2011 SABCS, San Antonio Breast Cancer Symposium, 6-10 December, San Antonio, USA

Analysis of the development of breast tissue

Professor Geoff Wahl – University of California, San Diego, USA

My name is Geoff Wahl, I am a basic scientist, I work at the Salk Institute in La Jolla California and I work on a variety of types of cancers but, over the past several years have focussed most of the lab on exploration of breast cancer. In addition, I’ve taken interest outside the lab because I feel that it’s a great responsibility for scientists to be able to explain their work to the public as well as to funders at Congress and elsewhere. So I’ve become very active in the American Association for Cancer Research and was their President in 2007.

In the 1800s, it was suggested by a man named Durante that in order to understand breast cancer we first had to understand how the normal breast develops and we had to understand the types of cells that are formed in the embryo that will eventually contribute to the formation of the mature gland. He felt that by understanding the processes associated with development, we would begin to see how those processes are perturbed in the abnormal development that creates a cancer. Now this is a very difficult problem; if you were to look at an American penny you would see the Lincoln monument and in the middle of the Lincoln monument, between two pillars, you would see Lincoln sitting there. And if you had very good eyesight you could see his head. If you multiply the penny by about ten times, the size of his head would magnify, would be about the size of one of the little buds in the embryo that we know will later form one of the ten mammary glands in the mouse. So to begin to understand the cells that contribute in the mammary gland during development, to the formation of the mammary gland in the adult, you have to have really good eyes, a really good microscope, a lot of patience and an incredible amount of dexterity. So five years ago we began this very difficult project to dissect every day of embryonic what we call mammogenesis, the generation of the mammary gland during formation of the embryo, to ask whether at any point during the formation of the gland in the embryo, there are cells that fulfil the characteristics of stem cells that would be generated. The two characteristics of stem cells are that a single cell should be able, when it’s transplanted into a suitable environment, to regenerate an entire functional mammary gland and then that upon formation of that mammary gland, if you take a piece of the mammary gland there should be a stem cell that remains there that upon re-transplantation that will also be able to regenerate another mammary gland and so on. That’s called the process of self-renewal. Those two characteristics, multi-lineage differentiation and self-renewal, are the two hallmarks of a stem cell.

Now, from work that was done by other people fifty years ago, we and everybody else in the mammary field felt that the stem cells that would contribute to the formation of the adult gland would be formed as soon as you could see the first evidence of a mammary rudiment forming in the embryo. So we reproduced the early work, just to make sure we could do that because we had never worked on the mammary gland before. We did that and we even did it better than people did it earlier because we took single rudiments to transplant and earlier people had taken multiple rudiments. So we knew that a single rudiment from the earliest day that we could see that the mammary gland was going to form, which is at about eleven or twelve days of embryogenesis, could generate a complete mammary gland upon transplantation. We thought, OK, the stems cells must be there and that’s great because that’s a very simple structure to work with.

And then we got a shock. And the shock was when we dissociated those rudiments into individual cells and we transplanted the individual cells into a recipient from which the mammary gland had been removed and asked whether those cells could reform the mammary gland, we found that we could not… the frequency was so low we couldn’t see it. And we took hundreds of mice and thousands or tens of thousands of cells and tried to transplant it and we never saw it. It was incredibly frustrating. But then what we did is we analysed every day of embryogenesis to find out, well if they aren’t there at the beginning, they must be there at some time, when do they start to appear? And that’s where we got the surprise. This is what you live for in science. Failure is constant but that one success is addictive because you’re going to see something that no-one else in the world has ever seen before. And what we saw was that at mid-gestation, about day fifteen or sixteen, all of a sudden we started to be able to measure the existence of cells that, when transplanted, could form a full mammary gland. And then by the end of embryogenesis, about day eighteen, the number of these cells had increased dramatically.

Why is this important? Well it turns out that that day of embryogenesis at which we saw the magic happening, that corresponds to a time when the rudiment that will ultimately form the mammary gland is not only undergoing… the cells are undergoing exuberant proliferation, they’re dividing rapidly, but also the bud that looks like a light bulb is elongating to start to look a tube and then it gains the capacity to start to eat through the cells underneath it that usually encapsulate the bud, called the mesenchyme. They eat through the mesenchyme until the cells lodge into what we call the primitive fat pad, that will be the fat pad that will surround the structure in the adult. That process of cell division, invasion and growing to a new spot looks like cancer but it’s under control and that’s why the gland will eventually form and form the beautiful alveolar structure that we see in the adult.

So then we started to ask the next question – if the process looks like cancer, are the genes that are expressed in these cells similar to those that are expressed in human cancer? We’re using mouse as a model but if the idea is right then the genes that are present should be conserved and we should see them during the abnormal processes that we call cancer. In order to do that we need to purify these stem cells. That was a job in itself because of the small size of the structure but we did it and we enriched the cells so highly that we had a more pure population of stem cells than anybody has ever obtained before. That was good because that would make the announcers of the gene signatures easier. In addition, the stem cells are maintained by association with cells that surround them that we call the niche. So we isolated the cells that are not stem cells that we presumed among them would contain the cells of the niche and we thought maybe signalling from there was important too. What we found then was that the genes that were expressed in these highly enriched stem cells, when we looked at a variety of kinds of human breast cancer that are divided into what are called the intrinsic molecular subtypes that Chuck Perou first reported ten years ago, that we could see that there were some cancers that were highly enriched for expression of these embryonic genes.

Now if you look at women with basal-like breast cancer, these are women who typically don’t have expression of the oestrogen receptor, the progesterone receptor or HER2, and the cells are highly proliferative and people have said they look immature, they look stem-like. I’m not sure what that means because I don’t know what a stem cell looks like, but if you look at the genes that they express, there were some cases, maybe 50-60% of women with this basal-like breast cancer, those cancers were highly enriched in these embryonic stem cell genes.

I just said something very important. Not all cancers of this basal subtype show the expression of these genes. This is an important problem in cancer biology. Not all breast cancers are the same; two women who present with what we call superficially the same type of breast cancer,  maybe both are called basal-like because of the criteria, those cancers are as individual in their genetic differences as the two people who present there, as individual as the finger prints are individual of the two patients. So we need to consider the genes and the genes that are expressed in individual patients separately.

So we have a way of now fractionating these basal-like cancers according to a new property – similarity to embryonic stemness. We look at whether cancers that are enriched in these genes, patients do… how do they behave, what is their prognosis going to be? So it turns out that we’ve divided the genes into different groups, and I won’t get into that, it’s too complicated here. But it turns out that if some of the groups of genes are enriched in the cancers of these patients, those patients are going to be seen to do much more poorly when we look at their performance over time; others won’t, they’ll do fine. So what this gives us is another tool to say, well maybe these patients with enrichment for those genes have to be treated more aggressively than these other patients. And that’s what we need to know – we need to know who to treat aggressively and we need to know what to treat them with that will be best. So that’s the next question - are there any genes in these embryonic signatures that would tell us what to treat patients with whose tumours are enriched for those genes?

Now, it turns out that these cells don’t have expression of the oestrogen receptor, our stem cells, they don’t have expression of the progesterone receptor, they have low expression of HER2. And so on those criteria, many doctors would say there’s no reason to treat them with therapies that we would otherwise use for patients who have high expression of something like HER2, which is a cell surface receptor that receives certain signals from the environment to cause the cells to grow. But as we were analysing, remember I told you there were two components that are important – there’s the stem cell and there are the cells with which it interacts, the stroma, as we analysed the genes that were expressed in both, we noticed that there is a strong likelihood that there’s an interaction that the stroma secreting factors that are interacting with cell surface receptors on the stem cells that are causing the stem cells to grow. And, low and behold, a lot of these factors look like they should be activating a pathway containing genes of which one member is HER2, not the only member, it’s one of them.

So, against all dogma, we decided to ask whether these cells were sensitive to inhibitors of a pathway called the EGF receptor family of receptors. HER2 is one member of that, HER2 means Human EGF Receptor 2, there are four receptors. So we used an inhibitor that inhibits other receptors in the family, HER1, HER3 and HER4, and sure enough the stem cells stopped growing.

This gives us the following idea, which is the important bottom line of the seminar. We need new prognosticators that will more accurately tell us how patients are likely to do and we need more reliable predictors to tell us what we should treat those patients with, that’s an important theme of this meeting. The genes that we’ve identified in the embryonic stem cells, which nobody has ever analysed before, if we compare those genes to all the other genes that have been used in all the other prognostic signatures, they’re virtually unlike them. 70% of them are new, never been thought about before. So what we would suggest is the following, that as we refine the signature we start to determine whether there are women whose tumours are enriched for these genes, and in those cases where we see stem-like tumours, true stem-like tumours defined by this functionally derived signature, because they were derived on the basis of the function of a particular group of cells, that we would then treat those people with antagonists of pathways that we know to be relevant to the growth of these cells. Cancer is a caricature of development, we’re understanding normal development better so let’s use this knowledge gain from basic science to create the wisdom to treat people more effectively.

A number of them were growth factor receptors but there are genes in many other categories, we’re just starting to scratch the surface of it. Literally, what you heard has never been heard before. The paper was just accepted so I’m now free to speak with you about it; it will be published soon and then we hope that the entire community will look at this carefully and critically and that we will all then work together for the benefit of the patients who suffer from these diseases. I also said that basal-like cancers, a subset of them, a significant subset, had showed enrichment for these foetal mammary stem cell genes, that isn’t the only class of tumours that show that. Some tumours in the HER2 class do as well, that’s important because we know that that pathway is important for the growth of these cells. But there are also tumours in the luminal classes and, although luminal class tumours are thought of as generally being responsive to anti-hormonal therapies, because they often express the oestrogen receptor, what I have learned is that there are many people with luminal tumours who do very, very poorly in the long run, so clearly we need better treatments for them too and those are the most numerous cases. Some of the most important targets are in people with triple negative breast cancers because some of them do very well with chemotherapy but those that don’t do well do extraordinarily poorly and we need to be able to treat them better. We hope that this is one inroad into that more successful treatment.