Big Think Interview With Michael Wigler

A conversation with the genetics professor at Cold Spring Harbor Laboratory.
  • Transcript


Michael Wigler: Michael Wigler.  Professor of genetics at Cold Spring Harbor Lab.

Question: What does your research consist of on a day-to-day basis?

Michael Wigler: Our lab studies the genome of organisms and also the genome of cancer cells.  And we work on two kinds of problems: the evolution and outcome of cancers, and also on genetic disorders of a spontaneous sort, that is, non-heritable genetic disorders.  And those are two very—it sounds like two very different things, but they’re related by our methodology, which is genomic analysis.

What we do is called difference analysis, for example, if we’re looking at a cancer, we’ll want to see where that cancer has mutated relative to the genome of the person who gave rise to that cancer.  That’s differential genomic analysis.  And it tells us where the cancer has mutated.  And from the types of mutations, the number of mutations, we can infer a lot about cancer etiology. 

Question: Is biology becoming a more quantitative than qualitative science?

Michael Wigler: Well, biology has always been influenced strongly by quantitative types.  Many physicists in the late ‘30s, early ‘40s, ‘50s, came into biology, strongly influenced it.  There was a period, I would say, from the time I was a graduate student in the mid-‘70s until the mid- to late-‘90s, where it was not particularly quantitative, and that was largely because of the revolution in recombinant DNA.  So, really all you needed to be a good biologist was a good sense of logic and a good imagination.  And mathematical and statistical skills weren’t really that necessary for much of biology.  And I was in that group actually.  I had studied earlier on as a mathematician but I used almost none of those mathematical tools when doing biological research.  Of course, the logic comes in handy, but the tools were not very valuable.  There was no place for them because the kind of data that we were getting was very individual data and I actually had a rule of thumb. I actually disliked statistics early on in my life and I felt that if I needed to do statistics to see what I was observing, then I wasn’t really observing anything. 

But that changed with the advent of the sequencing of the human genome.  That changed everything.  And the development of new high throughput methods of extracting data, it forced biologists to reconsider the value of statistics and mathematics in the analysis of their subject.  So, a number of biologists moved in that direction.  Not a lot, but quite a number did.  And I was one of those who moved in that direction.

Question: How has the sequencing of the genome “changed everything”?

Michael Wigler: You know, we are so close, historically, to that period, and the data that’s coming out of that effort is still being generated.  I think it’s very hard for any of us to really judge the impact that it has had.  It was a huge revolution in terms of the kinds of experiments one can conceive of doing.  The only thing comparable in my lifetime was the recombinant DNA revolution which changed entirely the kinds of experiments people did.  

Since sequencing methods are changing so fast, the cost of sequencing has dropped enormously.  And with each drop in the cost, it changes entirely how you think of attacking the problem.  So, in a few years from now we’ll be in a position to have DNA sequence of a very high quality for a million people and know the medical history of these million people.  And there’ll be—I don’t even think our computers are yet to a stage where they will be able to handle data of that type and the kind of analysis tools that will be needed to analyze that haven’t been developed yet.  So, we’re in a really a strange point in the history of biology where things are changing so rapidly, we can’t quite see the shape of the future yet. 

Question: What has your research revealed about the genetic causes of cancer?

Michael Wigler: Yeah.  Well, the first observation was that there was a very strong correlation between the extent to which the genome in a cancer cell has changed and the lethality of the cancer.  So that, if one’s looking at cancer and there’s lots of changes in the genome, that patient is less likely to survive than a patient whose genome has just begun to evolve.  That was the first major observation. 

There were a lot of particular details that emerged from those studies, that is, we found the locations of genes that are called uncA genes and tumor suppressor genes.  The individual genes at these places, many of the changes are what we call recurrent.  They happen over and over again in different people with the same cancer, and there are genes in those regions that one can show functionally alter the capacity of the cancer cell to grow, divide, or spread in the individual.  So this has been an engine also for the discovery of new cancer genes. 

We weren’t the first ones to do this.  People have been using these techniques for a while, including ourselves, for a period of 10 years or more.  Sometimes particular drugs that are given to a patient are determined by whether that patient has a particular gene amplification in their cancer.  The most well-known example of that is patients with amplification of the HER2 gene will likely respond to Herceptin.  So, our review has been that specific amplifications will correlate with drug sensitivity, we’re in the middle of exploring that, and we’ve also begun to look at single cells within cancer.  So that we can now actually look at the genome of an individual cell within the cancer and that’s giving us a much more detailed picture of how the cancer has evolved. 

So, we think we’ll be able to identify, for example, the earliest cells, the earliest mutations in a cancer that will tell us how the cancer began to grow in the first place.  It will also tell us what you might call the tribal, or population structure of the cancer, and that tells us about how the cancer is... how the individual cancer cells are interacting with each other, interacting with the host, and migrating through the cancer, and possibly migrating throughout the patient.  So that we think that by looking at the individual cells of the cancer, we’ll be able to improve clinical staging and drug treatment enormously.  But this is a long-term project.  This will take us five years, 10 years.

Question: How might this research impact clinical cancer treatments?

Michael Wigler: Well, I can give you two ways—there are many ways this research could impact the clinic.  I can give you two very concrete examples.  If a new drug is being tested in a population with a particular type of cancer, one might look for correlations between response to the drug and the genome profile.  That could tell you which patients are likely to respond to a drug so that patients don’t have to take a drug that’s not going to benefit them and don’t have to suffer the side effects of a drug that’s not going to benefit them.  And that will ultimately lead to the design of better drugs.  

A second way—and this next way is not quite science fiction, but we’re looking a little bit into the future—when we can examine the genome of individual cells, and can do that cheaply, we can develop early detection tests for cancer that are based on blood.  So, it’s now being appreciated widely that even cancers that perhaps have not yet metastasized release their cells into the bloodstream and do so in fairly large numbers so that you can collect cells from the blood and identify them as a kind of cell that shouldn’t be in the blood.  But people haven’t yet been able to look at the genomes of these individual cells.  So, some of the methodology that we are developing will enable us to do that.  So you can imagine that at some time in the future, you can draw blood in the doctor’s office and just like the doctors now do what’s called a blood count to determine how many white blood cells you have, whether it’s likely that you’ve got a fever, they’ll be able to sort out from the blood this small proportion of cells that might be being spun off by a cancer somewhere undetected in the body.  And by looking at the genome of those cells, and possibly by also looking at the RNA that those cells are making, I'll be able to say "This person has malignant bone cancer," and then you can look for that. 

So, this technology can ultimately lead to early detection for cancer.

Question: How did you become interested in autism?

Michael Wigler: My personal interest in autism dates from when I was a child, and I had a friend whose brother was quite strange.  And when I was in medical school, I realized that he had autism.  It was actually Asperger’s. He was a very bright kid, never looked you in the face, constantly was throwing his arms up like that as though he had made some great discovery; and knew everything about baseball statistics. And so it made an imprint on me at an early age. And it’s sort of a wonderful, it was sort of a wonderful thing to see this fellow who actually grew up to, I think he had a successful career as a disc jockey.  So, I was always interested in autism and because I come from a family that’s somewhat left-wing, always looking for ways I can do something that is a benefit to society.  And it struck me that autism was not a disorder that was studied by the scientific community very deeply.  But in the worst cases, it was tragic for the families that had an autistic child. 

So, I was motivated by both of those things to have an interest in autism.  And when we began to study cancer, which was in the early 1980’s, I knew at the time they were studying cancer that the tools that we were developing could later be applied to genetic disorders.  Not the kind of genetic disorders where you inherit something from your parents, but the kind of genetic disorders that arise spontaneously because of mutation in the parent’s germ line. 

An example of those kinds of mutations that everybody’s familiar with is Down syndrome; or Trisomy 21 I guess is the clinically correct way to refer to it.  These are new mutations.  You don’t inherit it in the classical sense, but it was obvious to people who thought about it that human genome is not static; it changes over time.  That’s how we evolve.  And most of those changes are not good.  They result in some disorder or another, but they’re hard to study.  Most people who study genetic disorders study inherited kinds of genetic disorders.  I was interested in the other kind of genetic disorders that result from new mutation.  And new mutations are what we study when we look at cancers.  When we’re comparing a cancer to the normal person’s genome, the cancers differ by new mutation.  That’s called somatic mutation. 

The same tools that find somatic mutation can find germ line mutations if you compare the child to the parents.  The incidence of autism being relatively high—and by and large, these children are so different from their parents—it seemed to me that it was likely, just a priori, that autism was the result of new mutation in the germ line possibly affecting many, many, many genes that result in the same end behavior, or similar end behaviors, and that was being ignored by the community. 

So, when we had the tools to go look at this, we did so.  And so it was a combination of opportunism because we had developed the tools, and intrinsic interest from both a social point of view, the social good, and also from a personal point of view.  That is, I had a personal interest in how does the brain go from being what we would recognize as belonging to a normal person to somebody who is, in wondrous ways, very different from us.

Question: What is autism?

Michael Wigler: Well, there are a triad of behaviors that are the earmarks of autism.  The include difficulty in social interactions, delay in the development of speech and communication.  And those are distinguishable and repetitive behaviors, almost obsessive-like behaviors.  

The recognition of this triad as a condition we call autism began only in the late ‘30s, and as the diagnostic criteria began to be more widely applied, more and more children were being called autistic.  And the definition, I think, I mean, when people now talk about autism spectrum disorders where a child has varying degrees of these abnormalities.  It is not, in fact, an extremely well-defined disorder.  It has sloppy boundaries to normal behavior.  We all know people that are awkward socially, there are many people who learn language late in life, and we all may know people that have stutters, or have obsessive behaviors, or even hang wringing.  So there is something of a continuum of all three of these things.  That’s not a condition whose boundaries are well-defined.  Yet, if you meet a child with autism, you can generally say that there is something profoundly wrong here. 

But it’s a hard disorder to define better than that.  And probably the reason it’s harder to define better than that is that the number of genes involved.  The number of underlying causes that can create this triad is very great.  For example, the syndrome itself is enormously varied.  And if you have listened to somebody who studies autistic children—children with autism, you’ll frequently hear them say that each child that they see is different than the next.  It’s not really a syndrome in the way that Down syndrome is a syndrome.  There are a variety of genetic disorders that are frequently—you can almost tell that the children who have these disorders have the same underlying cause, because they’ll actually look alike.  It’s not just Down syndrome that has that property, Progeria has that property.  There are a number of childhood disorders where the children who have these disorders actually look alike.  

That’s not the case in autism.  Each child has—is sort of wonderfully different than the next child, so there’s a huge amount of variability.  And I think this has confounded the general public because it appears that the rate of autism has been going up so dramatically.  In fact, I think that’s mainly due to increased diagnosis.

Question: What is the “unified theory of autism” that you’ve developed?

Michael Wigler: The unified theory of autism attempts to reconcile several observations.  The first observation is that having siblings with autism is more common than one would expect if each incidence of autism was random.  So, if a child is born has autism, a brother is born, the chances that that brother has autism are much higher than a male born to another family. 

And twins, identical twins have an extremely high concordance.  Something like 90%.  There is no other cognitive disorder whose concordance among identical twins is as high. 

So, those two facts tell you that there is a genetic component to autism.  However, there are families that have autistic children and there are large families and only one child will have autism.  So, the genetics would look to be complicated.  There’s an inherited component because siblings have a higher rate of concurrence, but there might also be a sporadic component.  So, the issue is how to reconcile that. 

I think that prior to our serious involvement in the field, people assumed there was what was called this complex inherited model.  That there are many genes that may be in the wrong state in the parents that come into some combination in the child, so the children of these parents have a higher chance of having autism.  But it’s not a classical Mendelian pattern where half of your kids have it, or a quarter of your kids have it.  Half will have it if it’s a dominant, a quarter if it’s recessive.  The pattern seems more complicated than that. 

What we did was come in and say, well, you know, it could be a combination of both.  In some families, it is perhaps simple Mendelian and in other families it’s spontaneous.  And if you assume that there are a large number of genes that can give you autism, then you could have a very large proportion of autism being generated by spontaneous mutation.  But if the mutations don’t all have complete, what’s called complete penetrance, that is, you can pass on the mutation and the child can carry it and not show the disorder, then his or her children could then be at risk in a Mendelian way of inheriting that gene.

So combining these two ideas that the sibling risks is really a combination of simple Mendelian in some families with other families being spontaneous mutation unifies these two observations and does so in a coherent model.  So, the coherent model is that humans are mutating, the rate of new mutation giving rise to autism is perhaps on the order of 1 in 200 kids, and something like half of those kids actually don’t come down with a diagnosis, they mature, they get married, they have children and those children are then at risk from the carriers. 

Now, one of the very important clues that is compatible with this model is that the risk of autism is much higher in boys than in girls.  If the model, almost any model would predict that whatever genetic abnormalities exist in the boy, those abnormalities will exist in the girl.  So girls have something that makes them resistant.  So girls, in fact, could be natural carriers of genes that in the boy would give the boy autism.  And that girl might grow up and be a healthy and desirable mate and have children and her children, particularly her male offspring might be at high risk because they might inherit the gene that she safely carries.  That’s the essence of unified theory.  It does not explain why autism, why boys are at higher risk than girls.  But it does suggest that you can have two forms of genetic involvement; an inherited involvement from a carrier parent and also those rare mutations that destroy a gene in the germ line. 

Now, I should say, and I really have to mention this, that in the model we’re not saying that only women are carriers.  In fact, there’s well-known example that’s been in the news of a male sperm donor who had something on the order of 20 male offspring and half of them had autism.  So, that’s clearly a case where the sperm donor, who I guess was judged to be normal, probably maybe even brilliant or even genius, was a carrier of a simple dominantly inherited Mendelian trait.

Question: Why do older parents tend to have more autistic children?

Michael Wigler:  The incidence of autism goes up with the age of the parent, and that’s entirely consistent with the new mutation idea.  Because it’s already well established in males that the number of point mutations found in the male’s offspring go up with the age of the father.  And there’s also a correlation with the age of the mother.  So, there may be a mild increase in the rate of autism in those cultures where having children is differed and delayed.  The magnitude of that effect is not going to explain the overwhelming explosion in the number of diagnoses, but there may be a mild increase in the rate of autism due to that.  And the age dependence on the parents is consistent with the new mutation hypothesis.

Question: Do you believe that environmental factors such as vaccines increase the autism rate?

Michael Wigler: Well, any genetic disorder is an interaction with the environment.  So, I don’t exclude environment.  I just don’t see yet any strong evidence for a particular environmental factor.  I think that one could do studies.  For example, one could go to third world countries and do a study and ask is the rate of autism there the same as it is in the developed countries.  No one has done a study, that I know of, of that type, but it certainly could be done.  That would answer that question. 

But certainly anything to have to do with the development of an organism has an environmental component to it, but you can only study that when there’s some evidence which enables you to isolate that environmental component.  I think the vaccine studies have been now largely discredited.  They took mercury out of the vaccines and the rates of autism didn’t change.  And now of the 12 authors of the original paper that got some people very excited, I think 11 of those 12 authors have now withdrawn their backing for that paper and the methods used in that paper are really in doubt.  

So, I don’t take it as there being any evidence that vaccines are such an environmental factor.  It’s unfortunate that at the age at which parents begin to recognize autism in their children often correlates with the age at which they receive vaccinations.  That’s an unfortunate thing. 

There is a portion of autism, probably the majority of autism that you can detect it very early once the child is... almost after the child is born.  Many parents of autistic children say they could tell very early on that there was something wrong with their child.  There are about a quarter of the cases of autism where it looks like the child, in the parents’ opinion, has been developing normally and then, to their mind, suddenly goes off course.  And I think at least five percent of cases, it’s been very well documented that actually, a child has begun to lose gains that they have made.  So, there is this component that’s very tragic when a parent feels relieved that they’ve gotten through, and I think every parent who has a child suffers through nightmares that, you know, hoping that their child will be healthy and they give birth to a healthy child and then at age two or three, the child suddenly stops developing.  That’s a tragedy of horrendous proportions, and it’s natural for the parents of such children to look around for the possible causes; something external. 

However, it should be borne in mind that our brains continuously are developing at that age, and it is well-known that there are genetic defects whose onsets can occur at almost any particular age.  For example, there is a class of disorders that are called Storage Disorders where the child develops normally, but because of the buildup of some compound due to the faulty metabolism of some essential thing that they eat every day, builds up to a point and then begins to poison the brain.  And in these cases, the child will develop normally up to a certain age, and then will often regress and sometimes will die.  So, the idea that you can’t have sudden onset of an illness when the child is two or three is just wrong. 

If there were a clear environmental signal, for example, sonograms, or too much television, or vaccinations, that would be something that one could study, but in the absence of evidence for that, you have to ask yourself, well what should we be looking for?  Should it be the plastic in bottles?  And I don’t think we can do that in our culture.  I don’t think we can look for these possible environmental insults.  There are just far too many.  But if you go to a place like Nepal, or Mongolia, or someplace whose environment is completely different, they don’t have television, they still have grandmothers raising the children, they don’t get sonograms.  You could begin to tease out and do what epidemiologists do.  They go and do cultural comparisons.  So, for example, cultural comparisons have told us the incidences of breast cancer in Japan is one-third the rate of the incidences of breast cancer in America, and when Japanese women grow up in America, their rate of breast cancer is the same is American women.  Okay.  You can say, the environment possibly including culture in some way, because the rate, or the age on which you undergo puberty is relevant to breast cancer.  Has a study like that been done for autism?  No.  That’s where you would start.  And none of that’s been done as far as I know.

Question: What will be the impact of your research on autism treatment?

Michael Wigler: Yes, well there are two ways in which our work could inform clinical treatment.  In the area of early diagnosis.  If there’s a child and it’s developing—it’s giving off developmental clues there might be something wrong, if we had a list of the kind of genetic lesions we could screen for, we might be able to determine early on that this child is going to develop a form of autism.  And if it’s correct—most disorders are correctable to the extent that they are correctable, are more correctable early than late, when we know how to correct or treat, we’ll be able to start that sooner.  So, early diagnosis is going to be important for any disorder.  That’s one way. 

Another way is children with a particular genetic abnormality, that is, those children who share genetic abnormality, may have one particular way of treating them that’s different than children who have a different abnormality.  We will only learn about that once we can separate these children according to their genetic abnormalities.  That’s going to take many, many years. 

The third way is that in some cases, we will be identifying genes, who by their very nature, tell us this is a correctable, treatable, syndrome.  For example, we find a gene that’s involved in metabolism.  This child is perhaps got really a storage disorder of some type, but altering the diet in those cases might be able to treat the child.  But unfortunately, we don’t yet know the identities of the autism genes.  We have regions and there’s a huge effort underway.  I would say, in particular by doing very exhaustive sequence comparisons of children to their parents, we will identify the actual culprit genes.  And that will take us two to four years.  And there may be, unfortunately, I’m estimating around 400 such genes that each one of which can cause autism.  But when we have those genes, we see what they do; we can see what pathways they are interacting with, some of those will suggest immediately treatments that can be tested.  We will be able to make animal models and test drugs in animals to correct these things. 

So, in general, the way to understand a disorder is to understand its causes and then address those causes.  In the case of autism, most people would agree, I think most scientists would agree the causes are genetic, and we have a pathway to discover the genes.  So it will be easier to diagnose, classify by diagnosis into behavioral and even drug treatments, and discover new drug treatments. 

Question: What made you choose science as a career?

Michael Wigler: Well, the first thing I remember wanting to be was a middleweight boxer.  And that was because I used to punch my older brother and he said, some day you’ll be middleweight champion.  That was my first ambition.  After that, I drifted to science.  I think because my father was a chemist and my mother had a great deal of respect for the social utility of the mind.  In that period, which was the late ‘40s, following World War II, early ‘50s, people were very optimistic about the impact of technology on quality of life. 

The life of an artist was generally considered to be one of suffering, and so my parents certainly didn’t wish that on me.  And those were my two choices.  It was either science or the arts.  We didn’t have any—my grandfather was a tailor, so anything involving the hands was out of the question.  One had to live the life of the mind, and there were really these two paths.  I choose science, but toyed with writing when I was in high school and college, ultimately settling on mathematics, which I really enormously enjoyed.  And actually began to develop a disdain for science because science depended on the empirical world as a source for the imagination, whereas in mathematics, you didn’t have to depend on the empirical world.  So, to me, I thought that mathematics was the highest enterprise of the mind. 

But I wasn’t good enough at it and it was taking me out of contact with humans, so I decided I had to do something socially useful, so I went into medicine.  And that was a disaster.  I really couldn’t deal with the uncertainty of medicine, so I started doing research instead.  And that’s how I ended up being a biologist and molecular biologist.  So, I didn’t finish medical school, I went into microbial research instead and came back much later in my life to utilize mathematics. 

But in my case, it was entirely the influence of my parents.  They had admiration for the life of the mind and they didn’t have admiration really for anything else.  I mean, I guess there might have been some athletes that they admired.  They admired people who had broken down cultural barriers.  So, they had some admiration for people that struck down political archetypes, social archetypes.  But mainly they felt that their kids should be active with their minds and do things that they enjoyed based on their own imaginations, their own training.  So, I never questioned that. 

Unfortunately, I didn’t realize what they had done.  So, when I had children—in case Ben and Josh find this—it didn’t occur to me that you actually had to imbue this.  I thought it would just be natural for a child to want to be either a scientist or an artist.  And neither of my children had an interest in science.  And I realized that when it was too late.  So, I missed out with my kids. 

I think to get, if one has as a goal to have a society with more scientists and engineers in it, then the culture has to respect people who do that.  And the way these people are depicted in the cultural media is not generally positive.  There were in the ‘30s a number of books that were written.  I don’t remember their names, in which scientists of one type, Marie Curie, Louie Pasteur, were depicted in dramas as heroes.  But you don’t see that at all anymore.  Instead, scientists are villains, they’re socially awkward, they’re not the kind of people you can cuddle up to. And I think that if popular culture does not reflect the value of science, people are not going to go into it.  And America will be dependent on people coming in from the outside to fulfill the positions of engineers and scientists.

Question: Have you ever been completely surprised by the outcome of your research?

Michael Wigler: Well, science is a very—it’s actually a very difficult field because you need probably above everything else, extraordinary patience.  And what keeps you going is discovery.  And sometimes in a lifetime, you may have one outstanding discovery.  Einstein used to say that he was unusual in that he had had two.  But any one would have been enough to have kept him going.  Most scientists are not in that league, but we’ve all had at some scale things that we’re really very proud of if discover them.  Often, we are looking for them.  The idea that a lot of discovery is serendipitous and accidental is tremendously, tremendously overplayed.  I think it’s much more likely that one sees something, almost in everyday life that puzzles you and you carry it around with you for some period of time, and then you see some way of connecting to it.  You could say our discoveries in autism as an example of that.  At a very early age, I was impressed by this child and later saw an opportunity and I struck when the opportunity was there to satisfy my curiosity.  So, most discovery is of that type. 

Sometimes you see things that you can’t explain.  And I shouldn’t say sometimes, a lot of times you see things that you can’t explain.  And sometimes you come up with explanations that are really exciting.  And 99% of the time, those are wrong and there’s really some trivial explanation of the thing that’s gotten you excited. 

Early in my career I used to hate those things and I used to say, only a manic depressive would love living like this.  You see something that’s weird, you come up with some great encompassing idea that will explain it, it’s going to change how people think, and then the next day you realize that you were really a dumbass.  Nowadays when those things happen, I actually really enjoy them because there are so few real "Eureka!" moments in one’s life that you have to almost have to enjoy the fake ones.  I mean, after all, the feeling is just as good.  So, I’ve actually gotten to enjoy those weird results that we can’t explain, come up with fanciful ideas, and then try to batter them.  And then you get double satisfaction because you end up destroying the idea and it’s satisfying to destroy the idea.  Almost as much fun to destroy an idea as to create one.

Recorded April 12, 2010