TWI838648B - Uses and methods of compositions for the detection of colorectal cancer - Google Patents

Abstract

The present invention provides uses and methods of compositions for the detection of colorectal cancer, wherein the peptide composition for detecting colorectal cancer is selected from the group of amino acid sequences of SEQ ID NO: 1, SEQ ID NO: 2 and SEQ ID NO: 3 any two or more thereof. The detection method comprising the steps of: (a) Using a detection method to detect the amount of the peptide composition after a test blood sample of a test individual is subjected to a purification method; (b) Comparing the amount of the peptide composition measured by (a) with the amount of the peptide composition obtained from a blood sample of a non-colorectal cancer individual after being subjected to the purification method, determine whether the test individual suffers from colorectal cancer.

Description

Uses and methods of compositions for detecting colorectal cancer

The present invention relates to a biomarker composition and a detection method for detecting cancer, and in particular to a detection method and a peptide composition for detecting colorectal cancer.

According to the data from GLOBOCAN 2018 (Bray F, Ferlay J, Soerjomataram I, Siegel RL, Torre LA, Jemal A. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2018 Nov; 68(6): 394-424. doi: 10.3322/caac.21492. Epub 2018 Sep 12. Erratum in: CA Cancer J Clin. 2020 Jul; 70(4): 313. PMID: 30207593.), colorectal cancer is one of the most prevalent cancers in the world, and its global mortality rate ranks third among malignant tumors. It is estimated that there will be 1.8 million new cases and 881,000 deaths each year.

According to the 2017 National Health Administration Cancer Registration Report, colorectal cancer currently ranks second in the incidence rate of the top ten cancers and third in the mortality rate of the top ten cancers in Taiwan. Colorectal cancer is more common in Taiwan between the ages of 40 and 50, and the probability of males suffering from colorectal cancer is 15% of all cancers, while that of females is 8%. According to the Ministry of Health and Welfare's cancer registration statistics, the standardized incidence rate has been 22.9 per 100,000 population since 1995. In 2006, the number of colorectal cancer cases surpassed liver cancer for the first time, becoming the most common cancer in Taiwan, with more than 15,000 cases. The standardized incidence rate in 2015 was 43.0 per 100,000 population, an increase of 87.8%.

Risk factors for colorectal cancer include age, gender, genetic factors, having a first-degree relative with colorectal cancer, lack of exercise, smoking, obesity, and unhealthy eating habits.

The cancer staging of colorectal cancer is classified according to the seventh edition of the TNM staging system issued by the Union for International Cancer Control and the American Joint Committee on Cancer (AJCC) in 2009: depth of tumor invasion (T), number of lymph node invasion (N), and whether there is distant metastasis (M).

There are no obvious symptoms in the early stages of colorectal cancer. By the time it is detected, it is usually in the late stages of the cancer. Common symptoms include abdominal pain, diarrhea, constipation, rectal bleeding, occult blood in the stool, or weight loss.

Current colorectal cancer detection methods include: fecal occult blood test (FOBT), which uses immunoassay or chemical method to detect hemoglobin, but the sensitivity of this detection method is about 50%~70%; palpation, which can only diagnose lesions near the rectum; sigmoidoscope, which is an invasive examination. Before the test, the patient needs to fast and take medicine to remove feces in the intestine. There are also other risks, such as perforation. Barium enema can detect the entire colon, but its sensitivity is poor and may miss lesions such as polyps. Colonoscopy can be used directly to evaluate the condition of the colon. It is an invasive examination and has other risks, such as perforation.

The blood tumor markers currently used in clinical diagnosis of colorectal cancer are carcinoembryonic antigen (CEA) and carbohydrate antigen 19-9 (CA 19-9), with sensitivities of 33.3~64.5% and 33.3~47.8%, and specificities of 89.2~90.9% and 90.5~95.9%, respectively.

Hermunend et al. proposed the use of carcinoembryonic antigen and glycoantigen 19-9 for the prognosis of colorectal cancer (Acta Oncol, 2020.59(12): p.1416-1423), and estimated their sensitivities to be 97.0% and 89%, and their specificities to be 31.0% and 16%, respectively. The combined use of carcinoembryonic antigen and glycoantigen 19-9: carcinoembryonic antigen (≤5 μ g/l) combined with glycoantigen 19-9 (≤26 kU/l) had a sensitivity of 88.0% and a specificity of 9.0%, carcinoembryonic antigen (>5 μ g/l) combined with glycoantigen 19-9 (≤26 kU/l) had a sensitivity of 88.0% and a specificity of 9.0%, and carcinoembryonic antigen (>5 μ g/l) combined with glycoantigen 19-9 (≤26 kU/l) had a sensitivity of 88.0% and a specificity of 9.0%. g/l) combined with glycoantigen 19-9 (≤26 kU/l) has a sensitivity of 100.0% and a specificity of 88.0%. Therefore, currently in clinical practice, carcinoembryonic antigen and glycoantigen 19-9 are not used as biomarkers for colorectal cancer screening, but as biomarkers for prognosis assessment of colorectal cancer.

Although the prognostic value of carcinoembryonic antigen and glycogen antigen 19-9 used in clinical practice for postoperative follow-up of colorectal cancer meets clinical needs, their sensitivity for screening colorectal cancer is low and does not meet clinical expectations. Therefore, there is still a lack of non-invasive biomarkers for colorectal cancer screening that are both highly sensitive and highly specific.

One of the purposes of the present invention is to solve the problem of insufficient performance of non-invasive biomarker detection of colorectal cancer.

According to the purpose of the present invention, a peptide composition for detecting colorectal cancer, wherein the peptide composition is selected from a group consisting of any two or more of the amino acid sequences of SEQ ID NO: 1, SEQ ID NO: 2 and SEQ ID NO: 3, wherein the peptide composition is present in a blood sample.

Among them, the peptide combination for detecting colorectal cancer is detected by mass spectrometry analysis to detect the content of the peptide combination.

Wherein, the blood sample is a serum sample, a plasma sample, or a whole blood sample.

Among them, the blood sample was purified by wheat germ agglutinin.

According to the purpose of the present invention, a detection method for detecting colorectal cancer is provided, the steps comprising: (a) detecting the content of a peptide composition obtained by a purification method in a blood sample of a test subject, wherein the peptide composition is selected from a group consisting of any two or more of the amino acid sequences of SEQ ID NO: 1, SEQ ID NO: 2 and SEQ ID NO: 3; (b) comparing the content of the peptide composition detected by (a) with the content of the peptide composition obtained by the purification method in a comparison blood sample of a non-colorectal cancer subject, to determine whether the test subject suffers from colorectal cancer.

Among them, the comparison method is selected from the group consisting of multivariate logical regression, decision tree, random forest and support vector machine.

Wherein, the detection method is selected from mass spectrometry analysis or immunoassay.

Wherein, the blood sample to be tested and the blood sample for comparison are both serum samples, or the blood sample to be tested and the blood sample for comparison are both plasma samples, or the blood sample to be tested and the blood sample for comparison are both whole blood samples.

Among them, the purification method is a wheat germ agglutinin purification method.

The mass spectrometry analysis is selected from the group consisting of radical-assisted laser desorption mass spectrometry, liquid chromatography-mass spectrometry, liquid chromatography-tandem mass spectrometry, and gas chromatography-mass spectrometry.

According to this research method, it is confirmed that by detecting the peptide combination, the problem of insufficient sensitivity and specificity of traditional biomarkers for detecting colorectal cancer can be overcome.

Figure 1 shows the TNM staging system for colorectal cancer.

Figure 2 shows the grouping information of the plasma samples of the individuals to be tested.

Figure 3 shows the multiple reaction monitoring ion pairs and mass spectrum parameters of the screened peptide fragments.

Figure 4 is a dot matrix diagram of the mass spectrometry signal intensity of the peptide fragment of SEQ ID NO: 1 in the non-colorectal cancer group, the early colorectal cancer group, and the advanced colorectal cancer group.

Figure 5 is a dot matrix diagram of the mass spectrometry signal intensity of the peptide fragment of SEQ ID NO: 2 in the non-colorectal cancer group, the early colorectal cancer group, and the advanced colorectal cancer group.

Figure 6 is a dot matrix diagram of the mass spectrometry signal intensity of the peptide fragment of SEQ ID NO: 3 in the non-colorectal cancer group, the early colorectal cancer group, and the late colorectal cancer group.

Figure 7 shows the multiple reaction monitoring ion pairs and mass spectrometry analysis parameters of the screened peptide fragments.

Figure 8 shows the retention time and signal intensity of the three ion pairs of SEQ ID NO: 1.

Figure 9 shows the retention time and signal intensity of the three ion pairs of SEQ ID NO: 2.

Figure 10 shows the retention time and signal intensity of the three ion pairs of SEQ ID NO: 3.

Figure 11 is a calibration curve diagram of SEQ ID NO: 1.

Figure 12 is a calibration curve diagram of SEQ ID NO: 2.

Figure 13 is a calibration curve diagram of SEQ ID NO: 3.

Figure 14 is a dot matrix diagram showing the concentration of each peptide in each sample of each group of SEQ ID NO: 1.

Figure 15 is a dot matrix diagram showing the concentration of each peptide in each sample of each group of SEQ ID NO: 2.

Figure 16 is a dot matrix diagram showing the concentration of each peptide in each sample of each group of SEQ ID NO: 3.

Figure 17 shows the efficacy of peptide combinations in distinguishing non-colorectal cancer groups from early colorectal cancer groups evaluated by multivariate logistic regression and random forest, wherein the peptide combination is selected from a group consisting of any one or more of the amino acid sequences of SEQ ID NO: 1, SEQ ID NO: 2, and SEQ ID NO: 3. It is represented by a receiver operating characteristic curve.

Figure 18 shows the efficacy of peptide combinations in distinguishing non-colorectal cancer groups from advanced colorectal cancer groups evaluated by multivariate logistic regression and random forest, wherein the peptide combinations are selected from a group consisting of any one or more of the amino acid sequences of SEQ ID NO: 1, SEQ ID NO: 2, and SEQ ID NO: 3. It is represented by a receiver operating characteristic curve.

Figure 19 shows the efficacy of peptide combinations in distinguishing non-colorectal cancer groups from full-stage colorectal cancer groups evaluated by multivariate logistic regression and random forest, wherein the peptide combinations are selected from a group consisting of any one or more of the amino acid sequences of SEQ ID NO: 1, SEQ ID NO: 2, and SEQ ID NO: 3. It is represented by a receiver operating characteristic curve.

Figure 20 shows the performance of detecting the individual levels of SEQ ID NO: 1, SEQ ID NO: 2, and SEQ ID NO: 3 for detecting colorectal cancer, evaluated by the sum of the areas under the curves, sensitivity, specificity, accuracy, Student’s t-test, and statistical power.

Figure 21 shows the performance of detecting the individual contents of SEQ ID NO: 1, SEQ ID NO: 2, and SEQ ID NO: 3, and any two of them as a group for detecting colorectal cancer, evaluated by the sum of the area under the curve, sensitivity, specificity, accuracy, Student's t test, and statistical power.

Figure 22 shows the performance of detecting the individual contents of SEQ ID NO: 1, SEQ ID NO: 2, and SEQ ID NO: 3, and combining the three to form a group for detecting colorectal cancer, which is evaluated by the sum of the areas under the curve, sensitivity, specificity, accuracy, Student's t test, and statistical power.

In order to more clearly understand the content of the present invention, the following is a detailed description of the specific embodiments of the present invention in conjunction with the attached figures. The "one" mentioned in this specification means one, at least one, one or at least one.

Please refer to the sequence table. The present invention provides a peptide composition for detecting colorectal cancer. The peptide composition is selected from a group consisting of any two or more of the amino acid sequences of SEQ ID NO: 1, SEQ ID NO: 2, and SEQ ID NO: 3, and the peptide composition is present in a blood sample of a subject to be tested. The peptide fragment composed of the amino acid sequence of SEQ ID NO: 1 belongs to the 54th to 62nd amino acids of the protein platelet factor 4 (PF4). The peptide fragment composed of the amino acid sequence of SEQ ID NO: 2 belongs to the 429th to 438th amino acids of the protein inter-alpha-trypsin inhibitor heavy chain 4 (ITIH4). The peptide fragment composed of the amino acid sequence of SEQ ID NO: 3 belongs to the 198th to 207th amino acids of the protein apolipoprotein E (APOE).

In one embodiment, first, peptide fragments with different signal intensities in mixed plasma are identified and screened. Next, the signal intensity differences of each screened peptide fragment in individual plasma samples are measured. Then, a method for detecting colorectal cancer based on the content of peptide fragments is established. Finally, a machine learning method is used to combine different peptide fragment contents for the performance evaluation of colorectal cancer detection.

In the step of identifying and screening peptide fragments with different signal intensities in mixed plasma: The plasma samples used in the present invention were approved by the Human Trial Committee of Taipei Medical University (case numbers: N20130822 and N202007061), and 286 plasma samples of colorectal cancer subjects and 120 plasma samples of non-colorectal cancer subjects were applied for from the Taipei Medical University Joint Human Biodatabase.

The staging of colorectal cancer in the present invention is based on the seventh edition of the TNM staging system issued by the Union for international cancer control and the American Joint Committee on Cancer (AJCC) in 2009, where TMN refers to: depth of tumor invasion (T), number of lymph node invasion (N), and whether there is distant metastasis (M).

The basic data of plasma sample grouping is shown in Figure 2. 10 cases of colorectal cancer stage I, colorectal cancer stage II, colorectal cancer stage III, and colorectal cancer stage IV were selected from the discovery group, and 20 cases of non-colorectal cancer were matched according to gender and age. The plasma samples of each group were quantitatively mixed to obtain mixed plasma samples of each group for identification and screening of peptide fragments with different signal intensities in non-colorectal cancer and colorectal cancer of various stages.

The mixed plasma samples of each group were first purified by using agarose-bound wheat germ agglutinin (WGA, Vector Laboratories, Burlingame, CA, USA) to obtain the purified protein samples of the mixed plasma of each group. Then, the purified protein samples of the mixed plasma samples of each group were subjected to in-solution digestion to decompose the proteins into peptide fragments, which were then analyzed by nanoscale liquid chromatography-tandem mass spectrometry (Orbitrap Elite hybrid mass spectrometer, Thermo Electron, Bremen, Germany). Next, the data collected by nanoscale liquid chromatography-tandem mass spectrometry analysis were compared with the Universal Protein Knowledgebase (UniProt, 18 January 2020) through proteomics mass spectrometry analysis software (PEAKS 7 software, Bioinformatics Solutions, Waterloo, ON, Canada) to identify the protein to which each peptide fragment belongs, the post-translational modification on the peptide fragment, and the label-free quantification method (label-free quantification) to quantify the peptide fragments with different signal intensities. The peptide fragments with different signal intensities were screened out according to the analysis data of PEAKS 7 software. The screening results are shown in Figure 3. Finally, three peptide fragments without post-translational modification were screened out, including: SEQ ID NO: 1, SEQ ID NO: 2, and SEQ ID NO: 3. Among them, the retention time of SEQ ID NO: 1 is 5.9 minutes, three ion pairs are selected, the charge to mass ratio (m/z) of the parent ion is 520.3, the charge to mass ratio of the daughter ion is 902.5, 789.4 and 251.1 respectively, the collision energy is 20.1 electron volts, and the acceleration voltage is 130 volts. Among them, the retention time of SEQ ID NO: 2 is 6.2 minutes, three ion pairs are selected, the charge to mass ratio of the parent ion is 500.2, the charge to mass ratio of the daughter ion is 815.4, 702.3 and 587.3 respectively, the collision energy is 16.5 electron volts, and the acceleration voltage is 130 volts. Among them, the retention time of SEQ ID NO: 3 is 4.4 minutes. Three ion pairs are selected, the charge-mass ratio of the parent ion is 484.7, the charge-mass ratio of the daughter ion is 588.3, 489.2 and 360.1 respectively, the collision energy is 22 electron volts, and the acceleration voltage is 130 volts.

Among them, an ion pair is a group consisting of a parent ion and a daughter ion, and the values are all expressed in charge-to-mass ratio, where the charge-to-mass ratio is the ratio of the charge carried by a charged particle to its mass.

Among them, wheat germ agglutinin is a lectin extracted from wheat plant, and has a highly specific binding ability to N-acetyl-D-glucosamine (GlcNAc) structure on glycoprotein. In this embodiment, the steps of purifying mixed plasma samples with wheat germ agglutinin-agglutinin particles are as follows: first, 100 μL of wheat germ agglutinin-agglutinin particles are added to a 600 μL microcentrifuge tube, and then 20 μL of mixed plasma sample is added, and the mixture is fully mixed at room temperature with a rocker shaker. Next, after centrifugation, the microcentrifuge tube was washed three times with 400 μL phosphate buffered saline (PBS) to remove the protein not adsorbed on the WGAG-AG particles, and then washed twice with 50 μL precipitation buffer (0.5M N-acetylglucosamine dissolved in 1mM acetic acid) to precipitate the protein adsorbed on the WGAG-AG particles. The collected protein was collected in another microcentrifuge tube and quantified by Bradford assay. 20 μL of each sample was taken out and the protein was quantified by Bradford assay. g of precipitated protein into a new microcentrifuge tube and then dried using a vacuum concentrator to obtain a protein purified sample of mixed plasma purified by WWAG-AG particles.

The decomposition steps in the solution are as follows: First, take 10 μg of purified protein samples of individual plasma of each group and add 20 μL of an aqueous solution containing 0.1% formic acid for re-dissolution. Then, add 1 μL of 550mM dithiothreitol (DTT) and react at 56°C for 45 minutes. Then, add 2 μL of 450mM iodoacetamide (IAM) and react at room temperature in the dark for 45 minutes. Next, add 0.5 μg of trypsin and react at 37°C for 16 hours. Finally, add an appropriate amount of 1% formic acid aqueous solution to make the final formic acid concentration of the whole solution 0.1% to terminate the enzyme reaction.

Among them, nano-liquid chromatography-tandem mass spectrometry analysis is used to analyze samples that have been decomposed into peptide fragments by proteases. The peptide mixture was separated by a chromatography column (inner diameter: 75μm, length: 25cm, C18 BEH column, pore size: 130Å, filler material particle size: 1.7 μm , Waters), mobile phase A: aqueous solution containing 0.1% formic acid, mobile phase B: acetonitrile solution containing 0.1% formic acid, the separation gradient is that mobile phase B increases from 5% to 35% within 60 minutes, and the mobile phase A gradient decreases from 95% to 65% as the percentage of mobile phase B changes, flow rate: 300nL/min, column temperature: 35℃. Full-scan MS spectra were obtained by nano-liquid chromatography-tandem mass spectrometry analysis. The screening settings included: the mass-to-charge ratio was between 350 and 1600, and the target value of automatic gain control (AGC) was 10 ⁶ . The system resolution of the mass spectrometer was set to 120K (the system resolution value was provided by the mass-to-charge ratio of 400). The first 20 ions with the strongest signals were separated in sequence into a linear ion trap and fragmented by collision-induced dissociation (CID) to obtain the daughter ion mass spectrum data. The target value of automatic gain control was 10 ⁴ and the dynamic exclusion time was 60 seconds.

PEAKS 7 software (Bioinformatics Solutions, Waterloo, ON, Canada) was used for analysis, and the parameters were set as follows: MS tolerance: 10 ppm, MS/MS tolerance: 0.6 Da, and false detection rate: 1%. In addition, the protein quantitative additional tool (PEAKS Q) of PEAKS 7 software can analyze the database search results and quantitatively analyze the peptide sequences of proteins with different signal intensities through total ion current (TIC) normalization and label-free quantitative method. The parameters were set as follows: MS tolerance: 10 ppm, and retention time shift tolerance: 60 minutes. The post-translational modifications (PTM) of PEAKS 7 software were set to identify the post-translational modifications on the peptide sequence, including glycosylation and methylation. The parameters were set as follows: Carbamidomethylation (C)/+57.0215Da was set as fixed, oxidation (M)/+15.9949Da, glycosylation-related PTMs and methylation-related PTMs were set as variables. Glycosylation-related PTMs included hexose modified CRKTW (+162.0528Da), fucose modified TS (+146.0579Da), O-GlcNac modified STN (+203.195Da), Hex1HexNAc1 modified N(+511.1901Da), Hex1HexNAc1NeuAc1 modified NTS(+656.2276Da), and Hex1HexNAc1NeuAc2 modified NTS(+947.3231Da). The methylation-related post-translational modifications include: methylation modified CDE-HIKLNQRST(+14.0156Da), dimethylation modified KNR(+28.0313Da), K(+32.0564Da), K(+34.0631Da), and trimethylation modified KAR(+43.0058Da).

Among them, peptide fragments with different signal intensities were screened, and the parameters were set as follows: the statistically significant difference value (-10 lgP) was less than 13, which was equivalent to a p value less than 0.05. And peptide fragments with different signal intensities were screened, wherein the signal intensity difference was that the signal intensity of individual peptide fragments in the plasma samples of any stage of colorectal cancer group must be greater than 1.5 times the signal intensity in the plasma samples of non-colorectal cancer, or less than 0.8 times the signal intensity in the plasma samples of non-colorectal cancer. The signal-to-noise ratio (S/N) was greater than 5. Finally, mis-cleavage peptides were excluded, and peptide fragments with more than 10 amino acids were excluded.

Determining the signal intensity differences of each screened peptide fragment in individual plasma samples: In the step of determining the signal intensity differences of each screened peptide fragment in individual plasma samples, six peptide fragments with signal intensity differences in individual samples were verified in individual plasma samples, including three peptide fragments without post-translational modification: SEQ ID NO: 1, SEQ ID NO: 2, SEQ ID NO: 3. The basic data of plasma sample grouping is shown in Figure 2. 10 cases of colorectal cancer stage I, colorectal cancer stage II, colorectal cancer stage III, and colorectal cancer stage IV were selected from the discovery group, and 20 cases of non-colorectal cancer were matched according to gender and age. Colorectal cancer stage I and colorectal cancer stage II were classified as early colorectal cancer group, and colorectal cancer stage III and colorectal cancer stage IV were classified as late colorectal cancer group.

The individual plasma samples of each group were first purified by wheat germ agglutinin-aggregate particles to obtain purified protein samples of individual plasma samples. Then, the purified protein samples of individual plasma samples were decomposed in solution to decompose the proteins into peptide fragments, which were analyzed by UPLC-MS/MS. The three peptide fragments in Figure 3 were analyzed by non-labeled quantitative method and one-way analysis of variance (one-way ANOVA) was used.

The results are shown in Figures 4 to 6, where Figure 4 shows that in the SEQ ID NO: 1 content detection, the signal strength of the non-colorectal cancer group is 953.92±1257.19, the signal strength of the early colorectal cancer group is 2245.91±2046.9, and the signal strength of the late colorectal cancer group is 2920.63±3134.69. Figure 5 shows that in the SEQ ID NO: 2 group, the signal strength of the non-colorectal cancer group is 3144.23±3285, the signal strength of the early colorectal cancer group is 1536.31±1373.44, and the signal strength of the late colorectal cancer group is 1427.26±1386.54. Figure 6 shows that in the SEQ ID NO: 3 group, the signal strength of the non-colorectal cancer group is 3937±2082.73, the signal strength of the early colorectal cancer group is 2164.65±863.8, and the signal strength of the late colorectal cancer group is 1550.17±1405.8.

Figure 4 shows that the signal intensity of SEQ ID NO: 1 in the early colorectal cancer group and the late colorectal cancer group are both higher than the signal intensity of the non-colorectal cancer group, wherein the signal intensity of the early colorectal cancer group is 2.35 times that of the non-colorectal cancer group, and the signal intensity of the late colorectal cancer group is 3.06 times that of the non-colorectal cancer group. Figure 5 shows that the signal intensity of SEQ ID NO: 2 in the early colorectal cancer group and the late colorectal cancer group are both lower than the signal intensity of the non-colorectal cancer group, wherein the signal intensity of the early colorectal cancer group is 0.48 times that of the non-colorectal cancer group, and the signal intensity of the late colorectal cancer group is 0.45 times that of the non-colorectal cancer group. Figure 6 shows that the signal intensity of SEQ ID NO: 3 in the early colorectal cancer group and the late colorectal cancer group are both lower than that in the non-colorectal cancer group. The signal intensity of the early colorectal cancer group is 0.54 times that of the non-colorectal cancer group, and the signal intensity of the late colorectal cancer group is 0.39 times that of the non-colorectal cancer group.

In the step of establishing a method for diagnosing colorectal cancer based on the content of peptide fragments: According to the present invention, three peptide fragments with different signal strengths are found, and the three peptide fragments are analyzed by stable isotope-labeled MRM assay. Among them, the three peptide fragments include: SEQ ID NO: 1, SEQ ID NO: 2 and SEQ ID NO: 3.

The basic data of plasma sample grouping is shown in Figure 2. From the validation group, 47 cases of colorectal cancer stage I, 53 cases of colorectal cancer stage II, 50 cases of colorectal cancer stage III, 56 cases of colorectal cancer stage IV, and 80 cases of non-colorectal cancer were selected and matched according to gender and age. Colorectal cancer stage I and colorectal cancer stage II were classified as early colorectal cancer group, and colorectal cancer stage III and colorectal cancer stage IV were classified as late colorectal cancer group. In addition, the results of colorectal cancer stage I, stage II, stage III and stage IV were combined and classified as full-stage colorectal cancer group.

The individual plasma samples of each group were first purified by wheat germ agglutinin-aggregate particles to obtain the protein purified samples of the individual plasma samples. Then, the extended peptide sequence was added to the protein purified samples of the individual plasma samples, and then decomposed in the solution to decompose the protein into peptide fragments. The fragments were analyzed by triple quadrupole mass spectrometry (1260 Infinity II Quaternary Pump LC system). Quantitative analysis of SEQ ID NO: 1, SEQ ID NO: 2 and SEQ ID NO: 3 using internally added standards was performed, and the statistical results were analyzed using one-way analysis of variance (one-way ANOVA).

Among them, in the stable isotope-labeled MRM assay analysis, the extended peptide sequence is first added, and then the steps of decomposition in the solution are as follows: First, take 10μg of the protein purified sample of each group of individual plasma and add 20μL of aqueous solution containing 0.1% formic acid for re-dissolution. Then, add 1μL of internal standard, and then add 1μL of 550mM dithiothreitol (DTT) and react at 56℃ for 45 minutes. Next, 2 μL of 450 mM iodoacetamide (IAM) was added to react at room temperature in the dark for 45 minutes. Next, 0.5 μg of trypsin was added to react at 37°C for 16 hours. Finally, an appropriate amount of 1% formic acid aqueous solution was added to make the final formic acid concentration of the entire solution 0.1% to terminate the enzyme reaction.

The quantitative analysis steps using the internally added standard are as follows: the peak area of the non-isotope-labeled extended peptide sequence 1.95ng/mL~1000ng/mL is divided by the peak area obtained by the isotope-labeled extended peptide sequence analysis to obtain the relationship between the ratio of the non-isotope-labeled extended peptide sequence/isotope-labeled extended peptide sequence and the concentration of the standard as a calibration curve. At the same time, the SEQ ID NO: 1, SEQ ID NO: 2, SEQ ID NO: 3 detected in the plasma sample are divided by the peak area of the isotope-labeled extended peptide sequence analysis obtained by the analysis, and the ratio obtained is used to obtain the concentration of SEQ ID NO: 1, SEQ ID NO: 2, SEQ ID NO: 3 in the plasma sample through the calibration curve.

Among them, in the three-stage quadrupole mass spectrometer analysis, the peptide mixture was chromatographically separated by a column (C18 column, Phenomex, Kinetex, filler material particle size: 2.6μm, pore size: 100Å, length: 50 cm, inner diameter: 2.1 mm), mobile phase A was an aqueous solution containing 0.1% formic acid, and mobile phase B was an acetonitrile solution containing 0.1% formic acid. The overall chromatographic time was 12 minutes, including: 0-1 minute: 95% mobile phase A, 5% mobile phase B; 1-8 minutes, 45% mobile phase A and 55% mobile phase B; 8-9 minutes, 100% B, and a post-run time of 3 minutes. The injection volume of each sample was 5μL, and the mobile phase flow rate was 0.4mL/min. Among them, the isotope-labeled multiple reaction monitoring technology uses the amino acid sequences of peptide fragments SEQ ID NO: 1, SEQ ID NO: 2 and SEQ ID NO: 3 as the basic sequence, and according to the amino acid sequence of the protein to which the basic sequence belongs, four amino acids are added back at the head and tail of the basic sequence to simulate the state of enzyme cutting, which is hereinafter referred to as the extended peptide fragment. The extended peptide sequence synthesized according to the screened SEQ ID NO: 1, SEQ ID NO: 2 and SEQ ID NO: 3 includes: an extended peptide sequence without isotope labeling and an extended peptide sequence with a stable isotope labeling. The extended peptide sequence is shown in Figure 7, which is hereinafter referred to as the internally added standard. Among them, the non-isotope-labeled extended peptide sequence was used to prepare a standard with a concentration of 1.95ng/mL~1000ng/mL as a calibration curve, and then an equal amount of the extended peptide sequence with a stable isotope label was added to the standard and plasma samples for subsequent analysis.

The isotope labeling is performed with carbon isotope ( ¹³ C) and nitrogen isotope ( ¹⁵ N). The results are shown in Figures 8 to 16. Figures 8, 9 and 10 represent the retention time and signal intensity of three ion pairs of each peptide fragment of SEQ ID NO: 1, SEQ ID NO: 2 and SEQ ID NO: 3, respectively.

Among them, Figures 11, 12 and 13 represent the calibration curves of SEQ ID NO: 1, SEQ ID NO: 2 and SEQ ID NO: 3, respectively, and the lower limit of quantification (LLOQ) and the upper limit of quantification (ULOQ) are determined by chicken serum, wherein the quantitative range of SEQ ID NO: 1 shown in Figure 11 is between 3.90 and 1000 ng/mL, wherein the quantitative range of SEQ ID NO: 2 shown in Figure 12 is between 1.95 and 250 ng/mL, wherein the quantitative range of SEQ ID NO: 3 shown in Figure 13 is between 1.95 and 250 ng/mL. Among them, Figures 14, 15 and 16 represent the dot matrix plots of the concentration results of each peptide in each sample of each group. Among them, in the quantitative results of each group, the concentration of SEQ ID NO: 1 in the non-colorectal cancer group was 9.59±27.14ng/mL, the concentration in the early colorectal cancer group was 25.41±33.36ng/mL, the concentration in the late colorectal cancer group was 27.81±11.55ng/mL, and the concentration in the all-stage colorectal cancer group was 26.61±33.82ng/mL. Among them, in the quantitative results of each group, the concentration of SEQ ID NO: 2 in the non-colorectal cancer group was 3.05±3.44ng/mL, the concentration in the early colorectal cancer group was 2.07±0.27ng/mL, the concentration in the late colorectal cancer group was 2.10±1.56ng/mL, and the concentration in the all-stage colorectal cancer group was 2.08±0.39ng/mL. Among them, the quantitative results of SEQ ID NO: 3 in each group showed that the concentration of non-colorectal cancer group was 24.79±13.58ng/mL, the concentration of early colorectal cancer group was 14.37±4.23ng/mL, the concentration of late colorectal cancer group was 16.85±12.87ng/mL, and the concentration of all-stage colorectal cancer group was 15.61±3.75ng/mL.

The quantitative results are shown in Figures 14, 15, and 16. Figure 14 shows that in the quantitative results of SEQ ID NO: 1 in each group, the concentration of the non-colorectal cancer group was 9.59±27.14ng/mL, the concentration of the early colorectal cancer group was 25.41±33.36ng/mL, the concentration of the late colorectal cancer group was 27.81±11.55ng/mL, and the concentration of the all-stage colorectal cancer group was 26.61±33.82ng/mL. Among them, the content of the early colorectal cancer group was 0.68 times that of the non-colorectal cancer group, and the content of SEQ ID NO: 1 in the late colorectal cancer group was 0.69 times that of the non-colorectal cancer group. Figure 15 shows the quantitative results of SEQ ID NO: 2 in each group. The concentration of the non-colorectal cancer group was 3.05±3.44ng/mL, the concentration of the early colorectal cancer group was 2.07±0.27ng/mL, the concentration of the late colorectal cancer group was 2.10±1.56ng/mL, and the concentration of the all-stage colorectal cancer group was 2.08±0.39ng/mL. Among the SEQ ID NO: 2 groups, the content of SEQ ID NO: 2 in the early colorectal cancer group was 2.65 times that of the non-colorectal cancer group, and the content of SEQ ID NO: 2 in the late colorectal cancer group was 2.90 times that of the non-colorectal cancer group. Figure 16 shows the quantitative results of SEQ ID NO: 3 in each group. The concentration of non-colorectal cancer group is 24.79±13.58ng/mL, the concentration of early colorectal cancer group is 14.37±4.23ng/mL, the concentration of late colorectal cancer group is 16.85±12.87ng/mL, and the concentration of all-stage colorectal cancer group is 15.61±3.75ng/mL. The content of SEQ ID NO: 3 in the early colorectal cancer group is 0.58 times that of the non-colorectal cancer group, and the content of SEQ ID NO: 3 in the late colorectal cancer group is 0.68 times that of the non-colorectal cancer group.

In the step of evaluating the performance of combining different peptide fragment contents for detecting colorectal cancer by machine learning: In the step of evaluating the performance of combining different peptide fragment contents for detecting colorectal cancer by machine learning, GraphPad Prism (vers.5.0; GraphPad Software, San Diego, CA, USA) was used to draw the receiver operating characteristic curve (ROC curve) and calculate the area under the curve (AUC), and the Youden index was used as the cutoff value for judging the diagnostic effect to display the sensitivity, specificity, accuracy, p value, and statistical power, where the p value was calculated using the Student's t test to determine whether there was a significant difference between the groups.

In addition, in order to evaluate the efficacy of any one or any two or more of the three individual peptides in the diagnosis of colorectal cancer, the results of the quantitative addition of the standard were integrated through machine learning methods, including: decision tree (DT), random forest (RF), support vector machine (SVM), and multivariate logistic regression (LR), and paired comparisons of any two different receiver operating characteristic curves were performed using biomedical statistical software (MedCalc Statistical Software, vers. 15.4; MedCalc Software, Ostend, Belgium).

Among them, the machine learning method uses scikit-learn (vers.0.21.3) to perform 10-fold cross validation, which is to split the data into ten equal parts, use nine-tenths of the data for machine learning model training, and the remaining one-tenth of the data for verifying the machine learning model training results. The parameter tuning is adjusted according to the area under the curve and the value obtained by the training set and validation set in each 10-fold cross validation. The final parameters of each machine learning model are as follows: The decision tree parameter settings are as follows: Decision tree depth: 10, the core model (kernel of the model) is set to Gini. The parameters in the random forest are set as follows: the number of decision trees (tree value number): 200, the core model is set to gini. The support vector parameters are set as follows: γ value (value of gamma): 1× ^10-10 , C value (value of C): 1× ^10-7 , and the kernel of the model is set to radial basis function (RBF). The multivariate logical regression uses the default settings. Finally, the performance of each machine learning model is evaluated by the sum of areas under the curve, sensitivity, specificity, accuracy, Student's t test, and statistical power. The results are shown in Figures 17 to 22.

FIG. 17 shows the efficacy of the peptide combination in distinguishing the non-colorectal cancer group from the early colorectal cancer group by multivariate logistic regression and random forest evaluation, wherein the sum of the areas under the multivariate logistic regression curve for differentiation based on the content of SEQ ID NO: 1 is 0.67, the sum of the areas under the multivariate logistic regression curve for differentiation based on the content of SEQ ID NO: 2 is 0.80, the sum of the areas under the multivariate logistic regression curve for differentiation based on the content of SEQ ID NO: 3 is 0.82, the sum of the areas under the multivariate logistic regression curve for differentiation based on the content of SEQ ID NO: 1, SEQ ID NO: 2 and SEQ ID NO: 3 combined is 0.90, the sum of the areas under the multivariate logistic regression curve for differentiation based on the content of SEQ ID NO: 1, SEQ ID NO: 2 and SEQ ID NO: 3 combined is 0.91, and the sum of the areas under the multivariate logistic regression curve for differentiation based on the content of SEQ ID NO: 1, SEQ ID NO: 2 and SEQ ID NO: 3 combined is 0.93. The sum of the areas under the random forest curve for NO:3 classification is 0.88. FIG. 18 shows the efficacy of the peptide combination in distinguishing the non-colorectal cancer group from the advanced colorectal cancer group by multivariate logistic regression and random forest evaluation, wherein the sum of the areas under the multivariate logistic regression curve for differentiation based on the content of SEQ ID NO: 1 is 0.63, the sum of the areas under the multivariate logistic regression curve for differentiation based on the content of SEQ ID NO: 2 is 0.72, the sum of the areas under the multivariate logistic regression curve for differentiation based on the content of SEQ ID NO: 3 is 0.70, the sum of the areas under the multivariate logistic regression curve for differentiation based on the content of SEQ ID NO: 1, SEQ ID NO: 2 and SEQ ID NO: 3 combined is 0.85, the sum of the areas under the multivariate logistic regression curve for differentiation based on the content of SEQ ID NO: 1, SEQ ID NO: 2 and SEQ ID NO: 3 combined is 0.86, and the sum of the areas under the multivariate logistic regression curve for differentiation based on the content of SEQ ID NO: 1, SEQ ID NO: 2 and SEQ ID NO: 3 combined is 0.94. The sum of areas under the random forest curve for NO:3 classification is 0.88. It is represented by the receiver operating characteristic curve and the sum of areas under the curve is calculated. Figure 19 shows the efficacy of the peptide combination in differentiating the non-colorectal cancer group and the full-stage colorectal cancer group by multivariate logistic regression and random forest evaluation, wherein the sum of the area under the curve for differentiation based on the content of SEQ ID NO: 1 is 0.66, the sum of the area under the curve for differentiation based on the content of SEQ ID NO: 2 is 0.77, the sum of the area under the curve for differentiation based on the content of SEQ ID NO: 3 is 0.79, the sum of the area under the multivariate logistic regression curve for differentiation based on SEQ ID NO: 1, SEQ ID NO: 2 and SEQ ID NO: 3 is 0.88, and the sum of the area under the random forest curve for differentiation based on SEQ ID NO: 1, SEQ ID NO: 2 and SEQ ID NO: 3 is 0.96.

Figure 20 shows the performance of detecting the individual contents of SEQ ID NO: 1, SEQ ID NO: 2, and SEQ ID NO: 3 for detecting colorectal cancer, wherein the sum of the area under the curve of SEQ ID NO: 1 is 0.67 (0.58-0.74), the sensitivity is 75.7% (67.1%-87.5%), the specificity is 56.5% (43.3%-69.0%), and the accuracy is 63.2% (55.9%-80.2%). The sum of the area under the curve of SEQ ID NO: 2 is 0.80 (0.77-0.83), the sensitivity is 64.3% (51.9%-75.4%), the specificity is 87.1% (76.1%-95.3%), and the accuracy is 76.3% (72.7%-82.3%). Among them, the sum of the areas under the curve of SEQ ID NO: 3 is 0.82 (0.74-0.88), the sensitivity is 78.6% (67.1%-87.5%), the specificity is 74.2% (61.5%-84.5%), and the accuracy is 77.5% (73.3%-83.9%).

Figure 21 shows the performance of detecting the individual contents of SEQ ID NO: 1, SEQ ID NO: 2, and SEQ ID NO: 3, and any two of them as a group for detecting colorectal cancer. The performance of the contents of SEQ ID NO: 1 and SEQ ID NO: 2 for detecting colorectal cancer was 0.76-0.85 with the decision tree, random forest, support vector machine, and multivariate logistic regression, the sensitivity was 66.8%-83.2%, the specificity was 62.3%-84.2%, and the accuracy was 72.9%-78.7%. Among them, the content of SEQ ID NO: 1 and the content of SEQ ID NO: 3 were used to detect the efficacy of colorectal cancer. The sum of the areas under the curves ranged from 0.76 to 0.86, the sensitivity ranged from 79.7% to 86.1%, the specificity ranged from 66.9% to 82.0%, and the accuracy ranged from 74.5% to 80.2% using decision trees, random forests, support vector machines, and multivariate logistic regression. Among them, the content of SEQ ID NO: 2 and SEQ ID NO: 3 were used to detect colorectal cancer. The sum of the areas under the curves ranged from 0.77 to 0.87, the sensitivity ranged from 72.9% to 85.9%, the specificity ranged from 75.0% to 85.6%, and the accuracy ranged from 75.6% to 82.5% using decision trees, random forests, support vector machines, and multivariate logistic regression.

Figure 22 shows the performance of detecting the individual contents of SEQ ID NO: 1, SEQ ID NO: 2, and SEQ ID NO: 3, and combining the three to form a group for detecting colorectal cancer. The sum of the areas under the curves of decision trees, random forests, support vector machines, and multivariate logistic regression ranged from 0.78 to 0.97, the sensitivity ranged from 73.9% to 89.2%, the specificity ranged from 77.8% to 97.9%, and the accuracy ranged from 76.9% to 92.3%.

From the above experimental results, it can be seen that the performance of a peptide combination composed of any two or more of SEQ ID NO: 1, SEQ ID NO: 2 and SEQ ID NO: 3 in detecting colorectal cancer is better than that of a single peptide. In addition to having comparable specificity, the sensitivity and accuracy are better in most combinations.

In summary, the peptide combination proposed in the present invention can provide high sensitivity and high specificity for detection, which means that the method proposed in the present invention can bring application value to clinical detection.

However, the above is only a preferred embodiment of the present invention, but it cannot limit the scope of implementation of the present invention; therefore, all simple equivalent changes and modifications made according to the scope of the patent application of the present invention and the content of the invention specification are still within the scope of the patent of the present invention.

<110> Taipei Medical University

<120> Use and method of the composition for detecting colorectal cancer

<160> 3

<210> 1

<211> 9

<212> PRT

<213> Homo sapiens

<400> sequence (3 English letters represent one amino acid)

<210> 2

<211> 10

<212> PRT

<213> Homo sapiens

<400> sequence (3 English letters represent one amino acid)

<210> 3

<211> 9

<212> PRT

<213> Homo sapiens

<400> sequence (3 English letters represent one amino acid)

Claims (8)

A use of a peptide composition for detecting colorectal cancer, which is used to prepare a composition for detecting colorectal cancer, wherein the peptide composition is composed of three peptide fragments, and the amino acid sequences of the peptide fragments are shown in SEQ ID NO: 1, SEQ ID NO: 2 and SEQ ID NO: 3, respectively. The peptide composition is obtained by obtaining a vacuum-dried protein purified sample after a blood sample is purified by wheat germ agglutinin, and then the protein purified sample is re-dissolved in a formic acid aqueous solution and subjected to a trypsin digestion; wherein the SEQ ID NO: 1, the SEQ ID NO: 2 and the SEQ ID NO: 3 do not have post-translational modifications; wherein the method for judging whether a test subject suffers from colorectal cancer is when the SEQ ID NO: 1 obtained after a comparison blood sample is purified by the method is equal to the SEQ ID NO: 2. When the concentration of SEQ ID NO:1 is less than the concentration of SEQ ID NO:1 obtained after the test blood sample is purified by the purification method, and the concentration of SEQ ID NO:2 obtained after the comparison blood sample is purified by the purification method is greater than the concentration of SEQ ID NO:2 obtained after the test blood sample is purified by the purification method, and the concentration of SEQ ID NO:3 obtained after the comparison blood sample is purified by the purification method is greater than the concentration of SEQ ID NO:3 obtained after the test blood sample is purified by the purification method, it is judged that the test individual may suffer from colorectal cancer; wherein the comparison blood sample is derived from an individual without colorectal cancer. The use of the peptide composition for detecting colorectal cancer as described in claim 1 is to detect the content of the peptide composition by a mass spectrometry analysis or an immunoassay. The use of a peptide composition for detecting colorectal cancer as described in claim 1, wherein the blood sample is a serum sample, a plasma sample, or a whole blood sample. A method for detecting colorectal cancer comprises: (a) detecting the content of a peptide composition obtained by a purification method in a blood sample of a test subject, wherein the peptide composition is composed of three peptide fragments, and the amino acid sequences of each peptide fragment are shown as SEQ ID NO: 1, SEQ ID NO: 2 and SEQ ID NO: 3 respectively; (b) comparing the content of the peptide composition detected by (a) with the content of the peptide composition obtained by the purification method in a comparison blood sample of a non-colorectal cancer subject, to determine whether the test subject suffers from colorectal cancer; wherein the SEQ ID NO: 1, the SEQ ID NO: 2 and the SEQ ID NO: 3 are respectively NO:3 is not modified after translation; wherein the method for determining whether the subject to be tested suffers from colorectal cancer is that when the concentration of SEQ ID NO:1 obtained after the comparison blood sample is purified by the method is less than the concentration of SEQ ID NO:1 obtained after the test blood sample is purified by the method, and the concentration of SEQ ID NO:2 obtained after the comparison blood sample is purified by the method is greater than the concentration of SEQ ID NO:2 obtained after the test blood sample is purified by the method, and the concentration of SEQ ID NO:3 obtained after the comparison blood sample is purified by the method is greater than the concentration of SEQ ID NO: NO: When the concentration is 3, it is judged that the individual to be tested may suffer from colorectal cancer; wherein the purification method is the wheat germ agglutinin purification method. A method for detecting colorectal cancer as described in claim 4, wherein the comparison method is selected from the group consisting of multivariate logistic regression, decision tree, random forest and support vector machine. A method for detecting colorectal cancer as described in claim 4, wherein the detection method is selected from a mass spectrometry analysis or an immunoassay. The method for detecting colorectal cancer as described in claim 4, wherein the blood sample to be tested and the blood sample for comparison are both serum samples, or the blood sample to be tested and the blood sample for comparison are both plasma samples, or the blood sample to be tested and the blood sample for comparison are both whole blood samples. A method for detecting colorectal cancer as described in claim 6, wherein the mass spectrometry analysis is selected from the group consisting of radical-assisted laser desorption/ionization mass spectrometry analysis, liquid chromatography-mass spectrometry analysis, liquid chromatography-tandem mass spectrometry analysis, and gas chromatography-mass spectrometry analysis.

Images