Proteome is a functional translation of the genome, directly regulating cancer phenotypes. Cancer proteomics will further our understanding of cancer biology, and lead to novel therapeutic strategies. The project is conducting cancer proteomics, the research goals of which are 1) a comprehensive understanding of the proteomic background of cancers, 2) identification of key proteins regulating the clinico-pathological features of cancer, and 3) the development of biomarkers for personalized medicine. The malignant tumors under proteomic investigation in the project include lung cancer, liver cancer, esophageal cancer, colon cancer, soft-tissue sarcomas (gastrointestinal stromal tumor, osteosarcoma, Ewing's sarcoma, and synovial sarcoma), and malignant mesothelioma.

The poor prognosis of patients with hepatocellular carcinoma (HCC) is attributed to intrahepatic recurrence. To understand the molecular background of early intrahepatic recurrence, comparison of the protein expression profiles of primary HCC tissues was conducted between 12 patients who showed intrahepatic recurrence within six months post surgery and 15 patients who showed no recurrence until two years post surgery. Two-dimensional difference gel electrophoresis (2D-DIGE) identified 23 protein spots, the intensities of which were highly associated with early intrahepatic recurrence. The prediction performance of the 23 identified proteins was validated using additional HCC tissues from 13 HCC patients (six with early intrahepatic recurrence and seven without recurrence). All but one of the 13 patients were grouped according to their recurrence status based on the intensity of the 23 protein spots. Mass spectrometry identified 23 proteins corresponding to the spots. Although 13 of 23 have been previously reported to be correlated with HCC, their association with early intrahepatic recurrence had not been established. The identified proteins may be candidates for prognostic markers and contribute to improvement of the existing therapeutic strategies.

Prediction of the response to treatment would allow personalized medicine. To identify predictive biomarker candidates, the protein expression profiles of primary tumors of lung adenocarcinoma were compared among the patients developing recurrence after curative operation and then treated with gefitinib, an anti-cancer drug for non-small cell lung carcinoma. In the comparison between the 31 responders and 16 nonresponders, a support vector machine algorithm selected nine proteins that best distinguished the two groups. The prediction performance of the selected nine proteins was validated by an external sample set including six responders and eight nonresponders, yielding positive and negative predictive values of 100% (six of six) and 87.5% (seven of eight), respectively. The differential expression of one of the nine proteins, the heart-type fatty acid-binding protein, was successfully validated by ELISA. Study of these proteins may contribute to the development of personalized therapy for lung cancer patients.

Pancreatic cancer is one of the most deadly of malignancies, and a biomarker for early diagnosis has long been awaited. To identify diagnostic plasma biomarker candidates, we investigated the plasma proteome of pancreatic cancer patients. High-abundance plasma proteins (albumin, transferrin, haptoglobin, alpha-1-antitrypsin, IgG and IgA) were depleted by an immuno-affinity column, and low-abundance ones were separated into five fractions by anion-exchange chromatography. The fractionated plasma proteins were subjected to 2D-DIGE with highly sensitive fluorescent dyes. The quantitative protein expression profiles obtained by 2D-DIGE were compared between two plasma protein mixtures, one from five non-cancer bearing healthy donors and the other from five patients with pancreatic cancer. Among the 1200 protein spots observed, 33 protein spots were differentially expressed between the two mixtures; 27 of these were upregulated and six were downregulated in cancer. Mass spectrometry and database searching allowed the identification of the proteins corresponding to the gel spots. Upregulation of leucine-rich alpha-2-glycoprotein (LRG), which has not previously been implicated in pancreatic cancer, was validated by Western blotting in an independent series of plasma samples obtained from healthy controls, patients with chronic pancreatitis, and patients with pancreatic cancer. Our results demonstrate the applicability of a combination of multi-dimensional liquid chromatography with 2D-DIGE to plasma proteomics and suggest the clinical utility of plasma LRG measurement.

A proteome database will facilitate comprehensive proteomic study across different types of malignancies. Certain malignant tumors of different tissue origin share identical proteomic aberrations, and in contrast, certain protein groups show the unique expression patterns depending on the type of malignant tumors. An integrative proteome database will identify the proteomic aberrations underlying the general and unique proteomic backgrounds of malignant tumors. A proteome database would also be useful for biomarker studies in the validation phase. The proteome database is a part of the Genome Medicine Database of Japan (GeMDBJ), and is thus named GeMDBJ Proteomics. A database software was constructed to deposit 2D-DIGE and mass spectrometry data. A beta-version of GeMDBJ Proteomics was published in the website http://gemdbj.nibio.go.jp. The published proteome data include 1) 2D-DIGE data of nine pancreatic cancer cell lines and two normal pancreatic duct cell lines, 2) annotation for approximately 1,100 protein spots by mass spectrometry. The proteomics system developed at our laboratory, which was used to generate the proteomic data in GeMDBJ Proteomics, has also been published ⁽¹⁵⁴⁾. The proteome data of tumor tissues in our project will be published in GeMDBJ Proteomics in the near future.