Era of Hope 2011 Plenary Sessions

Session Title: Morning Session – Roads Less Traveled: New Perspectives in Breast Cancer Research

Title of Presentation: Interplay Between the Immune System and Breast Cancer

Stephen Elledge, PhD, Harvard Medical School: Next speaker is Peter Lee. And the title of his talk is Interplay between the Immune System and Breast Cancer.

Peter Lee, MD, Stanford University: Good morning. I want to especially say good morning to the advocates and express my admiration for what you do and appreciation. So right now I’d like to take you outside of the cancer cell and talk about what role the immune system plays in breast cancer. So for years people really assumed that the immune system plays no role in cancer because cancer is self and the immune system is not supposed to respond to self.

So why does immunity matter in human cancers? So in fact, most human tumors are infiltrated by immune cells and there’s mounting evidence now that when tumors, breast tumors are infiltrated by T-cells patients have better prognosis. And this positive association between T-cell infiltration and favorable clinical outcome is seen in all cancers looked at including GI cancer, ovarian, prostate, lung, and really all others.

And more recently comprehensive meta-analysis has concluded that tumor infiltration by T-cells is an independent and very strong prognostic factor of survival in breast cancer, so all these findings strongly support the notion that the immune system can in fact respond against cancer.

So really when we think about clinical outcome, we really should be thinking about a balance between two forces—the growth of cancer cells on the one side and the host immune response on the other side. Or if you’re a sports fan, the balance between offense and defense. So of course a lot of our attention really is focused on trying to slow down the offense by targeting the cancer cell; there should also be benefit in enhancing the defense doing immunotherapy.

But one major hurdle that we face is that cancer can actually induce immune dysfunction of the host. And when that happens the immune response is diminished and the—the balance is tipped in the favor of cancer growth. So over the last 10 to 15 years, a number of labs including my own have been using different modern immune methods and uncovered multiple immune defects in cancer patients. These include energy and apoptosis of anti-tumor T-cells, immune signaling defects leading to suboptimal immune function, increases in regulatory T-cells to suppress other immune cells, and really abnormalities in all other immune cell types that one can look at including dendritic cells and macrophages.

So how do we put all this together? So we’ve been taking a more systems approach to try and understand the breast cancer immune interplay in patients by focusing on the three major compartments in which tumor cells and immune cells interact. Of course, that includes the tumor compartments, also the draining lymph nodes, the axillary nodes in breast cancer patients, and of course the blood. And we used multiple different experimental approaches including histology, functional assays such as high-dimensional flow-cytometry and phosphoryl flow and also combined genomics and bioinformatics approaches.

One compartment that I particularly focus on is the tumor-draining lymph node. Most people think of the draining lymph nodes as simply soil for two—breast cancer cells to invade, but in fact lymph nodes are very important immune organs. And so how can breast cancer cells survive in this what should be a hostile environment for cancer cells and in fact often take—take over in these lymph nodes. So I really see what happens in a lymph node to be a very important determinant as to the efficacy of the host immune response.

So I just have time to give you two very, very quick examples of what we’re doing to try to interrogate the biology of tumor-draining lymph nodes. The first example is from Combined Genomics and Bioinformatics Analysis.

So when tumor breast cancer invades the draining lymph nodes one usually thinks of evolution of the tumor cells and genetic changes but what if it’s also facilitated by changes of the immune response within the lymph nodes so the door becomes open. So we wanted to address that question and we did this by looking at patients that do not have lymph node metastasis so N0 patients; these are the green—green bars and—or patients with lymph node metastasis, so N1 to 3 patients and those are the red bars.

And we focused specifically on the tumor-free lymph nodes so that we could really ask what’s happening immunologically in the lymph nodes rather than look at the response to tumor cells within a lymph node. So we did gene expression analysis of immune cells within tumor-free lymph nodes and what we found was that in fact even tumor-free lymph nodes clustered together between N1 to 3 patients and away from N0 patients suggesting that biologically these immune cells have different biological properties.

So through bioinformatics we looked at the gene expression patterns of these different populations and on the bottom we looked at different gene pathway databases, and we found that in N0 lymph nodes the—the immune cells pretty much preserved the expression of a lot of the immune genes and pathways as you would expect in fairly healthy lymph nodes. But in contrast in N1 to 3 lymphocytes—lymph nodes—they have primarily down-regulated all of the immune-related genes and pathways. And so that’s summarized here in the—on the top.

So comparing both tumor invaded and tumor-free lymph nodes from patients that are N1 to 3, so they have lymph node invasion, their immune cells all have down-regulated their immune signaling pathways already and in addition, in tumor-invaded lymph nodes they’ve also up-regulated cell cycle genes.

So going across the compartments looking both upstream in the tumor and downstream at the blood, we see similar patterns. So when we focus on N0 patients when we look at the immune cells that infiltrate their tumors, we find that they are primarily preserved in their immune-related signaling and that’s the case as well in the blood and as I’ve shown in the lymph node.

In contrast, in patients that are N1 to 3, the immune cells from the tumor actually have down-regulated immune signaling and up-regulated the tumor promoting signaling pathways and that’s the same with the blood and in the lymph nodes. So they’re up-regulating the inflammatory signaling pathways and B-cell signaling pathways, so the picture that emerges is that the defense in patients N1 to 3 are not only ineffective but in fact they’re helping the other team’s offense.

So we’re using a lot of bioinformatics tools to try to dissect the gene expression patterns in these different situations to look for therapeutic targets.

So now staying with—within the tumor-draining lymph nodes, I want to switch gears and look—and show you a little bit about what we’ve found using a totally different approach of histology. So histology has primarily been very descriptive and very—and at the most semi-quantitative and what we want to do is to make it absolutely quantitative but also bring in the element of spatial analysis because after all the—the cells exist in tissues with definite architecture and spatial relationships so we want to capture that in a rigorous way.

And to do that we really had to devise a new strategy. First we developed high-dimensional immunohistochemistry staining, so most IHC staining is one and at the most two colors; we’ve now pushed that up to five—even six colors and that allows us to look at a lot more different cell types on the same slide.

By utilizing a high resolution automated spectra-imaging system, we’re able to scan a whole slide and also by spectra-imaging which I don’t have time to explain—can have much better resolution of different markers. We developed our own custom software, image analysis software called GemIdent which I—which helps us identify every single cell on a slide which could be well over 1,000,000 cells. So we get very accurate counts but also gives us an XY coordinate of every cell. And that information can be fed into spatial analysis algorithms to really for the first time ask—what are the spatial relationships between different cell types including different immune cell types and immune cells with cancer cells?

So one area that we’ve focused on is the dendritic cells; the dendritic cells are very important antigen presenting cells and they’ve been widely studied in cancer immunotherapies. What we’ve noticed is that in healthy lymph nodes, dendritic cells don’t just exist as individual cells but—but they exist in—in large clusters so these dendritic cells are the orange-staining cells.

But in most tumor-draining lymph nodes these clusters are pretty much broken up and so the dendritic cells exist in isolation. So to quantify this, we collaborated with a computer science group in Notre Dame led by Dr. Danny Chen and we developed algorithms to help us identify cell clusters in tissue. And using these clusters we’re able to measure the sizes and distributions of—of dendritic cell clusters and as consistent with what I said, in healthy lymph nodes the—the dendritic cells primarily exist in fairly large clusters consisting of around 200 cells or even more.

But in tumor-draining lymph nodes some breast cancer patients, these cluster sizes fall dramatically especially in tumor-draining—tumor-invaded lymph nodes and as a mirror to that when we look for unclustered dendritic cells in healthy lymph nodes only a very small fraction, less than 10% of dendritic cells exist as unclustered cells. And disproportion increases dramatically particularly in tumor-invaded lymph nodes.

So dendritic cells interact and—and stimulate with T-cells, so we wanted next to ask—does this change in clustering patterns lead to any change in the DC’s ability to interact with T-cells? And so we extended an algorithm to look for DC–T-cell interactions and here’s what we found.

So now let’s focus on the right side first; so when we looked at unclustered dendritic cells, we found that healthy lymph nodes as well as tumor draining lymph nodes behave similarly. They have approximately 50 T-cells within a certain radius around each dendritic cell and that’s probably just due to packing of cells within a lymph node, so probably that’s—says that there’s really no productive interaction between T-cells and unclustered dendritic cells.

Now when we looked at clustered dendritic cells, in the healthy lymph nodes, the average goes up from about 50 to about 80 to 90 suggesting that the density of T-cells increases significantly around clustered dendritic cells as evidence that the dendritic cells are influencing and interacting with T-cells. In contrast, in the tumor-invaded lymph nodes, this value remains about 50 which is about the same as unclustered DCs suggesting that even clustered dendritic cells in these tumor-invaded lymph nodes have little influence on surrounding T-cells.

So this really clued us into looking at the biology of dendritic cell clustering and perhaps strategies to enhance dendritic cell clustering and function within draining lymph nodes as an immune therapeutic opportunity.

So to summarize, cancer progression is the balance between cancer growth and host immune response. Cancer induces immune dysfunction which tips the balance to cancer progression. Tumor-draining lymph nodes represent key sites where cancer and immune cells interact. By using a combined genomics and bioinformatics approach, we found that lymph node metastasis accompanies down-modulation of immune pathways in the lymphocytes and even in tumor-free lymph nodes.

And using a quantitative spatial analysis approach of histology, looking at the tumor-draining lymph nodes we found that dendritic cell clustering is disrupted with reduced influence on T-cells.

So after probably a several decades of hope but failures in cancer immunotherapy, I believe that it’s finally a field that is coming to its own now. The FDA has recently approved two cancer immunotherapies over the last year and a half. The first one was Provenge which is a dendritic cell-based vaccine for prostate cancer; that was approved in April of 2010. And that was followed within a year by anti-CTLA4 antibody which can stimulate T-cell activation and this is a treatment for metastatic melanoma.

The clinical benefits of both of these treatments however has been quite modest with the extension of survival only in months—about 4 months. So we need to do better; I think we can do better and we certainly need to bring immunotherapy to breast cancer. And to do this I believe that we need to develop combination immunotherapy. Combination therapies has been somewhat of a theme in this meeting—and that needs to combine vaccination with immunologically validated breast cancer antigens which is as an aside the topic of a recently awarded DoD Multi-Team Award between myself, Dr. Joe Gray, and also Paul Spellman at Oregon and Dr. Joe [Salenski] and Dr. John [Keppler] in Colorado—in Denver.

But also important would be strategies to restore and maintain optimum immune function in vivo. And so hopefully by targeting—by disrupting cancer immune dysfunction we can once again favor the defense so that they can go and crush the offense.

So I’d like to end by acknowledging people in my lab and by important collaborators. Thank you.

Steve Elledge: Yes?

Question: Hi, Patient Advocate; my question is—is there anything in your research that leads you to believe that we will no longer have to interfere in the lymph node system because the lymph system is a pretty important part of the body function and we’re messed up for the rest of our lives?

Peter Lee: When you say interfere are you talking about surgically removing?

Question: Correct

Peter Lee: I—I certainly share that hope and goal. I believe that we have that opportunity. I can—all I can say is that I agree and I’m working towards that.

Question: Thank you.

Question: So your gene expression analysis kind of suggests that there’s a causal effect because even in the non-tumor containing lymph nodes there are the differences. What do you think is the relationship for the clustering? Is it a cause or is it an effect to the cancer?

Peter Lee: So are you talking about the dendritic cell clustering?

Question: Yeah.

Peter Lee: Good question. Right now it’s a correlation so we—we need to investigate further to—to actually establish cause and effect. My hunch is that something upstream within the tumor compartment is changing biology of dendritic cells. And as you know the dendritic cells traffic between the tumor and the—and the lymph nodes, so when the dendritic cells are disrupted perhaps in the tumor site that when they traffic back to the lymph node they’re not able to do what they’re supposed to do.

Question: So—so we all know that chemotherapy impacts pretty severely on the immune system, so do you have any idea how chemotherapy impacts on these clusters of dendritic cells?

Peter Lee: Another good question, something that we’ll definitely look into and I believe that we—we have that opportunity by looking at women that get new adjuvant therapy—chemotherapy first and then looking at the lymph nodes.

Question: So I think you have some very elegant powerful techniques but I would make two points. I think you would have more power if you can distinguish because you’re taking the whole lymph node, particularly these genes because the assumption you know the tumors and the immune cells share many pathways in genes. So looking at the whole thing together may be not as informative as if you can somehow use markers which separate them. The second thing is—is the correct control a healthy lymph node or is the correct control a lymph node from an infected infection of the draining lymph node?

Peter Lee: So thank you for your suggestions. We—I think for your first point, we do actually independently separate the immune cells from the tumor cells when we do our analysis. And the second point, yeah; the—the comparison, so we have also in addition to healthy lymph nodes which by the way has actually been very challenging to get, we also have inflammatory lymph nodes as comparators as well as of course non-cancer invaded lymph nodes from patients. So we—we want to be as—as comprehensive as we can in looking at different—the biology of different types of lymph nodes.

Question: Very straightforward question; do the cluster versus single dendritic cells differ with respect to activation markers, Class 2, CDADC, et cetera?

Peter Lee: Yeah; so that data that I didn't get a chance to show, yeah there are a lot of differences that we’re seeing, for instance, maturation status and activation status. Okay; the last one in the middle?

Question: I am an 18-year survivor of intra-cystic papillary carcinoma and it crossed my mind that a study of those antibodies which surrounded my tumor might be of interest or is science way past what they could teach?

Peter Lee: I’m sorry; could you repeat the question again—sorry?

Question: Yes; the—my tumor was surrounded by antibodies in—in the cyst. And it’s a little unusual but not totally unusual and I was wondering if those tissues from people with my diagnosis could be used for research?

Peter Lee: Oh okay; certainly—I’d be happy to talk to you after the session. Thank you.