The Good, the Bad and the Very Bad

A New Platform Strategy to Identify Drug Targets in Cancer Stem Cells

  • Fig. 1: Proteomic workflow. Two batches of cells will be grown, one in a medium that contains a “light” (normal) amino acid and the other in a medium that contains a “heavy” amino acid, with the heavy amino acid containing 13C instead of 12C and 15N instead of 14N. After at least 6 cell doublings, both batches will be subject to cell sorting, where the side population and the parental population will be collected. The heavy labeled population can be combined 1:1 with the non labeled (light) parent cells and subjected to proteolysis and mass spectrometry analysis on an LTQ-Orbitrap hybrid mass spectrometer (ThermoElectron). Peptides will be analysed using MuDPIT and data-dependent acquisition will be employed and consists of one full-scan mass spectra (in the Orbitrap) followed by four tandem mass spectra in the LTQ. The relative abundances of each peptide identified in the mixture will be determined using the Census software platform. Census extracts the ion chromatogram and calculate peptide ion intensity ratios for each pair of extracted ion chromatograms.Fig. 1: Proteomic workflow. Two batches of cells will be grown, one in a medium that contains a “light” (normal) amino acid and the other in a medium that contains a “heavy” amino acid, with the heavy amino acid containing 13C instead of 12C and 15N instead of 14N. After at least 6 cell doublings, both batches will be subject to cell sorting, where the side population and the parental population will be collected. The heavy labeled population can be combined 1:1 with the non labeled (light) parent cells and subjected to proteolysis and mass spectrometry analysis on an LTQ-Orbitrap hybrid mass spectrometer (ThermoElectron). Peptides will be analysed using MuDPIT and data-dependent acquisition will be employed and consists of one full-scan mass spectra (in the Orbitrap) followed by four tandem mass spectra in the LTQ. The relative abundances of each peptide identified in the mixture will be determined using the Census software platform. Census extracts the ion chromatogram and calculate peptide ion intensity ratios for each pair of extracted ion chromatograms.
  • Fig. 1: Proteomic workflow. Two batches of cells will be grown, one in a medium that contains a “light” (normal) amino acid and the other in a medium that contains a “heavy” amino acid, with the heavy amino acid containing 13C instead of 12C and 15N instead of 14N. After at least 6 cell doublings, both batches will be subject to cell sorting, where the side population and the parental population will be collected. The heavy labeled population can be combined 1:1 with the non labeled (light) parent cells and subjected to proteolysis and mass spectrometry analysis on an LTQ-Orbitrap hybrid mass spectrometer (ThermoElectron). Peptides will be analysed using MuDPIT and data-dependent acquisition will be employed and consists of one full-scan mass spectra (in the Orbitrap) followed by four tandem mass spectra in the LTQ. The relative abundances of each peptide identified in the mixture will be determined using the Census software platform. Census extracts the ion chromatogram and calculate peptide ion intensity ratios for each pair of extracted ion chromatograms.
  • Table 1: 41 proteins were found to upregulated at least 1.5 fold.
  • Kim D. Janda, PhD, Scripps Research Institute
  • Sebastian C.J.  Steiniger, PharmD, PhD, Scripps Research Institute

Interest in cancer stem cells (CSC) continues to grow and we are now at a stage where a fundamental understanding of their mechanism of development and disease progression can be addressed [1]. Observations that solid tumors are typically heterogeneous and contain only a small number of clonogenic cells has recently been supported by the isolation of putative cancer stem cells from a number of solid cancers. It is well known that tumors consist of a heterogeneous population of cells, containing tumor cells, fibroblasts, endothelial cells, and other cell types. One hypothesis is that cancer stem cells are one of the few cell types that are sufficiently long-lived to acquire the necessary number of sequential mutations to convert a cell from the normal to malignant state. It is thought that these cells are responsible for the development of tumors, which makes them a prime target of any given cancer therapy.

The increased awareness on cancer stem cells has led to the hypothesis that current cancer drugs do not target the causative cell population that initiates cancer, namely the cancer stem cell sufficiently. If these cells are not eradicated, the result will be the recrudescence of the tumor after the therapy is finished. The development of new therapies that target cancer stem cells will depend on further characterization of these cells and the development of assays that allow for the evaluation of ligands that bind to cancer cells. The further characterization of cancer stem cells has important implications for future therapies and we would like to present a new platform strategy for the identification of drug targets.

Flow Cytometry and Cell Sorting


For solid tumors, investigators have utilized FACS technology and magnetic bead sorting to isolate a subpopulation with unique surface markers. These cells have been transplanted into immunocompromised NOD-SCID mice and were able to form tumors. This methodology was also used to isolate a subpopulation of highly tumorigenic cells from several human breast cancers [2]. However, the origin of the cancer stem cell is not quite clear. Wherein, the markers that were used to identify cancer stem cells were CD24 and CD44.

Finally, the cell population that was CD44+/CD24- had a 10- to 50-fold increased ability to form tumors compared to the original tumor mass in mouse xenograft models.

Proteomics, Quantitative Mass Spectrometry and SILAC

Proteomics can be defined as a systematic approach to analyze both qualitative and quantitative mapping of the whole proteome [3, 4]. We have explored the properties of breast cancer stem cells and were able to identify a novel target related to drug resistance. The novel platform technology that we have developed combines quantitative mass spectrometry with metabolic labeling and cell sorting. We have used this platform to dissect cancer stem cells and in total we were able to identify a number of potential targets that were upregulated in breast cancer stem cells. Additionally, to gain further insights into the proteins that may be responsible for the selfpropagating and tumorigenic properties of breast cancer stem cells we applied proteomic tools. We were able to provide strong evidence that several important proteins in cell cycle and differentiation are upregulated in the stem cell population. Furthermore, we were also able to analyze the function and the involvement in drug resistance with one of the identified proteins.

Proteomic analysis may be investigated on many different levels; we selected SILAC analysis to examine the relative protein abundance in two distinct breast cancer cell populations, namely Hoechst33342high and Hoechst33342low. Figure 1 provides a schematic diagram of the quantification strategy. Our choice of using SILAC to examine the CSC proteome derives from the high throughput, fast speed, and easy manipulation of the proteome using SILAC, which is advantageous compared with two-dimensional PAGE or conventional Western blot analysis. Its major advantage is that one can control for variation in both sample preparation and peptide ionization because labeled and unlabeled samples can be mixed before fractionation, preventing the accumulation of systematic errors. However, this technique is more suitable for cultured cell lines than for primary cells and isobaric tag for relative and absolute quantitation (iTRAQ) or label-free MS workflows would be good alternatives for other cell types [5].

Using this technique, the quantification of nearly 900 proteins was accomplished. Table 1 shows the SILAC ratio distribution for all proteins upregulated or downregulated at least 1.5-fold in the stem cell population. Using SILAC, in total, 41 upregulated proteins were identified to be upregulated, and excitingly several of these proteins are known to play a role in differentiation and cell cycle control. The most interesting proteins identified include: (a) AHNAK - neuroblast-associated differentiation factor, whose expression and cellular localization has been dynamically regulated during cell cycle progression; and it has been localized to sites of major morphogenesis during mouse placentation [6]. (b) Proliferation-associated protein, also known as Ebp1, which is involved in the regulation of cell growth and differentiation in mammary epithelial cells through putative DNA-binding properties after translocation into the nucleus [7]. (c) E3 ubiquitin-protein ligase TRIM33, known to play a role in the control of cell proliferation; its association with SMAD2 and SMAD3 stimulates erythroid differentiation of hematopoietic stem/progenitor cells [8]. (d) SIAH-1-interacting protein was also discovered; SIAH-1-interacting protein (CacyBP/SIP), a target protein of S100, has been identified as a component of a novel ubiquitinylation complex leading to β-catenin degradation, which was found to be related to the malignant phenotypes of gastric cancer [9]. As stem cells are a continually self-renewing proliferative population, control of growth may be a crucial step in moving away from the stem cell phenotype, and some of these proteins may potentially serve as therapeutic targets.

In terms of downregulated proteins identified by SILAC, pyruvate kinases, whose activity has been shown to be generally increased in malignant tumors such as carcinomas, were found to be prominent [10]. This result is of interest from the standpoint of the tumorigenicity and malignancy of the SP cells. Another important molecule that was downregulated was peroxiredoxin 6, which seems to be involved in tumor progression and metastasis [11]. To further complement these findings, the differential profiles of several targets by RT-PCR and immunoblotting were characterized and found to possess similar patterns.

SiRNA Silencing of TB4 Results in Decreased Drug Resistance in MCF7 Cells

Proteomic analysis of the SP cells also revealed an increase in expression of Thymosin beta 4 (TB4). TB4 is a small, 43-amino acid, actin-binding, intracellular protein involved in cell motility [12]. Increased expression of this protein has been found in various tumors, for example, breast, ovary, uterus, colon, and thyroid cancers [13-16]. TB4 expression has been studied in breast cancer cells like MCF7 and MDA-MB231 cells, and the levels of β thymosins in tumor tissues have been proposed to constitute a potential diagnostic marker for the state of a tumor. Overexpression of TB4 has been associated with increased invasiveness and metastasis in colon cancer [17]; therefore, we postulated that decreasing the levels of thymosin expression in breast cancer cells may alter the growth rate and chemosensitivity of these cells. siRNA silencing of TB4 resulted in an increase in cell death upon paclitaxel and doxorubicin treatment in MCF7 and MDA-MB231 cells compared with the control siRNA, indicating that the chemoresistance of these cells had been decreased. This observation was echoed in a previous study that showed that paclitaxel-induced cell death was increased in HeLa cells upon siRNA silencing of thymosin [18]. As our findings demonstrate, the SP cell phenotype strongly contributes to drug resistance, and these results implicate TB4 in the drug resistance of CSCs. Furthermore, PKM2 was found to be downregulated 1.5-fold. This result is interesting, considering a recently published study wherein it was demonstrated that rapidly growing cancer cells have a high activity of pyruvate kinase M2 [19]. The activity of this key glycolytic enzyme can be activated and controlled by tyrosine kinase signaling pathways. Regulation of PKM2 activity may provide a direct link between cell growth signals using tyrosine kinase and thus, also, control of glycolytic metabolism. This regulatory pathway could be a key determinant of how CSCs modulate their metabolism and their cell growth.

Conclusion

Overall, our platform technology proved to be successful and we were able to show that there are differences in protein expression and tumorigenicity between the two analyzed populations. We are fully aware of the fact that there might be differences in protein expression between primary cancer cells and cancer cell lines, nevertheless, we are convinced that our study is of great scientific value and sheds more light into the character of cancer stem cells [20].

References
[1] Sell S.: Crit Rev Oncol Hematol, 51(1), 1-28 (2004)
[2] Al-Hajj M. et al.: Proc Natl Acad Sci USA 100(7), 3983-3988 (2003)
[3] Mann M.: Nat Rev Mol Cell Biol 7(12), 952-958 (2006)
[4] Everley P.A. et al.: Mol Cell Proteomics 3(7), 729-735 (2007)
[5] Unwin R.D. et al.: Blood 107(12), 4687-4694 (2006)
[6] Downs K.M. et al.: Gene Expr Patterns 2(1-2), 27-34 (2002)
[7] Yoo J.Y. et al.: Br J Cancer 82(3), 683-690 (2000)
[8] He W. et al.: Cell 125(5), 929-941 (2006)
[9] Ning X. et al.: Mol Cancer Res 5(12), 1254-1262 (2007)
[10] Yilmaz S. et al.: Arch Med Res, 34(4), 315-324 (2003)
[11] Chang X.Z. et al.: Breast Cancer Res 9(6), R76 (2007)
[12] Safer D.R. et al.: Proc Natl Acad Sci USA 87(7), 2536-2540 (1990)
[13] Xie D. et al.: Int J Oncol 21(3), 499-507 (2002)
[14] Clark E.A. et al.: Nature 406(6795), 532-535 (2000)
[15] Santelli G. et al.: Am J Pathol 155(3), 799-804 (1999)
[16] Hall A.K.: Int J Cancer 48(5), 672-677 (1991)
[17] Wang W.S. et al.: Oncogene 22(21), 3297-3306 (2003)
[18] Oh S.Y. et al.: Exp Cell Res 312(9), 1651-1657 (2006)
[19] Christofk H.R. et al.: Nature 452(7184), 230-233 (2008)
[20] Steiniger S.C. et al.: Stem Cells 26(12), 3037-3046 (2008)

 

 

Authors

Contact

Scripps Research Institute
10550 N. Torrey Pines Road
La Jolla, CA 92037-1000
USA

Register now!

The latest information directly via newsletter.

To prevent automated spam submissions leave this field empty.