Professor of genetics, neurobiology
University of Rochester Medical Center
Name: Mark Noble
Position: Professor of genetics, neurobiology and anatomy, department of biomedical genetics, University of Rochester Medical Center, Rochester, NY, since 2000
Background: Huntsman Cancer Institute, 1995-2000; Ludwig Cancer Institute, Middlesex Hospital Branch, 1986-1995; Institute of Neurology, London, 1981-1986; postdoc, University College London, 1977-1981; PhD, genetics, Stanford University, 1977
One of the earliest pioneers in the biology of stem cells, Mark Noble was part of a team that discovered the first progenitor cell in the central nervous system. This cell, the oligodendrocyte progenitor cell, is one of the most extensively studied progenitor cells in biology, in part due to the importance of myelin destruction in CNS diseases such as multiple sclerosis and Alzheimer’s disease.
He has also helped discover growth techniques that have enabled researchers to expand these cells on a large scale. This, in turn, allowed researchers to show for the first time how transplanting purified cells can repair CNS damage, and led to the discovery of several additional CNS progenitor cells.
Today, Noble and colleagues at the University of Rochester continue to study topics such as genetic diseases of the developing CNS; how CNS precursor cell function responds to environmental toxicants, chemotherapeutic agents, hormonal and nutritional deficiencies, and viral infections; and regulatory pathways involved in CNS repair.
Most recently, the researchers published a paper in the Nov. 30 issue of Journal of Biology that demonstrates how progenitor cells could be used to predict the potential toxic effects that chemotherapeutic agents can have on the CNS. Noble and colleagues are also in the process of developing a high-throughput assay that can evaluate chemotherapeutic toxicity on the CNS. This week, Noble took a few moments to discuss this work with CBA News.
Did your team develop the underlying technologies used in this study, or is this more of a method that brings together existing technologies?
Yes and no. I’m one of the founders of stem cell biology of the nervous system. I was part of the team that isolated the very first progenitor cell from the central nervous system back in 1983, and then my lab discovered the second one, an adult-specific cell, in 1989. Then my colleague Margot Mayer-Pröschel was responsible for the next two, which were isolated from the embryonic spinal cord. In the sense that we are the only team that has this extent of expertise in all of these diverse cell types, you could say that we did develop these technologies.
But the technologies that we are employing to study these problems – none of those are novel technologies. It’s the knowledge of how to work with the cells that’s so critical.
And using these cell lines …
Well, we use primary cells – never cell lines. That’s a critical distinction. Cell lines, such as cells isolated from tumors, cells immortalized in the laboratory, or cells that have been grown in vitro over many passages, have very different properties. In particular, such cells are less responsive to the kinds of stimuli that play important roles in normal and abnormal development and tissue maintenance. We have another paper that will be coming out soon in which we used similar approaches to discover what appears to be one of the broad organizing principles of toxicology that applies at environmentally relevant levels of methylmercury, lead, paraquat, thimerosal, and all sorts of chemicals. And we’ve discovered a biochemical pathway in which all these things converge. This pathway is activated at low levels of toxicant exposure, such as the levels of methylmercury that would be in your blood if you ate tuna two or three times a week. I don’t think this work could have been accomplished using cell lines.
So the critical distinction between what we do and what many companies do is, by using primary progenitor cells as our screen, we have a level of sensitivity that you just can’t achieve in other ways. We’re using cells whose job it is to respond to the environment, in that such responsiveness is required to enable them to differentiate in the right time at the right place.
In the Journal of Biology paper, to determine the viability and growth of these cells after treatment, you just examined them on a microscope slide. Do you have to manually count these cells with a microscope?
It’s easy to speed up, but it actually doesn’t take that long. Actually looking down a microscope also enables us to obtain valuable information. These cells change their shape. For example, the progenitors for oligodendrocytes, the cells that make myelin in the brain – when they differentiate, they change their shape. When we look down a microscope, we’re not only getting information on death, but we’re getting information on what the cells look like. Because we’re using cell-type specific antibodies to recognize different developmental stages, we’re also getting information on differentiation. We’re extracting a level of information that maybe nowadays you could get through computer-driven visualization, but frankly, from our point of view, I trust my eyes more than I trust a machine. There is no a priori reason not to [use automation]. We utilize a lot of general plate reader-type assays. But when it comes time for publication, I do trust what I see down a microscope more than I trust a plate reader. I’ve seen many ways in which plate reader assays can give you false information.
Can you explain a little bit more what you were doing when you analyzed immunolabeling and TUNEL staining of tissue sections from mice that had been treated with chemotherapy?
One of the striking findings of these studies, as we’ve seen so many times, is what I’ve come to call in vitro veritas, which means that what we see in the tissue culture dish predicts with enormous accuracy what we’re going to see in vivo. Those cells that we identified as being vulnerable in the tissue culture dish – neuronal progenitors, oligodendrocyte progenitors, oligodendrocytes themselves – are exactly the cells that we find to be vulnerable when we look in the animal.
Do you think that’s an overarching commentary on the ability of cell-based assays to predict in vivo results?
I don’t mean this answer to be arrogant, but I’ve consulted for a lot of companies, and I’ve trained people to do various types of things, and our concern for 20 years now has been with the development of tissue culture approaches that predict with high accuracy results in the animal. If one uses our approaches, then you get these kinds of outcomes. If you want to cut corners, to superficially make an assay seem faster, then you don’t have that high crossover. That’s essential, because people get in and do these assays and say, ‘Oh, well, maybe I’ll change the plating density, and I don’t want to spend quite that much money on growth factors, and we have these other media that I really like better.’ We’ve done all this work to be able to have an assay system where, if we see it in a tissue culture dish, the odds are that we’re going to see it in an animal. We can do that, and we’ve done it in other contexts. So, yes, cell-based assays are very predictive if you use assay systems rationally designed to mimic what goes on in vivo.
You wrote in the paper that there are a very few studies on the neurotoxicity of chemotherapy, despite growing evidence that there is a connection. However, the effect of chemotherapy on, say, liver cells and immune cells has certainly been heavily studied. Why do you think that is?
I think that there are multiple reasons for that. One reason is that the oncologist who treats the cancer patient is not always seeing them again at these long intervals, post-treatment. That may fall under the purview of a different investigator. Of course, the major concern when you’re treating someone with cancer is keeping them alive. And if somebody’s liver fails, or they lose immune function, they’re going to be dead pretty quickly. One attends to these effects because there is no choice.
Most of the effects that we’re looking at are delayed effects. If you look at, for example, [the chemotherapeutic agents] cytarabine and BCNU in our paper, in vivo the loss of dividing cells is getting worse as time goes on after treatment is completed. But if you’re done with your cancer treatment, then often you’re done with that particular doctor. You have to be followed up on in an appropriate way. Places like the Netherlands Cancer Institute have done a fantastic job of this. Pediatric oncologists have also done a fantastic job of this. One of the great amusing joys of science is that you’re always finding things where your reaction is, ‘Man, I could’ve done that 10 years ago!’ Those sometimes turn out to be very interesting discoveries, and I think that may be the case here. The reason we’re particularly interested in this is because we were one of the pioneering groups in identifying the concept of precursor cell dysfunction as a basis for disease in the central nervous system.
One of your major findings was that these drugs caused a greater reduction in viability of progenitor cells than in cancer cells. But you also stated that most of these actually had very little effect on the cancer cell lines examined. Why have these been used at all in cancer treatment?
Most cancer cell lines used in these types of studies come from people with malignant tumors and, the fact of the matter is, we have not made great strides in treating malignancies. Ninety-eight or so percent of the true advances in cancer treatment have come when treating earlier stage tumors – recognizing them and, in some cases, treating them. But in terms of truly spectacular advances in treating advanced malignancy, aside from new therapies for lymphoma, there is very little to celebrate.
Do you think your method might be promising for developing some sort of high-throughput assay for evaluating chemotherapeutic toxic effects?
Yes, and that is precisely what we’re doing now.
Would it be difficult to provide or work with large amounts of these cell types in a high-throughput screen?
We don’t have any problem doing so. It’s maybe like the difference between chess and checkers. Chess is a more complex game, but does that mean that you can’t export chess to millions of people? No – you just have to spend the time learning how to do it. We’ve trained dozens and dozens of people over the years, and we’ve used these approaches to study all sorts of molecular problems, and screen for drugs, and discover survival factors, and death regulatory strategies, with no problems.
Are you interested in commercializing this, and what needs to be done to ramp this up into a high-throughput screen?
There is no immediate sound-bite answer to that, because we’re pursuing multiple strategies, and there are multiple opportunities. These effects, for instance, are a branch of the field of toxicology. We’ve got 80,000 to 150,000 registered chemicals in the US about which we know nothing, and are therefore released freely into the environment. Because we are working with cells that are developmentally relevant and relevant to tissue repair and maintenance, we can rapidly screen these compounds in such a manner as to point to the kinds of questions one would ask in the animal in order to determine what kinds of toxicities exist. In terms of that area, whether we’re talking chemotherapy or toxicology or any other concerns about drugs we give to patients, one area that we are very interested in is being able to recognize the potential for adverse side effects very early.
A very important fact to remember in all this is that not all cancer patients have these effects. As with any side-effect profile, you get a cancer patient to whom you give a treatment, and they’re fine – their hair doesn’t fall out, and they feel terrific, and you get someone else who is a complete wreck. Why is that? We have some fairly compelling evidence now that we’re at least beginning to figure that out. We can take animal strains, and on the basis of asking a few questions, can make fairly accurate predictions about the intensity of their side-effect profiles. So another commercialization route we’re interested in is developing prognostic indicators that up front tell you: Here is an individual at risk, and you have to manage them differently. We’re pursuing that very actively right now. Because I’m more interested in physiology and proteomics, we’re pursuing it through those routes. This is also essential in terms of clinical trial design. I’m broadly involved in the field of stem cell medicine, and one of the important interests is in using transplantation for tissue repair. Some of these clinical trials are very tough to design and interpret. When we do these clinical trials and transplantation, if we could stratify these patients in whom stem cells may function better, or less well, then we have a much higher probability of running a successful clinical trial.