A study published in Cell provides an unprecedented look at the dozens of molecular steps that occur to bring about endometrial cancer, commonly known as uterine cancer. The study offers insights about how physicians might be able to better identify which patients will need aggressive treatment and which won’t, and offers clues about why a common treatment is not effective with some patients. The study, funded by the National Cancer Institute, also suggests a potential role for already-approved drugs that target proteins known as CDK12, SMARCA4 and PML in other types of cancer. One could say that the work is a detailed molecular snapshot of endometrial cancer – except that the information is so vast, touching upon tens of thousands of molecular actors taking part in thousands of interactions at different times, that the work is more like a frame-by-frame video, documenting steps that play out over years in patients’ bodies.
Welcome. I’m your host, Jess Wisse. On today’s episode we’ll talking about how scientists are taking a new approach to better understand and fight cancer.
Stay tuned to learn more.
JW: New research published in Cellshows a never-before look the steps that happen when a woman develops endometrial cancer. This type of cancer affects the uterine lining and it can be deadly. PNNL researchers are using their expertise in mass spectrometry and cancer biology to better target this disease.
Meet one of them: Karin Rodland
KR: I’m Karin Rodland. I am a PNNL laboratory fellow and I'm one of the lead cancer biologists at PNNL. The main thing that I do is provide expertise about cancer biology to the mass spectrometry group that does proteomics and metabolomics measurements of lots of different tumors.
As I got into my 30s and I started really doing this, it's like got my PhD and I thought “Where am I going to postdoc? And what am I going to do in my own research lab?” The number of people I knew who were friends who had cancer—it was just mind-boggling. And I would go to Lions Clubs and Rotary Clubs to do this kind of lay outreach and I would start by saying, “If you or someone you know has had cancer raise your hand.” And every single arm in the room would go up. That's why I do this.
JW: Karin has studied cancer biology since the 80s and she’s one of the top experts in the field.
KR: KR: I ended up at Oregon Health & Science University as an assistant professor in the mid 80s and I was there for 17 years. And I earned the right to go on sabbatical and I came to Pacific Northwest National Lab to learn proteomics because they were the world's best at proteomics and I thought I was going to need that technology for my research. And I came on sabbatical for a year and I really enjoyed the team research philosophy and culture that they do at PNNL. I was totally impressed with the mass spectrometry technologies and the computational biology expertise and I saw a great opportunity to apply all these capabilities to biomedical research and particularly to cancer.
JW: For years doctors and scientists have known that cancer is a genetic disease. Our genes control much of what happens in our bodies, including the way our cells function, grow, and divide. Because of this, cancer researchers have spent a lot of time studying the DNA and RNA of cancer cells.
But Karin and her team looked one step closer. They studied the proteins synthesized by these cells.
KR: What we call the central dogma of molecular biology, is you have the genes and they're the blueprint. And they send out kind of a Xerox copy and that's the messenger RNA. And then the message gets made into proteins and the proteins actually do the work. And the way that they do the work is by being modified with phosphorylation which turns them on and off, or by acetylation which opens them up or closes them down.
So it's easy to measure DNA and it's easy to measure RNA. The technology has been very well developed, it's inexpensive, and it's easy. So scientists and doctors measure what is easy and convenient to measure. So we can measure genes and that's the only tool that most doctors have for doing precision medicine. It's the genes and maybe they can do the RNA. And so all these models have been built up trying to predict disease outcome based on the genes and the RNA. What we found is that when you add the proteins you get much, much more information. And sometimes the information from the RNA is a little bit misleading. It's a little bit different than the information that you get from the proteins, but if we correlate the proteins with the known clinical features we find that the protein modifications are more powerful in this study.
JW: In this most recent study, Karin and her team studied nearly 150 uterine tissue samples. And they did so by using a tool called a mass spectrometer. This is an incredibly sensitive instrument that can measure the smallest parts of a sample. Using the mass spectrometer, the team took many different types of measurements—so many, that they actually took more than 12 million measurements – the most ever taken of proteins for cancer research. They tried to measure everything that they possibly could.
KR: For cancer research, the type of research that we do at PNNL with our great mass spec is discovery research. We're not trying to test the hypothesis. We're trying to study what it is and describe it in in great detail. And then we hand that information off to the basic scientists at OHSU say, and they tease out parts of it and they do a very specific experiment to see what the relationship is. And so that's how the science grows and grows and grows.
JW: This type of research was only possible with an amazing team of collaborators. Like much of the research done at PNNL, this research was done by a multidisciplinary team where each team member is an expert in something unique.
KR: We have a great team at PNNL Tao Lu runs the mass spec and he runs it very well. He knows how to design experiments to make things work on the mass spec. And he just knows how to get people to work together and work well in everything. There's a very large mass spec team that's very great. There's Paul Piehowski who works on the sample processing. And Marina Gritsenko who solves the problems in sample processing. She's a very prominent author on this paper because doing the sample processing was so important.
There's Ron Moore who keeps the instruments running well. Then there are the people who help us interpret the data. Jason McDermott and his team, they take all that mass spec data and they start to make sense out of it and build the pieces of the mass spec machinery. Sam Payne is on our team he was at PNNL he's now at Brigham Young University. Bing Zhang is the lead of the Baylor team and he's been a collaborator with us for ten years. And then the consortium, the CPTAC consortium has brought together a number of high-power labs that do nothing but genomic analysis. PNNL is not as strong in genomic analysis as we are in proteomic analysis, so we've been teaming with the folks at Washington University in the lab of Li Ding and the folks at New York University in the lab of David Fenyo.
JW: Obviously, it’s bad to get cancer. But there are different types of cancer. And the type of cancer a patient has will determine how aggressively is spreads. Karin describes the differences as “bad actors” vs “good actors.” By doing their in-depth protein analysis, her team can now better identify if a cancer is a bad actor or good actor.
KR: When somebody has a tumor the first things that the doctor does is to sample the tumor by a biopsy or removing the tumor if it's small and localized. And then you give it to a specialized kind of doctor called a pathologist who looks at it under the microscope. And for over a hundred years we have a lot of observational data about if the tumor looks like this it's going to behave bad. If it looks like this it's likely to behave well, but we don't know why. We just know that there's an association between what it looks like and how it behaves.
And then there are tumors that we know behave badly when they look like they should be tame tumors. Okay, so there's a type of appearance that we call serous endometrial cancer and it's a bad actor because it doesn't look like a well-developed uterus. And there's a type of cancer that we call endometrioid endometrial cancer and it's a good actor. It's pretty much doing what it's supposed to do—it’s just growing faster than it's supposed to and you can whip it into shape pretty easily. But there's a small percentage of those endometrioid endometrial cancers that become bad actors and that metastasize and kill the woman. And you can't tell it by looking under the microscope. And you can't really tell it by looking at the DNA. And so what we found was the protein behaviors in those bad actors that look like the proteins in the serous type that we know are going to be bad actors.
So, we can look for the common features that define a tumor that's going to be aggressive and nasty and a bad actor. So not only does that allow us to make a prediction about, you know, “You can rest assured you can be comfortable the surgery is going to cure you.” But then if you're not in that nice reassuring category, we can start to do a better job of attacking the problem, of developing targeted therapeutics that are going to attack precisely what is broken in those tumors that are the bad actors. No matter what they look like, it's whether proteins are good or bad.
JW: Not only did the team find protein data to be so rich, they were able to use this information to learn more about immune cells. Tumors attract immune cells. They are a big part of the problem when it comes to the spread of cancer because they can trick the body into thinking tumor cells aren’t dangerous. Think of an intruder wearing a disguise to mask their true identity.
KR: When we look at the tumor, we're not just looking at the tumor cells themselves. We're also looking at the immune cells that have been attracted to the tumor. And immune cells, their job is to kill anything that's foreign. And a tumor cell is a foreign cell—it has changed and mutated. So, it should look foreign to your immune system and the immune system should attack it and kill it. But many tumors make immunosuppressive molecules that tone down the immune system.
So we can actually measure how much of the immunosuppressive nature is there. So one of the hottest therapies in cancer these days is immunotherapy where we stimulate the immune cells to kill the cancer cell. We remove the suppressive factors and we stimulate the aggressive factors and they kill the tumor cells. So with the proteins we can identify how much tumor suppression is there and whether the immunotherapy will work. But even when we stimulate the tumor cells, the immune cells to be active—they have to have certain machinery that allows them to actually reach out and touch the tumor cell and recognize that it's a tumor cell. And so we can also tell whether the tumor cells that are there have enough of this machinery to actually do their job.
So this is going to help us determine whether immunotherapy will work for that patient or not. Because immunotherapy right now is only working in 40 to 60 percent of people. We don't know why it works in some and not in others, but when it does work it's practically a cure. When it doesn't work it can also make you very sick it can stimulate your immune system to attack your healthy cells. So we don't want to tune up your immune system if it's not going to work against the cancer. So this allows us to be more precise in how we use immunotherapy.
JW: There’s one huge benefit to learning more about the proteins of immune cells. With this knowledge, doctors might be able to spare patients unnecessary side effects.
KR: Well most immune therapies make you feel like if you've had the worst flu you've ever had. The early days of immunotherapy used the same molecule that your immune cells make when you have the flu. And patients that I've worked with said we don't want you to research immunotherapy because it makes you feel really horrible and it's not working often enough to be worthwhile.
So, we had to understand the biology enough that we could make immunotherapy successful and that we could also use different strategies that didn't make you quite so sick. Almost everybody who gets immunotherapy feels like they have a really crappy flu. But you know, I'll go through the flu if it will cure my cancer. Some people get what's called an immune storm and the whole immune system just flares up like a thunderstorm and it can attack the heart muscle and that's obviously very dangerous.
JW: But Karin didn’t just find a better way to do immunotherapy. She and her team almost accidentally discovered something that could be a live-safer for patients in the future. With the protein data they were able to identify an alternative use for a pre-existing, FDA approved drug. This other use? Cancer treatment.
KR: Going back to the genome data that was available: So we had a p53 mutant cancer. A lot of endometrial cancers have a mutation or a fault in the p53 gene. That is a gene that normally suppresses growth. So it's what we call a tumor suppressor, so if it's broken it doesn't work—the break is off and the cells grow. But there's no drug that treats p53. Because it does so many different things. It's just difficult to drug.
So by using the studying the protein data instead of just the gene mutation we can see the proteins downstream of p53 that are activated. Their activity is increased when p53 is broken. Okay, so it's like a Rube Goldberg machine, you know. And if you drop the ball into the bucket the chute kicks the mouse. And so if we can't stop the ball from dropping in the bucket, maybe we can inhibit the shoe from kicking the mouse. So by doing the proteins, we can outline the whole Rube Goldberg machine.
And so in this case we identified that a protein downstream of the p53 that was activated when p53 is broken is called cyclin dependent kinase 12. And there is a drug out for that that has been approved by the FDA. Now without our data you would never have thought of using that drug and endometrial cancer, but now that we can see the whole Rube Goldberg machine we can see that maybe the drug against cdk 12 will work in endometrial cancer.
JW: This is exciting because this means that a clinical trial could begin soon. All of this is because of the advanced protein measurements done at PNNL.
KR: To me the big advantage of doing the protein measurements and the phosphoprotein measurements is that we're actually able to track the flow of information in a cancer cell from the external environment that is supporting the growth of the tumor cell to the DNA in the nucleus. So that we're making more cancer cells, and more cancer cells, and more cancer cells so that whole pathway of information is very important. We can't get that from the gene mutations. We can only get that for measuring the proteins and the phosphoproteins so that's the big takeaway.
The second big takeaway is that when you add in information about the phosphorylated proteins it really tells you not only what roads are there, but which roads have the most cars, which are having the most traffic, which is really driving the disease. And that's the information that you need to have to do the targeted therapies that people are working.
JW: Now with improved insight into what the proteins are doing, how they are behaving, and changing over time, patients can receive life-saving medicine, that’s tailored to them, before it’s too late.
KR: Nothing happens overnight, but you know we licked infectious diseases. Maybe we can lick cancer. You know as Brian Druker says, “We want cancer to be something you die with, not of.”
JW: A big thanks to Karin and other researchers like her who are a part of the Precision Medicine Innovation Collaboratory led by PNNL and Oregon Health State University. And with that I’ll let Karin wrap up our latest episode of Pods of Science:
KR: And we’re done!
Music
JW:Thanks for listening to Pods of Science. Want to learn more? Follow us on social media at PNNLab. We're on Twitter, Instagram, Facebook, and LinkedIn. You can also visit our website at pnnl.gov. Thanks for listening.