Overview of Proteomics with Applications to Ovarian Cancer
Summer Gibbs
Genetics 144, Oncogenomics (Winter 2005, Dr. Charles Brenner)
I. Proteomics: An Overview
Proteomics involves global analysis of protein function.
The study of proteomics is instrumental in our understanding of cellular changes that result in cancer. As pointed out by Seller and Yates, except for the water in our bodies, almost everything is made of or by proteins (1). Following completion of the human genome project, it was estimated that there might be 1.5 million proteins, which translates to about 50 protein products per gene (incorporating alternative splice forms and alternatively post-translationally modified proteins). Thus, although the sequence of the human genome is of unquestioned scientific importance, its true value may lie in the proteins. The field of proteomics may allow for a new outlook on the human genome that will enable research to design more informed studies that can be more easily interpreted (1).
Proteomics based discovery methods can be grouped into three main categories; two-dimensional polyacrylamide gel electrophoresis (2D-PAGE), protein arrays, and mass spectrometry (2).
2D-PAGE separates proteins by both their charge (isoelectric points) and molecular weights (2) Proteins are detected by addition of a marker to the gel, where markers range from silver ions and dyes to fluorescent molecules (1). With the current technology it is possible to separate protein mixtures into their individual polypeptide components, compare expression profiles of different samples, or to perform experiments where a condition of interest is applied and the protein response is observed (2). 2D-PAGE is somewhat limited by its dynamic range as well as difficulties separating certain classes of proteins (1).
Protein arrays are a protein microarray based technology that allows for detection of specific proteins or protein function. The main differences between these two types of arrays are the molecules that are immobilized on the surface of the array. For the protein detection array, a set of probes is spotted onto the surface of the array that the protein will interact with, where the most common probe is the antibody. For the protein function type array a number of different proteins are spotted onto the array surface and these proteins are then used to test for binding with proteins, antibodies, hormones, nucleotides, nucleic acids, etc (2). An illustration of this concept can be viewed in figure 1 of a review of proteomics by Seller and Yates (1).
Mass spectrometry has also emerged as a useful tool in proteomics and proves to be most functional in identification of specific proteins and sequencing of proteins of interest discovered during experimentation (1). Two types of mass spectrometry that are of particular use as ionization techniques are electrospray ionization (ESI) and matrix-assisted laser desorption ionization (MALDI). These ionization techniques can be used in combination with mass analysis techniques such as time-of-flight (ToF), single or triple quadrapole, ion trap and others to analyze samples. MALDI-ToF-MS or ESI-MS are used when comparison is made between samples to determine if the same proteins are present. Both methods are dependent upon gene and protein bioinformatics to determine the precise mass of the protein so that it may be compared to the mass obtained from the MALDI-ToF-MS or the ESI-MS (1). Another application of mass spectrometry is termed peptide mass fingerprinting. This technique can be used to identify a protein using its sequence as determined by mass spectrometry and a database containing all of the possible protein sequences in the organism of interest. The basic method for peptide mass fingerprinting involves digestion of an intact protein with a site-specific enzyme. The molecular weights of this set of peptides are measured via mass spectrometry and then compared against a known set of peptide molecular weights (following cleavage with the same enzyme) in a database for the test organism. When most of the molecular weights of the sample match those in the database it indicates that the particular protein has been identified.An illustration of the steps involved in peptide mass fingerprinting can be seen in figure 3 of the review of proteomics by Seller and Yates (1).
II. Epithelial Ovarian Cancer: Macroscopic and Microscopic Views
Ovarian cancer is the leading cause of gynecologic malignancies among American women. There are approximately 26,000 cases diagnosed each year and about 16,000 women die of ovarian cancer per year. Over 70% of those diagnosed are not diagnosed before the disease has spread beyond the pelvis, where the five year survival rate for these patients is only about 15% (3).
Macroscopic View:Most primary neoplasms in the ovary arise from the epithelial layer. There are three main types of tumors that arise from the epithelium which are termed serous, endometrioid and mucinous tumors. Serous tumors account for about 30% of all ovarian tumors, where about 75% are benign or borderline malignant and the other 25% are in fact malignant. Mucinous tumors account for about 25% of all ovarian neoplasms, where about 80% are benign or borderline malignant, while the rest are malignant. Endometrioid tumors account for about 20% of all ovarian tumors. Endometrioid tumors are usually carcinomas; benign forms are much less common (4).
Microscopic View:Currently the exact etiology of ovarian cancer is unknown, but under extensive study. Recent research suggests involvement of various genetic alterations in the path to ovarian cancer. Both p53 and retinoblastoma (Rb), which are known to be important tumor-suppressor genes, are frequently inactivated in human ovarian cancers. It has been determined that in more that half of ovarian carcinomas, mutations of p53 is observed. Allelic loss of the Rb locus is less common; only 30% of ovarian carcinomas show this loss. However through analysis of the signaling molecules upstream of Rb it was determined that at least 80% of ovarian tumors exhibit some type of disruption in this general pathway. Activation of hTERT is observed in all ovarian malignancies where as telomerase activity is rarely seen in benign ovarian tumors. Other mutations and amplifications that are often found in ovarian carcinoma cells are found in the proto-oncogenes c-myc, pkb/akt, and Ras. Amplification of c-myc has been reported in about 30% of ovarian cancers, where as activation of the pkb/akt pathway has been reported in over 30% of ovarian cancers. Alterations in the Ras signaling pathway are also common in ovarian carcinomas and can occur by several mechanisms (5).
III. Epithelial Ovarian Cancer: Molecular Mechanisms and Cellular Models
It is well established that the involvement of different tumor suppressor genes and oncogenes is important in the development of ovarian cancer; however the molecular mechanisms by which these changes take place have yet to be determined. One difficulty in the study of the molecular mechanisms was lack of an appropriate ovarian carcinoma model. In 2004, Liu, et al published an ovarian cancer cell model developed in their lab which produces tumors morphologically similar to high-grade, undifferentiated human ovarian carcinoma (6). This model was developed by transforming human ovarian surface epithelial (HOSE) cells. The HOSE cells had their p53 gene, Rb gene and ectopic expression of the telomerase catalytic subunit disrupted as well as their Ras signaling pathway activated (6).
IV. Use of Functional Proteomics in Ovarian Cancer Model Study
In a paper by Young, et al published in Cancer Research (July, 2004) a study was completed which identified several major cellular pathways that are activated following ras transformation of ovarian cancer cells. Young, et al showed how these pathways contribute to initiation and development of oncogenic transformations in the ovarian epithelium (5).
The epithelial ovarian cancer model discussed in section three was used in this study. An immortalized T29 HOSE cell line was compared to a T29H HOSE cell line. The T29 cell line was generated from IOSE29 cells using hTERT. It formed few colonies on soft agar and was not tumorogenic in nude mice. The T29H cell line was generated by infection of the T29 HOSE parent cell line with a retrovirus that contained the oncogenetic H-RasV12. The transformed T29H cells formed colonies well in soft agar and were tumorogenic in nude mice. As can be seen in figure 1 of this publication, Ras levels are significantly higher in the T29H cell line then in the T29 parent cell line. These two cell lines were compared for the purpose of this study (5).
To determine the proteomic changes associated with Ras-mediated oncogenic transformation 2D gel elecrophoresis was run on both cell lines. More than 2200 distinct protein spots between the pH of 4 and 7 and within the molecular weight range of 6,000 - 200,000 could be resolved to be different between the T29 and T29H cell lines. Quadruplicate gels were analyzed for spots with average normalized volume that had at least a 50% different between the T29 and T29H cell lines. These spots of interest were processed by in-gel digestion and analyzed by peptide mass fingerprinting followed by database matching in the human nonredundant database. Young, et al were able to positively identify thirty proteins associated with Ras-mediated transformation by this method (5).
Young, et al discovered that five of the thirty positively identified, significant proteins are involved either directly in metabolizing reactive oxygen species (ROS) or in maintaining the redox balance of the cell. The five identified enzymes and their function are listed below:
1. Peroxiredoxin 3 - Part of the antioxidant family of thioredoxin peroxidases. Peroxiredoxin 3 exists in the mitochondria.
2. Thioredoxin Peroxidase (Peroxiredoxin 4) - Antioxidant enzyme that uses thioredoxin to scavenge hydrogen peroxide from the cell.
3. Selenophosphate Synthetase - A rate-limiting enzyme for the incorporation of selenium into key redox metabolizing enzymes.
4. NADH Dehydrogenase Ubiquinone Fe/S 3 - 30 kilodalton subunit of the mitochondrial complex I of the electron transport chain.
5. Glyoxyalase I - Key enzyme in the detoxification of methylglyoxal, a reactive side product of glycolysis.
The gel that shows these upregulated proteins in the T29H cell line vs. the T29 cell line can be seen in figure 2 of the publication by Young et al (5).
Following the observation of this upregulation of proteins involved in metabolizing ROS and maintaining the redox balance of the cell in the T29H cell line, experiments were performed to test if the T29H cell line was more resistant to oxidative stress then the T29 cell line. These oxidative stress experiments were performed by exposing both the T29H cell line and the T29 cell line to varying concentrations of hydrogen peroxide (25 - 400 micromolar) for varied time periods (5).
At 100 micromolar hydrogen peroxide treatment the T29 cells showed marked cell death after 6 hours, while the T29H cells were largely unharmed under the same conditions. A phase-contrast microscopy image illustrating the T29H cell line's resistance to the hydrogen peroxide treatment can be seen in figure 3a of the Young et al publication. The effect of the hydrogen peroxide treatment on cell survival and proliferation was also assessed. The growth of the T29 cell line was dramatically repressed after the 6 hour, 100 micromolar hydrogen peroxide treatment, where the T29H cell line was not significantly affected by the same treatment (5).
Caspase 3 activity and nuclear DNA status were assessed to determine if the T29 cells were succumbing to apoptosis following the hydrogen peroxide treatment. Caspase 3 activity is often used as a marker of apoptosis since it is one of the major downstream effectors of the apoptotic cascade. Following a 12 hour treatment with 100 micromolar hydrogen peroxide, Hoescht 22658 staining was used to assess the nuclear DNA condensation and fragmentation. The T29 cells exhibited significantly more DNA condensation and fragmentation as compared to the T29H cells. The T29 cell line also had significantly higher caspase 3 activity then was seen in the T29H cell line. These data taken together suggest that Ras-transformed HOSE cells are resistant apoptosis to following oxidative stress (5).
To verify that the Ras activation was in fact responsible for the resistance to hydrogen peroxide mediated apoptosis two methods to inhibit Ras were used.
The first method was through the use of farnesyl transferase inhibitor FTI-277 to block Ras activity. Using a Ras activation assay it could be seen that FTI-277 significantly inhibited Ras activation in the T29H cell line. T29H cells that were pretreated with FTI-277 24 hours prior to hydrogen peroxide treatment showed a marked decrease in resistance to cell death. Phase-contrast microscopy images illustrating this point can be seen in figure 5b of the Young, et al publication. FTI-277 treatment was also observed to increase the level of caspase 3 activity significantly in FTI-277 T29H pretreated cells over untreated T29H cells when both the pretreated and untreated cells were subject to hydrogen peroxide (5).
Although farnesyl transferase inhibitors are often used to inhibit Ras activation in vivo it is well established that these inhibitors have the potential to inhibit other cellular processes that are also dependent upon farnesylation and thus is not necessarily specific to the inhibition of Ras activation. As a second method of verification of Ras activation responsibility for the hydrogen peroxide mediated apoptotic resistance T29H cells had cellular Ras levels selectively suppressed through the use of retroviral mediated siRNA. The levels of active Ras in T29H cell line were significantly reduced through the expression of two siRNA vectors; one to target the wild-type H-Ras and one to target the H-RasV12 mutation. Stable expression of the two siRNA vectors in T29H cells lead to increased sensitivity to hydrogen peroxide treatment as compared to the T29H parent cells. A phase-contrast microscopy image that illustrates this can be seen in figure 6c of the Young, et al publication. Additionally, suppressing Ras activation through the siRNA vectors significantly increased caspase 3 activity following hydrogen peroxide treatment (5).
When the data collected with both types of Ras inhibition are considered the results suggest that Ras activation is essential in the protection of T29H cells from oxidative stress mediated apoptosis induced by hydrogen peroxide treatment (5).
Table 1 - Summary of Results by Young et al (5).
V. Reactive Oxygen Species: Applications to Cancer
The mitochondria are important sites for the production of ROS in cells. During glycolysis the mitochondrial electron transport chain consumes oxygen through oxidative phosphorylation. This process produces a large quantity of the necessary cellular energy in the form of ATP. This process is also responsible for a large quantity of ROS that are produced in the cell. During the oxidative phorphorylation process it is estimated that between 0.4% and 4% of the oxygen consumed is release in the mitochondria as ROS. ROS include the following compounds; hydrogen peroxide, superoxide, singlet oxygen, and the hydroxyl radical (7).
Hydrogen peroxide is the only one of these ROS that is stable enough to diffuse out of the mitochondria to have a cytoplasmic effect on the cell. It is known that low levels of hydrogen peroxide are important for proper cellular function. Low levels of hydrogen peroxide play a role in receptor-mediated cell signaling pathways, normal cell proliferation and transcriptional activation. However, high levels or significantly increased levels of mitochondrial hydrogen peroxide are harmful to the cell, inducing apoptosis and thus killing the cell (7).
Oncogenetically transformed cells have an increased growth rate and thus produce more ROS than normal cells due to their increased cellular respiration rate. High levels of ROS are in fact though to be an important contributor to the genetic changes which push a cell from normal to cancerous. However, even in these oncogenetically transformed cells levels of ROS that are too high will induce apoptosis. Thus, as seen in the comparison of the T29 and T29H cells in the publication by Young et al, some oncogenes may upregulate antioxidant proteins to allow for the protection of fast growing cancerous cells (5).
VI. Discussion and Conclusions
Proteomics is an important tool in the study of cancer. Epithelial ovarian cancer while accounting for a small percentage of the total incidence of cancer has a very high mortality rate for those who are diagnosed. The study of proteomic changes of epithelial ovarian cancer cell models may lead to an understanding of the molecular mechanisms that are behind this disease.
The publication by Young et al illustrated that HOSE cells that have been transformed to have activated Ras mutations have upregulated proteins that are responsible for the metabolism of ROS and for maintaining redox balance in the cell. Figure 7 of the Young et al publication shows the pathways and mechanism of the Ras-mediated protection from ROS-induced apoptosis. Experimentation with hydrogen peroxide showed that the cells with actived Ras had significantly higher resistance to apoptosis. When the Ras transformed cells had their Ras inactivated they were once again resensitized to hydrogen peroxide treatment (5).
There is further study to be completed on this topic, but it is conceivable that an enhanced antioxidation capability may constitute a common mechanism for tumor cells to evade apoptosis induced by oxidative stresses at high ROS loads (5).
References:
(1) Seller TA, Yates RJ. Review of Proteomics With Applications to Genetic Epidemiology. Genetic Epidemiology, 2003. 24(2): 83 - 98.
(2) Celis JE, Gromova I, Rank F, Gromov P. Proteomics in Bladder Cancer. Oncogenomics: Molecular Approaches to Cancer, edited by Brenner C, Duggan D. Oncogenomics. Wiley-Liss, Hoboken, New Jersey, 2004, pgs. 157 - 175.
(3) Posadas EM, Davidson B, Kohn EC. Proteomics and ovarian cancer: implications for diagnosis and treatment: a critical review of recent literature. Current Opinion in Oncology, 2004. 16(5): 478 - 484.
(4) Kumar V, Abbas AK, Fausto N. Robbins and Cotran: Pathological Basis of Disease. Elsevier Saunders, Philadelphia, Pennsylvania, 2005. Pgs. 1093 - 1098.
(5) Young TW, Mei FC, Yang G, Thompson-Lanza JA, Liu J, Cheng X. Activation of Antioxidant Pathways in Ras-Mediated Oncogenic Transformation of Human Surface Ovarian Epithelial Cells Revealed by Functional Proteomics and Mass Spectrometry. Cancer Research, 2004. 64(13): 4577 - 4584.
(6) Liu J, Yang G, Thompson-Lanza JA, Glassman A, Hayes K, Patterson A, Marquez RT, Auersperg N, Yu Y, Hahn WC, Mills GB, Bast RC Jr. A Genetically Defined Model for Human Ovarian Cancer. Cancer Research, 2004. 64(5): 1655 - 1663.
(7) Nonn L, Berggren M, Powis G. Increased Expression of Mitochondrial Peroxiredoxin-3 (Thioredoxin Peroxidase-2) Protects Cancer Cells Against Hypoxia and Drug-Induced Hydrogen Peroxide-Dependent Apoptosis. Molecular Cancer Research, 2003. 1(9): 682 - 689.