Home,  Research,  People,  Contacts,  Downloads,  Links,  Pubs/Presentations/Posters

Differential proteomics using isotope-assisted tandem mass spectrometry to study colon cancer (collaboration with Willliam F. Dove , Depts. Of Genetics and Oncology at McCardle Laboratories).  Wisconsin is developing an extensive series of murine (mouse and rat) genetic models for human diseases.  These models include animals mutated in particular tumor suppressor genes or carrying particular activated oncogenes. One of the many mouse model systems currently studied on campus is the Min mutant (Moser et al., 1990), which develops multiple intestinal neoplasms.  Through mouse genomics the causative molecular lesion for this mutant has been identified (Su et al., 1992).   In collaboration with members of the Sussman laboratory and the UW Biotechnology Center Mass Spectrometry Facility, the Dove laboratory is applying state of the art proteomics techniques including 15N metabolic labeling to better understand the physiology of the Min mouse and to uncover possible peptide and protein biomarkers. 

The mammalian intestine is highly proliferative and is continually exposed to a genotoxic environment. As such, it is a common site of cancer; cancers of the gastrointestinal tract account for ~80,000 deaths per year. Clearly, it is important to understand gastrointestinal cancer pathology at the molecular, cellular, and histologic levels. In the United States, the most common gastrointestinal malignancy is colorectal cancer. The majority of colorectal cancer cases are sporadic, without a known hereditary component. Yet rare inherited forms of this disease exist and have been highly informative.  For example, Familial adenomatous polyposis (FAP), or Gardner’s syndrome, is an autosomal dominant disease caused by a germline mutation in the adenomatous polyposis coli (APC) tumor suppressor gene. Individuals with FAP can develop several thousand benign polyps throughout the colon. Although the benign lesions themselves are not dangerous, 5% of these adenomas will progress to malignancy.

The laboratory mouse provides a model to study the many genetic pathways that regulate intestinal tumorigenesis.  Multiple intestinal neoplasia (Min) mice, a model of FAP, are heterozygous for ApcMin, a germ-line truncating mutation at codon 850 of the Apc gene, and develop multiple intestinal neoplasms. The biological phenotype of the Min mouse is strongly affected by the genetic background.  In the sensitive C57BL/6 background, ca. 100 adenomas are commonly found, of which 95 lie in the small intestine and 5 in the colon.  Such animals are obviously sick by about 100 days of age.  By contrast, when the Min mutation is introgressed onto the AKR background, animals develop only 1 or 2 adenomas, each in the small intestine.  Thus the Min mouse must be studied under genetically homogeneous conditions.  To this end, the Min mutation has been introgressed onto the sensitive C57BL/6 genetic background for more than 60 backcross generations.  Because the mutation was induced in this same background, it is believed that C57BL/6 and C57BL/6-Min/+ differ only by a single heterozygous basepair out of their 3 billion basepair genome.  The “molecular phenotype” that can be analyzed by differential mass spectrometry between sets of Min/+ and +/+ mice that have been matched for gender, age, and environment will therefore minimize biological variability.

              Metabolic labeling of the Min mouse by 15N and 14N is currently being applied at Wisconsin to compare the proteomes of Min mice to their wildtype litter-mates.  Using light (14N) and heavy (15N) isotope-labled Spirulina (an algae that is commercially available with either 99.6%-[14N] (i.e., natural abundance) or heavy isotope labeled 98%-[15N]) a diet was formulated with the help of a nutritionist, Barbara Mickelson at Harland Teklad Inc. Madison, WI.  For these experiments, we obtained 14N and 15N-enriched Spirulina from Spectra Stable Isotopes (Columbia, MD). After enrichment with nitrogen-depleted oil and vitamins a formulation was created that supports normal, healthy growth in weaned mice.  This powdered green material is fed to mice in lieu of their normal food, starting at weaning, ca. two weeks after birth (see picture above). 

In our first experiment, we set up two cages, each containing four mice. In each cage two of the mice were Min, and two were wildtype (the Min mutation is dominant and these eight mice were the progeny of 2 mothers crossed with the appropriate male).  The mice in one cage were fed exclusively the natural abundance formulation (14N), while those in the other cage were fed the 15N formulation.  After two months of growth, ca. 20 different organs and tissues were isolated from each of the eight mice, and were homogenized in 14N and 15N pairs.  This experimental design allows us to create BOTH sets of protein extracts, one with 14N wildtype and 15N Min, and the second with 15N wildtype and 14N Min.  In addition, each of these combinations is done in duplicate, for a total of 4 paired samples. Thus our experimental design controls for any unintended effects of the isotopic labeling procedure and begins to address biological variability. Tissue samples from these mice are being used for two independent proteomic comparisons: first, samples from a host of different organs will be compared to explore widespread physiological differences between wildtype and Min mice at an intermediate stage in disease development.  Second, blood samples were drawn from these mice at several times throughout disease development.  The peptide and protein complements of these blood samples will be compared to identify potential markers of disease progression and to monitor their relative changes over time. 

Our first goal for the metabolic labeling experiment, comparison of organs from the Min and wildtype mice, has the potential to substantially improve our understanding of the physiology of the Min mouse.  In general several technical challenges need to be addressed to gain maximum insight from these types of comparisons.  First, tissue homogenization and protein extraction are major sources of sample preparation variability.  Second, as is true for most proteomic samples, the complexity of each sample and the wide range of abundances for various proteins within each sample make complete characterization difficult.  Fractionation is essential to improve our sensitivity, yet tends to introduce sample to sample variability.  Both of these issues are addressed by metabolic labeling, in that both the Min and wildtype tissue are combined so that all stages of sample preparation and analysis are internally controlled; thus metabolic labeling allows us to employ more extensive fractionation and improves our sensitivity. 

The second goal of the metabolic labeling project is to screen blood drawn at various times throughout disease progression for potential colon cancer biomarkers. As anyone reaching the age of 50 can attest, there would be tremendous benefits to providing a blood analysis for colon cancer in clinical settings.  The current investigation using metabolic labeling is an extension of previous work from the Dove and Sussman laboratories using proteomic approaches to identify candidate biomarkers.   Previously using non-quantitative two-dimensional LC-based mass spectrometry survey approaches we identified several peptides in serum of Min mice that seemed to be markers of disease.  As shown in Figure 1, one of these peptides was shown by comparison with a synthetic, isotopically labeled standard to be slightly increased in Min mice (Huttlin et al 2005).  Due to the extensive sample preparation required for these experiments, the physiological significance of these results was difficult to assess.   Use of metabolic labeling for the current quantitative proteomic survey will control for sample handling and should allow us to identify higher quality leads for future characterization as potential biomarkers. 

A large problem with the proteomic analysis of blood is that protein abundances vary over several orders of magnitude and a small number of especially abundant proteins tend to dominate each sample.  In order to address this problem, we have successfully employed various fractionation techniques, including a commercial immunoaffinity column that depletes mouse serum for albumin. For example, as shown in Figure 2, we are able to deplete mouse plasma of albumin, which allows us to detect changes in the abundance of several proteins in Min vs. wildtype mice (sample 1 vs. sample 2) using Sypro Ruby stained 2D gel electrophoresis. These samples are from preliminary experiments using mice that are 90-100 days old, and which are already showing signs of disease. Our metabolic labeling experiment uses much younger mice, which show no differences in health or behavior between MIN and WT.  These inexpensive ‘pilot’ 2D gel experiments however have helped us to optimize our fractionation and other techniques such that our samples are processed most effectively.

We are currently identifying and quantifying 14N and 15N peptides and proteins in samples from both of these experiments using our ESI-QTOF.  Whether working with tissue samples or serum, extensive sample preparation and fractionation is essential to get maximum proteome coverage.  The internal control provided by metabolic labeling is essential, since by separating the proteins extensively via a variety of different techniques, one can identify a far greater number of proteins and peptides in each sample while still making reliable quantitative comparisons. Metabolic labeling is relatively expensive in mammals, with the isotopically labeled diet costing $5,000 for the four mice.  Thus, while metabolic labeling allows us to perform more extensive fractionation, it practically limits the amount of tissue we may use for the experiment.