The Dyson Laboratory studies the role of the retinoblastoma tumor suppressor (pRB). pRB is expressed in most cell types and its functions enable cells to stop dividing. pRB is inactivated in many types of cancer; a change that is thought to be an important step in tumor progression. We have three main goals: we want to understand the molecular details of how pRB acts, we want to know how the inactivation of pRB changes the cell, and we are using these insights to target tumor cells.
My group investigates the mechanisms that limit cell proliferation in normal cells and the ways that these controls are eroded in cancer cells. Our research focuses on the protein product of the retinoblastoma susceptibility gene (RB1) and its target, the E2F transcription factor. E2F controls the expression of a large number of target genes that are needed for cell proliferation. This transcriptional program is activated when normal cells are instructed to divide, but it is deregulated in tumor cells, providing a cellular environment that is permissive for uncontrolled proliferation. pRB has multiple activities but one of its most important roles is to limit the transcription of E2F targets. As a result, most tumor cells select for changes that compromise pRB function. Our current research program spans four different aspects of pRB/E2F biology.
a. Dissecting the molecular functions of pRB
pRB’s mechanism of action is an enigma. Ironically, this is not because little is known; instead pRB has been linked to hundreds of proteins and implicated in many cellular processes, and this has created a confusing picture. Currently, the major issues are: which of pRB’s activities and interactions are functionally important, how are all of these activities regulated, and how is the pool of pRB divided been all of its potential targets?
A key technical obstacle that has prevented resolution of these issues is that it has not yet been possible to purify endogenous pRB complexes from human cells and to profile the components. Because of this, it has been unclear precisely which proteins are targeted by pRB at any given moment. Very recently, we have solved this problem. We are now able to isolate the endogenous pRB complexes from normal cells and, in collaboration with the Haas lab we are using Mass Spectrometry to obtain detailed snapshots of pRB in action. In addition, we have taken advantage of an shRNA resource at the MGH Cancer Center and have built a library of constructs that target each of the 230 proteins that have been reported to physically interact with pRB. Together, these tools give us a great opportunity to identify the proteins that interact with pRB and to dissect the processes that are the molecular basis for pRB function.
b. Proteomic profiles give a new perspective on the effects of RB1 mutation
The activity of pRB, and changes in E2F regulation have traditionally been measured by quantifying the levels of RNA transcripts synthesized from genes that are directly controlled by pRB/E2F proteins. pRB inactivation changes the transcription of a vast number of genes (between hundreds and thousands) and it has not been feasible to ask whether most of these changes in mRNA levels affect protein levels. For over two decades it has been assumed that the changes in levels of RNA transcripts in RB1 mutant cells are generally followed by similar changes in protein synthesis; and that the RNA signatures associated with pRBB loss/E2F activation give us a meaningful picture of the consequences of RB1 inactivatedmutation.
Taking advantage of the latest developments in quantitative proteomics we have been able to move beyond this stumbling block. We generated proteomic profiles of tissues shortly after ablation of the mouse Rb1 gene. Remarkably, when the protein changes were compared with the changes detected by RNA-sequencing we discovered that the two types of profiles give strikingly different answers. This new data shows that mutation of Rb1 has effects on protein levels that are very different from, and far more extensive than, the changes predicted from RNA data. Taken together results indicate that the transcripts upregulated by Rb1/RB1 loss are subject to extensive post-transcriptional control. This mechanism of regulation is not well understood but is clearly an important aspect of E2F biology.
One of the surprising features of the proteomic data is that the most consistent change in different Rb1 mutant tissues is a decrease in mitochondrial proteins. Accordingly we discovered that RB1 mutant cells have a proliferation disadvantage when they are grown in low-glucose conditions that put extra demands on mitochondrial function. In such conditions, pRB-deficient cells are more sensitive to mitochondrial poisons. These results are exciting because they show that the mutation of Rb1/RB1 changes the cell in ways that had not previously been suspected. The protein signatures may provide useful biomarkers in tumor samples and, most importantly, they may reveal new ways to target tumor cells.
c. Targeting tumor cells with RB1 mutations.
The ultimate goal of all RB research is to use the information gleaned from molecular and mechanistic studies to improve cancer treatments. The activity of pRB can be altered in several different ways and pRB is functionally compromised in most types of cancer. However, in three types of tumor (retinoblastoma, osteosarcoma and small cell lung cancer) RB1 is almost always mutated. We infer from this that the complete elimination of pRB activity is especially significant in these tumors.
We have compiled a list of changes that may represent vulnerabilities that can be targeted in RB1 mutant tumors. This analysis has been greatly enhanced by a collaboration with the Broad Institute that has enabled us to examine the results of high-throughput screens that have systematically profiled cancer cell lines for shRNAs that affect cell proliferation. By classifying the cell lines according by their RB1 status we have identified proteins that are more important for the proliferation of RB1 mutant cells than other cells. Our results suggest that there may not be a single weakness that is universal to all RB1 mutant cancers, but that different types of RB1 mutant tumors need to be targeted in specific ways. Perhaps most importantly, we now have well-justified lists of candidate genes to assay in different types of RB1 mutant tumors.
d. The biological consequences of eliminating E2F activity.
The deregulation of E2F is an important consequence of pRB-inactivation and inhibition of E2F activity has been widely discussed as a potential therapeutic strategy. To understand the consequences of global inhibition of E2F activity we have taken advantage of the relative simplicity of the Drosophila E2F/RB network and are carrying out a detailed analysis of dDP mutant animals, in which loss of dDP eliminates E2F function. dDP mutation causes extensive transcriptional changes. Proteomic profiles of dDP mutant animals reveal changes in protein levels that are different from, and even more extensive than, the transcriptional changes. By integrating the RNA and protein profiles with ChIP data showing the genome-wide distribution of E2F and RBF proteins, we have identified a small set of direct dE2F/dDP target genes that display strong changes in both RNA and protein levels in dDP mutant tissues. These candidates are currently being tested in genetic studies to identify the direct targets of dE2F/dDP proteins that are functionally significant drivers of dDP mutant phenotypes.