Phenotyping of Lewis and secretor HBGA from saliva and detection of new FUT2 gene SNPs from young children from the Amazon presenting acute gastroenteritis and respiratory infection

The Histo-blood group antigens (HBGA) are host genetic factors associated with susceptibility to rotavirus (RV) and human norovirus (HuNoV), the major etiological agents of viral acute gastroenteritis (AGE) worldwide. The FUT2 gene expressing the alpha-1, 2-L- fucosyltransferase enzyme is important for gut HBGA expression, and also provides a composition of the phenotypic profile achieved through mutations occurring in populations with different evolutionary histories; as such, it can be considered a genetic population marker. In this study, Lewis and secretor HBGA phenotyping was performed using 352 saliva samples collected from children between three months and five years old born in the Amazon (Brazil, Venezuela and English Guyana) presenting AGE or acute respiratory infection (ARI), the latter considered as control samples. The total of children phenotyped as secretors was 323, corresponding to 91.80%. From these, 207 (58.80%) had a Le (a + b+) profile. The HBGA profiles were equally found in children with AGE as well as with ARI. The rs1047781 of the FUT2 gene was not detected in DNA from saliva cells with a Le (a+b+) profile. However, mutations not yet described in the FUT2 gene were observed: missense 325A>T, 501C>T, 585C>T, 855A>T and missense substitutions 327C>T [S (Ser) > C (Cys)], 446 T>C [L(Leu) > P(Pro)], 723C>A [N(Asn) > K(Lys)], 724A>T [I(Ile) > F(Phe)], 736C>A [H(His) > N(Asn)]. The SNP distribution in the FUT2 gene of the analyzed samples was very similar to that described in Asian populations, including indigenous tribes.

FUT3 and FUT2 genotyping and glycoconjugate profile Lewisb as a protective factor to Toxoplasma gondii infection

The interaction between the ABO, FUT2 and FUT3 genes results in the synthesis of different glycoconjugates profiles expressed in gastrointestinal tract. Moreover, the protozoan Toxoplasma gondii, which causes toxoplasmosis, utilizes this organ as an infection route. We analyzed the frequencies of the different glycoconjugate profiles which were determined by phenotyping ABO and genotyping the status secretor (FUT2; substitution G428A) and Lewis (FUT3; substitution T202C and C314T) histo-blood systems, assessed by PCR-RFLP and PCR-SSP, respectively. A total of 244 pregnant women (G1: Seropositive; G2: Seronegative) for IgG T. gondii antibodies were enrolled. IgG anti-T. gondii antibodies were determined by ELISA. G1 was composed of 158 (64.8%) sample and G2 by 86 (36.2%). The glycoconjugate profile was accessed in 151 seropositive and 85 seronegative samples by the combination of ABO and Lewis phenotyping as well as FUT2 and FUT3 genotyping. In G1, 36 (22.8%) presented the glycoconjugate profile ALeb, 5 (3.3%) A, 13 (8.6) BLeb, 1 (0.6%) B, 41 (27.1%) Leb, 13(8.6%) H, 38(25.2%) Lea and 4 (2.6%) Lec. G2 was composed of 13 (15.3%) of ALeb, 15 (17.6%) BLeb, 1 (1.2%) B, 42 (49,4%) Leb and 14 (16.5) Lea. H and Lec glycoconjugate profiles were not found in G2. The frequencies of the glycoconjugates profiles Leb (p = 0.001) and H (p = 0.005) were significantly different compared between G1 and G2. The glycoconjugate profile H inferred from the ABO phenotyping and FUT3 and FUT2 genotyping is associated with infection by T. gondii in pregnant women and the Leb profile appears to protect the infection by this parasite.

Distribution and impact of erythrocyte Lewis-system antigens on patients with ischemic heart diseases in the west of Georgia

The aim of the research is to reveal the cases of Lewis System Antigens Phenotype in West Georgia and setting the connection between antigens expressivness and IHD. Therefore, we have phenotypically tested 393 people (236 healthy donors; average age 42±7,5 and 157 patients ill with ischemic heart diseases; average age 62,5±7,5). In accordance with the findings, the number of Lewis -antigens among healthy population is 46,6% (110 ±4,8; p<0.05) with Lea-b+ phenotype; 30,9% - with Lea-b- phenotype(73±2,9; p<0.03); 19% - with Lea+b- phenotype-(47±1,7; p<0.03) Only 2,6% cases of phenotype Lea+b+ (6±0,2; p<0.02),were revealed among healthy population. As for the patients with ischemic heart diseases we got the following results: 41% cases of Lea-b- phenotype (65±3,9; p<0.05); 32,8% - Lea-b+ (51±3,2; p<0.03); 21,1% - Lea+b- - phenotype 33±2,9; p<0.03) and 5,6% - Lea+b+ phenotype (8±1,2; p<0.02). On the whole, in the West Georgia the most frequent phenotype is Lea-b+ among healthy population and Lea-b- phenotype among people with ischemic heart diseases. Research was carried out in a control group according to Lewis antigen phenotype. People were separated in two groups; I group -healthy people with Lea-b- phenotype and II group - healthy people with Lea-b+ and Lea+b- phenotypes. On the basis of the research we concluded that people in the first group (with Lea-b- phenotype) had a high BMI, arterial hypertension and lower indexes of high density lipoprotein and triglyceride than the people in the second group(with Lea-b+ and Lea+b- phenotypes. These kinds of changes (characterised to the people with Lea-b- phenotype) are associated with a high risk of ischemic diseases and atherosclerosis. To sum up, people with Lea-b- phenotype have a high risk of ischemic heart disease. In accordance with the findings, Lewis phenotype research can be carried out to detect HID and other diseases as well (hypertension, ischemic insult and insulin-dependent diabetes mellitus).

Investigation of unexpected serum CA19-9 elevation in Lewis-negative cancer patients

BACKGROUND

Cancer patients with a Lewis (a-b-) phenotype have no carbohydrate antigen 19-9 (CA19-9) in their serum. However, we found a small but distinct elevation in the serum CA19-9 level in three cancer patients with the Lewis-negative phenotype. Here, we investigated the reason of such phenomena.

METHODS

Six cancer patients with a Lewis-negative phenotype were selected by very low CA19-9 concentrations: three showed a small elevation (Group A) and the other three showed no elevation (Group B) in the serum CA19-9. We investigated the difference by analyzing the Lewis/Secretor genotypes.

RESULTS

All of the six patients with a Le (a-b-) phenotype were genuine Le-negative genotypes: four individuals were homozygous for le1 (le(59,508)), one patient was compound heterozygous for le1 (le(59,508)) and le2 (le(59,1067)) and one patient was compound heterozygous for le1 and le(202,314). As for the Secretor gene, the three patients in Group B were homozygous for Se2 (one patient) or compound heterozygous for Se2 and sej (two patients), while the patients in Group A were all homozygous for sej genotypes.

CONCLUSIONS

Even genuinely Le-negative patients, who genetically lack the Le enzyme and theoretically never produce CA19-9, occasionally show a slight increase in serum CA19-9 level when they are homozygous for Se-negative genotypes and suffer from advanced cancer with overproduction of glycans as precursors of CA19-9. Although such cases are not frequent, we should be acquainted with the correlation between serum CA19-9 values and genotypes of Lewis and Secretor genes.

Reduced frequency of blood group Lewis a-b- in female Type 1 diabetes patients

AIMS

To examine a disputed association between the Lewis(a(-)b(-)) phenotype and Type 1 diabetes (T1D).

METHODS

Lewis red blood cell phenotyping was performed for 97 T1D White patients and 100 control subjects using monoclonal antibodies. Two historical cohorts were also included as a control population.

RESULTS

T1D patients had a lower frequency (4.1%) of Lewis(a(-)b(-)) blood group compared with simultaneously tested healthy control subjects (10.0%) and the historical control group (11.1%, P = 0.02). Male T1D patients showed a Lewis(a(-)b(-)) frequency of 8.0%, which was similar to both matched healthy male donors (9.8%) and historical (9.5%) male control subjects. Unexpectedly, none of the female T1D patients displayed Lewis(a(-)b(-)) phenotype, vs. 10.3% and 10.8% of female control subjects (P = 0.039 and 0.017).

CONCLUSIONS

The Lewis(a(-)b(-)) phenotype occurs less frequently in T1D compared with healthy control subjects with a strong female gender bias.

Involvement of intestinal alkaline phosphatase in serum apolipoprotein B-48 level and its association with ABO and secretor blood group types

Serum levels of intestinal alkaline phosphatase (IAP), a protein implicated in transcellular transport of chylomicrons, vary among ABO blood groups. In rat enterocytes, IAP is associated with chylomicron secretion, but the rat expresses only blood group A. It is not known whether chylomicron secretion may be affected in humans who express multiple blood group types. Serum samples from 40 healthy subjects were obtained after overnight fast and 3h after a high-fat meal, and assayed for IAP and apolipoprotein B-48 (apoB-48), both proteins exclusive to intestine, although only apoB-48 is found in chylomicrons. The two proteins were greater in subjects without blood antigen A (B and O) than in those with this antigen (A and AB); 2.4- and 4.7-fold for IAP and 1.5- and 2.0-fold for apoB-48 before and after the meal, respectively. Moreover, IAP and apoB-48 levels were strongly correlated in the subjects with the secretor phenotype (r > 0.81). These results indicate that IAP is strongly involved in chylomicron formation and fatty acid metabolism might change among ABO blood type. In addition, ABO blood type classification in apoB-48 measurement would improve the diagnostic value in the evaluation of metabolic syndrome.

ABH and Lewis antigen distributions in blood, saliva and gastric mucosa and H pylori infection in gastric ulcer patients

AIM: To investigate the ABH and Lewis antigen expression in erythrocytes, saliva and gastric epithelium, as well as the association between H pylori and the presence of gastric epithelial lesions.

METHODS: The distribution of ABH and Lewis blood group antigens in erythrocytes, saliva and gastric mucosa of H pylori-infected gastric ulcer patients was analyzed. Forty-two patients with gastric ulcer were studied, and fifty healthy individuals were used as control group. The blood group antigens were determined by direct hemagglutination, dot-ELISA and immunohistochemical methods in erythrocytes, saliva and gastric mucosa specimens, respectively. Diagnosis for H pylori infection was performed by conventional optical microscopy and ELISA.

RESULTS: A higher seroprevalence of IgG H pylori specific antibodies was observed in gastric ulcer patients (90%) compared to the control group (60%). We observed a significant increase of phenotypes O, A₂ and Lewis b in H pylori-infected patients. The expression of these antigens had progressive alterations in areas of ulcerous lesions and intestinal metaplasia.

CONCLUSION: ABH and Lewis blood group antigens are a good indicator for cellular alterations in the gastric epithelium.

Combined role of the Lewis antigenic system, Chlamydia pneumoniae, and C-reactive protein in unstable angina

OBJECTIVES

The goal of this study was to assess the prognostic role of the Lewis antigenic system, Chlamydia pneumoniae (CP) seropositivity (CP+), and C-reactive protein (CRP) levels in unstable angina (UA).

BACKGROUND

The role of CP infection in acute coronary syndromes is contradictory. The Lewis antigenic system, a genetically determined blood group system associated with infections and several disorders, including ischemic heart disease, might influence the susceptibility to CP infection, inflammatory response, and risk of cardiac ischemic events.

METHODS

The CRP levels, Lewis antigens, and CP+ were measured in 54 patients with Braunwald's class IIIB UA. All patients were followed up for one year, and the occurrence of new coronary events (coronary death, myocardial infarction, and recurrence of instability) were recorded.

RESULTS

Twenty-five coronary events occurred during follow-up. At univariate analysis CRP >3 mg/l (CRP+) (p = 0.0056), Lewis antigen b (Leb+) (p = 0.028), and the combination of Leb+ and CP+ (p = 0.006) and of CRP+ and Leb+ (p = 0.003) were associated with new coronary events, while CP+ alone was not. At multivariate analysis, CRP+ (p = 0.008) and combined Leb+CP+ (p = 0.03) were independent predictors of worse outcome. The event rate was 64% in CRP+ patients, 67% in Leb+CP+ patients, and 86% in CRP+Leb+CP+ patients. Combined Leb+CP+, but not Leb+ and CP+ alone, was related to CRP levels (p = 0.03). Among CP+ patients, CRP levels were higher in Leb+ than Leb- (p = 0.028).

CONCLUSIONS

Our data demonstrate that in UA the Lewis antigenic system plays an important role, probably determining individual susceptibility to the detrimental effects of CP infection and by determining an enhanced inflammatory response.

Elevated carbohydrate antigen 19-9 (CA 19-9) in patients with Echinococcus infection

The carbohydrate antigen 19-9 (CA 19-9), a determinant (sialylated lacto-N-fucopentaose 119) of a circulating oligosaccharide antigen, is a frequently used tumor marker. Echinococcus spp. infects humans throughout the world and may be able to synthesize closely related molecules which could interfere with the measurement and interpretation of CA 19-9 concentration. The main objective of the present study was to determine the range of CA 19-9 levels in the sera of patients infected by E. granulosus (cystic hydatide disease; CYSHD) or E. multilocularis (alveolar hydatide disease; ALVHD). Serum samples were collected from patients (aged 10-85 years) over a period of 5 years: from 19 patients with CYSHD and from 20 patients with ALVHD. Infection was confirmed by positive Echinococcus serology and clinical evidence provided by imaging and/or histopathological findings. CA 19-9 was detectable in 13 patients with CYSHD (13.5 +/- 8.5 kU/l) and 13 patients with ALVHD (30.0 +/- 21 kU/l; p < 0.05). Thus ALVHD patients exhibited a significantly higher plasma level of CA 19-9 than CYSHD patients. The serum level of CA 19-9 assessed with an increased cut-off value (> 22 kU/l) was elevated in nine (45%) of 20 ALVHD patients compared to two (11%) of 19 CYSHD patients (p < 0.05). Sera from patients with Echinococcus multilocularis infection contain substances which cross-react with CA 19-9. These substances originate either from the parasite or are synthesized by the host in response to the infection, and possibly bear the Lewis-a antigen or closely related structures which are recognized by anti-CA 19-9 antibodies. Our findings are relevant to the investigation of patients presenting with cystic lesions for which the differential diagnosis includes an infectious or neoplastic origin.

Molecular analysis of plasma alpha 1,3-fucosyltransferase deficiency and development of the methods for its genotyping

Four patients with mental illness were found to be deficient in plasma alpha1,3-fucosyltransferase for the first time in Japan [Exp Clin Immunogenet 1999;16:125-130]. Complete sequencing of FUT6 genes in these individuals revealed the presence of two point mutations, i.e., G739 to A (Glu-->247 to Lys) and C945 to A (Tyr-->315 to stop). In addition to two reported alleles having G739 to A (pf1) and G739 to A and C945 to A (pf3), a new mutated allele having C945 to A (pf2) was found to be present and all the individuals who lack alpha1,3-fucosyltransferase activity in plasma were found to possess pf genes homozygously (pf/pf). In order to detect such lethal mutations in FUT6 genes easily, PCR-RFLP methods have also been developed and applied for the screening of FUT6 deficiency in a large number of samples which resulted in the demonstration of three additional FUT6-deficient individuals. The absence of alpha1,3-fucosylated molecules on alpha(1)-acid glycoprotein in plasma from all the 7 individuals was confirmed to result from the plasma alpha1,3-fucosyltransferase deficiency.

Typing for the human lewis blood group system by quantitative fluorescence-activated flow cytometry: large differences in antigen presentation on erythrocytes between A(1), A(2), B, O phenotypes

BACKGROUND

Lewis phenotyping by hemagglutination is an unreliable routine method for Lewis antigen designation. Now genomic typing of the Lewis gene is available. Additionally, flow cytometry has been used for typing. We wanted to compare the results of Lewis typing in healthy individuals using the three methods.

MATERIALS AND METHODS

Ninety-three randomly selected plasma donors were genotyped for inactivating Secretor (FUT2) G428A and Lewis (FUT3) T59G, T202C, C314T, G508A and T1067A point mutations. All Le(a+b-) individuals (nonsecretors) were homozygous for the FUT2 G428A mutation and all Le(a-b-) individuals had inactivating mutations on both FUT3 alleles. Fixed erythrocytes were analyzed by fluorescence-activated flow cytometry and the results were compared with hem- agglutination and genotypic data. Antigen availability was expressed as median fluorescence intensity and as percentage positive cells with fluorescence intensities > or =10(2).

RESULTS

Using an anti-Le(a) reagent a mean of 99% of erythrocytes from Le(a+b-) individuals and 1% of erythrocytes from Le(a-b-) or Le(a-b+) individuals were stained positive. Using an anti-Le(b) reagent, a mean of 71% of erythrocytes from A(1), 95% from B and 99% from O and A(2) Le(a-b+) individuals and less than 10% of erythrocytes from Le(a-b-) or Le(a+b-) individuals were stained positive. After papain treatment 100% of the erythrocytes from A(1) and A(1)B Le(a-b+) individuals stained positive without increase in background staining. The flow cytometric technique revealed large differences in staining intensities, within each ABO Le(a-b+) subgroup which was not directly correlated to plasma donation frequencies nor to Secretor or Lewis genotypes.

CONCLUSION

Flow cytometry may prove valuable as a Lewis blood group typing technique but also as a research tool when investigating Lewis phenotypes of human erythrocytes.

Critical increment of Lewis blood group antigen in serum by cancer found by photothermal immunoassay

Lewis blood group antigen levels in human sera were assayed with a highly sensitive photothermal immunoassay which is based on laser-induced photothermal detection. Comparison of 32 colon cancer patients' sera and 34 healthy persons' sera showed that cancer patients' sera contained more Lewis antigens than healthy persons' sera. Le(a) antigen level in Le(a-) type persons and Le(b) level in Le(b-) type persons differentiated healthy persons and colon cancer patients. Furthermore, it was found that in Lewis blood phenotype (a-) several cancer patients' sera specimens changed to (a+). Many reports demonstrated that Lewis phenotype of erythrocytes changed with various conditions, including carcinomas, but they dealt mostly with erythrocytes and salivas or showed immunohistochemical evidence, and there are no reports on the quantitative analysis of ordinary (noncancerated) Lewis antigen levels in human sera. This is because Lewis antigens in sera, unlike those found in saliva, are too small to quantify with conventional immunoassay and there has been no highly selective method to measure Lewis antigens in sera. The increase of Lewis antigen in cancer patients' sera is presumed to antecede the blood type change. Our assay presented here, a highly sensitive assay of Lewis antigens, will greatly contribute to an early detection or diagnosis of cancers.

Predictive value of preoperative serum sialyl Lewis X-i antigen levels in non-small cell lung cancer

BACKGROUND AND STUDY OBJECTIVE

Serum sialyl Lewis X-i antigen has been shown to be one of the most specific markers for monitoring non-small cell lung cancer(NSCLC). We evaluated the correlation between serum levels of sialyl Lewis X-i antigen and mediastinal lymph node involvement in NSCLC patients.

METHODS

One hundred patients with untreated NSCLC were included in this study. All the patients received surgical treatment and the lymph nodes were evaluated pathologically. Pretreatment sera from these patients examined for the levels of sialyl Lewis X-i antigen, using a radioimmunoassay.

RESULTS

Serum levels of sialyl Lewis X-i antigen levels in patients with N2 were higher than those of N0-1(p = 0.0049). According to the receiver characteristic curve analysis, the recommended cut-off level of the antigen between N0-1 patients and N2 patients was 38.0 U/ml.

CONCLUSION

Measurement of serum levels of sialyl Lewis X-i antigen might provide preoperative information in patients with NSCLC.

Serum sialyl lewis X-i antigen levels in non-small cell lung cancer: correlation with distant metastasis and survival

To evaluate the correlation between serum levels of sialyl Lewis X-i antigen and distant metastasis and survival in patients with non-small cell lung cancer (NSCLC), we measured the serum levels of the tumor marker in 371 patients with untreated NSCLC. The sialyl Lewis X-i antigen level was measured using a RIA kit. In patients with adenocarcinoma or other NSCLC subtypes, there was a correlation between serum sialyl Lewis X-i antigen and stage of the disease (P = 0.0001 and P = 0.0015, respectively). Levels of the marker varied significantly depending on the number of metastatic organs in adenocarcinoma (P = 0.0089) and in other NSCLC subtypes (P = 0.002). Univariate analysis showed that survival of NSCLC patients with high (more than 100 units/ml) sialyl Lewis X-i antigen levels was significantly poorer than that of patients with low antigen levels (P = 0.0001). Multivariate analysis using Cox's proportional hazard model showed that high sialyl Lewis X-i antigen levels correlated significantly with poor survival (P = 0.004). Our data suggest that a high serum level of sialyl Lewis X-i antigen seems to be an indicator of the presence of metastasis and might indicate the need for a careful investigation of all putative metastatic sites. The serum levels of sialyl Lewis X-i antigen may reflect the extension of metastasis and would be helpful in considering treatment options.

[The Lewis antigen system as a marker of IHD risk]

Prevalence of Lewis antigens and phenotypes has been investigated in 60 healthy donors and 74 patients with ischemic heart disease. It was found that phenotype Le(a-b-) marks high risk and phenotype Le(a-b+) is a resistance marker for development of ischemic heart disease. It would be valid to elucidate relationships between Lewis system and pathogenesis of internal diseases.

A missense mutation in the FUT6 gene results in total absence of alpha3-fucosylation of human alpha1-acid glycoprotein

The major alpha3-fucosyltransferase activity in human plasma is encoded by the gene for fucosyltransferase VI (FUT6). A missense mutation (Gly-739 --> Ala) in this gene is responsible for deficiency of enzyme activity in plasma. To examine whether this fucosyltransferase is the sole enzyme responsible for the alpha3-fucosylation of serum glycoproteins in the liver, we studied the fucosylation of three glycoproteins in sera of individuals with or without inactivated FUT3 and/or FUT6 gene(s) but with a functional FUT5 gene. alpha1-Acid glycoprotein was used as the principal reporter protein for liver alpha3-fucosyltransferase activity, because of its high fucose content. In all individuals with the FUT6 missense mutation Gly-739 --> Ala in double dose, no fucosylation of alpha1-acid glycoprotein was found. This alpha1-acid glycoprotein was not intrinsically resistant to fucosylation, since it was susceptible to in vitro fucosylation using an alpha3/4-fucosyltransferase isolated from human milk. The same result was found for alpha1-antichymotrypsin and alpha1-protease inhibitor. On the other hand in all individuals with alpha3-fucosyltransferase activity in plasma, alpha3-fucosylated glycoforms of the glycoproteins studied were found. The degree of fucosylation of alpha1-acid glycoprotein was correlated with alpha3-fucosyltransferase activity (Rs = 0.82). These data indicate that the product of FUT6, but not of FUT3 or of FUT5, is responsible for the alpha3-fucosylation of glycoproteins produced in liver and suggest that this organ is a major source of alpha3-fucosyltransferase activity in plasma.

Hepatic neoexpression and increased plasma levels of Lewis Y, a carbohydrate antigen, in chronic inflammatory liver diseases

The clinicopathologic relevance of the hepatic expression of Lewis Y (Le(y)), a carbohydrate antigen, and its plasma levels was studied in benign and malignant liver diseases. Tissue and plasma antigens, respectively, were determined with an avidin-biotin-peroxidase complex method and a radioimmunoassay using monoclonal antibody AH6. Normal liver cells and bile ductules did not express Le(y). In the inflammatory tissues, the liver cells and proliferated bile ductules expressed Le(y). The strongest expression by the liver cells was observed in chronic active hepatitis with severe activity and that by the ductules in liver cirrhosis. Only 1 of 16 hepatocellular carcinomas expressed Le(y). The plasma levels of Le(y) increased significantly but nonspecifically in chronic persistent hepatitis, chronic active hepatitis, liver cirrhosis, and hepatocellular carcinomas. It was concluded that (1) Le(y) is an inflammation-associated but not a cancer-associated antigen; (2) the more the tissue damage advances, the more the antigen is expressed; and (3) hepatic and plasma Le(y) are, however, nonspecific markers of necroinflammatory liver diseases.

ABH secretor status, as judged by the Lewis phenotypes, in Norwegian survivors from meningococcal disease

Survivors from meningococcal disease (serogroups B and C) and a control series (blood donors) were examined for their ability to secrete ABH blood group substance. The examination was done indirectly by determining their Lewis phenotypes. There was no significant difference in the secretor status between the two groups.

Clinical relevance of tumour markers CA 19-9 and CA-50 in sera from patients with pancreatic duct carcinoma

Serum levels of the tumour antigens CA 19-9 and CA-50 and their relation to the extent of the disease were studied in 97 patients with carcinoma of the pancreas. Of 13 patients with normal serum concentrations of CA 19-9, 11 (84.6%) had irresectable disease, whereas 87.5% of the patients with resectable disease expressed antigen levels above cut-off. Following attempted radical surgery, preoperatively elevated serum levels decreased in eight patients (50%). Unchanged and high levels were associated with residual disease and early death. Clinical signs of recurrence were preceded by elevated serum levels of both tumour antigens.