Serological markers for detecting colorectal cancer and their application for inhibiting colorectal cancer cells

Abstract

Embodiments relate to serological markers for detecting the colorectal cancer and applications of the serological markers. A phospholipid scramblase1 (PLSCR1), a stomatin-like protein 2 (STOML2) or a transport protein Sec61β (SEC61β) increases expression in the blood at the earlier stage of the colorectal cancer. Detecting the expression of the PLSCR1, STOML2 or SEC61β protein or an induced autoantibody of each protein in a blood sample is used to diagnose the colorectal cancer. Moreover, the serological marker improves the detection efficiency and the sensitivity in detecting the colorectal cancer and is used to predict the prognosis. The serological markers are applied in preparing a detection device or inhibiting the growth of the colorectal cancer cells.

Claims

1. A method for detecting and treating colorectal cancer, comprising: obtaining a blood sample from a human; detecting PLSCR1 and STOML2 proteins in said collected blood via an immunoassay selected from the group consisting of Western blot assay and enzyme linked immunosorbent assay (ELISA); quantifying protein expression levels of PLSCR1 and STOML2 in said collected blood via liquid chromatography-tandem mass spectrometry (LC-MS/MS); comparing the protein expression levels of PLSCR1 and STOML2 with a cutoff value to determine whether said human is at increased risk of having colorectal cancer or not; wherein if the levels are above the cutoff value said human has colorectal cancer, and administering anti-PLSCR1 antibody to said human having colorectal cancer.

2. The method as claimed in claim 1, wherein said detecting further includes detecting SEC61β in said collected via the immunoassay.

3. The method as claimed in claim 2, wherein protein expression level of SEC61β is quantified by the immunoassay on SEC61β or an induced autoantibody of SEC61β.

4. The method as claimed in claim 3, wherein said comparing further includes comparing the protein express level of SEC61β with the cutoff value.

Description

BRIEF DESCRIPTIONS OF THE DRAWINGS

(1) The patent or application file contains at least one drawing executed in color. Copies of this patent or patent application publication with color drawing(s) will be provided by the Office upon request and payment of the necessary fee.

(2) FIG. 1 shows the expression of PLSCR1 protein in colorectal cancer tissues and normal tissues;

(3) FIG. 2 shows the expression of PLSCR1 protein in blood samples of colorectal cancer patients and healthy controls;

(4) FIG. 3A is a photograph showing the expression of PLSCR1 protein in a normal tissue (normal colorectal mucosa);

(5) FIG. 3B is a photograph showing the expression of PLSCR1 protein in a benign colorectal polyp tissue;

(6) FIG. 3C is a photograph showing the expression of PLSCR1 protein in a colorectal adenocarcinoma tissue;

(7) FIG. 4 shows the detection efficiency of PLSCR 1 protein, the current used CEA marker and a combination of both markers;

(8) FIG. 5. shows a relationship between the PLSCR1 expression and the overall survival rate of the colorectal cancer patients;

(9) FIG. 6. shows an in vitro inhibition of the colorectal cancer cell growth by treating with a PLSCR1 antibody;

(10) FIG. 7. shows an in vivo inhibition of the colorectal cancer by treating with a PLSCR1 antibody;

(11) FIG. 8. shows the expression of SEC61β protein in the colorectal cancer tissues and the matching normal tissues;

(12) FIG. 9. shows the SEC61β autoantibody response in blood samples of colorectal cancer patients and healthy controls;

(13) FIG. 10. shows the SEC61β autoantibody response at different stages of the colorectal cancer;

(14) FIG. 11. shows the detection efficiency of the SEC61β autoantibody, the current used CEA marker and a combination of both markers;

(15) FIG. 12. shows the expression of STOML2 protein in blood samples of colorectal cancer patients and healthy controls;

(16) FIG. 13. shows the detection efficiency of STOML2 protein, the current used CEA marker and a combination of both markers;

(17) FIG. 14. shows the STOML2 autoantibody response in blood samples of colorectal cancer patients and healthy controls; and

(18) FIG. 15. shows the STOML2 autoantibody response in different cancers.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT

Embodiment 1: Screening and Selecting a Serological Marker for Detecting the Colorectal Cancer

(19) The embodiment selects a phospholipid scramblase 1 (PLSCR1) as a serological marker for detecting the colorectal cancer. The PLSCR1 is over-expressed in the colorectal cancer tissue and is tested the efficiency as a serological marker to detect the colorectal cancer.

(20) 28 colorectal cancer patients are enrolled. A colorectal cancer tissue (tumor tissue) and a normal tissue away from the colorectal cancer tissue for at least 10 centimeters collected from each patient is respectively analyzed the proteomic profile. All tissues are cut into small pieces, washed by using 0.9% sodium chloride to remove blood, homogenized in STM solution (5 ml STM solution/g tissue, 0.25 M sucrose, 10 mM Tris(hydroxymethyl)aminomethane hydrochloride (Tris-HCl) and 1 mM magnesium chloride) to form a homogenous solution.

(21) The homogenous solution is centrifuged at 260×g for 5 minutes to remove the precipitated nuclei and tissue debris. A separated supernatant solution is further centrifuged at 1500×g for 10 minutes to precipitate the crude membrane proteins. The crude membrane proteins are homogenized in two-thirds volume of the original homogenate volume (contains 0.25M STM solution and protease inhibitor) and centrifuged at 16000×g for an hour to purify the membrane proteins. The membrane pellet is washed by 0.1M sodium carbonate for overnight and is centrifuged at 16000×g for an hour to re-collect the membrane proteins. The purified membrane proteins are homogenized in 90% formic acid and stored in −80° C.

(22) The purified membrane proteins are resuspended in 6M urea, 5 mM ethylene diamine tetraacetic acid (EDTA), and 2% sodium dodecyl sulfate (SDS) in 0.1M triethylammonium bicarbonate (TEABC) and sonicated at 4° C. for 10 minutes. A bovine serum albumin (BSA) is used as an internal standard (0.1% BSA in the purified membrane proteins). The purified membrane proteins were reduced by 5 mM Tris(2-carboxyethyl)-phosphine (TCEP) and alkylated by 2 mM methyl methanethiosulfonate (MMTS) at room temperature for 30 mM. A 40% acrylamide solution (acrylamide:bisacrylamide is 29:1, v/v), 10% (w/v) ammonium persulfate (APS) and 100% N,N,N′,N′-Tetramethylethylenediamine (TEMED) are added to the purified membrane protein mixture to polymerize as a gel.

(23) The gel is cut into small pieces and subjected to tryptic digestion in 25 mM TEABC solution at 37° C. to form peptides. The peptides are extracted from the gel using sequential extraction with 25 mM TEABC, 0.1% (v/v) trifluoroacetic acid (TFA) in water, 0.1% TFA in acetonitrile (ACN), and 100% ACN. The extracted peptides are concentrated, desalted and analyzed by a liquid chromatography-tandem mass spectrometry (LC-MS/MS).

(24) Information collected from LC-MS/MS is searched by Mascot v2.2 against International Protein Index (IPI) human database from the European Bioinformatics Institute. The selection constraints: (a) only tryptic peptides with up to two missed cleavage sites were allowed; and (b) 0.3-Da mass tolerances for MS and 0.1-Da mass tolerances for MS/MS fragment ions. Only unique peptides with scores ≧35 (p<0.05) were confidently assigned. The amino acid sequence of enrolled peptide is further compared to identify the corresponding protein.

(25) Results from the LC-MS/MS analysis are used to compare the expression of a specific protein in the colorectal cancer tissue (tumor tissue) and the normal tissue. For example, peak area of protein A in the colorectal cancer tissue and the normal tissue is R1 and R2 respectively. The peak area of the internal standard in the colorectal cancer tissue and the normal tissue is S1 and S2 respectively. Comparing the ratio of the protein A and the internal standard in the colorectal cancer tissue and the normal tissue, a difference in abundance of 1.5-fold to indicate a statistically significance of higher or lower expression, respectively.

(26) PLSCR 1 is selected from the LC-MS/MS analysis due to the significant overexpression in the colorectal cancer tissue than the matching normal tissue, and is further tested the expression in the western blotting assay and the immunohistochemistry (IHC) assay.

(27) Western Blotting Analysis

(28) The colorectal cancer tissue and the matching normal tissue are suspended in a lysis buffer (0.25 M sucrose, 10 mM Tris-HCl, pH 7.6, 1 mM MgCl.sub.2, 1% SDS) with protease inhibitor mixture (20 μg/μL aprotinin, 20 μg/μL leupeptin, and 1 mmol/L phenylmethanesulfonyl fluoride (PMSF); protein:protein inhibitor was 100:1, v/v) and homogenized on ice to form a protein suspension. 40 μg of protein suspension was subjected to the western blot analysis.

(29) With reference to Table 1, Table 1 illustrates the expression of the PLSCR1 in the colorectal cancer tissue compared with their normal counterpart. 18 of 28 paired tissues show at least 1.5 folds overexpression in the colorectal cancer tissue compared with their normal counterpart.

(30) TABLE-US-00001 TABLE 1 The expression of the PLSCR1 in the colorectal cancer tissue compared with their normal counterpart. Fold (colorectal cancer No. tissue/normal tissue) 1 1.92 2 3.91 3 4.02 4 0.90 5 1.38 6 0.98 7 4.03 8 2.88 9 2.13 10 3.04 11 0.85 12 2.16 13 1.09 14 3.09 15 12.59 16 11.85 17 10.87 18 2.86 19 6.33 20 1.37 21 2.04 22 2.17 23 1.40 24 0.98 25 8.85 26 0.72 27 6.62 28 5.19

(31) The PLSCR1 of the embodiment has an UniProt accession number 015162 with sequence shows as following (SEQ ID NO: 1):

(32) TABLE-US-00002         10         20         30         40         50         60 MDKQNSQMNA SHPETNLPVG YPPQYPPTAF QGPPGYSGYP GPQVSYPPPP AGHSGPGPAG          70         80         90        100        110        120 PPVPNQPVYN QPVYNQPVGA AGVPWMPAPQ PPLNCPPGLE YLSQIDQILI HQQIELLEVL        130        140        150        160        170        180 TGPETNNKYE TKNSPGQRVY PAAEDTDCCT RNCCGPSRPF TLRIIDNMGQ EVITLERPLR        190        200        210        220        230        240 CSSCCCPCCL QEIETQAPPG VPIGYVIQTW HPCLPKFTIQ NEKREDVLKI SGPCVVCSCC        250        260        270        280        290        300 GDVDFEIKSL DEQCVVGKIS KHWTGILREA FTDADNFGIQ FPLDLDVKMK AVMIGACFLI        310 DEMFFESTGS QEQKSGVW

(33) However, one of ordinary skill in the art will realize that any sequence has more than 90% similarity with above-mentioned sequence is capable to apply in the present invention.

(34) In order to understand whether the PLSCR 1 is significant presented in a blood specimen or has capability to be a serological marker, blood samples respectively from colorectal patients and healthy controls are collected and electrophoresised to separate proteins in blood samples by SDS-PAGE. Proteins in the SDS-PAGE are transferred to a polyvinylidene fluoride (PVDF) membrane. The PVDF membrane is blocked with 5% skim milk in Tris-buffered saline-Tween buffer (25 mmol/L Tris, 190 mmol/L NaCl, and 0.5% [v/v] Tween 20, pH 7.5) and then incubated with primary rabbit antihuman PLSCR1 polyclonal antibody (1:1000 dilution) at 4° C., overnight. After being washed, the membranes were incubated at 25° C. for 1 h with peroxidase-conjugated mouse antirabbit IgG antibody (1:5000 dilution), and then were developed with a chemiluminescence reagent kit and photographed. Immunoblot images were analyzed by an Imagemaster analyser, and the band intensities are presented in arbitrary units (AU).

(35) With reference to FIGS. 1 to 3C, respectively show the PLSCR1 expression in the colorectal tissues and blood samples from CRC patients or healthy volunteers. For the tissue samples, the western blot assay shows the PLSCR1 expression in the colorectal cancer tissue and in the normal tissue is 63.2±41 and 29±28.1 AU (p<0.001), respectively. The PLSCR1 expression is at least two folds in the colorectal cancer tissue than in the normal tissue. The IHC assay shows the PLSCR 1 expression increases with the malignant grade of the colorectal cancer in the tissue biopsy. For the blood samples, the western blot assay shows the PLSCR1 is significant expression in the blood samples of the colorectal cancer patients (44 AU) than in the healthy controls (18.5 AU). Therefore, the PLSCR1 has potential to be a serological marker to detect the colorectal cancer.

(36) One of ordinary skill in the art realizes the western blot assay for detecting the PLSCR1 expression might be replaced by an immunoassay such as an enzyme linked immunosorbent assay (ELISA).

(37) One of ordinary skill in the art also obviously understands when the PLSCR1 is highly expressed in the colorectal cancer, a corresponding antibody (autoantibody) induced by the PLSCR1 will be increased. Therefore, the autoantibody of PLSCR1 is able to be a serological marker for detecting the colorectal cancer.

(38) Specificity and Sensitivity of the PLSCR1 for Detecting the Colorectal Cancer

(39) With reference to FIG. 4 and Table 2, show the comparison between the PLSCR1 and a current used CEA marker.

(40) TABLE-US-00003 TABLE 2 Sensitivity of CEA and PLSCR1 in CRC patients with different stage. Stage of the Numbers Sensitivity colorectal of Combined cancer sample CEA.sup.a (A) PLSCR1.sup.b (B) (A) + (B) Early stage 25  4 (16%) 20 (80%) 20 (80%) Advanced stage 34 18 (53%) 27 (79%) 30 (88%) All CRC patients 59 22 (37%) 47 (80%) 50 (85%) .sup.aCutoff value ≧ 5 ng/mL. .sup.bCutoff value ≧ 13.9 AU.

(41) The PLSCR1 alone has better detection efficiency while applying in detecting the early stage and advanced stage colorectal of cancer than the CEA does. The overall detection rate of the PLSCR1 is at least 2 folds than the CEA so that the PLSCR1 has great capability for detecting the colorectal cancer. Moreover, combination of the PLSCR1 and the CEA will increase the efficiency in CRC detection.

(42) Relationship Between the PLSCR1 Expression and the Overall Survival Rate

(43) With reference to FIG. 5, a Cox proportional hazards model is applied to evaluate the relationship between the PLSCR1 expression and the overall survival rate. The CRC patients were divided into two groups (high PLSCR1 and low PLSCR1) according to a cutoff point of the IHC SI value (SI) by 1. Patients in low PLSCR1 group have longer survival period of 49.8 months than in high PLSCR1 group of 39.2 months. The PLSCR1 expression and the overall survival rate obviously exists a significant relationship.

(44) Further, an anti-PLSCR1 antibody that is produced from the PLSCR1 of the embodiment is used to be identified the capability of CRC inhibition in a mouse model. The anti-PLSCR1 antibody is directly contacted with a colorectal cancer HT29 cells. With reference to FIG. 6, the MTT (3-(4,5-cimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide) assay shows the anti-PLSCR1 antibody with concentration more than 0.35 μg/well (μg/mL) is able to inhibit the colorectal cancer cells after two days.

(45) With reference to FIG. 7, after transplanting the HT29 cells (1×10.sup.6 cells) to the subcutaneous region of female nude mice (7 weeks old) for 29 days, respectively injecting an affinity purified anti-PLSCR1 monoclonal antibody (+PLSCR1 Ab) or an IgG (−PLSCR1 Ab) into the mice every two days for 6 days. The anti-PLSCR1 monoclonal antibody significantly inhibits the growth of HT29 cells, but same dose of IgG does not effect the growth of HT29 cells. While observing an H&E staining of the colorectal cancer tissue, the amount of the cancer cells of the colorectal cancer tissue decreases after the mice treated with the anti-PLSCR1 monoclonal antibody. Therefore, the anti-PLSCR1 monoclonal antibody affects and inhibits the cell growth both in vitro and in vivo.

(46) Table 3 shows selected proteins by using the above mentioned methods which might has potential to be colorectal cancer detection markers. The markers are overexpressed in at least 20 tissue pairs from a total of 28 colorectal tissue pairs.

(47) TABLE-US-00004 TABLE 3 Proteins have potential to be colorectal cancer detection markers. No. of CRC cell line.sup.c Patients with Expressed in Large COLO Protein Upregulation Location.sup.a CRC Tissue.sup.b Intestine.sup.c 205 HCT-116 HCT-15 HT29 KM-12 SW-620 CEACAM5 25 PM x x x STOML2 24 PM x x x x x x x CEACAM6 24 PM x DEFA1 24 Extracellular x region RPL36 23 Cytoplasm x x x x x x TSPO 23 OM x x x x x x x x OCIAD2 22 OM x x x x x x IKIP 22 OM CYBA 22 PM x x x x x x x GOLT1B 22 OM x x x x x x x ELA2 22 Extracellular region CYB5B 21 OM x x x x x x x SEC61B 21 OM x x x x x x TOMM40 21 OM x x x x x x TMEM63A 21 PM x x x x x TSPAN8 21 PM x x x x x x x SLC16A3 21 PM x x x x x x x x LAMP1 21 OM x x x x x x x ANXA4 21 PM x x x x x x x x TOMM5 21 Unknown x x x x x x x x SLC2A1 20 PM x x x x x x TMCO1 20 OM x x x x x x OPRS1 20 PM x x x x x x x .sup.athe location is annontated by Gene Onotology and Ingenuity Pathway Analysis Knowledge Base. PM indicates plasma membrane and OM indicates organelle membrane .sup.bthe expression of proteins in colorectal cancer tissues are obtained in Human Protein Reference Database .sup.cthe data are obtained from Ingenuity Pathway Analysis

(48) Accordingly, the PLSCR1 is able to be applied in manufacturing a detection device for detecting the colorectal cancer. The detection device comprises a sample container, at least an anti-PLSCR1 antibody and a substrate. Blood samples are collected and applied in the sample container, and reacts with the anti-PLSCR1 antibody. The substrate might a chromophoric reagent or a fluorescence reagent that is combined with another anti-PLSCR1 antibody and develop a signal related to the binding between the PLSCR1 and the anti-PLSCR1 antibody, which is compared with a database or a cutoff value. The database collects the values of PLSCR1 expression from healthy controls. The cutoff value is calculated by estimating the values of PLSCR1 expression from healthy controls.

Embodiment 2: The Capability of a SEC6β Protein as a Serological Marker for Detecting Colorectal Cancer

(49) This embodiment uses a SEC61β protein listed in the Table 3 or an autoantibody induced from the SEC61β protein as a marker to detect the colorectal cancer.

(50) In this embodiment, a comparison of SEC61β expression between the colorectal cancer tissue and the matching normal tissue is established by using western blot assay. Differences between the embodiment 1 and 2 shows as following:

(51) The colorectal cancer tissue and the matching normal tissue is respectively mixed with an electrophoresis buffer contains 2% SDS and 5% 2-mercaptoethanol and heated at 100° C. for 5 minutes. Proteins in the tissues are fractioned by a 12% denatured polyacrylamide gel and then are transferred to the PVDF membrane. The PVDF membrane is blocked with 5% skim milk. Proteins in the PVDF membrane react with a rabbit anti-human SEC61β polyclonal antibody at room temperature for 2 hours and then react with a peroxide-conjugated second antibody at room temperature for 1 hour.

(52) The SEC61β of the embodiment has an UniProt accession number P60468 with sequence shows as following (SEQ ID NO: 2):

(53) TABLE-US-00005         10         20         30         40         50         60 MPGPTPSGTN VGSSGRSPSK AVAARAAGST VRQRKNASCG TRSAGRTTSA GTGGMWRFYT         70         80         90  EDSPGLKVGP VPVLVMSLLF IASVPMLHIW GKYTPS

(54) However, one of ordinary skill in the art will realize that any sequence has more than 90% similarity with above-mentioned sequence is capable to apply in the present invention.

(55) With reference to FIG. 8, the western blot assay indicates that the SEC61β expression in the colorectal cancer tissue is higher than in the normal tissue. A similar result is also shown in an IHC assay.

(56) With reference to FIGS. 9 and 10, shows a relationship between the autoantibody induced from the SEC61β and the colorectal cancer. In the blood samples of colorectal cancer patients, the expression of SEC61β autoantibody is higher than in the blood samples of healthy controls. Also, in the earlier stage of the colorectal cancer, the SEC61β autoantibody in the blood is significantly increased.

(57) With reference to FIG. 11 and Table 4, shows a comparison of the detection efficiency between the SEC60 autoantibody and the current used CEA marker. The SEC61β autoantibody has the sensitivity and the specificity in detection the colorectal cancer by 79% and 75% respectively, whereas for CEA, these values were 40% and 87%, respectively. Combination of the SEC61β autoantibody and the CEA marker show a higher detection capacity than either marker alone (AUC=0.838).

(58) TABLE-US-00006 TABLE 4 A comparison of the detection efficiency in the colorectal cancer between the SEC61β autoantibody and the current used CEA marker. Sample anti-SEC61βantibody CEA anti-SEC61βantibody + CEA Source No. positive negative P value positive negative P value positive negative P value Patient 86 68(79%) 18(21%) <0.001 34(40%) 52(60%) <0.001 61(71%) 25(29%) <0.001 Healthy 72 18(25%) 54(75%)  9(13%) 63(87%)  8(11%) 64(89%) volunteer

(59) With reference to Table 5, shows the detection efficiency of the SEC6β autoantibody and the current used CEA in detecting CRC patients with different stages. The SEC61β autoantibody has great capability to detect the colorectal cancer in early stage and has detection efficiency more than 75%.

(60) TABLE-US-00007 TABLE 5 The detection efficiency of the SEC61β autoantibody and the current used CEA in CRC patients with different stages SEC61β Stage of Sample CEA autoantibody CEA + CRC No. positive positive SEC61β I 10 0(0%)  8 (80%)  7(70%) II 25 11(44%) 19(76%) 18(72%) I + II 35 11(31%) 27(77%) 25(71%) III 35 12(34%) 29(83%) 26(74%) IV 16 11(69%) 12(75%) 10(63%) III + IV 51 23(45%) 41(80%) 36(71%) All 86 34(40%) 68(79%) 61(71%)

(61) This embodiment indicates the SEC61β autoantibody has high capability to be a serological marker to detect the colorectal cancer; especially in the early stage, the SEC61β autoantibody has the detection efficiency more than 75%, which is helpful to promote the early detection and the prognosis evaluation.

Embodiment 3: The Capability of a STOML2 Protein as a Serological Marker for Detecting the Colorectal Cancer

(62) A STOML2 protein is selected and listed in the Table 3 to be identified the capability to detect the colorectal cancer.

(63) This embodiment has a difference with the embodiment 2, that is, this embodiment uses an ELISA method to detect the expression of the STOML2 protein in blood samples of colorectal cancer patients or healthy controls.

(64) The STOML2 of the embodiment has an UniProt accession number Q9UJZ1 with sequence shows as following (SEQ ID NO: 3):

(65) TABLE-US-00008         10         20         30         40         50         60 MLARAARGTG ALLLRGSLLA SGRAPRRASS GLPRNTVVLF VPQQEAWVVE RMGRFHRILE         70         80         90        100        110        120 PGLNILIPVL DRIRYVQSLK EIVINVPEQS AVTLDNVTLQ IDGVLYLRIM DPYKASYGVE        130        140        150        160        170        180 DPEYAVTQLA QTTMRSELGK LSLDKVFRER ESLNASIVDA INQAADCWGI RCLRYEIKDI        190        200        210        220        230        240 HVPPRVKESM QMQVEAERRK RATVLESEGT RESAINVAEG KKQAQILASE AEKAEQINQA        250        260        270        280        290        300 AGEASAVLAK AKAKAEAIRI LAAALTQHNG DAAASLTVAE QYVSAFSKLA KDSNTILLPS        310        320        330        340        350 NPGDVTSMVA QAMGVYGALT KAPVPGTPDS LSSGSSRDVQ GTDASLDEEL DRVKMS

(66) However, one of ordinary skill in the art will realize that any sequence has more than 90% similarity with above-mentioned sequence is capable to apply in the present invention.

(67) With reference to FIGS. 12 and 13, the expression of the STOML2 has significant difference in the blood samples of the colorectal cancer patient and the healthy control, which is similar to the results of the embodiment 1 and 2. Also, combination of the STOML2 protein and the current used CEA marker increases the detection capability of the STOML2 protein.

(68) With reference to FIGS. 14 and 15, shows the detection efficiency and the specificity of the STOML 2 autoantibody in detecting the colorectal cancer. The STOML 2 autoantibody in the blood sample of the colorectal cancer is significantly higher than in the healthy control. The STOML 2 autoantibody level in the colorectal cancer is also higher than other cancers.

Embodiment 4

(69) In this embodiment, the sensitivity and the specificity of different antigens is calculated by a backward elimination method with multivariate logistic regression model. All antigens have excellent specificity and are able to be applied alone or combined each other to improve the efficiency of the colorectal cancer detection.

(70) TABLE-US-00009 TABLE 6 The sensitivity and the specificity of different antigens in detecting the colorectal cancer sensitivity specificity antigen (%) (%) AUC.sup.1 p value.sup.2 RPH3AL 13.83 90.67 0.554 0.33 RPL36 22.34 96 0.597 <0.05 STOML2 36.17 90.67 0.666 <0.001 P53 31.91 94.67 0.655 <0.001 Survivin 14.89 93.33 0.574 0.07 ANXA4 11.7 94.67 0.523 0.51 SEC61β 18.09 97.33 0.612 <0.001 STOML2 + P53 + 52.1 88 0.754 Survivin + ANXA4 .sup.1AUC: area under ROC curve .sup.2p-value is calculated by t-test