A proteomic study by researchers at the University of Missouri and Harry S. Truman Memorial Veterans' Hospital has identified signaling networks that are potentially involved in the metastasis of breast cancer to the brain, suggesting the need for new treatment paradigms.
The research, published this month in PLoS One, used 2D DIGE followed by LC-MS/MS to identify 12 proteins that differentiated brain-targeting metastatic cells from a parental cell line. When authors compared their results to a pair of previous proteomic and genomic studies, they found that the signaling networks associated with the genes and proteins in all three studies were highly related.
This suggests these larger networks may provide a better target for drug therapy than individual genes or proteins, Vladislav Glinskii, the paper's lead author, told ProteoMonitor this week.
"I was always interested in organ specificity of cancer metastasis," Glinskii said. "Several years ago, I wouldn't have spent my time on proteomics. [It was] not ready to answer questions we were trying to ask as related to cancer metastasis."
Now, however, "the mass spectrometry-based proteomics approaches offer so much potential," he said. "We decided to go this way to try to look what is actually driving organ specificity."
Glinskii said breast cancer brain metastasis is not only a devastating complication, but is also a somewhat underexplored area, because the metastatic process is hard to follow and study.
In their analysis, the Missouri researchers used 2D-DIGE followed by LC-MS/MS with an Applied Biosystems-MDS-Sciex 4000 Qtrap mass spectrometer to identify the proteins over- or under-expressed in a brain-targeting sub-line of metastatic breast cancer cells, MB231-B, relative to the MDA-MB231 parental cell line, finding a signature of 12 differentially expressed proteins.
Glinskii said the team took advantage of similar research by a group led by Spain's Hospital Duran i Reynals, which also used a 2D-DIGE approach, but a different experimental system with another metastatic cell line, MDA-MB-435. That study was published in 2008 in the Journal of Proteome Research.
"We were very interested to look at the differences and similarities between the two sets. Because with two different experimental systems and the same proteomic application, they came up with 19 differentially expressed proteins and we came up with 12, and only two proteins were common between the two," Glinskii said.
This overlap, he said, "is statistically significant." But the group also wanted to compare the findings on the level of associated signaling networks. Using the Ingenuity Pathway Analysis software package, they found that the 12-protein signature was associated with two major signaling networks involving TNFα/TGFβ-, NFϰB-, HSP-70-, TP53-, and IFNγ-associated pathways.
IPA was also used to evaluate associated networks for the Hospital Duran i Reynals study, as well as a genomic analysis by researchers at the Memorial Sloan-Kettering Cancer Center. Comparing the three IPA associations, the Missouri researchers found that networks found for each were "highly related," and all involved the five pathways cited above.
"In fact," Glinskii said, "the similarity was so high that the networks could actually be merged."
This suggests the networks could be an important factor in successful colonization of the brain by metastatic breast cancer, with the implication, the authors wrote, that these networks as a whole may be a more effective potential therapy target than individual genes or proteins.
"We showed that seemingly different sets of data are pointing out the same major signaling pathways associated with breast cancer metastasis," said Glinskii. This suggests that over-expression of different proteins and genes can all lead to activation of the same pathways, he added, noting that is interesting given that drugs are routinely targeted to individual genes or proteins.
Some drugs "do have success," Glinskii said, attacking more narrow targets. "But this success is temporary."
"Picture a key pathway as a plaza," he explained. "There are many streets that lead to that plaza. You shoot down one street or another [and] cancer cells are so robust and so adaptable they inevitably… use other roads, other avenues, to overcome this road block."
"They will emerge, they will get advantage. That's why it's so hard to treat metastases," he said.
It is an important outcome of the study, according to Glinskii, that these results may call for the development of new treatment strategies. "We need to shut down simultaneously major focal points of the pathways, not peripheral proteins that are popping up as over-expressed," he said.
Because the roles of these pathways in a variety of diseases are well established, many drugs targeting them already exist and others are being developed, Glinskii said. He cited several currently available drugs used in inflammatory conditions that target TNFα, such as iInfliximab (sold under the trade name Remicade by Centocor in the US) or etanercept (marketed as Enbrel by Pfizer) as well as a TGFβ2 inhibitor, trabedersen, currently involved in clinical studies of high-grade glioma and advanced pancreatic carcinoma metastasis.
"The point we are trying to make is that [it may be necessary] to develop new paradigms. The means that are available to target the NFϰB pathway or the TNFα/ TGFβ pathway, they might be used together and in combination with cytotoxic therapies in order to achieve success."
"Right now," he added, "this is on a level only of very plausible speculation or hypothesis. It will remain a hypothesis until it is tested experimentally and clinically. But we thought it [was] important to try to make this point and try to initiate the process of developing new treatment paradigms that may require more complex combinations of different agents that are being used now."
Joan Massagué and Manuel Valiente, the MSKCC researchers behind the genomic study that Glinskii and his colleagues compared their findings in their paper, said in an e-mail to ProteoMonitor that the Missouri group's results and the implications for treatment are interesting.
However, they wrote, "Their comments about the implication of these results to develop future and more efficient therapy are interesting, but one would need to know which of these many nodes and common pathways should be targeted. Would a proteomic approach be useful to that end?"
In an e-mail to ProteoMonitor, Glinskii replied, "I do not think that there is a single approach (genetic, genomic, or proteomic) or a single study that could answer this question definitively." He called the question of developing properly targeted treatments a "very complex problem" that may require rigorous investigative efforts of many research groups.
"What makes this problem even more challenging is that these major signaling pathways are highly interconnected and interrelated and control fundamental cellular processes not only in cancer, but also in normal cells," he said.
Glinskii said he and his collaborators have two goals moving forward. On one hand, they hope to try to develop more efficient therapy techniques for breast cancer brain metastasis. "This experimental system has been developed in vivo, so we have the system in hand where we can induce experimental brain tumors, and we… will try different combinations of drugs, probably at first even without chemotherapy, to prove further that these pathways are really critical for the development of brain metastasis," he said.
At the same time, he added, the researchers also need to develop sufficient understanding of the processes involved on the molecular level.
"Without that, every trying with this or that therapeutic agent is a shot into the sky," he said. "If you know how the process works, then we can identify targets, and it's is likely that we will have to go after multiple targets simultaneously."
"We hope this study will induce more interest," he said. "This is just one approach that we've used. But there are many other approaches … and hopefully it will spark additional interest and many other groups will get involved."
Have topics you'd like to see covered in ProteoMonitor? Contact the editor at abonislawski [at] genomeweb [.] com.