|
||||||||||||||||||||||||
Dear ICBC Team Members:
We are continuing to make great progress within the consortium. CPAS 2.0 is now available and includes several major updates, and the Singapore team has released a tool for the ICBC community to share candidate lists. As a follow-up to successful regional workshops last year, Dr. Myeong-Hee Yu and Prof. Eunok Paek, of the Functional Proteomics Center, are organizing a workshop on Bioinformatics, October 5-6, 2007 at Korea University.
The scientific update for this newsletter was provided by Drs. Julian Watts and Ruedi Aebersold on an integrated, multi-tracked approach to the identification of novel candidate breast cancer biomarkers. In addition, we have included a link to a long list of recent publications from Consortium members.
The team highlight this issue is the ovarian cancer team led by Dr. Nicole Urban.
Best regards,
Lee Hartwell
President and Director
Fred Hutchinson Cancer Research Center
An integrated, multi-tracked approach to the identification of novel candidate markers for breast cancer.
Julian Watts¹ and Ruedi Aebersold¹,². ¹Institute for Systems Biology, Seattle and ²Swiss Federal Institute of Technology (ETH) Zurich and Faculty of Natural Sciences, University of Zurich.
As the burden of cancer and other diseases grows worldwide, early detection is becoming become an increasingly important strategy for intervention and treatment, as well as for economic reasons. Among the markers with the highest potential for alleviating suffering of patients are plasma protein biomarkers that indicate changes in the health status of patients in a minimally invasive manner. Unfortunately, to date, discovery-based proteomic approaches to identify novel candidate markers in a sample as complex as plasma have proven largely inadequate. Therefore we believe there is a need to develop and implement alternative proteomic workflows aimed at such problems in order to gain the necessary improvement in performance.
Our approach to this problem has been to take a step back from mass spectrometry (MS)-based marker discovery from blood plasma and instead rely on more sensitive and well established MS approaches for directed analysis of known targets in complex samples. This concept has been around for a while and is often referred to as a 'targeted' or 'candidate-based' approach to marker discovery. In broad terms, armed with a list of known targets (in our case, peptides derived from candidate proteins), utilization of a tandem MS (MS/MS) protocol called multiple reaction monitoring (MRM) allows for identification of the targets, if present, in a complex mixture at a level of sensitivity several orders of magnitude higher than possible in a standard discovery-based MS/MS workflow. The problem with such an MRM-based approach is that you need prior knowledge of your candidates, which is why earlier research efforts have mostly followed the discovery-based approach. However, a couple of recent developments have made the possibility of shifting over to an MRM-based approach for candidate marker identification possible.
Firstly, recent improvements in triple quadrupole spectrometers (the hardware architecture required to perform an MRM experiment) have made it possible to profile hundreds of different target proteins at high sensitivity and specificity in a single analysis and in a relatively high throughput fashion. This now makes it possible to screen the hundreds of plasma samples that would come out of population-based or longitudinal disease studies for the same several hundred proteins in an acceptable time frame. Secondly, the explosion of publicly available biological data and databases, and tools for their interrogation, now provide powerful resources that enable data mining for candidate markers not identified via our experimental workflows. Additionally, such data resources also facilitate the parsing of our own cancer-specific datasets against them to better determine which candidates are more 'interesting' and thus potentially better targets for MRM-based profiling studies.
In order to take advantage of these new tools and resources, we are now focusing our research efforts in two primary ways. First, we are concentrating our cancer candidate marker identification efforts specifically on breast cancer, which will allow us to conduct our searches and screens far more thoroughly, hopefully improving our chances of discovering truly useful markers. Second, we are implementing a three-pronged, yet integrated approach to the determination of the best breast cancer candidate makers worthy of being taken forward to the subsequent confirmation and validation stages of the process. This new workflow is summarized in Figure 1.
|
Our first track for candidate breast cancer marker protein identification is focused along the same lines as we have generally been pursuing such efforts recently, i.e. the comparison of N-glycosylated proteins enriched from both affected and unaffected breast tissue. Protein identifications are ultimately made via MS/MS analysis of their N-glycosylation site-containing tryptic peptides (1,2). The yielding of this additional information, i.e. the N-glycosites for each protein is also useful since it is known that aberrant protein glycosylation has also been linked to cancer. Finally, restricting our analyses to N-glycosylation sites not only helps reduce sample complexity, but likely enriches for candidate plasma marker proteins for cancer since glycosylated proteins are typically extracellular (thus more likely to be deposited in blood), are frequently differentially expressed as a result of cell transformation, and from the observation that most known and approved cancer markers are glycoproteins.
The second track for candidate breast cancer marker protein identification is to similarly profile cell surface N-glycoproteins on various known and otherwise well-characterized cell culture models of breast cancer. Initial studies are focusing on the estrogen receptor (ER) positive cell lines MCF-7 and T-47D, the ER negative line HS 578T and the Her2 positive line SK-BR-3. These will allow us to measure alterations in N-glycoprotein expression and secretion in response to various cancer relevant treatments, such as response to estrogen or Herceptin (an approved chemotherapy drug for breast cancer) as well as over time. Expansion of this workflow could additionally include comparison of cell lines representing different stages of cancer (e.g. invasive vs. non-invasive). Finally, cell culture models are also amenable to in vivo stable isotope labeling (SILAC) that allow for controlled quantitative proteomic measurements of changes in protein abundance, an experiment which is difficult at best to perform in tissue, especially tissue that is limited in supply.
The third track for candidate breast cancer marker protein identification is purely an in silico workflow. Here we are using informatic approaches to mine available public data sources, such as literature databases and micro array data that pertain to cancer in general, as well as breast cancer specifically (such as from the Harvard Institute of Proteomics breast cancer 1000 project). Such mined datasets can then be further curated via functional and biological data (e.g. GO annotations, protein interaction data, etc.) to try and assign additional relevance on the basis of whether subsets of proteins in the data set, for example, map to common biological pathways or functional relationships. These approaches should facilitate the generation of small enough candidate protein marker lists to be amenable to the MRM-based profiling workflows.
Another piece of the puzzle, and one that is particularly valuable for the in silico-derived candidate lists, will be the application of our PeptideAtlas and UniPep databases (3-5). PeptideAtlas is, in essence, a searchable repository for peptide identifications verified by MS/MS. UniPep is somewhat similar, but instead contains only identified glycopeptides. The PeptideAtlas/UniPep data is helpful in a number of ways. It tells us which proteins/peptides have previously been observed via MS and how often. This is very useful information when trying to identify the best proteotypic peptides for a candidate protein of interest for monitoring in MRM experiments, since it is typical that some and not all peptides from the same protein are routinely detected when that particular protein was in the original sample under analysis (proteotypic peptides are peptides that are unique in the sequence database for the protein/gene of interest and that are readily detectable by MS). PeptideAtlas also stores MS/MS data for all the peptide identifications in the database. This is again very useful for planning an MRM experiment where we don't yet know what transitions to measure in the MS for the target peptides, where a good guess can be made by examining previously obtained MS/MS spectra for the peptides. Currently there are four public builds of the PeptideAtlas: human, human plasma, yeast and Drosophila. Additionally, we have an in-house beta build of a mouse PeptideAtlas. We hope to make that public once we accumulate sufficient mouse data in the database.
The final piece of the puzzle, using the empirical data collected via the research tracks above and parsed through the plasma PeptideAtlas and predictive models to generate a list of proteotypic peptides that will uniquely and unambiguously identify the candidate proteins of interest in the plasma samples obtained from individuals both affected or unaffected by breast cancer, will be to synthesize these target peptides in an isotopically heavy form. These peptides, which will be individually quantified and quality controlled, will serve as external standards for identification and quantification of the target peptides/proteins in the MRM-based screening experiments. This will not only allow us to confirm or refute the presence of the target peptides in the plasma samples on the basis of observation of the appropriate isotopic pairs, confirmed through MRM observation of the expects fragment transitions, but will also provide information on absolute target peptide/protein concentration in the original sample on the basis of heavy:light isotope peak ratios. Such information will be invaluable as we progress from the profiling phase to the marker confirmation and validation phases of the project. Finally, regardless of our successes or failures in identifying new markers for breast cancer, the result of this part of the project will be the delivery to the research community of a substantial library of proteotypic peptides for human plasma proteins, along with their optimized MRM transitions and isotopically labeled and accurately quantified reference peptides for the quantitative analysis of plasma samples at high throughput.
- Zhang et al. (2003) Identification and quantification of N-linked glycoproteins using hydrazide chemistry, stable isotope labeling and mass spectrometry. Nat. Biotechnol., 21, 660-6.
- Zhang et al. (2007) Mass spectrometric detection of tissue proteins in plasma. Mol. Cell. Proteomics, 6, 64-71.
- Desiere et al. (2005) Annotation of the human genome with peptide sequences obtained by high-throughput mass spectrometry. Genome Biol., 6, R9.
- Deutsch et al., (2005) Human Plasma PeptideAtlas. Proteomics, 5, 3497-500.
- Zhang et al. (2006) UniPep - a database for human N-linked glycosites: a resource for biomarker discovery. Genome Biol., 7, R73.
CPAS 2.0 now available!
Several major updates to CPAS have been released in version 2.0.
Sequest Support: CPAS can now control a Sequest pipeline.
Performance Improvements, including ProteinProphet runs up to 100x faster, FASTA loading is approximately 10x on SQLServer, 2-3x faster on Postgres, and other improvements to performance.
- Protein Search: You can now search for a protein by name (or gene name, if protein annotations have been loaded) in all the MS2 runs in the current folder and subfolders.
In addition, CPAS has now been extended to include other components, which are disseminated through Labkey.org, a software foundation supporting extensions to the platform. Additions include capacity for analyzing flow cytometry, a observational study module enables researchers to define a study in terms of the intervention, the cohorts, a and an assay plan (some specializations for Vaccine research trials).
| [ Back to Top ] |
You may access the list of 2007 publications from Consortium members here.
| [ Back to Top ] |
None at this time.
| [ Back to Top ] |
Regional Workshop: Bioinformatics Workshop at Korea University, October 5-6, 2007
Dr. Myeong-Hee Yu and Prof. Eunok Paek of Korea Institute of Science and Technology (KIST) are hosting a Bioinformatics Workshop October 5-6, 2007 at Korea University. If you would like to give a talk on Analysis Modules or Recent Developments in TPP, please submit an abstract to Prof. Paek (E-mail: paek@uos.ac.kr).The workshop will be held at Korea University. Arrangements for lodging have been made with the Seoul Royal Hotel (www.royal.co.kr), located in the heart of downtown Myung-dong. For workshop attendees, 50% of room costs will be covered by the FPC for up to 3 nights. (Attendees will be responsible for $80 per room per night.) The financial support provided by FPC will be for up to 3 rooms (3-6 persons) per night per each ICBC team. There is no limitation on the number of participants from each team, and no registration charge. Workshop speakers will be fully reimbursed for the cost of lodging for up to 3 nights. Workshop attendees should contact Dr. Yu (mhyu@kist.re.kr) or Professor Paek (paek@uos.ac.kr) in order to get the proper hotel room rate.
Transportation between the Seoul Royal Hotel and Korea University will be provided.
There is limousine bus service (hotel and commercial) from Incheon Airport to the hotel.
Sharing Candidate Lists
The Singapore team has set up a cancer biomarker candidate lists system (CLUB) for the ICBC community to contribute their list of candidate genes or proteins directly. For more information, you may access the system via the following link: http://club.bii.a-star.edu.sg.
Users may upload and compare their lists with others' lists, maintain a list of their own candidate lists and share these lists with other users. It also provides the annotations from various publicly available databases, and lists of candidates from various experiments curated from literature, including the list of candidate biomarkers paper by Drs. Malu Polanski and Leigh Anderson. The Singapore team is currently working on features which allow users to filter their list of candidates based on criteria they could define.
ICBC Meetings
Presentations from the December ICBC meeting are now posted on CPAS. Minutes will be posted soon.
The 2008 ICBC meeting is tentatively scheduled for February 21-22 in sunny Hawaii, home to warm, tranquil waters and breathtaking beauty. More details to come.
APEC Update
Dr. John Potter represented the Fred Hutchinson Cancer Research Center at the recent Life Science Innovation Forum of the Asia Pacific Economic Cooperation (APEC) meetings in Adelaide, Australia. The recommendation of the participants was to re-endorse APEC economies' commitment to supporting projects such as the International Cancer Biomarker Consortium and the Asia Cohort Consortium.
| [ Back to Top ] |
| Project Title: | Pacific Ovarian Cancer Research Consortium | |
| Cancer Site(s): | Ovary | |
| Principal Investigator(s): |
Nicole URBAN, Sc.D. |
|
| Participating Institutions: |
Fred Hutchinson Cancer Research Center (FHCRC - Coordinating Center) |
|
| Clinical Samples: | Human tissues, sera, plasma, and urine | |
| Technical Approaches: | Transcript and Proteomic Technologies ELISA Assay ScFv Yeast Display Library Bead-based Multiplexing Assay System Protein/peptide quantitation using visible isotope-coded affinity tags (VICAT) and 16/18O-labeling |
Brief Description of Project:
The Pacific Ovarian Cancer Research Consortium (POCRC) SPORE in ovarian cancer brings together a group of dedicated and enthusiastic scientists from 8 institutions, each one contributing substantially to the intellectual activity of the SPORE. Institutions in the POCRC include the University of Washington (UW), Swedish Medical Center (SMC), Cedars-Sinai Medical Center (Cedars), Institute for Systems Biology, (ISB), Marsha Rivkin at Virginia Mason as well as the FHCRC and the Vancouver Island Cancer Center (VICC) in British Columbia. The Center for Ovarian Cancer Research (MRC), and Benaroya Research Institute POCRC has identified four areas where opportunities exist for us to make a difference, including the identification of tumor markers and the subsequent development of blood tests for risk assessment and early detection of ovarian cancer. Early detection, a key to improved outcomes for women with ovarian cancer, is a major focus of our research program. Our general strategy in meeting translational research goals includes 1) exploitation of emerging molecular technologies to identify biologically relevant genes, proteins and antigens as candidate markers and targets for translation, 2) a systematic approach to prioritizing markers and targets for evaluation, 3) a collaborative approach to evaluating candidate markers and targets that includes evaluation of candidates identified by colleagues at other institutions, and 4) use of novel statistical methods to use markers to predict biologic phenotype.
Our primary objective is to recommend a set of markers and an algorithm that can be used in a clinical trial for ovarian cancer screening to prevent ovarian cancer from escaping early detection. Achieving this goal requires the development of a novel statistical model to identify promising markers, as well as the evaluation of several currently existing and novel markers using serum specimens from two currently funded studies for ovarian cancer screening.
We have developed statistical methods for combining markers in a panel [1], and for using biomarkers longitudinally to improve their performance [2]. These methods are useful for evaluation of the markers for inclusion in the panel, as well as for using the panel clinically. We are evaluating the protein products of several genes that we found earlier to be over-expressed in malignancy [3], for their presence in serum and for their ability to identify women with ovarian cancer. Selection of the genes for study was based on statistical analysis of their joint contribution to a panel able to discriminate between malignant and normal tissue. Top genes include HE4, Mesothelin, SPINT2, SLPI, CD24, IFI27, MUC1, Keratin 8, FOLR and GPR39. Using mouse monoclonal antibodies, sandwich ELISA's were developed for HE4 and Mesothelin. To validate their performance singly and in a panel, proteins were measured by ELISA in an independent set of blinded samples. HE4 was shown to be more specific than CA125 in discriminating women with malignant tumors from those with benign tumors [4]. Mesothelin in combination with CA125 was shown to perform better than either used alone in discriminating women with malignant tumors from healthy women [5].
To complete the panel, we are using novel affinity reagents. Antibodies are the classical reagents used to detect disease markers in a large variety of diagnostic tests, from IHC to ELISA tests. Single chain Fragment variable (scFv) antibodies contain only the heavy and light chain variable sequences of the antibody recognition sequences. To isolate such antibodies, we are using the yeast display scFv library developed by Feldhaus and colleagues [6], which contains 2x109 scFv derived from a naïve human B cell population. The yeast display scFv are tagged with HA and c-myc epitopes, and recognize proteins without regard to their immunogenicity. We use the library both to identify scFv against known proteins such as SPINT2 and SLPI, and for discovery of novel proteins. For the latter application, we proceed by several rounds of subtraction as for cDNA subtractive libraries. The yeast display scFv library has several advantages over other available sources of antibodies such as the phage display libraries. First, it features the absence of any growth-related bias, due to a galactose promoter that prevents scFv expression during library amplifications. In addition, it provides the ability to sort the specific antigen-binding scFv by flow cytometry, and the efficient production by the yeasts of the scFv for future use. We have isolated one scFv candidate that binds preferentially to a pool of ovarian carcinoma sera versus a pool of control sera, and have identified the corresponding protein by immunoprecipitation and mass spectrometry. As in any proteomics approach, the source of patient material is critical, as it determines the discriminatory power of the antigen(s) identified.
- McIntosh MW and Pepe M, Combining Several Screening Tests: Optimality of the Risk Score. Biometrics, 2002. 58: p. 657-664.
- McIntosh MW and Urban N, A Parametric Empirical Bayes Method for Screening using Longitudinal Observations of a Biomarker. Biostatistics, 2003. 4(1): p. 27-40.
- Schummer M, Ng WV, Bumgarner RE, Nelson PS, Schummer B, Bednarski DW, Hassell L, Baldwin RL, Karlan BY, and Hood L, Comparative hybridization of an array of 21,500 ovarian cDNAs for the discovery of genes overexpressed in ovarian carcinomas. Gene, 1999. 238(2): p. 375-85.
- Hellstrom I, Raycraft J, Hayden-Ledbetter M, Ledbetter JA, Schummer M, McIntosh M, Drescher C, Urban N, and Hellstrom KE, The HE4 (WFDC2) protein is a biomarker for ovarian carcinoma. Cancer Research, 2003. 63(13): p. 3695-700.
- McIntosh M, Drescher C, Karlan B, Scholler N, Hellstrom K, and Hellstrom I, Combining CA 125 and SMR serum markers for diagnosis and early detection of ovarian carcinoma. Gynecologic Oncology, 2004. 95(1): p. 9-15.
- Feldhaus MJ, Siegel RW, Opresko LK, Coleman JR, Feldhaus JM, Yeung YA, Cochran JR, Heinzelman P, Colby D, Swers J, Graff C, Wiley HS, and Wittrup KD, Flow-cytometric isolation of human antibodies from a nonimmune Saccharomyces cerevisiae surface display library. Nat Biotechnol, 2003. 21(2): p. 163-70.
Team Members and Expertise:
Martin MCINTOSH, Ph.D. (FHCRC — Statistician and Project Principal Investigator) [mmcintos@fhcrc.org]
Nicole URBAN, ScD. (FHCRC — POCRC Principal Investigator and Early Detection Project Co-Investigator) [nurban@fhcrc.org]
Garnet ANDERSON, Ph.D. (FHCRC — Project Co-Investigator) [arnet@whi.org]
Natalie SCHOLLER, M.D. (FHCRC — Co-Investigator) [nscholle@fhcrc.org]
| [ Back to Top ] |
 Home page
|
||
| Faculty & Staff Directory
|
||
| Scientific Project Pages
|
This Web page lists scientific projects that have developed information for the public, scientists or clinicians. |
|
| Scientific Division Pages
|
These Web pages describe our four scientific divisions. | |
| - Basic Science | ||
| - Clinical Research | ||
| - Human Biology | ||
| - Public Health Sciences |
||
| Shared Resources
|
This Web page provides information about scientific services at FHCRC. |
Fred Hutchinson Cancer Research Center, home of three Nobel laureates, is an independent, nonprofit research institution dedicated to the development and advancement of biomedical technology to eliminate cancer and other potentially fatal diseases.
| [ Back to Top ] |
