AudioHelicase podcast: Whitehead's Robert Weinberg on cancer research's past, present, and future

Whitehead Member Robert Weinberg

22nd January, 2018

Tags: Weinberg Lab Stem Cells + Therapeutic CloningCancerGenetics + Genomics

From his discovery of the first oncogene and tumor suppressor, to his work revealing critical aspects of the mutational basis of cancer, and more recently its metastatic behavior, Whitehead Founding Member Robert Weinberg’s work has been foundational in our understanding of cancer biology. 





Lisa Girard: Welcome to AudioHelicase, the podcast of Whitehead Institute, unwinding the science and the people behind some of the Institute’s most exciting discoveries. I’m Lisa Girard, Director of Communications at Whitehead Institute. In this episode, I’m talking with Robert Weinberg, a Founding Member of Whitehead Institute and a professor of biology at MIT. From his discovery of the first oncogene and tumor suppressor, to his work revealing critical aspects of the mutational basis of cancer, and more recently its metastatic behavior, Weinberg’s work has been foundational in our understanding of cancer biology. For his many contributions, Bob has received among his numerous awards, the National Medal of Science, the Breakthrough Prize in Life Sciences, and the Wolf Prize in Medicine. Speaking with Bob recently, I asked him to tell me a little bit more about some of his earliest work in the field.

Robert Weinberg: Our work in this area started a very long time ago, in the mid-1970s, when people in my lab were working on tumor viruses – those that actually affect rodents and create among other things sarcomas. And in 1978, we undertook an experiment that ultimately would prove to have been, I think, the most important work in my own career, which addressed the following question: how does a chemical carcinogen actually induce a normal cell to become a cancer cell?

And so in 1978, we began a series of experiments that demonstrated within a year that within the DNA—within the genes of cancer cells—there lay mutant genes that were mutated because of the actions of the carcinogens. And once they were mutated, began then to orchestrate the abnormal, malignant behavior of the cancer cells. And this proved for the first time that cancer is truly a genetic disease—a disease of cells that have suffered different types of mutational damage to their genomes and thereafter begin to grow abnormally.

Several years later, actually in the mid-1980s, we were involved in the isolation of another kind of gene, the first gene in the early work being called an oncogene, the other kind of gene being termed a tumor suppressor gene. In this case, the first gene was a RAS oncogene. The gene isolated in the mid-1980s was a tumor suppressor gene involved in the development of the rare eye disease retinoblastoma. The first set of genes, the oncogenes, work like an accelerator pedal stuck to the floor of a car, whereas the second gene, the retinoblastoma gene, works as a brake lining. And as we soon found out, and others found out as well, cancer cells carry different kinds of mutations that hyperactivate the oncogenes and inactivate the tumor suppressor genes—in other words, activate the accelerators and inactivate the brake linings. And this led very shortly thereafter to experiments in our hands, which demonstrated indeed, already starting in the early 1980s actually, that cancer cells carry multiple mutated genes in their genomes and these multiple mutated genes must collaborate with one another to create a malignant, actively growing cancer cell.

This work continued in the 1990s. By then, we had become involved in isolating yet another gene involved in the pathogenesis, the development of human cancer, called telomerase. And we exploited that in 1999 to demonstrate for the first time that we could actually take normal human cells and convert them into cancer cells through the introduction of a series of genes—as it turned out, five in all. And this demonstrated in a direct and experimentally irrefutable way that cancer cells carry multiple mutated genes in their genomes and without this suite of multiple mutated genes, such cells really cannot grow and form tumors.

This led to the illusion in my mind that we kind of understood how primary cancers are formed. That is to say, by introducing, for example, this group of five genes into a normal human cell, one could make a cell that was capable of launching the outgrowth of a primary tumor. But there still was one embarrassing shortfall in our overall body of knowledge, and that is how metastases form. That is to say, how cancer cells in a primary tumor are able to spread to distant sites in the body where they are able to found new tumor colonies, the so-called metastases.

Girard: And why is understanding metastasis so important?

Metastases are more than just an academic side issue in the context of cancer research because about 90% of cancer-associated mortality derives from the growth of these metastases rather than from the actions of the primary tumor. And so in 2002, we began to look for genes that could propel primary tumor cells to distant sites in the body that could enable them to invade and to metastasize. And that led within several years to the discovery of a gene –called Twist – which empowers the breast cancer cells to invade the site of primary tumor formation into adjacent tissues and thereafter spread, for example, to form new colonies in the lungs—metastatic colonies.

What was particularly striking about this new gene was that it had been associated earlier with critical steps in the early development of the Drosophila embryo, more specifically, it was involved in a cell biological program that was called the epithelial mesenchymal transition (EMT), a program that converts cells that grow much like skin cells—epithelial cells in a layer—into connective tissue cells that have much different properties. And as we soon learned, simply by activating this EMT program, one could turn a primary tumor cell into one that had acquired invasive and metastatic properties. And so since that time, which is now more than 15 years, we have been working very intensively on precisely how carcinoma cells are able to spread to distant sites in the body by focusing on this EMT program. Soon others joined in that as well, and this is now a very actively growing field in the whole area of cancer research.

Girard: There has been a lot of attention in recent years on cancer stem cells. What are these, and what role do they play in the creation of tumors?

Weinberg: In 2008 we discovered yet another aspect of the EMT program that we, at least, had previously not anticipated. And that was if one takes a more epithelial carcinoma cell and activates within that carcinoma cell the EMT program, this cell also acquires the ability to become a cancer stem cell. Cancer stem cells are subpopulations of cells within a given tumor that are capable of forming new tumors in a mouse. In other words, that if one looks at the populations of cells within a given tumor, they may actually all share the exactly the same set of genetic mutations—the same mutant genomes—but they behave quite differently—some being capable of seeding new tumors, so-called tumor initiating cells—the great majority of the cells in the same tumor being incapable of doing so. And these tumor-initiating cells became called cancer stem cells, and their involvement, in among other things metastasis, seems to be critical. After all, when a cell disseminates, travels, from a primary tumor to a distant tissue, once it’s there, ostensibly it would require some type of tumor-initiating capability in order to function, to serve, as the founder of a new colony of cancer cells. And so this study of cancer stem cells also became rapidly connected with the whole understanding of metastasis.

One of the main problems these days in clinically treating metastases via various kinds of anticancer therapies is the properties of these growths. Often there may be dozens of metastases scattered throughout the body, many in sites where they cannot be safely resected from the tissue. The presence of multiple metastases that one can see suggests as well the presence of myriad other metastases that somehow fly under the radar screen. Unanswered by what I have just told you is the question of how different the cells within the metastasis are from the cells in the founding mother tumor. And there the answer is in part still unknown. It’s clear that the genomes of metastases are very similar to the genomes of the primary cancer cells. And therefore it’s currently plausibly the case that when primary carcinoma cells leave a primary tumor, they don’t change the sequences of DNA in their genomes, instead they activate this non-genetic program—the epithelial mesenchymal transition, or EMT—and the actions of this program then empowers them to invade and to get into the bloodstream, and to disseminate to distant tissues.

Girard: I’ve heard trying to remove a tumor without addressing cancer stem cells has been likened to pulling out a weed while leaving the roots behind. So why are they so difficult to eliminate from tumors?

Weinberg: The cancer stem cells are dangerous, not only because they can be the engines of metastatic spread, but also because their quasi-mesenchymal attributes render them more resistant to elimination by existing chemo- and radiotherapeutic treatments. In the long run, it would be highly desirable to eliminate the cancer stem cells from within tumors or even from within metastases. How to do so is not yet totally clear. One way to do so would be to develop drugs that preferentially kill cancer stem cells. Another alternative strategy is to induce the cancer stem cells to exit the stem cell state and to become more differentiated and thereby become non stem cells that no longer exhibit the resistance to chemotherapeutic compounds. None of them as yet has identified a compound that is brilliantly successful in inducing these changes, but there is reason to be optimistic that such compounds will indeed be developed in coming years.

Girard: So tumors are comprised of cancer stem cells and non stem cells. Are there more types of cells within a tumor, and how does this affect how cancers are treated?

Weinberg: The major problem that confronts cancer researchers comes from the fact that as cancers develop, the cells within the cancer become quite diversified often and different subpopulations of cells within a primary tumor, for example, or even within a metastasis, acquire different types of traits. It means that for example whereas an initially highly responsive tumor that wilts away under the attack of chemotherapy, may be followed six months later by the eruption of a closely related tumor cell that somehow has evaded, eluded destruction by the initial chemotherapy and now represents the engine that drives metastatic relapse or even primary tumor relapse.

This highlights yet another unsolved problem for cancer biology, and that is, how do disseminated, unapparent metastatic deposits in different sites of the body persist for many years, only after some delay, to suddenly erupt and to create an entirely new macroscopic tumor that is readily detectable by available clinical diagnostic procedures. Right now my group is increasingly working on how these metastatic deposits, which are sometimes termed dormant, are able to suddenly start growing again.

There are yet other problems, which some of our work has addressed. For example, we found that if one has dormant cancer cells implanted on one side of the mouse, and one creates a wound on the other side of the mouse, often the dormant cancer cells, which would otherwise be kept under the control of the immune system, suddenly erupt. And the take home lesson for us from that is, that surgery is itself not an unalloyed benefit for the patient because there must be instances where the post-surgical wound healing response actually is responsible for provoking the outgrowth of a tumor that otherwise would never show itself.

Girard: So recently, there’s been a lot of focus on personalized medicine—using a person’s individual genetics to tailor their cancer treatment. And it seems like designing treatment based on the apparent drug sensitivity of someone’s particular tumor would be a huge advantage in therapy. So is it really the “magic bullet”?

Weinberg: Only a minority of patients have mutations in their tumors that actually render those tumors susceptible to be treated by a targeted drug. Even more important is the fact that the response that that patient gets from the initial, highly successful targeted therapy is followed up six, eight, ten months later by the resurgence of the tumor because there are certain subsets of cells in the primary tumor that were never successfully eliminated by the targeted drug. And so the question of whether sequencing tumor genomes is going to be the magic bullet is to my mind something that has been oversold by its proponents.

There is another problem as well, and that is if you confront the EMT program that I described before—the epithelial mesenchymal transition program—which confers on carcinoma cells many critical attributes their behavior. That’s simply because of the fact that the EMT program gets activated and confers malignant traits on cells without there being any change whatsoever in the DNA sequences of the cancer cell genome. And so I believe the majority of the attributes of many cancer cells are governed not by mutant DNA sequences, which clearly are playing a role, but rather by nongenetic programs, such as the EMT program that I mentioned before. That creates a major pickle, because we’ve not yet figured out using bioinformatics how to combine information about the DNA sequences and their mutated form in people’s tumors together with the operations of these nongenetic programs. The two obviously exert very profound effects in governing the behavior of cancer cells, but how these two types of regulators of cancer cell behavior interact with one another is still elusive.

Girard: Where do you see the biggest advances in cancer treatments in the near term?

Weinberg: To be sure the use of various arms of clinical oncology, therapy, and diagnostics has reduced in a real way breast cancer mortality in this country by between 30 and 35%. Incidence of a disease like breast cancer is very difficult to know because our current diagnostic tools are so powerful that the great majority of breast cancers that we diagnose are unlikely to develop into life threatening malignancies.

Another point to make is that the major reductions in cancer-associated mortality that will come over the next decades will not come from people like me and people who develop drugs, but rather will come from prevention. That is to say, people who change their lifestyle, wishing to minimize the appearance of cancers, will have a far greater effect on reductions in cancer mortality than trying to treat cancers that have already appeared. And we have to grapple with that as a reality since the progress in treating quite a few highly aggressive cancers over the last decades has been rather limited.

Girard: So how do you envision cancer research a decade from now?

Weinberg: I think cancer research in ten years will be much more driven by bioinformatics, by very complex algorithms, which attempt to figure out why cancer cells behave the way they should. It will be driven by a lot of research, which describes how drugs can be applied in combination, rather than as a single agent, to patients, where the combinatorial agents act with great efficacy. There also may be a resurgence of interest in, for example, the biochemistry of cancer-causing proteins, an area which has lagged over the last decade as people have become more and more enamored with sequencing cancer cell genomes.

We realize now that within cancer cells that suffer mutant genes, these mutations wreak havoc on the cell by perturbing complex signaling networks that have been developed literally over the past one and a half billion years. We understand in rough outline how they work, but the devil is in the details. And we don’t really understand how these circuits process many kinds of signals. And this is a problem of such complexity that people have stayed away from it, but I think that with increasing urgency, one will need to revisit this problem in the coming years so that one can actually look at what is the decision making circuitry in cells, the signal processing proteins within cells that ultimately decide whether a cell will grow or not grow, and whose decision-making is effectively short-circuited by the mutations that create oncogenes or inactivate tumor suppressor genes.


Girard: That was Robert Weinberg, a founding member of Whitehead Institute and a professor of biology at MIT. You can learn more about Whitehead science on our website at And you can listen to other AudioHelicase episodes on SoundCloud and iTunes. For Whitehead Institute, I’m Lisa Girard. Thanks for listening.


Produced by Nicole Giese Rura

Original music by Chocolat Billy. CC BY-NC-ND 4.0

© Whitehead Institute for Biomedical Research              455 Main Street          Cambridge, MA 02142