human Choriogonadotropin (hCG): comparison betweendeterminations of intact hCG, free hCG /J-subunit, and "total" hCG + ßinserum during the first half of high-risk pregnancy. Clin. Chem., 36:651-655.1990.
7. Alfthan, H., and Stenman, U-H. Pregnancy serum contains the ii-core fragment of human Choriogonadotropin. J. Clin. Endocrinol. Metab., 70: 783-787, 1990.
8. Wchmann. R. E., Blithe, D. L., Akar, A. H., and Nisula, B. C. Disparitybetween rf-core levels in pregnancy urine and serum: implications for theorigin of urinary 0-core. J. Clin. Endocrinol. Metab., 70: 371-378, 1990.
9. Schroeder. H. R.. and Halter, C. M. Specificity of human /i-choriogonadot-ropin assays for the hormone and for an immunoreactive fragment present inthe urine during normal pregnancy. Clin. Chem.. 29: 667-671. 1983.
10. Kalo, Y., and Braunstein, G. D. rf-Core fragment is a major form of immu-noreaetive urinary chorionic gonadotropin in human pregnancy. J. Clin.Endocrinol. Metab.. 66: 1197-1201. 1988.
11. Stenman. U-H., Alfthan, H.. Ranta. T.. Vartiainen, E., Jalkanen, J., andSeppälä,M. Serum levels of human chorionic gonadotropin in nonpregnantwomen and men are modulated by gonadotropin releasing hormone and sexsteroids. J. Clin. Endocrinol. Metab., 64: 730-736, 1987.
12. Vaitukaitis, J. L., and Ebersole, E. R. Evidence for altered synthesis of humanchorionic gonadotropin in gestational trophoblastic tumors. J. Clin. Endocrinol. Metab.. 42: 1048-1055, 1975.
13. Cole, L. A., Hartle, R. J., Laferla, J. J.. and Ruddon. R. W. Detection of freeli subunit of human chorionic gonadotropin (hCG) in cultures of normal andmalignant trophoblastic cells, pregnancy sera, and sera of patients with cho-riocarcinoma. Endocrinology. 113: 1176-1178. 1983.
14. Stenman. U-H.. Alfthan. H.. and Halila, H. Determination of chorionicgonadotropin in serum of nonpregnant subjects and patients with trophoblastic cancer by time-resolved immunofluorometric assay (abstract). In: XIIAnnual Meeting, International Society for Oncodevelopmental Biology andMedicine, Houston, 1984, Abstract 97.
15. Fan, C., Goto. S., Furuhashi, Y., and Tomoda, Y. Radioimmunoassay of thefree d-subunit of human chorionic gonadotropin in trophoblastic disease. J.Clin. Endocrinol. Metab., 64: 313-318, 1987.
16. Mann. K., and Karl, H-J. Molecular heterogeneity of human chorionic gonadotropin and its subunits in testicular cancer. Cancer 11Miila. I. 52:654-660,1983.
17. Light, P. A., Foster, J. P., Felton. T.. Eckert, H., and Tovey, K. C. Molecularheterogeneity of chorionic gonadotropin in some testicular cancer patients.Lancet. /: 1284, 1983.
18. Braunstein, G. D., Vaitukaitis, J. L., Carbone, P. P., and Ross, G. T. Ectopieproduction of human chorionic gonadotrophin by neoplasms. Ann. Int.Med., 78: 39-45. 1973.
19. Dexeus, F., Logothetis, C.. Hossan, E., and Samuels, M. L. Carcinoembryonic antigen and .>Inumili chorionic gonadotropin as serum markers foradvanced urothelial malignancies. J. Urol.. 136: 403-407. 1986.
20. Gropp. C., Havemann. K.. and Scheuer. A. Ectopie hormones in lung cancerpatients at diagnosis and during therapy. Cancer (Phila.). 46:347-354, 1980.
21. Kahn. C. R., Rosen. S. W.. Weintraub. B. D.. Fajans. S. S., and Corden, P.Ectopie production of chorionic gonadotrophin and its subunits by islet cell
4632
Research. on February 9, 2021. © 1992 American Association for Cancercancerres.aacrjournals.org Downloaded from
rfhCG AND CORE /ihCG IN MALIGNANT PANCREATIC AND BILIARY DISEASE
tumors; a specific marker for malignancy. N. Engl. J. Med., 297: 565-569,1977.
22. Oberg. K.. and Wide, L. hCG and hCG subunits as tumour markers inpatients with endocrine pancreatic tumours and carcinoids. Acta Endocrinol..9«:256-260, 1981.
23. Rutanen. E-M., and Seppala. M. The hCG-ff subunit radioimmunoassay innontrophoblastic gynecological tumors. Cancer (Phila.), 41: 692-696, 1978.
24. Carenza, L., di Gregorio, R.. Mocci, C., Moro, M., and Pala, A. Ectopiehuman chorionic gonadotropin: gynecological tumors and nonmalignantconditions. Gynecol. Oncol.. 10: 32-39, 1980.
25. Papapetrou, P. D., Sakarelou, N. P., Braouzi. H., and Fessas. P. H. Ectopieproduction of human chorionic gonadotropin (hCG) by neoplasms: the valueof measurements of immunoreactive hCG in the urine as a screening procedure. Cancer (Phila.), 45: 2583-2592. 1980.
26. Papapetrou, P. D.. and Nicopoulou. S. C. The origin of human chorionicgonadotropin /i-subunit-core fragment excreted in the urine of patients withcancer. Acta Endocrinol.. 112: 415-422, 1986.
27. Cole, L. A., Schwartz, P. E.. and Wang. V. Urinary gonadotropin fragments(UGF) in cancers of the female reproductive system. I. Sensitivity and specificity, comparison with other markers. Gynecol. Oncol.. 31: 82-90, 1988.
28. O'Connor. J. F., Schlauerer, J. P., Birken, S., Krichevsky, A., Armstrong, E.
G., McMahon. D.. and Canfield, R. E. Development of highly sensitiveimmunoassays to measure human chorionic gonadotropin. its fi-subunit. andii core fragment in the urine: application to malignancies. Cancer Res., 48:1361-1366.1988.
29. Wang. V.. Schwartz. P. E.. Chambers, J. T., and Cole, L. A. Urinary gonadotropin fragments (UGF) in cancers of the female reproductive system. II.Initial serial studies. Gynecol. Oncol.. 31: 91-100, 1988.
30. Hattori, M., Yoshimoto. Y.. Matsukura. S.. and Fujita. T. Qualitative andquantitative analyses of human chorionic gonadotropin and its subunits produced by malignant tumors. Cancer (Phila.). 46: 355-361, 1980.
31. Helger, R., Rindfrey, H., and Hilgcnfcldt, J. Eine Methode zur direktenBestimmung des Creatinins in Serum und Harn ohne Entwei/iung nach einermodifizierten Jaffé-Methode.Z. Klin. Chem. Klin. Biochem.. 12: 344-349.1974.
32. Birken. S.. Armstrong, E. G.. Kolks, M. A. G., Cole, L. A.. Agosto, G. M.,Krichevsky. A., Vaitukaitis. J. L., and Canfield, R. E. Structure of the humanchorionic gonadotropin fi-subunit fragment from pregnancy urine. Endocrinology. 123: 572-583. 1988.
33. Metz, C. E. Basic principles of ROC analysis. Semin. NucÃ.Med.. S: 283-298, 1978.
34. Storring, P. L.. Gaines-Das, R. E., and Bangham. D. R. International reference preparation of human chorionic gonadotrophin for immunoassay: po
tency estimates in various bioassay and protein binding assay systems; andinternational reference preparations of«and fi subunits of human chorionicgonadotrophin for immunoassay. J. Endocrinol., 84: 295-310, 1980.
35. Morgan, F. J., Birken, S., and Canfield. R. E. The amino acid sequence ofhuman chorionic gonadotropin: the «subunit and the •¿�subunit. J. Biol.Chem., 250: 5247-5258, 1975.
36. Wehmann. R. E., and Nisula, B. C. Characterization of a discrete degradationproduct of human chorionic gonadotropin fi-subunit in humans. J. Clin.Endocrinol. Metab.. SI: 101-105, 1980.
37. lies, R. K., Purkis, P. E., Whitehead, P. C, Oliver, R. T. D., Leigh, I., andChard, T. Expression of beta human chorionic gonadotropin by nontrophoblastic non-endocrine "normal" and malignant epithelial cells. Br. J. Cancer,61: 663-666. 1990.
38. Kardana. A., and Cole, L. A. Serum hCG fi-core fragment is masked byassociated macromolecules. J. Clin. Endocrinol. Metab., 71: 1393-1395,1990.
39. Puisieux, A., Bellet, D., Troalen. F., Razafindratsita. A.. Lhomme, C., Bo-huon. C.. and Bidart. J-M. Occurrence of fragmentation of free and combinedforms of the fi-subunit of human chorionic gonadotropin. Endocrinology,126: 687-694, 1990.
40. Haglund. C., Roberts. P. J., Kuusela, P., Scheinin, T. M., Mäkelä,O., andJalanko. H. Evaluation of CA 19-9 as a serum tumour marker in pancreaticcancer. Br. J. Cancer. 53: 197-202. 1986.
41. Ingham, K. C.. Weintraub. B. D.. and Edelhoch, H. Kinetics of recombination of the subunits of human chorionic gonadotropin. Effect of subunitconcentration. Biochemistry, 15: 1720-1726, 1976.
42. Norman. R. J.. Poulton. T.. Gard, T.. and Chard. T. Monoclonal antibodiesto human chorionic gonadotropin: implications for antigenic mapping, ¡m-munoradiometric assays, and clinical applications. J. Clin. Endocrinol.Metab., 61: 1031-1038, 1985.
43. Stenman, U-H., Alfthan. H., and Turpeinen, U. Method dependence of interpretation of immunoassay results. Scand. J. Clin. Lab. Invest., 51 (Suppl.206): 86-94, 1991.
44. Gupta, M. K., Arciaga. R.. Bocci, L., Tubbs, R., Bukowski, R., and Deodhar,S. D. Measurment of a monoclonal-antibody-defmed antigen (CA 19-9) inthe sera of patients with malignant and nonmalignant disease. Cancer (Phi-la.). 56:277-283. 1985.
45. Haglund, C. Tumour marker antigen CAI25 in pancreatic cancer: a comparison with CAI9-9 and CEA. Br. J. Cancer. 54: 897-901. 1986.
46. Haglund. C., Kuusela. P.. Jalanko. H., and Roberts, P. J. Serum CA 50 as atumor marker in pancreatic cancer: a comparison with CA 19-9. Int. J.Cancer, 39: 477-481, 1987.
4633
Research. on February 9, 2021. © 1992 American Association for Cancercancerres.aacrjournals.org Downloaded from
1992;52:4628-4633. Cancer Res Henrik Alfthan, Caj Haglund, Peter Roberts, et al. Diseaseand Urine of Patients with Malignant Pancreatic and Biliary
Fragment of Human Choriogonadotropin in the SerumβCore Subunit of Human Choriogonadotropin andβElevation of Free
Updated version
http://cancerres.aacrjournals.org/content/52/17/4628
Access the most recent version of this article at:
E-mail alerts related to this article or journal.Sign up to receive free email-alerts
Subscriptions
Reprints and
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Permissions
Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)
.http://cancerres.aacrjournals.org/content/52/17/4628To request permission to re-use all or part of this article, use this link
Research. on February 9, 2021. © 1992 American Association for Cancercancerres.aacrjournals.org Downloaded from