Nutrition and cancer Charles Kaufman


CHARLES KAUFMAN: Thanks so much. Thanks for
having me today. I’m excited to hear the talks today and share with you what we worked
on. So, I’m an oncologist and a developmental biologist, and I study cancer biology. So,
I sort of became, in a nutshell, intrigued by this idea of caloric restriction, which
was introduced earlier today, and as was pointed out, it’s been studied across many organisms.
I found this interesting review of it—refer back to this Italian nobleman, who at the
age of 35 decided to start cutting down calories and lived to be 103, and he became like the
proponent, of sort of, intermittent fasting. Clearly, this—this idea—it seems like
it’s been around for a long time, but it doesn’t seem like you fully understand what’s
going on there. As you talk about … fasting, at least, seems to be associated with improved
metabolic parameters and a number of studies with less cancer, and it’s just in a very
broad sense. How does all this work? So, I also became really intrigued with … I’m
interested in studying transcription, epigenetic changes. In this recent review, I’m talking
about just one little aspect of this notion of oncometabolites are in some cancers. The
cancer cell produces a metabolite that doesn’t exist normally, and it can actually interfere
with DNA methylation and lead to altered gene expression and altered differentiation. So,
just the idea that one can draw a line from this metabolic pathway disruption down to
alterations in gene expression, in my mind, was sort of where I’d love to be someday.
Maybe 20 years from now. So, this is sort of a first foray into that. What I’m interested
in is understanding the emergence of cancer, and so the way I think of this is, you have
a normal tissue, and that normal tissue undergoes some change; usually, an oncogene gets activated
loses a tumor suppressor, and those cells are now prone to forming cancer. Often, they
call this a “cancerized field” of cells. What we don’t really understand is why,
if all these cells that are at risk form cancer, why it is only one cell, we think, start to
form cancer and then expand? So, the model I study that in is melanoma where one has
a B-RAF mutation. I have many moles, I have many B-RAF mutations, but thankfully, only
very rarely does that nevus undergo additional changes and form melanoma. So, this is sort
of … two things that are interesting to me is that maybe this is modifiable. I don’t
know that I can keep these B-RAF mutations from forming, but maybe I can prevent those
additional changes from occurring to go on to actually form melanoma. In our zebrafish,
we have a model developed by Liz Patton in Len Zon’s lab, in which human B-RAF V600E
oncogene is expressed in all melanocytes. So, you have this fish with tens of … let’s
say 10,000 or so melanocytes, and those fish only go on to get roughly one to three melanomas,
and it takes months. So, there are other things that must be happening in the intervening
time that leads to melanoma forming, and I wanted to try to understand that. So, this
fish really is … the whole fish is our field at risk. This is what the model looks like.
So here’s a fish with a rather large tumor on its back here, and these are identifiable
pretty readily by the raised … the raised appearance. They’re responsive to B-RAF
inhibitor therapy, so if you know about … anything about human melanoma treatments, B-RAF inhibitors
are a huge part of that. Our fish are also responsive to these, and then they look … they
look like human melanomas, histologically and at the gene expression level. So, briefly,
I’ll just give you a quick background with one of the tools that we have available and
then at the end sort of how I moved one of our recent forays into how nutrition diet
might figure in. So, neural crest is a embryonic lineage that forms a number of different cell
types, including, importantly, melanocytes here. Melanocytes are the pigment cells that
form on our bodies, and, if they are growing inappropriately, form melanoma. And there’s
a gene in the fish called crestin; here’s our zebrafish embryo, crestin in purple marking
the neural crest. One of the postdocs in Len’s lab, Rich White, a good friend, discovered
that in a normal adult, crestin turns off, but when the fish forms melanoma, crestin
turns back on again. So, this suggested to us that crestin might be a good marker of
tumor formation. So, I developed a crestin EGFP reporter line that reproduced the expression
pattern of crestin in the embryo and then, importantly, in the adult fish. Here’s a
fish swimming around with a shimmer on its back there, and when we put on the EGFP filter,
you can see that shimmer lighting up. It is very bright, very easy to see; this is filmed
with an iPhone 5. So, it’s pretty robust, and it was a very happy day in my postdoc
career when I saw that fish. So, you know, what this allows us to do then is track tumor
formation. This is something one might be able to see grossly just holding the fish
up, but you know it’s preceded 5 or more weeks by this early patch of cells that was
going to form a tumor. So now we can look at the events earlier on. In fact, it comes
out a little bit bright here, but I’m able to spot even a single cell that can go on
to form melanoma, and so it allows us to see the earliest events, we think, in tumor formation.
The idea then is that one of the early barriers to tumor formation is acquiring an oncogene
activation and tumor suppression loss, and I think that one of the other steps is this
re-emergence of neural cross-progenitor identity. So, it involves, whether it be de-differentiation
reprogramming—I know these are very loaded terms, so you don’t really know—but you
think probably some sort of de-differentiation. Just as a quick nutshell of what we were able
to see is that when we look at gene expression and epigenetic landscape near important drivers
of melanoma, such as SOX10, you see these regions called super-enhancers. These are
highly active regions on the chromatin that sort of incorporate many cell type-specific
and oncogene-specific transcription factors. We think this is really the regulatory center
that‘s driving this process, and so we’ve been focusing in and understanding what signaling
pathways are impinging on this … this regulatory element. This has actually been shown, as
well, in the mouse model more recently where the fully pigmented melanocytes seems to de-differentiate
and go on and form the melanoma. So, it was nice. In the fish world, I think we’re always
trying to say that it’s relevant to human cancer, so we see the human … excuse me,
in the mouse model, as well, so that was a nice overlap there. Really, I want you to
look at this emergence of melanoma as an epigenetic phenomenon on what are the sort of upstream
events. Something we’re working on now is to look more deeply at the changes in the
melanocytes from the melanocytes to the melanoma and … sorry this got overlapped here … but
we’re now making red melanocytes, green tumors, and we’re able to look at each step
along the way as one goes from a normal B-RAF-mutated melanocyte to a melanoma and really look at
the changes in gene expression and epigenetic landscape more thoroughly. Moving back, then,
what sort of upstream events are going on? Are there environmental probations to this
transition from precancer to cancer? When I moved from my postdoc In Len’s lab with
Chris Lawrence’s, our zebrafish facility administrator at Wash U, I just sort of wanted
to see how things were going to behave in this new system. So, we have the Tritone feeders
at Wash U, and very simply, it just changed feeding labels on our adult fish from four
times, two times, one time a day to sort of establish what we‘re going to be the parameters
tumor onset in the lab. We started out with this feeding protocol, which is pretty standard
thing where the fish are in E3 for 6 days and go on rotifers first for a couple of weeks,
and then—I’ll expound on this a little bit—then sort of increasing amounts of GEMMA
plus rotifer, and then ultimately just GEMMA feed at about 42 days onward. As shown here,
as they get a little bit older every week, we’re adding more and more feedings into
where, by 6 to 7 weeks they’re getting fed about 12 times a day. Then I just adjusted
the number of times a day, so in principle it’s same amount of feed, just different
numbers of time today, mostly just logistical, because it was easy to change QR labels and
see what would happen. It‘s such … what we saw, it was very reproducibly that when
you feed the fish four times a day versus two versus one, you have a large change in
the median onset. Here is about 40 … 42-day difference between four times versus two times
a day. A couple things we did was when they were—we had the embryos, we, as I say, grew
them all together, but we mixed them all up before distributing them to different tanks.
I was worried about sometimes tests have the “jackpot effect” where one group of larvae
got, you know, more died off or they didn’t grow as well, and then also we would replace
fish when we saw a tumor with casper fish so that we kept the numbers the same. Maybe
this isn’t perfect, but it felt like it was better than just … we typically would
just remove a fish when you see a tumor, and so you kind of accelerate the process because
you have less and less fish as they get more tumors. So, this same number of fish throughout
the experiment, as well. And then this was a cohort that was born in, let’s say, week
1, and we did it on several different cohorts, and the median here are within a week or two
of each other, and this sort of was true across three different cohorts of fish all grown
over the course of about two months. So, it felt like, you know this was reassuring to
us, that this was a real phenomenon and that it was reproducible our in our facility. Then
more … I’m sorry … when I (indiscernible) it now looking back, we should have done more
measurements and weights and things like that, but we sort of did at least a snapshot that
when we measured snout-to-tail length of these fish at 155 days and 211 days, clearly random
days, and their life span, that there was a pretty consistent and stable difference
in the four versus two times a day and one time a day feeding. Fish that were fed more
were bigger; those that were fed less were smaller. That seemed to be consistent across
the board for those groups. So now with this tool that we have to not only look at, you
know, tumor progression, I was interested to see—can we see what’s going on earlier
in the process and is this affecting tumor initiation? This is a little bit trickier
but what I see now preliminarily is that when we increase feeding, we also see increased
appearance, or I should say earlier appearance, of these crestin, these melanoma-positive
tumor patches. You can see this even earlier in the process. So, you know, seeing a raised
tumor means you got millions of tumor cells, but now we’re just looking for the appearance
of you know a single EGFP-positive cell, and we can already see that earlier on the process.
This got me really excited. Maybe now we’re getting some initial glimpse into some of
the metabolic rewiring that could be going on during this epigenetic switch from the
last site to melanoma. Now this is just one example cohort, but I’ve seen this over
multiple cohorts, and we can even make the … can even see that the effect is more pronounced
when one adjusts the amount of feeding that’s going on during the juvenile stages. I didn’t
put that data up here because it’s still preliminary, but you know I think that there’s
been mention that what sort of feeding we need to do at different steps of life cycle.
Clearly, my sense is that some this is being sort of imprinted in—that‘s a specific
term—some of this is being determined even during the juvenile phase of growth, and so
that may have an important impact on in what one sees. So from here with this tool in hand,
we have a lot of different directions to go, and I just put out a few that we’re sort
of intrigued by doing intermittent fasting with the fish, looking at what at … just
at a basic level, what part of the metabolic machinery is involved. Is hyperglycemia enough,
let’s say, to induce this change? There’s been a lot of talk about looking at specific
amino acids or other nutrients, and so we’ve sort of been interested in—generally speaking—in
the types of foods that would be available where we can cut out a specific amino acid
or use genetic approaches to alter metabolic pathway of interest. And then start to tie
that into changes in epigenetic parameters that we can study. I put Gary Patty … I
know he was intrigued by this and wasn’t able to be here today, but you have a shirt
guide with a student who was in Steve Johnson’s lab, as well, and was very interested in this
process, and so we’re looking for some collaborations and starting to look at true metabolomics.
He’s a chemist; I am not a chemist. And so, you know, we can start to work together
to sort of address some of these questions going forward. Then just a few … a lot has
been mentioned, and I would say just some of my thoughts on pros and cons. I guess I’d
call this an intensive feeding protocol. We do see this rapid generation time with about
6 weeks of sexual maturity, but I would say, as well, we definitely have a short reproductive
cycle. So, for those who are doing genetic screens in our lab, we want to produce gazillions
of embryos very quickly, that’s awesome, but for my tumor fish it’s less helpful
because they get sicker sooner. We have to cycle through generations, and then we’re
trying to inbreed and so we have to outcross and inbreed, and so that actually is maybe
less efficient for us. I think there’s definitely great consistency across experiments, but
I think this has … been brought up about these potential increases and expense and
staffing when you have 20 Tritones running in a facility. As one who recently went through
the job search process, proposing that may not be feasible at all places. So, what are
the cost of entry becoming a zebrafish researcher in a new place? I think that that’s something
that I’m sure will be considered in the recommendations going forward. Just a number
of folks to thank briefly: my group here—at Len’s group, I should say in Boston— and
then my group here at St. Louis, and I wanted to particularly thank Vadim, who helped with
the feeding experiment, he’s now a master at spotting melanoma tumors, and my graduate
student, in particular, Jon, who’s been working on the metabolism stuff, and Eva,
who generated a lot of epigenetic data for us. So, I’m happy to take questions.
UNIDENTIFIED MALE: So when you … I noticed on your length curve, and since all cells
are … you’re not producing hypertrop … like, they’re not bigger cells, so you’re producing
more cell divisions, right, with the higher food? So, the fish are proportionately bigger
because there are more cells? So, can you normalize your curve to number of cells that
basically you’re seeing and sooner because there are more cells there because they’re
… cycling more quickly? CHARLES KAUFMAN: Yeah, that’s a great … I
think … I’ve never … we haven’t counted the number of … are there more cells at
risk, either more melanocytes at risk to develop melanoma or, as you say, the number of cell
divisions that increase. Yeah, I haven’t done that, but that’s a good thought. There’s
that literature about if you, you know, count the number of cell divisions for a given organ
type, it seems to correlate with your risk of getting cancer in tissues that divide more,
you get more cancer in—that kind of a notion for sure. It’s a good thought. Chris?
UNIDENTIFIED MALE: So just to clarify, the only thing you varied was the frequency of
feeding, not the total amount of food going into the tank?
CHARLES KAUFMAN: It was the 60 milligrams per dump from the feeder, and that was … we
just did it four times, two times, one time. And the reason, also, why we picked those
is that, in theory, in our facility, when the fish achieve a size that you want, you’re
supposed to switch down to two and one. So, I just went with what was on our protocol
and what … I just started at the one just to do a slower amount from the get-go.
UNIDENTIFIED MALE: So you did have a total … your amount of food going into each tank
did vary. CHARLES KAUFMAN: Yeah, it varied across the
… yeah. Thanks. UNIDENTIFIED MALE: Not just frequency.
CHARLES KAUFMAN: Yeah. Thanks.

Leave a Reply

Your email address will not be published. Required fields are marked *