Large‐scale quantitative clinical proteomics by label‐free liquid chromatography and mass spectrometry

Ayako Negishi; Masaya Ono; Yasushi Handa; Hidenori Kato; Kohki Yamashita; Kazufumi Honda; Miki Shitashige; Reiko Satow; Tomohiro Sakuma; Hideya Kuwabara; Ken Omura; Setsuo Hirohashi; Tesshi Yamada

doi:10.1111/j.1349-7006.2008.01055.x

. 2008 Dec 22;100(3):514–519. doi: 10.1111/j.1349-7006.2008.01055.x

Large‐scale quantitative clinical proteomics by label‐free liquid chromatography and mass spectrometry

Ayako Negishi ^1,², Masaya Ono ¹, Yasushi Handa ³, Hidenori Kato ³, Kohki Yamashita ³, Kazufumi Honda ¹, Miki Shitashige ¹, Reiko Satow ¹, Tomohiro Sakuma ⁴, Hideya Kuwabara ⁴, Ken Omura ², Setsuo Hirohashi ¹, Tesshi Yamada ^1,^✉

PMCID: PMC11159643 PMID: 19154406

Abstract

We previously reported the development of an integrated proteome platform, namely 2‐Dimensional Image Converted Analysis of Liquid chromatography and mass spectrometry (2DICAL), for quantitative comparison of large peptide datasets generated by nano‐flow liquid chromatography (LC) and mass spectrometry (MS). The key technology of 2DICAL was the precise adjustment of the retention time of LC by dynamic programming. In order to apply 2DICAL to clinical studies that require comparison of a large number of patient samples we further refined the calculation algorithm and increased the accuracy and speed of the peptide peak alignment using a greedy algorithm, which had been used for fast DNA sequence alignment. The peptide peaks of each sample with the same m/z were extracted every 1 m/z and displayed with along the horizontal axis. Here we report a precise comparison of more than 150 000 typtic peptide ion peaks derived from 70 serum samples (40 patients with uterine endometrial cancer and 30 controls). The levels of 49 MS peaks were found to differ significantly between cancer patients and controls (P < 0.01, Welch's t‐test and interquartile range [IQR] of >40), and the differential expression and identification of selected three proteins was validated by immunoblotting. 2DICAL was is highly advantageous for large‐scale clinical proteomics because of its simple procedure, high throughput, and quantification accuracy. (Cancer Sci 2009; 100: 514–519)

Abbreviations:

2DICAL: 2‐dimensional image converted analysis of liquid chromatography and mass spectrometry
CC: correlation coefficient
HPLC: high‐performance liquid chromatography
IHRP: inter‐α‐trypsin inhibitor family heavy chain‐related protein
IQR: interquartile range
LC‐MS: liquid chromatography and mass spectrometry
MS/MS: tandem mass spectrometry
m/z (M/Z): mass‐to‐charge ratio
nano‐ESI: nano‐electrospray ionization
QTOF‐MS: quadruple time‐of‐flight mass spectrometry
RT: retention time

The incidence and morbidity of endometrial cancer have been rising in Western countries and Japan. In 1970, endometrial cancer constituted only 3% of all uterine cancers, but the proportion had increased to 40% by 1998.⁽ ¹ ⁾ The contemporary human lifestyle is characterized by excessive fat consumption, obesity, physical inactivity, high energy intake, hypertension, and a high serum glucose concentration, which are risk factors for endometrial cancer,⁽ ² , ³ , ⁴ , ⁵ , ⁶ ⁾ and therefore its incidence is predicted to increase further. Abnormal uterine bleeding is the most frequent initial symptom of endometrial cancer, but many other disorders including endometriosis, metroptosis, myoma uteri, and uterine sarcoma can also produce this symptom. Endometrial cancer is usually diagnosed by histological examination of endometrial tissue. However, endometrial biopsy is often associated with complications such as infection, bleeding and perforation of the uterus, and the development of an alternative safe test is therefore necessary for diagnosis of endometrial cancer.

The circulating serum proteome holds great promise as a reservoir of information that will be applicable for diagnosis of various diseases. Various gel‐based and MS‐based quantitative analyses of plasma/serum proteins have been conducted actively to identify serum biomarkers that can be applied to blood tests.⁽ ⁷ , ⁸ , ⁹ , ¹⁰ ⁾ HPLC with a flow rate of the nanoliter‐per‐minute order coupled with high‐speed MS/MS scanning has especially attracted considerable attention because of its comprehensive protein identification capacity.⁽ ¹¹ , ¹² , ¹³ ⁾ However, MS/MS is essentially not a method for quantification, and several attempts have been made to provide a quantitative dimension to nano‐LC‐MS/MS. Semi‐quantification is possible to some extent by counting the number of peptides sequenced.⁽ ¹⁴ , ¹⁵ , ¹⁶ ⁾ To achieve more accurate quantification, several labeling methods have been developed,⁽ ¹⁷ , ¹⁸ , ¹⁹ ⁾ but the number of samples that can be compared with these isotope labeling methods is limited, and the efficacy of labeling cannot be controlled from one experiment to another. Furthermore, the long‐term stability of nano‐LC‐MS can be troublesome. The sensitivity, resolution, and profile of MS instruments change over time, and more seriously the flow condition of LC may fluctuate even from sample to sample. All of these factors prevent the use of nano‐LC‐MS/MS for quantitative clinical studies in which comparison of a large number of samples is necessary.

We previously developed a labeling‐free MS‐based quantitative proteome platform known as 2DICAL.⁽ ²⁰ ⁾ Unlabeled protein samples were digested with trypsin and separated by nano‐flow (200 nL/min) HPLC. MS spectra were obtained every 1 s (3600 scans/h) by using the fast scan mode of quadrupole time‐of‐flight (QTOF)‐MS. Because we eliminated time‐consuming MS/MS in the biomarker discovery phase, 60 000–160 000 MS peaks could be detected and quantified during a 1‐h run of LC‐MS. Using 2DICAL we reported comprehensive analyses of proteins that were expressed differently in non‐clinical samples between metastatic and non‐metastatic cancer cell lines. In order to apply 2DICAL to clinical studies that require the comparison of a large number of patient samples, we further refined the calculation algorithm and improved the accuracy and speed of the peak alignment of 2DICAL (Ono et al. manuscript in preparation). In this short technical report, we describe a large‐scale quantitative comparison of the serum proteome of endometrial cancer patients using 2DICAL. This method can be applied to clinical studies of any kind, especially those in which isotope labeling is not applicable.

Materials and Methods

Clinical samples. Serum samples (n = 70) were collected at the National Hospital Organization Hokkaido Cancer Center Hospital (Sapporo, Japan) between 2000 and 2004 with the informed consent of the donors.⁽ ²¹ ⁾ The subjects included 40 patients with untreated endometrial cancer (age 54.0 ± 10.6 years) and 30 controls (50.3 ± 16.3 years) comprising healthy volunteers (n = 15) and patients with metroptosis (n = 15). The samples were collected in glass tubes, and after allowing them to clot, the serum was separated and cryopreserved at –80°C until analysis. The protocol of the study was reviewed and approved by the ethics committees of the National Hospital Organization Hokkaido Cancer Center (Sapporo, Japan) and the National Cancer Center (Tokyo, Japan).

Sample preparation. Twelve highly abundant serum proteins (albumin, IgG, α1‐antitrypsin, IgA, IgM, transferrin, haptoglobin, α1‐acid glycoprotein, α2‐macroglobulin, apolipoprotein A‐I, apolipoprotein A‐II, and fibrinogen) that constitute ~95% of the total protein mass of human serum, were depleted using an IgY‐12 spin column (Beckman Coulter, Fullerton, CA, USA) as instructed by the supplier. Briefly, a 10‐µL serum sample was diluted with 490 µL dilution buffer (1:50 dilution) and applied to an IgY‐12 spin column. After 15 min of incubation with constant rotation on an end‐to‐end rotator, the flow‐through fraction was recovered into 2‐mL microcentrifuge tubes by centrifugation at 300 g for 1 min. The half volume (250 µL) was precipitated with 250 µL of acetonitrile, dried with a SpeedVac concentrator (ThermoFisher Scientific, Waltham, MA, USA), and dissolved in 30 µL of milliQ water, to which 10 µL of 5 M urea, 2.5 µL of 1 M NH₄HCO₃, and 3.3 µg of Sequencing Grade Modified trypsin (Promega, Madison, WI, USA) were added. After 20 h of digestion at 37°C, peptides were precipitated with 50‐µL acetonitrile, dried with a SpeedVac concentrator, and then dissolved in 80 µL of 0.1% formic acid.

LC‐MS. LC‐MS and data acquisition were performed as reported previously.⁽ ²⁰ ⁾ Briefly, the peptide samples were separated in triplicate with a linear gradient from 0 to 80% acetonitrile in 0.1% formic acid at a speed of 200 nL/min for 60 min using a splitless nano‐flow HPLC system (DiNa; KYA Technologies, Tokyo, Japan, or NanoFrontier; Hitachi High‐technologies, Tokyo, Japan). MS spectra were acquired with a nano‐ESI‐QTOF‐MS (QTOF Ultima; Waters, Milford, MA, USA) every second for 60 min in the 250–1600 m/z range.

Peak display and statistical analyses. MS peaks of each sample with the same m/z were extracted every 1 m/z and aligned with the retention time of LC along the horizontal axis. A proper identification (ID) number was applied to each MS peak. Statistically significant differences were detected using Welch's t‐test and IQR. The statistical tests were performed using tools available in the R statistical package (Version 2.0.1) (http://www.r‐project.org/).

Protein identification by MS/MS. Peak lists were created using the MassNavigator software package (Version 1.2; Mitsui Knowledge Industry, Tokyo, Japan). Target MS/MS data acquisition for candidate peaks was performed on the same samples. MS/MS data were aligned with the reference data, and the target peak was extracted with a tolerance of ±0.5 m/z and ±0.4 min of RT. The MS/MS data were analyzed with the Mascot software package (Version 2.2.1; Matrix Sciences, London, UK) against the NCBInr database (NCBInr_20070419.fast) using the following parameters. Initial peptide tolerances in MS and MS/MS modes were ±0.05 Da and ±0.1 Da, respectively. Trypsin was designated as the enzyme, and up to one missed cleavage was allowed. The score threshold to achieve P < 0.05 is set by the Mascot algorithm, and is based on the size of the database used in the search.

Immunoblot analysis. Anti‐human apolipoprotein A‐IV mouse monoclonal antibody (A4‐18A3) was purchased from BML (Saitama, Japan). Anti‐complement component C3 mouse monoclonal antibody (GAU 017‐01) and anticomplement component C4A mouse monoclonal antibody (HYB 162‐02) were purchased from AntibodyShop (Gentofte, Denmark). Anti‐human α2‐macroglobulin mouse monoclonal antibody was purchased from R&D Systems (Minneapolis, MN, USA) and used as a loading control.

Serum samples were separated by sodium dodecylsulfate–polyacrylamide gel electrophoresis (SDS‐PAGE) and electroblotted onto polyvinylidene difluoride (PVDF) membranes. The blots were incubated with one of the above primary antibodies and relevant horseradish peroxidase‐conjugated secondary antibody and detected using an enhanced chemiluminescence kit.⁽ ²² ⁾

Results

High‐throughput LC‐MS analysis of clinical samples. Seventy serum samples (from 40 uterine endometrial cancer patients and 30 controls) were blinded, randomized, and measured in triplicate. We performed an average of 10.5 LC‐MS runs every day (3.5 samples/day), and the entire 210 runs were finished in a month without any serious problems. Fig. 1(a) is a plot depicting the difference of RT in the 209 runs from a reference run selected at random. The new nano‐flow HPLC system with a degassing function showed much smaller RT jittering (0.32 min on average) than our previous HPLC system (1.68 min). However, if MS spectra of particular m/z were displayed along the RT, there was still significant time jittering among runs (before alignment, Fig. 1b). However, the small RT jittering was readily adjustable (after alignment, Fig. 1b). The CC value for the entire MS peaks of triplicate runs was at least 0.96 throughout the experiment (Supporting Fig. S1a–c). We also ran a standard serum mixture at least once a week to ensure consistency of the LC‐MS during the same period (Supporting Fig. S1d).

Fine alignment of mass spectrometry (MS) peaks by retention time (RT) adjustment of 210 LC‐MS runs. (a) Comparison of RT (in the range 30.0–70.0 min) between a reference liquid chromatography and mass spectrometry (LC‐MS) run (plotted on the horizontal axis) and the other 209 LC‐MS runs (plotted on the vertical axis). Seventy serum samples (from 40 uterine endometrial cancer patients and 30 controls) were run in triplicate. One of the 210 LC‐MS runs was selected at random as the reference. MS peaks having the same m/z but different LC‐flows were selected using a greedy algorithm and aligned by adjustment of RT.⁽ ²⁰ ⁾ The differences of RT from the reference run could be assessed by the deviation from the central 45‐degree line, but no significance RT difference was apparent among the 210 LC‐MS runs. (b) 622.0 (621.5–622.5)‐m/z MS peaks of the 210 runs (70 samples in triplicate) displayed with RT (39.5–50.5 min) along the horizontal axes. Although no significant RT variation was observed in the global analysis (a), it was difficult to compare MS peaks among runs due to small variations of RT before RT alignment (top). A comparison became possible only after RT adjustment (bottom).

After RT alignment, a total of 154 992 MS peaks were detected across the 70 serum samples in the 250 m/z to 1600 m/z and 20–70 min range, and these were numbered serially from ID 1 to ID 154992. The entire MS peaks detected in a representative run are displayed in Fig. 2(a), with m/z‐values along the horizontal axis and RT along the vertical axis.

Location of the selected 49 peaks on the 2‐dimensional gel‐like view. (a) Two‐dimensional display of >150 000 mass spectrometry (MS) peaks of a representative sample with the m/z‐values (250–1600 m/z) along the horizontal axis and retention time (RT) (20.0–70.0 min) along the vertical axis. 49 MS peaks whose mean intensity differed significantly between 40 endometrial cancer patients and 30 controls (*P <* 0.01, Welch's t‐test and interquartile range (IQR) of >40) are highlighted in red. (b, c) A representative peak differentially expressed between a healthy control (U301) (b) and an endometrial cancer patient (U078) (c) is highlighted by red arrows.

Difference between uterine endometrial cancer patients and controls. After normalizing to the total ion intensity of all of the peaks, the maximum intensity of each peak was averaged. To eliminate peaks that had little or no variation across samples (peaks that were not working well), we selected 1043 peaks whose intensity had an IQR of >40 across the 70 samples. There were 49 MS peaks showing a statistically significant difference between the cancer patients and controls (P < 0.01, Welch's t‐test) (highlighted in red, Fig. 2a). Fig. 2(b,c) depicts a representative MS peak (indicated by an arrow) differentially expressed between a cancer patient (Fig. 2b) and a control (Fig. 2c).

Validation of differential expression. The 49 MS peaks were subjected to MS/MS for protein identification. A database search revealed that the MS/MS spectra of 11 peaks matched the amino acid sequences of proteins deposited in the NCBInr database (NCBInr_20070419.fast) with significant confidence (P < 0.05), including apolipoprotein A‐IV, IHRP, and complement components C3, C4A, and C4B (Table 1 and Supporting Figs. S2–5). It is noteworthy that apolipoprotein A‐IV, complement component C4A, complement component C3, and IHRP were repeatedly detected two, four, two, and two times, respectively (Table 1). The identification and differential expression were confirmed by immunoblotting using sera from five endometrial cancer patients and five healthy controls (Fig. 3a). Apolipoprotein A‐IV was down‐regulated in the cancer patients, whereas complement components C4A and C3 were up‐regulated, confirming the results of 2DICAL (Fig. 3b).

Table 1.

Detected peptides that differed between patients with uterine endometrial cancer and controls

ID	M/Z	Charge	Accession number	Protein description	Peptide sequence	Mascot score	Average peak intensity		P‐values^*
ID	M/Z	Charge	Accession number	Protein description	Peptide sequence	Mascot score	Cancer (n = 40)	Control (n = 30)	P‐values^*
8627	552	2	gi\|178757	Apolipoprotein A‐IV precursor	IDQTVEELR	76.09	101.3 ± 41	138.1 ± 58.6	0.005815
9150	558	2	gi\|179674	Complement component C4A	VGDTLNLNLR	51.60	204.4 ± 102.6	142.3 ± 55.6	0.002147
10500	573	2	gi\|1041907	GP120, IHRP = ITI heavy chain‐related protein {internal fragment}	GPDVLTATVSGK	51.33	131.3 ± 68.6	88.7 ± 44.8	0.002911
13516	605	2	gi\|179665	Complement component C3	YYTYLIMNK	40.95	201.1 ± 87.2	157.4 ± 45.8	0.009564
14166	611	2	gi\|1483187	Inter‐α‐trypsin inhibitor family heavy chain‐related protein (IHRP)	ETLFSVMPGLK	49.85	155.0 ± 50.1	123.4 ± 31.1	0.002124
14947	619	2	gi\|179674	Complement component C4A	KYVLPNFEVK	37.50	90.6 ± 42.0	68.7 ± 21.0	0.006617
17519	644	2	gi\|178757	Apolipoprotein A‐IV precursor	TQVNTQAEQLR	59.77	71.5 ± 44.9	127.9 ± 81.3	0.001617
25565	715	3	gi\|179674	Complement component C4A	MRPSTDTITVMVENSHGLR	48.54	301.4 ± 147.6	225.7 ± 77.1	0.008078
25827	717	2	gi\|87191	Complement C4B‐human (fragment)	AEMADQAAAWLTR	47.75	134.2 ± 83.9	93.5 ± 38.6	0.009756
28122	736	2	gi\|179665	Complement component C3	IPIEDGSGEVVLSR	74.91	582.2 ± 293.0	405.6 ± 185.3	0.003440
37184	814	2	gi\|179674	Complement component C4A	EELVYELNPLDHR	89.65	477.2 ± 228.5	348.2 ± 122.3	0.003940

Open in a new tab

Welch's t‐test.

Confirmation of differential expression. (a) Confirmation of differential expression of apolipoprotein A‐IV, complement component C4A, and complement component C3 in serum samples from five endometrial cancer patients and five controls by immunoblotting. (b) Gel‐like view with retention time (RT) along the horizontal axis (top) and distribution of peak intensity (bottom) of apolipoprotein A‐IV (left), complement component C4A (middle), and complement component C3 (right) in 40 cancer patients (red) and 30 controls (blue).

Discussion

The entire protein content (or proteome) of each human sample varies significantly from individual to individual, and 100 biomarkers with P‐values of <0.01 might be “discovered” in a large data set consisting of 10 000 variables. The accidental identification of mere personal heterogeneity as a biomarker candidate should be avoided by comparing a sufficient number of cases and controls. “Biomarkers” identified in a small sample set rarely pass through the pipeline connecting the biomarker discovery phase and the preclinical phase. According to our calculations, when the experiment involves a comparison of 10 cases and 10 controls, a factor that has a 90% discrimination rate can have a confidence interval that ranges from 64% to 98%, and comparison of at least 40 cases and 40 controls would be necessary to discover a biomarker with a diagnostic accuracy of >90% even with the use of high generalization capability algorithms such as support vector machine (Kuwabara et al. manuscript in preparation).

In order to apply 2DICAL to clinical studies that require comparison of such a scale of patient samples, we refined the calculation algorithm and increased the speed and accuracy of the MS peak alignment. We exploited the fact that the order of appearance of the peptide peaks is constant even though the LC RT fluctuates. We first selected MS peaks having the same m/z every 1 m/z and displayed their MS spectra along the LC RT. Peptide peaks having the same m/z but different LC‐flows were aligned by adjusting RT, and the corresponding peaks were selected using a greedy algorithm that had been used for fast DNA sequence alignment. Using the new algorithm we compared over 150 000 MS peaks of 40 uterine cancer patients and 30 controls by aligning more than 10 000 000 data points, and identified 49 MS peaks whose intensity differed significantly between cancer patients and controls (P < 0.01, Welch's t‐test and IQR of >40). We performed MS/MS analysis for protein identification (Supporting Figs. S2–5), and the differential expression of representative proteins was confirmed by immunoblotting (Fig. 3a).

Direct mass spectrometric analysis of plasma/serum samples is considered to be most immediate to the discovery of protein biomarkers of various diseases, but it has been hampered by the marked dominance of a handful of particularly abundant plasma/serum proteins, including albumin, immunoglobulins, and transferrins, and thus removal of these abundant proteins and reduction of sample complexity are necessary for comprehensive protein profiling.⁽ ²³ , ²⁴ ⁾ The highly abundant proteins present in quantities of mg/mL make up less than 0.1% of the total number of proteins, yet they constitute more than 95% of the bulk mass of total protein in serum. Proteins of interest as potential biomarkers for malignant or non‐malignant diseases (e.g. prostate‐specific antigen, various interleukins and cytokines) are usually present in plasma/serum only at ng/mL to pg/mL levels. Because the dynamic range of MS is limited, it is difficult to detect these biomarker candidates among highly abundant proteins. We used an Ig‐Y immunoaffinity column to increase the concentration of serum proteins with lower abundance. This procedure removed 12 highly abundant proteins with satisfactorily high reproducibility.⁽ ²⁵ ⁾ However, only proteins with relatively high abundance such as apolipoprotein A‐IV and complement component C4A (Table 1), which have been implicated in human malignancies other than uterine endometrial cancer,⁽ ²⁶ , ²⁷ ⁾ were identified, indicating that our LC‐MS is still not looking deeply enough into candidate biomarker proteins with low abundance. It will be necessary to develop a more comprehensive prefractionation method for plasma/serum samples to exploit the full capacity of our method.

Although high‐quality MS/MS spectra were obtained from nearly all peptides (data not shown), MS/MS data for only 11 peptides matched the amino acid sequences of proteins deposited in the database with high confidence (P < 0.05). The Human Proteome Organization (HUPO) recently completed the first large‐scale collaborative study to characterize the human serum and plasma proteomes. Although 9504 proteins identified with one or more peptides, and 3020 proteins identified with two or more peptides, only 889 proteins were identified with a confidence level of at least 95%. The high rate of false identification may be attributable not only to novel exons in alternatively spliced variants of known proteins or previously nonannotated gene sequences,⁽ ²⁸ ⁾ but also to post‐translational modification of proteins. Plasma/serum biomarker candidates have been reported to be invariably proteins modified with aberrant glycosylation, cleavage, or dimerization.⁽ ²⁹ , ³⁰ , ³¹ ⁾ A low frequency of completely matched peptide sequences in the database is inevitable at this point. However, the reproducible quantification of marker peptides by MS could be directly applicable to clinical use without the need for actual protein identification.

We have demonstrated that large‐scale quantitative proteomic comparison is readily possible using this new version of 2DICAL. 2DICAL is a quantitative proteome platform characterized by its simplicity and throughput and is now applicable to any kinds of large‐scale biomarker discovery studies.

Supporting information

Fig. S1. Reproducibility of 2‐Dimensional Image Converted Analysis of Liquid chromatography and mass spectrometry (2DICAL). (a–c) Two‐dimensional plots indicating the high reproducibility of mass spectrometry (MS) intensity (in log scale) between corresponding >150 000 peaks of a representative sample. The sample was run three times (Runs 1, 2, and 3). The average correlation coefficient (CC) values were 0.99 between Runs 1 and 2 (a), 0.96 between Runs 1 and 3 (b), and 0.97 between Runs 2 and 3 (c). (d) The median intensity of the entire MS peaks (>150 000 peaks) of triplicates was compared between two different days one week apart. The average CC value was 0.96.

Figs. S2–5. High‐speed tandem mass spectrometry (MS/MS) analysis. Labeled MS/MS spectra and peak lists of ID 17519, ID 37148, ID 28122, and ID 14166, which matched the sequences of apolipoprotein A‐IV, complement component C4A, complement component C3, and inter‐α‐trypsin inhibitor family heavy chain‐related protein (IHRP), respectively.

Please note: Wiley‐Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

Supporting info item

CAS-100-514-s001.ppt^{(1.2MB, ppt)}

Acknowledgments

We thank Ms. Ayako Igarashi for her excellent technical assistance. This work was supported by the ‘Program for Promotion of Fundamental Studies in Health Sciences’ conducted by the National Institute of Biomedical Innovation of Japan; the ‘Third‐Term Comprehensive Control Research for Cancer’ conducted by the Ministry of Health, Labor, and Welfare of Japan; and generous grants from the Naito Foundation, the Princess Takamatsu Cancer Research Fund, and the Foundation for the Promotion of Cancer Research.

References

1. Kuwabara Y, Susumu N, Banno K et al . Clinical characteristics of prognostic factors in poorly differentiated (G3) endometrioid adenocarcinoma in Japan. Jpn J Clin Oncol 2005; 35: 23–7. [DOI] [PubMed] [Google Scholar]
2. Armstrong B, Doll R. Environmental factors and cancer incidence and mortality in different countries, with special reference to dietary practices. Int J Cancer 1975; 15: 617–31. [DOI] [PubMed] [Google Scholar]
3. Zeleniuch‐Jacquotte A, Akhmedkhanov A, Kato I et al . Postmenopausal endogenous oestrogens and risk of endometrial cancer: results of a prospective study. Br J Cancer 2001; 84: 975–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
4. Anderson KE, Anderson E, Mink PJ et al . Diabetes and endometrial cancer in the Iowa women's health study. Cancer Epidemiol Biomarkers Prev 2001; 10: 611–16. [PubMed] [Google Scholar]
5. Calle EE, Rodriguez C, Walker‐Thurmond K, Thun MJ. Overweight, obesity, and mortality from cancer in a prospectively studied cohort of U.S. adults. N Engl J Med 2003; 348: 1625–38. [DOI] [PubMed] [Google Scholar]
6. Furberg AS, Thune I. Metabolic abnormalities (hypertension, hyperglycemia and overweight), lifestyle (high energy intake and physical inactivity) and endometrial cancer risk in a Norwegian cohort. Int J Cancer 2003; 104: 669–76. [DOI] [PubMed] [Google Scholar]
7. Rai AJ, Zhang Z, Rosenzweig J et al . Proteomic approaches to tumor marker discovery. Arch Pathol Lab Med 2002; 126: 1518–26. [DOI] [PubMed] [Google Scholar]
8. Petricoin EF, Ardekani AM, Hitt BA et al . Use of proteomic patterns in serum to identify ovarian cancer. Lancet 2002; 359: 572–7. [DOI] [PubMed] [Google Scholar]
9. Honda K, Hayashida Y, Umaki T et al . Possible detection of pancreatic cancer by plasma protein profiling. Cancer Res 2005; 65: 10613–22. [DOI] [PubMed] [Google Scholar]
10. Kakisaka T, Kondo T, Okano T et al . Plasma proteomics of pancreatic cancer patients by multi‐dimensional liquid chromatography and two‐dimensional difference gel electrophoresis (2D‐DIGE): up‐regulation of leucine‐rich α‐2‐glycoprotein in pancreatic cancer. J Chromatogr B Anal Technol Biomed Life Sci 2007; 852: 257–67. [DOI] [PMC free article] [PubMed] [Google Scholar]
11. Aebersold R, Mann M. Mass spectrometry‐based proteomics. Nature 2003; 422: 198–207. [DOI] [PubMed] [Google Scholar]
12. Takahashi N, Kaji H, Yanagida M, Hayano T, Isobe T. Proteomics: advanced technology for the analysis of cellular function. J Nutr 2003; 133: 2090S–6S. [DOI] [PubMed] [Google Scholar]
13. Ishihama Y. Proteomic LC‐MS systems using nanoscale liquid chromatography with tandem mass spectrometry. J Chromatogr A 2005; 1067: 73–83. [DOI] [PubMed] [Google Scholar]
14. Gao J, Opiteck GJ, Friedrichs MS, Dongre AR, Hefta SA. Changes in the protein expression of yeast as a function of carbon source. J Proteome Res 2003; 2: 643–9. [DOI] [PubMed] [Google Scholar]
15. Liu H, Sadygov RG, Yates JR 3rd. A model for random sampling and estimation of relative protein abundance in shotgun proteomics. Anal Chem 2004; 76: 4193–201. [DOI] [PubMed] [Google Scholar]
16. Ishihama Y, Oda Y, Tabata T et al . Exponentially modified protein abundance index (emPAI) for estimation of absolute protein amount in proteomics by the number of sequenced peptides per protein. Mol Cell Proteomics 2005; 4: 1265–72. [DOI] [PubMed] [Google Scholar]
17. Gygi SP, Rist B, Gerber SA, Turecek F, Gelb MH, Aebersold R. Quantitative analysis of complex protein mixtures using isotope‐coded affinity tags. Nat Biotechnol 1999; 17: 994–9. [DOI] [PubMed] [Google Scholar]
18. Ong SE, Blagoev B, Kratchmarova I et al . Stable isotope labeling by amino acids in cell culture, SILAC, as a simple and accurate approach to expression proteomics. Mol Cell Proteomics 2002; 1: 376–86. [DOI] [PubMed] [Google Scholar]
19. DeSouza L, Diehl G, Rodrigues MJ et al . Search for cancer markers from endometrial tissues using differentially labeled tags iTRAQ and cICAT with multidimensional liquid chromatography and tandem mass spectrometry. J Proteome Res 2005; 4: 377–86. [DOI] [PubMed] [Google Scholar]
20. Ono M, Shitashige M, Honda K et al . Label‐free quantitative proteomics using large peptide data sets generated by nanoflow liquid chromatography and mass spectrometry. Mol Cell Proteomics 2006; 5: 1338–47. [DOI] [PubMed] [Google Scholar]
21. Kikuchi S, Honda K, Yamada T. Serumalbmin‐associated peptides of patients with uterine endometrial cancer. Cancer Sci 2007; 98: 822–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
22. Idogawa M, Yamada T, Honda K, Sato S, Imai K, Hirohashi S. Poly (ADP‐ribose) polymerase‐1 is a component of the oncogenic T‐cell factor‐4/β‐catenin complex. Gastroenterology 2005; 128: 1919–36. [DOI] [PubMed] [Google Scholar]
23. Anderson NL, Anderson NG. The human plasma proteome: history, character, and diagnostic prospects. Mol Cell Proteomics 2002; 1: 845–67. [DOI] [PubMed] [Google Scholar]
24. Tirumalai RS, Chan KC, Prieto DA, Issaq HJ, Conrads TP, Veenstra TD. Characterization of the low molecular weight human serum proteome. Mol Cell Proteomics 2003; 2: 1096–103. [DOI] [PubMed] [Google Scholar]
25. Liu T, Qian WJ, Mottaz HM et al . Evaluation of multiprotein immunoaffinity subtraction for plasma proteomics and candidate biomarker discovery using mass spectrometry. Mol Cell Proteomics 2006; 5: 2167–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
26. Kawakami T, Hoshida Y, Kanai F et al . Proteomic analysis of sera from hepatocellular carcinoma patients after radiofrequency ablation treatment. Proteomics 2005; 5: 4287–95. [DOI] [PubMed] [Google Scholar]
27. Miguet L, Bogumil R, Decloquement P et al . Discovery and identification of potential biomarkers in a prospective study of chronic lymphoid malignancies using SELDI‐TOF‐MS. J Proteome Res 2006; 5: 2258–69. [DOI] [PubMed] [Google Scholar]
28. States DJ, Omenn GS, Blackwell TW et al . Challenges in deriving high‐confidence protein identifications from data gathered by a HUPO plasma proteome collaborative study. Nat Biotechnol 2006; 24: 333–8. [DOI] [PubMed] [Google Scholar]
29. Hakomori S. Tumor malignancy defined by aberrant glycosylation and sphingo (glyco) lipid metabolism. Cancer Res 1996; 56: 5309–18. [PubMed] [Google Scholar]
30. Ehmann M, Felix K, Hartmann D et al . Identification of potential markers for the detection of pancreatic cancer through comparative serum protein expression profiling. Pancreas 2007; 34: 205–14. [DOI] [PubMed] [Google Scholar]
31. Villanueva J, Shaffer DR, Philip J et al . Differential exoprotease activities confer tumor‐specific serum peptidome patterns. J Clin Invest 2006; 116: 271–84. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supporting info item

CAS-100-514-s001.ppt^{(1.2MB, ppt)}

[b1] 1. Kuwabara Y, Susumu N, Banno K et al . Clinical characteristics of prognostic factors in poorly differentiated (G3) endometrioid adenocarcinoma in Japan. Jpn J Clin Oncol 2005; 35: 23–7. [DOI] [PubMed] [Google Scholar]

[b2] 2. Armstrong B, Doll R. Environmental factors and cancer incidence and mortality in different countries, with special reference to dietary practices. Int J Cancer 1975; 15: 617–31. [DOI] [PubMed] [Google Scholar]

[b3] 3. Zeleniuch‐Jacquotte A, Akhmedkhanov A, Kato I et al . Postmenopausal endogenous oestrogens and risk of endometrial cancer: results of a prospective study. Br J Cancer 2001; 84: 975–81. [DOI] [PMC free article] [PubMed] [Google Scholar]

[b4] 4. Anderson KE, Anderson E, Mink PJ et al . Diabetes and endometrial cancer in the Iowa women's health study. Cancer Epidemiol Biomarkers Prev 2001; 10: 611–16. [PubMed] [Google Scholar]

[b5] 5. Calle EE, Rodriguez C, Walker‐Thurmond K, Thun MJ. Overweight, obesity, and mortality from cancer in a prospectively studied cohort of U.S. adults. N Engl J Med 2003; 348: 1625–38. [DOI] [PubMed] [Google Scholar]

[b6] 6. Furberg AS, Thune I. Metabolic abnormalities (hypertension, hyperglycemia and overweight), lifestyle (high energy intake and physical inactivity) and endometrial cancer risk in a Norwegian cohort. Int J Cancer 2003; 104: 669–76. [DOI] [PubMed] [Google Scholar]

[b7] 7. Rai AJ, Zhang Z, Rosenzweig J et al . Proteomic approaches to tumor marker discovery. Arch Pathol Lab Med 2002; 126: 1518–26. [DOI] [PubMed] [Google Scholar]

[b8] 8. Petricoin EF, Ardekani AM, Hitt BA et al . Use of proteomic patterns in serum to identify ovarian cancer. Lancet 2002; 359: 572–7. [DOI] [PubMed] [Google Scholar]

[b9] 9. Honda K, Hayashida Y, Umaki T et al . Possible detection of pancreatic cancer by plasma protein profiling. Cancer Res 2005; 65: 10613–22. [DOI] [PubMed] [Google Scholar]

[b10] 10. Kakisaka T, Kondo T, Okano T et al . Plasma proteomics of pancreatic cancer patients by multi‐dimensional liquid chromatography and two‐dimensional difference gel electrophoresis (2D‐DIGE): up‐regulation of leucine‐rich α‐2‐glycoprotein in pancreatic cancer. J Chromatogr B Anal Technol Biomed Life Sci 2007; 852: 257–67. [DOI] [PMC free article] [PubMed] [Google Scholar]

[b11] 11. Aebersold R, Mann M. Mass spectrometry‐based proteomics. Nature 2003; 422: 198–207. [DOI] [PubMed] [Google Scholar]

[b12] 12. Takahashi N, Kaji H, Yanagida M, Hayano T, Isobe T. Proteomics: advanced technology for the analysis of cellular function. J Nutr 2003; 133: 2090S–6S. [DOI] [PubMed] [Google Scholar]

[b13] 13. Ishihama Y. Proteomic LC‐MS systems using nanoscale liquid chromatography with tandem mass spectrometry. J Chromatogr A 2005; 1067: 73–83. [DOI] [PubMed] [Google Scholar]

[b14] 14. Gao J, Opiteck GJ, Friedrichs MS, Dongre AR, Hefta SA. Changes in the protein expression of yeast as a function of carbon source. J Proteome Res 2003; 2: 643–9. [DOI] [PubMed] [Google Scholar]

[b15] 15. Liu H, Sadygov RG, Yates JR 3rd. A model for random sampling and estimation of relative protein abundance in shotgun proteomics. Anal Chem 2004; 76: 4193–201. [DOI] [PubMed] [Google Scholar]

[b16] 16. Ishihama Y, Oda Y, Tabata T et al . Exponentially modified protein abundance index (emPAI) for estimation of absolute protein amount in proteomics by the number of sequenced peptides per protein. Mol Cell Proteomics 2005; 4: 1265–72. [DOI] [PubMed] [Google Scholar]

[b17] 17. Gygi SP, Rist B, Gerber SA, Turecek F, Gelb MH, Aebersold R. Quantitative analysis of complex protein mixtures using isotope‐coded affinity tags. Nat Biotechnol 1999; 17: 994–9. [DOI] [PubMed] [Google Scholar]

[b18] 18. Ong SE, Blagoev B, Kratchmarova I et al . Stable isotope labeling by amino acids in cell culture, SILAC, as a simple and accurate approach to expression proteomics. Mol Cell Proteomics 2002; 1: 376–86. [DOI] [PubMed] [Google Scholar]

[b19] 19. DeSouza L, Diehl G, Rodrigues MJ et al . Search for cancer markers from endometrial tissues using differentially labeled tags iTRAQ and cICAT with multidimensional liquid chromatography and tandem mass spectrometry. J Proteome Res 2005; 4: 377–86. [DOI] [PubMed] [Google Scholar]

[b20] 20. Ono M, Shitashige M, Honda K et al . Label‐free quantitative proteomics using large peptide data sets generated by nanoflow liquid chromatography and mass spectrometry. Mol Cell Proteomics 2006; 5: 1338–47. [DOI] [PubMed] [Google Scholar]

[b21] 21. Kikuchi S, Honda K, Yamada T. Serumalbmin‐associated peptides of patients with uterine endometrial cancer. Cancer Sci 2007; 98: 822–9. [DOI] [PMC free article] [PubMed] [Google Scholar]

[b22] 22. Idogawa M, Yamada T, Honda K, Sato S, Imai K, Hirohashi S. Poly (ADP‐ribose) polymerase‐1 is a component of the oncogenic T‐cell factor‐4/β‐catenin complex. Gastroenterology 2005; 128: 1919–36. [DOI] [PubMed] [Google Scholar]

[b23] 23. Anderson NL, Anderson NG. The human plasma proteome: history, character, and diagnostic prospects. Mol Cell Proteomics 2002; 1: 845–67. [DOI] [PubMed] [Google Scholar]

[b24] 24. Tirumalai RS, Chan KC, Prieto DA, Issaq HJ, Conrads TP, Veenstra TD. Characterization of the low molecular weight human serum proteome. Mol Cell Proteomics 2003; 2: 1096–103. [DOI] [PubMed] [Google Scholar]

[b25] 25. Liu T, Qian WJ, Mottaz HM et al . Evaluation of multiprotein immunoaffinity subtraction for plasma proteomics and candidate biomarker discovery using mass spectrometry. Mol Cell Proteomics 2006; 5: 2167–74. [DOI] [PMC free article] [PubMed] [Google Scholar]

[b26] 26. Kawakami T, Hoshida Y, Kanai F et al . Proteomic analysis of sera from hepatocellular carcinoma patients after radiofrequency ablation treatment. Proteomics 2005; 5: 4287–95. [DOI] [PubMed] [Google Scholar]

[b27] 27. Miguet L, Bogumil R, Decloquement P et al . Discovery and identification of potential biomarkers in a prospective study of chronic lymphoid malignancies using SELDI‐TOF‐MS. J Proteome Res 2006; 5: 2258–69. [DOI] [PubMed] [Google Scholar]

[b28] 28. States DJ, Omenn GS, Blackwell TW et al . Challenges in deriving high‐confidence protein identifications from data gathered by a HUPO plasma proteome collaborative study. Nat Biotechnol 2006; 24: 333–8. [DOI] [PubMed] [Google Scholar]

[b29] 29. Hakomori S. Tumor malignancy defined by aberrant glycosylation and sphingo (glyco) lipid metabolism. Cancer Res 1996; 56: 5309–18. [PubMed] [Google Scholar]

[b30] 30. Ehmann M, Felix K, Hartmann D et al . Identification of potential markers for the detection of pancreatic cancer through comparative serum protein expression profiling. Pancreas 2007; 34: 205–14. [DOI] [PubMed] [Google Scholar]

[b31] 31. Villanueva J, Shaffer DR, Philip J et al . Differential exoprotease activities confer tumor‐specific serum peptidome patterns. J Clin Invest 2006; 116: 271–84. [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Large‐scale quantitative clinical proteomics by label‐free liquid chromatography and mass spectrometry

Ayako Negishi

Masaya Ono

Yasushi Handa

Hidenori Kato

Kohki Yamashita

Kazufumi Honda

Miki Shitashige

Reiko Satow

Tomohiro Sakuma

Hideya Kuwabara

Ken Omura

Setsuo Hirohashi

Tesshi Yamada