Simultaneous Inhibition of the Raf/MEK/ERK and PI3K/Atk/mTOR Pathways has Significant Synergistic Effects on Proliferation and Migration in C6 Glioma Cancer Cells
William Henry Bloxham1*, Christine Marshall-Walker2
Student1, Teacher2: Phillips Academy, 180 Main Street, Andover, MA 01810
The effect of combining Raf/ERK pathway inhibitors with PI3K/mTOR pathway inhibitors in C6 glioma brain cancer cells was studied. The Raf/ERK and PI3k/mTOR signaling pathways are both activated by similar sets of growth factor receptors and promote proliferation and migration. Both pathways have been shown to be significantly upregulated in glioma cancers, as compared to healthy cells. Although these pathways have been the focus of much glioma research, limited results have been achieved through inhibition of just one pathway. In this study, migration through a 3D transwell filter and proliferation, as evidencde by Ki-67 immunoreactivity, was measured. The migration and proliferation of C6 cells with simultaneous inhibition of both pathways was compared to migration and proliferation with inhibition of each pathway alone. In almost every case, statistically significant reductions to proliferation and migration were observed when combining inhibitors of both pathways. Furthermore, the data collected supports the hypothesis that each pathway may function as a backup for the other, such that the inhibition of one increase activity in the other.
Glioma cancers are both the most common brain cancers in humans and the most deadly. Gliomas are tumors originating from glial brain cells and are characterized by rapid and uncontrolled proliferation, resistance to immune responses, high motility and invasiveness, and low patient suvival.1,2 As many as 1 in 10,000 persons suffer from glioma cancers.1 Unfortunately, gliomas are very difficult to treat and median survival after diagnosis is less than 12 months.3 Even if 98% of the tumor volume is removed through surgery, median survival only increases to 13 months.4
The two most important characteristics of glioma cancers that contribute to the low survival rate in patients are rapid proliferation and unrestricted migration. Like virtually every other cancer, gliomas feature uncontrolled proliferation, or growth and division, as their direct and primary danger to their host. Glioma tumors can triple in volume in as little as two weeks in a patient. Migration is the next most significant characteristic of glioma cancers and is the main reason surgery does not see the success that it does in other cancers.4 It is essentially impossible to completely remove a glioma cancer from the brain via surgery because the cancer disperses itself across the brain as individual cells migrate through the extracellular matrix. Rapid proliferation of these diffuse cells virtually guarantees reemergence of the disease from other parts of the brain.1,5 A glioma cancer where the original tumor has spread to other parts of the brain through migration is shown in Figure 1.6 Rapid proliferation and migration would need to be addressed by any successful glioma cancer therapy.
The World Health Organization ranks glioma cancers from Grade I to Grade IV.1 Under this system, Grade I is the least deadly with a medium survival of 5 to 10 years, and Grade IV, glioblastoma, is the most deadly with a medium survival around 10 months.1 Glioblastoma is also the most common form of glioma, contributing to the overall deadliness of the desease.7 Although gliomas vary greatly even within the same grade, certain mutations have been associated with different grades of glioma. For example, both high and low grade gliomas almost always have at least one mutated growth factor receptor.1 In healthy cells growth factor receptors allow the body to control cell growth, but in gliomas these receptors are essentially always on, constantly telling the cell to grow. Biochemically, these growth factor receptors activate various signal pathways, including the Raf/MEK/ERK and PI3K/Akt/mTOR pathways.2,3,8 Higher grade gliomas often feature the deletion of natural inhibitors and regulators of these two pathways and mutations of multiple growth factor receptors.1,2
The Raf/ERK Pathway includes Raf, mitogen-activated protein kinase kinases 1 and 2 (MEK), and extracellular signal-regulated protein kinases 1 and 2 (ERK). The pathway is activated by growth factor receptors, aided by other proteins, phosphorylating Raf at serine-338. Raf phosphorylates MEK at serine-217 and serine-221, and MEK phosphorylates ERK at Threonine-202 and Tyrosine-204.2,9,10 Phosphorylated ERK then promotes proliferation, growth, DNA replication, migration, and other effects by activating over a hundred different proteins throughout the cell, such as cyclin D1, p90RSK, and myosin light chain kinase.2,3,10-12 This is a very important pathway in the study of glioma cancers as nearly every case of glioma features overactive ERK and MEK.2
The PI3K/Akt/mTOR pathway is another biochemical pathway associated with migration, proliferation, survival, and DNA replication in glioma cells.2 This pathway includes phosphatidylinositol 3-kinase (PI3K), protein kinase B (Akt), and mammalian target of rapamycin (mTOR). In this pathway, the same growth factor receptors that activate the Raf/ERK pathway activate PI3K, which activates Akt, which then activates mTOR. mTOR activates a variety of proteins for effects similar to those of ERK.2 Thus, the PI3K/mTOR pathway behaves very similarly to the Raf/ERK pathway. The Raf/ERK and PI3K/mTOR pathways are summarized in Figure 2.
Due to the up regulation of the PI3K/mTOR and Raf/ERK pathways in glioma cancers and their connection to some of the most important and most dangerous properties of glioma cancer, inhibition of either pathway was expected to be a highly effective glioma treatment. Phase II clinical studies, however, have shown that only a minority of patients respond to either PI3K/mTOR pathway inhibition or Raf/ERK pathway inhibition.14-16 These unexpected results have been a significant setback to glioma treatment development.
One explanation for why inhibition of a signalling pathway has been largely unsuccessful is that the two pathways may interact in some manner. One mechanism of interaction would be that the two pathways compete for activation by the growth receptors such that the inhibition of one pathway allows the other to achieve greater activation. Another mechanism would be a shared negative feedback loop, such that a signal reaching the end of either pathway prompts the production of an inhibitor of the other pathway or of both pathways. There is evidence that p70S6K plays a role in a shared feedback loop.2,17 Another more complicated mechanism of interaction would be that the signal is able to crossover between pathways midway. This crossover might happen in both directions or in just one direction and at one point or at multiple points. In all these mechanisms, the interaction allows a signal to reroute around any signal inhibitor.
Due to these interactions, it is possible that simultaneous inhibition of the Raf/ERK and PI3K/mTOR pathways might produce more effective cancer treatments. This simultaneous inhibition would block both pathways and could therefore prevent a signal blocked by inhibition of one pathway from utilizing the other pathway. Recently, multiple studies have tested similar hypotheses in other forms of cancer and seen encouraging results.18-21
This study examined the simultaneous inhibition of the Raf/ERK and PI3K/mTOR pathways in rat C6 glioma cells to determine whether combining inhibitors produces significantly greater effects than the sum of the effects of the inhibitors acting alone. In almost all cased, statistically significant reductions to proliferation and migration were observed when combining inhibitors of both pathways. The data may refine the current model of biochemistical interactions between the two pathways.
Materials and Methods
C6 glioma cells were a kind gift from Peter Carroll (Columbia University). These cells are a common model system for glioma research and have been shown to have similar properties to human glioblastomas.22,23 The cells were originally harvested from a chemically induced rat astrocytoma (grade III glioma)23,24 and are grown at 37.5°C, 5% CO2, and high humidity in a DMEM medium with 5% Fetal Bovine Serum (FBS) and a Penicillin/Streptomycin antibiotic.
In these experiments, activation of Raf is inhibited by Sorafenib from SelleckChem (S1040). Activation of ERK is inhibited by U0126 from Cell Signaling (#9903). Activation of PI3K is inhibited by Wortmannin from Invitrogen (PHZ1301). Activation of mTOR is inhibited by Rapamycin from Calbiochem. Inhibitors where tested both on their own and in combinations of one Ras/ERK pathway inhibitor and one PI3K/mTOR pathway inhibitor. The inhibitors used in this study and the proteins they inhibit are noted in Figure 2. All experiments used concentrations of 1μM for Sorafenib, 2μM for U0126, 3μM for Wortmannin, and .25μg/mL for Rapamycin.
For proliferation assays, 30,000 cells were plated onto 18mm glass coverslips (coated with poly-L-lysine) and grown in 1mL of DMEM with 5% FBS and Penicillin/Streptomycin antibiotic. After 24 hours of cells growth, inhibitors were added, and after another 6 hours cells were fixed with 4% paraformaldehyde. Cells were stained with Hoeschst and an anti-Ki67 antibody from Spring Bioscience (M306, 1:200), with a DyLight 549-conjugated anti-Rabbit IgG secondary from Jackson ImmunoResearch (711-505-152, 1:100). Hoechst bonds to the phosphate backbone of DNA, staining the nuclei of all cells blue. Ki67 is a nuclear protein that appears exclusively during the cellular division cycle while being completely absent during interphase.25 The fraction of cells expressing Ki67 is, therefore, used here to indicate the rate of proliferation in a sample of cells.
For migration assays Falcon HTS FluoroBlok Inserts were used. .5mL of DMEM with FBS was added to each well below the filter and .2mL of DMEM without FBS was added above each filter. 500,000 cells were added onto the top of each filter. FBS here acted as a chemo-attractant, prompting cells to attempt to migrate down through the filter. Inhibitors were added both above and below each filter at the same time the cells where added. The cells were given 6 hours to grow and migrate, after which they were fixed with paraformaldehyde and stained with Hoechst.
Cells were viewed under 20x magnification on a fluorescent microscope, which was a kind gift from Harvard Medical School. Images were captured using OpenLab Software as 8-bit black-and-white Tiff files. All reproductions of microscope images in this study have been false colored using iPhoto ‘11 and Microsoft Word 2008 for Mac, with blue representing signals from Hoechst staining and red representing signals from anti-Ki67 antibody staining.
Linear multiple regressions were run in Minitab 16 for Windows 7. P-values were used to determine statistically significance, which was defined as a P of less than .025. All graphs in this study were created in Microsoft Excel 2008 for Mac using data either from Minitab or directly from the experiments.
Four proliferation assays and three migration assays were run and multiple fields of view were captured from each plate or filter. In total, 425 photomicrographs were captured and 83,241 cells were counted. Figures 3 and 4 are example photomicrographs from proliferation assays and migration assays, showing the control, U0126, Wortmannin, and U0126 and Wortmannin treatments, respectively. The mean values and standard errors for each of the data outputs from each assay are shown in Figure 5.
Figure 3. Microscope images from proliferation assay 1. Proliferating cells appear red. Cells not proliferating appear blue.
Figure 4. Microscope images from migration assay 1.
Figure 5. Mean proliferation assay cell count, proliferation assay percent proliferation, and migration assay cell count by assay with standard errors.
Statistical regressions were used to simplify and normalize the data. The cell count data was fit to the following equation using a linear multiple regression: ln(Cell Count) = 5.95A1 + 4.90A2 + 4.57A3 + 5.11A4 + .334BSora + .276BU0126 - .276BWort - .137BRapa - .842BS&W - .625BS&R - .703BU&W - .550BU&R, where An equals 1 if assay number n produced a datum and equals 0 if not, where BX equals 1 if reagent x was used (either alone or in combination with another reagent) and equals 0 if not, and where BX&Y equals 1 if both reagent X and reagent Y were used and equals 0 if not. The p-values for the coefficients on the BX&Y variables are all less than 0.025. This regression equation is mathematically equivalent to Cell Count = (383*A1 + 134*A2 + 97*A3 + 166*A4) * (1.41Sora) * (1.32U0126) * (.984Wort) * (.872Rapa) * (.431S&W) * (.535S&R) * (.495U&W) * (.577U&R), where the variables referring to the presence of inhibitors are shown as just their subscripts. The proliferation rate data was fit to the following equation: ln(Proliferation Rate) = - .287A1 - .314A2 - .297A4 + .0921BSora + .0327BU0126 - .0297BWort - .0235BRapa - .133BS&W - .197BS&R - .135BU&W - .0625BU&R, where variable are assigned values in the same manner as the cell count regression. The p-values for the coefficients on the BX&Y terms are all less then 0.025 with the exception of BU&R. This equation is equivalent to Proliferation Rate = (75.1%*A1 + 73.1%*A2 + 74.3%*A4) * (1.10Sora) * (1.03U0126) * (.971Wort) * (.977Rapa) * (.875S&W) * (.821S&R) * (.874U&W) * (.939U&R), where variables are similarly abbreviated.
The migration data was fit to the following equation: ln(Migration) = 5.17A1 + 6.78A2 +5.60A3 - .399BSora + .035BU0126 - .400BWort - .154BRapa - 1.02BS&W - .264BS&R - 1.80BU&W - 1.07BU&R. The p-values for the coefficients on the BX&Y terms are all less than 0.025 with the exception of BS&R. The equation is mathematically equivalent to Migration = (176*A1 + 880*A2 + 270*A3) * (.671Sora) * (1.036U0126) * (.670Wort) * (.857Rapa) * (.361S&W) * (.768S&R) * (.165U&W) * (.343U&R). Normalized forms of the proliferation assay and migration assay regression results are shown in Figure 6A (cell count regression), Figure 6B (proliferation rate regression), Figure 6C (migration regression), and Figure 6D (migration divided by cell count) as the blue bars in the first three graphs.
Figure 6. Normalized regressions of cell count(6A), percent proliferation(6B), migration count(6C), and migration divided by cell count(6D). The significance of the migration divided by cell count graph is explained in the discussion. White bars represent the sum of the effects of inhibitors acting alone. P-values are for the differences between the white bars and blue bars.
In all three regressions, all the coefficients on the terms for the extra-additive, synergistic effects of combining inhibitors (i.e. the effects beyond what would be predicted from simply adding the effects of each inhibitor alone) were negative. Furthermore, ten out of twelve had P-values less than 0.025. It can therefore be concluded that, in most circumstances, combining inhibitors from the Raf/ERK and PI3K/mTOR pathways has reductions to cell count, proliferation, and migration that are statistically significantly greater than the sum of the effects from each inhibitor alone. This can be seen visually in Figure 6, where for the combination trials the white bar represents the hypothetical scenario of no extra-additive effects and the blue bar reflects the actual observations.
One of the most surprising set of results from this research is the effect of Raf/ERK inhibition on cell count and proliferation rate. Alone, Sorafenib and U0126 increased cell count and proliferation rate. This might suggest that the Raf/ERK pathway does not promote survival and proliferation but instead limits these characteristics. In all the inhibitor combinations, however, the addition of Sorafenib or U0126 greatly enhanced the effects of Wortmannin and Rapamycin. At first these results seem somewhat contradictory, but they can actually be explained by a relatively simple model of pathway interaction. Although both pathways are known to directly promote survival and proliferation, there is no reason to assume they do so with equal intensities. Thus, if the PI3K/mTOR pathway were stronger or more efficient at promoting survival and proliferation, the only interaction necessary to explain the proliferation assay results is a simple crosstalk by which inhibition of one increases activity in the other. Under the proposed model, Raf/ERK inhibition causes increased activity in the stronger PI3K/mTOR pathway, which more than cancels out any reduction to the proliferation dependent on the Raf/ERK pathway. Meanwhile, PI3K/mTOR pathway inhibition reduces the stronger proliferation and survival promotion but is limited in effectiveness because the Raf/ERK pathway becomes more activated and makes up for some of the difference. Together, inhibition of both pathways limits the proliferation promotions at the end of each pathway to achieve significant reductions to cell count and proliferation rate. It should be noted that the proliferation assay data fit this proposed model very well.
An important consideration when looking at the proliferation assay data is whether reductions in cell count are the result of decreases in proliferation or increases in cell death. Although further testing is necessary to come to a definitive conclusion on this question, that fact the cell count and proliferation rate are highly correlated across the entire data set suggests that cell death is not a significant factor.
The first thing that must be considered before analyzing the migration assay data is whether migration counts should be divided by the fractional changes in the proliferation assay cell counts. The argument for this adjustment would be that if there were fewer cells one would see fewer cells migrating regardless of whether migration is being specifically blocked. The argument against would be that most cells may not have time to divide and then migrate and that this adjustment is, therefore, an over-compensation. Both data sets are shown in Figure 6. This discussion will assume the migration-specific effects lie somewhere between what is shown in the two data sets.
The most surprising thing observed in the migration assay data is that, especially in the cell count adjusted data set, combinations involving Sorafenib do not lower migration much more than Sorafenib alone. This appears to upend the model derived from the proliferation assay data, and cannot be explained by unequal promotion of migration. The question is how to adjust the model for migration assay data without affecting its very good fit for the proliferation assay data. The simplest solution is the following: somewhere between Raf and ERK the Raf/ERK pathway has an offshoot pathway, such that the migration promotion at the end of the PI3K/mTOR pathway is dependent on activation by both the PI3K/mTOR pathway and this offshoot pathway. Figure 7 shows a diagram of the Raf/ERK and PI3K/mTOR pathways that has been refined according to the analysis in this study. Raf inhibition by Sorafenib, therefore, blocks both migration promotions making PI3K/mTOR inhibition redundant. U0126 meanwhile only blocks one migration promotion giving it the expected extra-additive effects when combined with PI3K/mTOR inhibition. There are, of course, possible variations on this model. This offshoot pathway could also be a separate pathway that uses Raf at a different site, or Sorafenib might inhibit other proteins within the cell, possibly including one of the proteins connecting mTOR to the promotion of migration.
A better understanding of the interactions of the Raf/ERK and PI3K/mTOR pathways has the potential to greatly advance treatments for glioma cancers, which remain difficult to treat. The data in this study suggest that combining inhibition of each pathway has large effects on migration, proliferation, and survival that are usually statistically significantly greater than the sum of the effects of each inhibitor alone. Understanding of this phenomenon and better describing the biochemistry of the Raf/ERK and PI3k/mTOR pathways has the potential to transform ineffective cancer drugs into successful treatments. The data here suggests a strong crosstalk between the two pathways as well as a potential bridge between the middle of the Raf/ERK pathway and the migration effects promoted by mTOR. Further research into these interactions between these pathways could highlight new targets for preventing glioma cell migration and proliferation.
1. Rich, J. & Bigner, D. (2004) Development of novel targeted therapies in the treatment of malignant glioma. Nature Reviews Drug Discovery vol. 1, 430-446.
2. Sunayama, J. et. al. (2010) Crosstalk between the PI3K/mTOR and MEK/ERK pathways involved in the maintenance of Self-Renewal and tumorigenicity of glioblastoma stem-like cells. Stem Cells vol. 28, 1930-1939.
3. Zohrabian, V. et. al. (2009) Rho/ROCK and MAPK signaling pathways are involved in glioblastoma cell migration and proliferation. Anticancer Research vol. 29, 119-123.
4. Lacroiz, M. et. al. (1995) A multivariate analysis of 416 patients with glioblastoma multiforme: prognosis, extent of resection, and survival. Journal of Neurosurgery vol. 95, 190-198.
5. Holland, E. (2000) Glioblastoma multiforme: the terminator. Proceedings of the National Academy of Sciences of the United States of America vol. 97, 6242-6244.
6. Bhat, A. et. al. (2010) Brain cancer and pesticide relationship in orchard farmers of Kashmir. Indian Journal of Occupational and Environmental Medicine vol. 14, 78-86.
7. Wrensch, M. et. al. (2002) Epidermiology of primary brain tumors: current concepts and review of the literature. Neuro-oncol vol. 4, 278-299.
8. Lind, C. et. al. (2006) The mitogen-activated/extracellular signal regulated kinase kinase 1/2 inhibtor U0126 induces glial fibrillary acidic protein expression and reduces the proliferation and migration of C6 glioma cells. Neuroscience vol. 141, 1925-1933.
9. Marais, R. et al. (1995) Ras recruits Raf-1 to the plasma membrane for activation by tyrosine phosphorylation. EMBO Journal vol. 14, 3136-3145.
10. Cell Signal. U0126 Datasheet. http://www.cellsignal.com/pdf/9903.pdf.
11. Thompson, N. & Lyons, L. (2005) Recent progress in targeting the Raf/MEK/ERK pathway with inhibitors in cancer drug discovery. Current Opinion in Pharmacology vol. 5, 350-356.
12. McCubrey, J. et. al. (2007) Roles of the Raf/MEK/ERK pathway in cell growth, malignant transformation and drug resistance. Biochimica et Biophysica Acta vol. 1773, 1263-1284.
13. Mukasa, A., Ligon, K., et. al. (2010) Mutant EGFR is required for maintenance of glioma growth in vivo, and its ablation leads to escape from receptor dependence. Proceedings of the National Academy of Sciences of the United States of America vol. 107, 2616-2621.
14. Sabatini, David M. (2006) mTOR and cancer: insights into a complex relationship. Nature Reviews Cancer vol. 6, 729-734.
15. Galanis, Evanthia et al. (2005) Phase II Trial of Temisirolimus (CCI-779) in Recurrent Glioblastoma Multiforme: A North Central Cancer Treatment Group Study. Journal of Clinical Oncology vol. 23, 5294-5304.
16. Cloughesy, Timothy F. et al. (2006) Phase II Trial of Tipifarnib in Patients With Recurrent Malignant Glioma Either Receiving or Not Receiving Enzyme-Inducing Antiepileptic Drugs: A North American Brain Tumor Consortium Study. Journal of Clinical Oncology vol. 24, 3651-3656.
17. Zimmermann, S. & Moeling, K. (1999) Phosphorylation and regulation of Raf by Akt. Science vol. 286, 1741-1744.
18. Zhang, Yan-Jie et al. (2009) Combined Inhibition of MEK and mTOR Signaling Inhibits Initiation and Progression of Colorectal Cancer. Cancer Investigation vol. 27, 273-285.
19. Chang, Qing et al. (2009) Effects of combined inhibition of MEK and mTOR on downstream signaling and tumor growth in pancreatic cancer xenograft models. Cancer Biology and Therapy vol. 8, 1893-1901.
20. Kawaguchi, Wakae et al. (2007) Simultaneous inhibition of the mitogen-activated kinase kinase and phosphatidylinositol 3’-kinase pathways enhances sensitivity to paclitaxel in ovarian carcinoma. Cancer Science vol. 98, 2002-2008.
21. Schreeder, M.T., et al. (2008) Phase I multicenter trials of perifosine in combination with sorafenib for patients with advanced cancers including renal and cell carcinoma. Journal of Clinical Oncology vol. 26, 16024.
22. Grobben, B. et. al. (2002) Rat C6 glioma as an experimental model system for the study of glioblastoma growth and invasion. Cell and Tissue Research vol. 310, 257-270.
23. Amberger, V. et. al. (1998) Spreading and migration of human glioma and rat C6 cells on central nervous system myelin in vitro is correlated with tumor malignancy and involves a metalloproteolytic activity. Cancer Research vol. 58, 149-158.
25. Scholzen, T et Gerdes, J. (2000) The Ki-67 Protein: From The Known to The Unknown. Journal of Cellular Physiology vol. 182, 311-322.
The C6 cells were a kind gift from Dr. Peter Canoll of Columbia University.