Serological Diagnosis And DNA Repair Gene XRCC1 Polymorphism In Patients With Primary Liver Cancer

[Background and aims]Most patients with primary liver cancer(PLC) have well defined risk factors. Screening and surveillance in individuals at high risk intermittently remains the key way to detect this malignant disease in an early stage. Radiology plays an important role in diagnosis for PLC, but it is still difficult to discriminate benign and malignant lesions smaller than two centimeters in diameter. Alpha-fetoprotein is widely used as a accessorial diagnostic test for PLC, but with insufficient sensitiveity and specificity. Ideal serological tumor markers specifically for PLC is urgently needed to detect this deadly disease in clinic. Glypican-3 was reported to be over-expressed in tissues of 80%-90% PLCs, and was detectable in the serum of patients with PLC. It was considered as a promising novel maker for PLC diagnosing. GGT(Gamma-glutamyl tranferase) and ALP(alkaline phosphatase) as tumor markers for PLC were controversial. CA19-9 (carbohydrate antigen19-9), CA242(carbohydrate antigen 242) and CEA (carcino-embryogenic antigen) were used as makers for tumors in gastroenterology, each of them had limited value for PLC diagnosing. The heterogeneity of PLC made it impossible for a single maker to detect the majority of patients with PLC, but the parallel combination of two or more markers had shown some promise in inceasing sensitivity. Although most patients with PLC accompanied with definite pathogenies, but not all the exposer will develop PLC in their lifetime. This suggested that genic factors and the interaction with environment may participate in the pathogenesis of PLC. XRCC1(X-ray cross complementing group1) encodes an DNA repair protein mainly involved in BER (base excition repair) process. Researches had shown the association between XRCC1 exon 10 Arg399Gln polymorphism and susceptibility of chronic HBV infection and aflatoxinB1contamination related PLC. The current research was designed to discuss the value of serological tumor markers GPC3, AFP, GGT, ALP, CA19-9, CA242 and CEA each alone and the combination with each other in the diagnosis for PLC. And to investigate the potential relationship between XRCC1 Arg399Gln polymorphism and PLC risk.[Patients]1. Patients for serological tumor markers tests:Fifty-five patients with PLC were studied, the ratio for male to female was 5:1, the average age was 57.5±12.82 years. All were inpatients of Beijing you'an hospital in 2006, and were clinically diagnosed with PLC according to the guidelines for PLC diagnosing established by the Chinese anti-cancer association in 2001. Control was fourteen patients diagnosed with liver cirrhosis by medical history, biochemical tests and radiology. The ratio for male to female was 4:1, the average age was 52.14±12.54 years.2. Patients for XRCC1Arg399Gln polymorphism analysis:Another fifty patients diagnosed with PLC in the same way described above were accepted as the cases.The ratio for male to female was 4:1, the average age was 55.5±11.88 years. The control group was consisted of sixty-one patient with cirrhosis, The ratio for male to female was 2.5:1, the average age was 49.33±10.83 years. All of the cirrhosis were followed up for six months without any evidence suggested for PLC and other tumors. ninety- two blood donors were regarded as the healthy control group. Aged from twenty to forty, without any severe chronic diseases.[Methods]1. Serum GPC3 was quantitated by a commercial human GPC3-ELISA kit, AFP was detected using automated electrochemiluminescene immunoassay, GGT and ALP was determined by automated biochemical analysis system OLYMPUS-AU640, andCA19-9, CA24-2 , CEA were measured by ELISA performed in our hospital routinely. Clinical data for matching were collected. The comprehensive diagnosis value of each maker alone and the combination of two or more of them were analysed using ROC(receiver operator characteristic) curve.2. Genomic DNA was extracted from peripheral blood leukocytes of each sample using phenolic and chloroform traditionally. All DNA samples were genotyped for XRCC1 Arg399Gln polymorphism by means of polymerase chain reaction-restriction fragment length polymorphism(PCR-RFLP) analysis. Clinical data may have modulation on PLC risk were collected, all information were restricted from no more than a week near the time each sample gained.[Results]1. Serum GPC3 in patients with PLC(317.89±206.02pg/ml) was higher than cirrhotic patients(240.8±110.140 pg/ml). No statistical significance was observed. The level of serum GGT was significantly higher in patients with PLC than in cirrhosis. There were no remarkable difference of serum ALP, CA19-9, CA24-2 and CEA levels between the two groups separately. The area under the ROC curve of each marker were:AFP (0.794,95%CI:0.676-0.911) >GGT (0.685,95%CI:0.554-0.815) >ALP (0.603,95%CI:0.454-0.753) >GPC3 (0.597,95%CI:0.453-0.741) >CA19-9 (0.502,95%CI:0.335-0.669) >CA242 (0.474,95%CI:0.314-0.633) >CEA (0.473,95%CI:0.281-0.594). The sensitivity and specificity of GPC3 were 31.5%(17/54) and 93.3% respectively, when the cutoff point was at 400pg/ml. GPC3 was more than 400pg/ml in 22.9% Patients with AFP-negative(<200ng/ml) PLC. The combination of GPC3 and AFP had a significant higher sensitivity(50%) than AFP alone for diagnosing PLC, P<0.001. The level of serum GPC3 exhibited a plus-correlation with that of AFP in some degree, but had no relationship with tumor size, macroscopic portal veinous invasion and metastasis. The sensitivity and specitivity of AFP were35.2% and 100% separately,when the cutoff point was at 200ng/ml.GGT had a sensitivity of 44.4%(24/54) and a specificity of 92.9% at the cutoff point of 100U/L. ALP had a sensitivity of 18.5%(10/54) and a specificity of 100% at the cutoff point of 200U/L. The level of serum AFP, GGT and ALP were all showed some plus- correlation with tumor size. The level of ALP was significantly higher in patients with portal veinous embolus than those without. CA19-9 had a sensitivity of 64.8%(35/54) and a specificity of 35.7% at the cutoff point of 37U/ml. CA24-2 and CEA shared in the same sensitivity of 11.1% (6/54) and specificity of 92.9%-100% at the cutoff point of 20U/ml and 5ng/ml separately. Serum CA19-9 level had a strong plus-coralation with ALP and TBIL levels. CA19-9 level in child-pughB or C patients was significantly higher than in child-pughA patients, the P value were 0.046 and 0.028 separately. The sensitivity of CA19-9 in child-pughB(73.1%) or C(88.9%) patients with PLC were significantly higher than in child-pughA (42.1%), the P value were 0.036 and 0.039 separately.2. AFP+GGT had the biggest area under the ROC(0.827,95%CI:0.725-0.928) when two markers were combinated for PLC diagnosing. It was bigger than AFP+GPC3, and was bigger than AFP in turn, but no significance were observed (P=0.543 and 0.473,separately). According to the cutoff points described above, AFP+GGT had a sensitivity of 59.3% (32/54) and a specificity of 92.9%. This was significantly higher than AFP alone, P<0.001. GGT+CA242(>20U/ml) showed a much more higher sensitivity of 64.8%(35/54) with the same specificity. When three makers were combinated, AFP+GGT+CA242 held the biggest area under the ROC curve. It was a little bigger than AFP+GGT, but with no significance(P=0.67). The sensitivity and specificity of AFP+GGT+CA242 were 61.1% and 92.9% separately. When four or more makers were combinated, the area under the ROC curve increased slightly, never with significance. The sensitivity gently improved concomitantly, butwith sacrifice of much specificity.3. The frequency of 399Gln allele genotype was 22% in the PLC group, 32% in the liver cirrhosis group and 27.7% in the healthy blood donors. The average age of patients with PLC(55.5±11.88) was significantly older than that of the cirrhosis (49.33±10.83), P=0.005. When taking the heathy blood donors as controls, 399Gln allele didn't show any influence on PLC susceptibility compared to the Arg/Arg wild homozygosis(OR 0.563,95%CI:0.277-1.141), and the same for taking liver cirrhosis as controls(OR 0.544,95%CI:0.253-1.141). In total patients with HBV infection, 399Gln allele genotype also exhibited no relation with PLC risk(OR 0.693,95%CI:0.300-1.141). While we observed a significant protective effect for PLC of the 399Gln allele genotype in patients older than 50 years(OR 0.235,95%CI:0.081- 0.687), P = 0.007. All of the gender, diabetes, heavy smoking and alcohol consumption showed no relationship with the Arg399Gln polymorphism. Patients with serum AFP higher than 20ng/ml accompanied with about eight times higher risk for developing PLC(OR 8.127,95%CI:3.065-21.546,P<0.001). There was no significant difference of the Arg399Gln polymorphism genotype frequency between AFP-positive and AFP-negative patients with PLC.[Conclutions]1. Serum GPC3 had limited value for diagnosing PLC alone. GPC3 was higher than 400pg/ml in the serum of about 22.9% patients with AFP-negtive PLC. The combination of GPC3 and AFP significantly improved sensitivity(50%) for diagnosing PLC than AFP alone. GPC3 may contribute to detect PLCs in an early stage. GGT also showed some promise for PLC diagnosing, The combination of GGT(≥100U/L) and AFP also significantly improved sensitivity without any decreasing of specificity. All of CA19-9, CA24-2 and CEA had dismal value for diagnosing PLC. Serum CA19-9 strongly associated with basic liver function.AFP+GGT have the best efficiency for PLC diagnosing when two markers were combinated. The combination of Three or more markers could improve the diagnostic sensitivity slightly but with sacrifice of much specificity.2. Patients with chronic liver disease in Beijing and circumjacent area holding the XRCC1-399Gln allele showed decreased risk for developing PLC. Patients with abnormal serum AFP levels(≥20ng/ml) were at higher risk for developing PLC. XRCC1 Arg399Gln polymorphism had no correlation with the level of serum AFP in patients with PLC.