Personalizing the Effects of Therapy
By: Carol A. Rouzer, VICB Communications
Published: January 24, 2012
Identifying a signature pattern of protein expression changes offers a new approach to monitor the effects of targeted cancer therapy.
Recent major advances in understanding cancer biology have revealed that cancer is not one, but many diseases. In fact, considering the vast number and variety of mutations present in malignant cells, one may conclude that each cancer patient suffers from his/her own unique disease. These discoveries have led to the realization that successful cancer therapy will require a personalized approach. Following evaluation of a patient’s tumor to determine the exact mutations that drive the malignant phenotype, the physician can identify the therapy that will best target those specific abnormalities. The result is maximal therapeutic efficacy with minimal toxicity to the patient. To achieve this goal, however, requires accurate and efficient methods to evaluate tumor genomes, and we need effective ways to monitor the patient’s response to therapy. Now VICB member Dan Liebler and his student Matthew Myers propose a new approach to assess the effects of targeted cancer chemotherapeutic agents. [M.V. Myers et al. (2011) Mol. Cell. Proteomics, published online Dec. 5, DOI: 10.1074/mcp.M111.015222].
Many cancers carry mutations in growth factor receptor signaling pathways that lead to abnormalities in protein tyrosine phosphorylation. Although directly measurable by immunoblot analysis, protein phosphorylation patterns change rapidly in living cells, and often the target proteins are present in very low amounts. Consequently, the direct monitoring of protein phosphorylation as an indicator of growth factor signaling is technically difficult in a clinical setting. The Leibler group hypothesized that a more reliable approach would be to monitor the changes in protein expression that are controlled by the signaling pathway. Specifically, they proposed that abnormal growth factor signaling should lead to a distinct pattern of protein expression and that therapy designed to block the abnormal signaling should return protein expression levels to more normal levels.
To test their hypothesis, Myers and Liebler focused on the epidermal growth factor (EGF) signaling pathway. When EGF binds to its receptor, the EGFR, the receptor dimerizes and autophosphorylates on multiple tyrosine residues. This activates the EGFR’s tyrosine kinase activity, leading to the phosphorylation and activation of downstream kinases, including extracellular signal-regulated kinase (ERK), phosphatidylinositide 3-kinase (PI3-K), and mammalian target of rapamycin (mTOR) (Figure 1).
Figure 1. EGF signaling pathway. Following binding of EGF, the EGFR dimerizes and authophosphorylates at multiple tyrosine residues. This activates the tyrosine kinase activity of the receptor, leading it to phosphorylate and activate other kinases and transcription factors. Image courtesy of Wikimedia Commons under the GNU Free Documentation License.
Abnormally EGF signaling is common in many tumor cells, including the A431 epidermoid carcinoma cell line. The Liebler group began their investigation by showing that these cells respond to EGF with increased tyrosine phosphorylation of the EGFR and of numerous cellular proteins. This response to EGF was blocked by gefitinib and cetuximab, two EGFR signaling inhibitors that are currently used in cancer therapy. Having established that A431 cells respond to both EGF and the drugs, the Liebler group combined isoelectric focusing with mass spectrometry in a shotgun proteomics approach to identify proteins exhibiting an altered expression in response to EGF that was reversed by the inhibitors. They found 191 proteins that exhibited changes in expression in response to EGF treatment alone, 237 that exhibited changes in expression from gefitinib treatment in the presence of EGF, and 133 that exhibited changes in expression from cetuximab treatment in the presence of EGF (Figure 2). However, when the data were combined, they found only 13 proteins for which EGF-induced expression changes were reversed by both drugs (Figure 2).
Figure 2. Shotgun proteomics analysis of changes in protein expression of control proliferating A431 cells versus cells treated with EGF alone, and between cells treated with EGF alone versus cells treated with EGF plus gefitinib or cetuximab. The results revealed 13 proteins for which EFG-induced expression changes were reversed by both drugs. Reproduced with permission from Myers et al. (2011) Mol. Cell. Proteomics, published online Dec. 5, DOI: 10.1074/mcp.M111.015222. Copyright 2011, ASBMB.
Shotgun proteomics provides a global survey of the protein composition of a complex sample but does not provide precise quantitation. Therefore, the Liebler group used multiple reaction monitoring (MRM) by mass spectrometry to obtain a more precise measurements of the changes in 12 of the 13 EGF inhibitor-responsive proteins identified by shotgun proteomics. The results confirmed that EGF treatment caused a significant change in expression for 10 out of 12 of these proteins. More importantly, the MRM results completely confirmed the shotgun proteomics results with regard to the response to the inhibitors. Only one protein failed to show a significant response in the case of just one of the two inhibitors.
These results suggested that the expression changes in this group of 12 proteins could be used as a “signature” for EGF-dependent signaling and, more significantly, the response to EGF-directed therapy. To further test this hypothesis, the Liebler lab applied MRM analysis to determine the effects of EGF and the two drugs on the expression of the 12 signature proteins in two additional human cell lines. DiFi, a rectal carcinoma cell line, and HCT116, a colorectal carcinoma cell line, had previously been shown to be sensitive and resistant, respectively, to EGF-targeted therapy. The investigators could detect 11 of the 12 proteins in DiFi cells, and all but two of these gave the same responses as A431 cells to EGF-targeted therapy. In contrast, of the 8 proteins detected in HCT116 cells, only 3 gave responses consistent with those observed in A431 cells when the cells were treated with EGF plus the drugs. These results support the hypothesis that the response to EGF-directed therapy can be monitored by a characteristic pattern of changes in this group of 12 proteins.
To test the value of the protein signature approach in a more clinically relevant setting, Myers and Liebler collaborated with Drs. Robert Coffey and Charles Manning to analyze DiFi and HCT116 solid tumors that had been grown in athymic nude mice. The tumor-bearing mice had been treated for one week with either saline or cetuximab before the tumors were removed, fixed in formalin, and embedded in paraffin for microscopic evaluation. Despite this sample processing, which is typical handling for clinical biopsies, the investigators found that they could successfully detect 8 out of 12 of the proteins in both tumor types. For DiFi tumors, 5 out of 8 of the detectable proteins responded to cetuximab treatment in the same way as A431 cells and DiFi cells in culture. In contrast, in HCT116 tumors, only 2 of the 8 detectable proteins exhibited the response to cetuximab observed in A431 cells, and this response was not the same as was observed for HCT116 cells in culture (Figure 3).
Figure 3. Protein expression responses of cultured cells (c) and tumor xenografts (x) of the DiFi and HCT116 cell lines in comparison to A431 cells following EGF-targeted therapy with gefitinib and cetuximab. Green and red colors indicate reduced and increased expression, respectively, in response to the drugs, while black indicates no change, and white indicates that the protein was undetected. Reproduced with permission from Myers et al. (2011) Mol. Cell. Proteomics, published online Dec. 5, DOI: 10.1074/mcp.M111.015222. Copyright 2011, ASBMB.
These highly promising results led Myers and Liebler to extend their collaboration with Coffey and evaluate the protein expression pattern in biopsies from a patient suffering from Menetrier’s disease, a hyperproliferative disorder of the lining of the stomach that results from abnormal EGF signaling. These patients respond well to EGF-directed therapy, which in this case was cetuximab. MRM analysis revealed that 9 of the 12 signature proteins could be detected in biopsy samples from the Menetrier’s disease patient. Of these, the response of 3 to cetuximab therapy were consistent with those observed in A431 cells, but 3 others were discordant, and the remaining 3 showed no significant changes. Although this effect was less striking than in the xenograft tumors, the Liebler group was encouraged by the fact that the 3 concordant proteins in this experiment had been the most consistent in response to EGF-directed therapy in all of the other test systems. They point out that Menetrier’s disease is not a cancer, and may therefore not respond in exactly the same way to EGF-targeted treatment as a malignancy. They remain optimistic that this approach can lead to an effective way to evaluate the response to targeted cancer chemotherapy. However, success will ultimately depend on use of the most relevant discovery models, and a very thorough and systematic statistical validation to define the best possible protein signature for each therapeutic modality.