Lewis: So, of course cancer can be many diseases but many cancers start out just as cells in the body that are doing their normal jobs and they may acquire a mutation that causes them to begin to divide out of control in many tissues. Prostate is a good example, I study prostate cancer now. You’ll have tissues in your prostate that started divided of control and actually, at this point they’re not really that dangerous. So, I’m gonna just create a lump and your immune system may clear out some of it and your body can usually deal with it. But, the issue is that every time a cell divides it has to, you know, faithfully replicate all its DNA information, and every time you replicate it there’s a chance you can have an additional mistake. So, the more times you divide a cell you sort of have this accumulating chance of having more mistakes, more mutations, and some of these mutations will cause the cells to do other dangerous things like break down the tissue around them, and maybe start moving out of the tissue, and that’s when they become very dangerous. So, you know, all the things the cancer cell does are things that cells have the intrinsic programming to achieve, right? And really in your biology of your body we have the challenge that we have the sort of single template of code in our DNA that has to do, you know, tons of different things, right? So, it has to create an epithelial cell that sits in a tissue that provides structure, and sort of sits around and may do very little. And then, you have immune cells that are highly plastic, they’re changing, they’re moving around, they’re adapting to their situations, and so all of the genes that make an immune cell be able to transit the entire body and invade different tissues, all those genes are in an epithelial cell that’s sort of sitting around creating structural integrity for a tissue. So, these mutations can basically change the programming of an epithelial cell, you know, and potentially hijack some of the programming of an immune cell so that they basically have properties of both. And because these cells are dividing, some of them may die and some of them may survive and persist. That’s an evolutionary process that, you know, basically selects for cells that can survive and get out of this environment. So, cells that are dividing out of control can have years and years of selection, right? So, they may divide for ten years without being an issue and then one day they all accumulate a mutation that allows them to break down the membrane in the basement membrane that’s surrounding that tissue and then suddenly they can escape.

Leigh: John and his team discovered eleven genes that are widely involved in the metastasis of cancer cells, but which are also not unique to any one type of cancer. So, Ryan and I were curious how this approach built on previous advances in cancer research.

Lewis: The way cancer develops and is diagnosed typically, it’s either diagnosed as metastatic already or not metastatic. So, before we got involved, guy named Peter Brooks, in the lab of James Quigley, in 1993 actually did a really cool experiment. He had the same idea that we had later, you know, what makes a cell that is able to successfully get out of a tumor and go somewhere else in the body and survive different from a cell that’s actually a cancer, but sort of sit still and isn’t deadly. So, he did this really neat experiment in mice where he was able to basically take a cell that could create tumors every time in a mouse. So, you put a few of these cells in the mice and it would make a tumor every time, and then he created a clone of that tumor that would make a tumor every time but also spread in every experiment. So, when he actually immunized mice against the non-metastatic tumor and he was able to basically tolerize that Mouse to that tumor, so the mouse is producing antibodies against it to try to get rid of it. So, he was able to tolerize the mouse against that non-metastatic tumor, and then he went and challenged the mouse again with the metastatic tumor. So, now the mouse is seeing this sort of the same tumor cells but is creating antibodies only against those unique features that are different from the original cancer cell. So, we came on the scene when we’d identified one particular antibody that recognizes a protein called tetraspanin CD151. It’s not a particularly remarkable protein except that when you give this antibody to a mouse with metastatic cancer, it completely blocked the spread of that cancer. So, one of the antibodies on the screen recognized a unique protein that was only on metastatic cells, this tetraspanin CD151. And, when you put that antibody, you know, in a mouse that had a metastatic tumor, the spread was completely blocked. So, we did an initial experiment doing imaging on these tumors in this model system where we realized that the antibody wasn’t actually killing the cells and it wasn’t even preventing them from dividing or, you know, multiplying, but what was doing a great job of is actually just creating stickiness between the cells, so increasing the adhesive properties of the cells which prevented them from escaping, and actually completely shut down metastases.

Watkins: John likens this approach of blocking the spread of cancers by increasing their adhesiveness to it being a sort of “tumor glue.” He and his team, however, ran into some challenges inhibiting the CD151 gene, as he describes next.

Lewis: So, we got excited about that, we published a nice paper about that, and we thought about great! we have a target now, this CD151, we have an antibody that recognizes the target. Now, maybe we can actually take a drug and bring it to clinic and, you know, the first drug to block metastases. So, unfortunately we know CD151 is on the tumor cells, but what we learned sort of pretty quickly after that was there’s a lot of CD151 in other sells. So, it’s on the cells that line your blood vessels, in the cells it’s on, and this is important platelets. So, the coagulating cells that circulate around your bloodstream, and when you put the antibody with platelets across links them and creates this sort of catastrophic coagulation event. So, as a potential clinical drug for cancer it was pretty much a no-go from the beginning. So, at the time we thought, oh geez, so what are we gonna do? Maybe we can engineer the antibodies sort of get around this. The platelet makes it specific for tumor cells over platelets and we didn’t really have any great ideas. What we did think of though, and this happened all the way back in 2006, we thought that okay, so when we image these cells, when the antibody is there, these cells look completely different, right? So, instead of sort of reaching out and invading into all the surrounding tissues, it’s quite remarkable the way they do that, when the antibodies there, they sort of form a really tight compact ball and just by looking at, its really obvious the difference between the two. So, we thought, at that time, if we had used a method to screen the entire genomes for every gene in the genome to look for other genes when we targeted them, that would create the same sort of phenotype, the same compactness in the cells. Maybe we could identify other therapeutic targets or other drug targets that could produce the same effect as what we saw with this antibody. And so, that’s sort of the project that was conceived at that time that we brought forward and it took us about 12 years to realize the potential of it.

Leigh: Before testing the ability of these therapeutic targets to block cancers in lab mice, John and his team first carried out their experiments on a much faster growing animal model: fertilized chicken eggs whose shells had been removed in order to access the embryo during the three weeks in which they develop. Ryan and I wondered where these eggs come from, as well as how researchers in John’s lab learn to de-shell the eggs without disturbing the fragile membranes inside.

Lewis: Currently so we’re fortunate now. We actually have a large agricultural research faculty and they have a whole cohort of chickens and produced these fertilized eggs on a regular basis. So, we have a constant supply of fertilized eggs these days. And, the shells for us are dispensable, so we actually use what we call an x-ova version of the chicken embryo. So, to put it in context, when an egg is fertilized it takes about 21 days for the embryo to develop inside the egg, and then for the chicken to hatch, and we basically work in that three week window to do our cancer metastasis experiments. So, we take the freshly fertilized eggs, we have to put them in sort of an agricultural incubator that sort of rocks them back and forth, similar to what the, you know, the mother hen would do, moving the eggs constantly, and after four days though they can survive outside the shell. So, actually every single new trainee that comes into my lab we sit them down with a hundred and twenty fertilized eggs, a lab coat protective goggles, and a dremel tool with a cutting wheel, and we say look, you have to work through these 120 eggs and and try to get them as perfect as possible. So, most trainees at this point have pretty good hands and they’re nervous and they want to do a good job. So, maybe out of the first ten they might explode maybe two or three, but by the time they get to ten they’ve usually got the hang of it and after that, you know, maybe out of 120 eggs they get 90 or 100, but every once in a while. We had one trainee in particular who was very excited about doing research wanted to do a research on cancer. He was actually a high school student who was doing a project in the lab. And, he and his partner, I sat them down with 120 eggs and gave them 60 each and asked them to use the dremel tool to open them up, and his partner went through and did the typical thing maybe broke one or two and got through the eggs. He sat down and the very first egg, before it even touched the dremel tool, he exploded it in his hands. It just went everywhere and all over his goggles and then the next one he, you know, he jolted in the dremel right through the egg and anyway. So, we get to about ten eggs and every single area had exploded, and we thought, you know, do we stop them or do we let him keep going? And yeah, he didn’t get a single one. So, we stopped him at about 25 and we said okay that’s good you need you tried your hardest, just never ever touch an egg again.

Watkins: Doug and I were interested in learning how John and his team maintain the de-shelled chicken embryos as they mature, as well as what parts of the chicken embryo they test their anti-metastasis therapies on.

Lewis: We cracked these of four day old chicken embryos into a dish, and what they need to survive in that environment is humidity, so over 85% humidity and they’re happy. So, the membranes don’t dry out and about 37- 38 degrees Celsius, which is about body temperature, and if you keep them under these circumstances they’ll continue to develop and grow just like they would inside the shell all the way to the point where they would hatch. So, what the membrane is interested in is not the one that you see out sort of on the when you’re peeling a hard-boiled egg for instance, and you have to peel off that membrane in a fertilized egg they create, what’s called a chorioallantoic membrane. This is a membrane that you’ll never see in an egg they get from a store, but in a fertilized egg it basically forms on the inside surface of the shell. It’s full of blood vessels and it’s basically needed like the placenta. The lung it’s taking an oxygen and that’s feeding the embryo, and so now you crack the egg into the dish, and so this chorioallantoic membrane that normally forms on the inside surface of the shell now forms on the top surface of this embryo. So, you can imagine we have maybe a three inch square dish and as this embryo develops it’s now getting about ten days the chorioallantoic membrane spreads out of the entire top surface of this embryo. So, just to describe it, it’s really almost completely transparent membrane, it’s full of blood vessels, so it’s got a network of blood vessels, it’s it’s got a direct connection to the heart, so it’s pumping blood through constantly and it’s really thin. So, it’s, you know, a fragment of a millimeter thin, you know, a fraction of an inch thin and transparent, and what’s nice about the model at this point, before it becomes a chicken, is that it’s immunocompromised. So, it’s immune system hasn’t fully developed yet, it doesn’t have antibody producing cells, it doesn’t have, you know, these killer T cells. And so, what that means is that we can inject human tumor cells into this membrane and they’ll grow and form tumors, and spread and metastasize just like they would in a human.

Leigh: After injection, many of these cells make their way throughout the embryo then back to the chorioallantoic membrane. We asked John to describe this process in more detail,. as well as how they go about tracking the development of the cancer cells that they injected into the chicken embryos.

Lewis: Between day ten and day twelve, this chorioallantoic membrane is sort of fully covered. The top surface it’s transparent, it’s very thin, and so we come in with a micro injection needle. For the purpose of these kinds of experiments, we can inject single cells into a vein in this membrane and typically will inject 20 to 25,000 cells at a time. When you do that they basically go to the the venous blood vessels which go straight into the heart and all of those 25,000 cells get pumped through the entire body and back into the chorioallantoic membrane of this chicken embryo. And each one of these individual cells, once they get back into the capillaries of this membrane, they’ll get stuck basically. So, the cancer cells are quite large compared to some other cells and a lot of the chicken cells. So, they’ll get stuck on these really tiny capillaries that sort of are all over this membrane as individual cells, and then over about an 8 to 12 hour period after that, they’ll sort of start moving around, and then they’ll do what we call extravasation. So, they’ll extravasated out of the blood vessels and into the surrounding tissue so they escape the blood vessels, and then you have basically, if you look at it, 12 hours after you inject these cells and in this case we’re actually videotaping these and using a fluorescent microscope. So, it’s like a Christmas tree and they’re labeled with a green fluorescent protein so we can see them. And so, under the right wavelength of light the chicken embryo lights up with all these green spots, and every one of these green spots is a single cancer cell that sort of circulated around, got stuck, and then escaped the blood vessel.

Watkins: Doug and I were struck by how extensively cancers can divide and migrate within as little as two day’s time. We wondered how it is that they’re able do so, as well as how John and his team went about inhibiting a single gene within each embryo in order to test the effect of that gene on metastasis.

Lewis: So, cells naturally use a variety of mechanisms to both migrate and invade into tissues. So, typically cells are polarized and sort of ones the business-end that’s reaching out and the back end is sort of letting go from the cells around it. So, the front end is creating these finger-like protrusions, and depending on what they’re made up of they can be called filopodia, they can be called lamellipodia, and if these structures happen to have enzymes at their tips, that break down anything they touch and sort of create a hole for them to move, then they’re called invader podía. And so, the cell is basically sticking out all these arms ahead of it sort of searching the surrounding area looking for a permissive place to sort of migrate through and in some cases will be sticking out these invader podía that will actually break a hole through the whatever’s in front of them and then the cell will slowly slither its way through that hole and get out. And so, going back to sort of that Christmas tree thing, so when we inject all these single cells and say five to six thousand of them gets stuck in this coil untuk membrane, so over a few days those will divide and depending on how sort of invasive and migratory they are, they’ll either spread out or if they’re not migrating they’ll just stick in one place and create sort of a dense spot. So, what we decided to do is use a library of constructs of tools, genetic tools, to allow us to inhibit every single gene of the genome and we mix them with the cells such that each individual cell would have a single gene that was affected. And so, you can imagine you have 5,000 cells and in each one of these embryos each one has a single gene that’s been knocked down or inhibited, and then we’re asking the question what is the effect of that gene on that individual cell as it divides? Does it affect its ability to spread out? And if it does, it’s really obvious. So, those cells that divide and and are inhibited from invading and migrating, produce a very compact really bright spot. And so, we basically surveyed the membranes of all these animals and picked out all the bright compact spots.

Leigh: Just because the genes worked to halt the spread of cancers in animal models doesn’t necessarily mean that they’re equally relevant to humans as well. So we asked John how he and his team went about identifying which of the genes that they investigated might be most likely to be associated with cancers in humans.

Lewis: So, obviously we’re doing this in a semi artificial system. We’re testing human cells screening in a chick embryo but validating in mice. So, there is a little bit of human in there but it’s a little bit removed from the human condition. So, one thing we did to try to get at, you know, how physiologically relevant are each of these targets. So, we looked at human cancers particularly those cancers where we had information from, both the primary tumor and a metastatic tumor, in that same patient and ask the question, are these targets up-regulated in the metastasis? And, we found significant associations between several of the targets and different cancers so KIF3B and in prostate cancer for instance, and our 2F1 was more prevalent in ovarian cancer. So, we think probably the different targets will have cancers where they are particularly relevant. The best way to do that is experimentally try to figure it out, but because KIF3B has such a strong phenotype, we think it might be probably more appropriate to go after first and should have relevance in multiple cancers. So, just to give you a bit of background, KIF3B is what we call a kinesin motor protein. So, your cells basically have these transit routes called microtubules that sort of go from the inside of the cells at the outside of the cell and they transport things along those routes, and these kinesins are actually a motor that walks along these highways and things to the outside of the cell and back in again. So, we think KIF3B is important for trafficking or taking things to the outside of the cell that are important in cell migration.

Watkins: John and his team’s methods were over 99% effective in blocking the metastasis of cancers. We were curious to learn what methods he and his team used to quantify this.

Lewis: I’ve got a great story about that 99% as well, but I’ll tell you the broad strokes first. So, the premise of the study was that if we really wanted to learn about how to block metastasis we really had to do it in a living system, because the way cells behave in a dish, you know, on plastic or glass when they grow typically in a lab can be completely different from how they behave in a complex three-dimensional system. And of course, in a dish you can’t replicate the fact that there are maybe 20 to 30 to 40 cell types, all sort of mingling together in a tissue. And so, the goal for this study was to use as complex system as we could. In this case, this EXO vote chicken embryo to model cancer cell metastasis in a way that, you know, if we screened and discovered these new targets that we had a good chance that they would have a significant effect on metastasis. We used two different methods to determine that. So, one of those was using fluorescent imaging of the lungs. So, we basically took the whole lung, put it under wide fields fleshlings microscope, and looked for those green cells at its breadth, and by visually, by eye, there was a significant difference between if you had those gene targets in there or if you had the control. So, it was a night and day. But, the second method we used was this really sensitive genetic method called PCR. And, PCR is an amplification reaction that amplifies really rare pieces of DNA, it can pick up, you know, as few as 10 cells and tell you very precisely how many cells are in each organ. So, the idea here was to amplify the human DNA out of the chicken, all the chicken tissue in the background. And sort of as a little technical detail of PCR, basically PCR uses a fluorescent chemical that’s in the reaction and as soon as that fluorescent chemical sort of starts to increase in your reaction, you basically say okay that’s positive.

Leigh: We followed up by asking John to share the “eureka” moment when he realized that the eleven genes he and his team identified could block over 99% of metastasis among cancer cells. We’ll hear from him about this revelation after this short break.

Ad: We Share Science

Leigh: Before the break, John was about to describe the when he realized that the methods he and his team developed could block over 99% of metastatic cancer cells.

Lewis: so I’m looking at the raw data and plotting it out and, you know, there was a significant difference. But, it looked like it was only about a 20% to 30% reduction in metastasis, so we were excited, but we weren’t sort of super enthused. And then, one day I’m looking at the data in lab meeting and I realized that we’re actually not plotting the number of cells that have metastasized, but the number of cycles in the PCR machine that it needed to basically get to that threshold. But, you can imagine this is an exponential reaction. So, every time you go through a cycle you’re exponentially increasing the number of copies of that, you know, whatever cell, so what in fact looked like a two-fold increase or a three-fold increase was actually, you know, a thousandfold increase. So, when we went back and ran the numbers, this is like oh my god, every single one inhibits over 99% and so that completely changed the temperature of the lab and our enthusiasm about getting the information out there. So, I’m glad we reanalyzing them, you know, I’m glad we finally discovered it because for us this is, you know, unprecedented. So, now we have, you know, 11 genes identified using a similar method five of them we’ve characterized a pretty serious detail and we tested the same cells in this model as we did in the screen. So, we know obviously that’s the best-case scenario is 99.5 percent reduction but we tested in some other cell lines when we got over 80 percent reduction in metastasis in those. So, I think, you know, these are sort of 11 shots on goal to be able to develop drugs that should block metastasis. And so, you know, I conceived of potentially the future where we identify a sort of aggressive cancers and give a prophylactic treatment to prevent any spread or any further spread but I think what’s really interesting and though is the study of metastatic cancers that have already spread. So, there’s sort of a thought out there that once the cat’s out of the bag, you know, what is an anti metastatic drug is gonna do? And there’s some really really neat studies out from a number of different labs showing this, but typically though what happens is that once sort of the initial metastases form say it’s a single metastasis those metastatic cells in the metastasis continue to evolve and actually seed other parts of the body. So, the spread during metastasis can happen both into a metastatic site back from the metastases into the primary tumor and then from that metastases to other sites in the body. So, it’s a very dynamic process and if you can if you think of it that way blocking the spread at any time should have a significant effect on survival.

Watkins: Given the potential of the methods that John and his team identified, we were eager to learn how long it might be until we can expect human trials to take place, as well as what that process will entail.

Lewis: I’ll say that you know it’ll probably take longer than we would like, but there’s some really positive aspects of sort of the list of genes that we’ve come up with. So, there’s a variety of different genes, a couple of them are enzymes, and enzymes have an enzymatic activity that can be blocked by sort of well-known drugs, a couple of the targets have had drugs or small molecules developed against them already. So, we have a place to start in developing a drug that could be used in humans. So, that’s really promising. A couple of the proteins we identified our cell surface proteins, and cell surface proteins can be relatively easily addressed using an therapeutic antibody, and therapeutic antibodies are used in the clinic all the time for many different diseases and sort of the development and creation of drugs around antibodies. This is relatively straightforward and then there’s a couple of proteins in there that we don’t really have a good idea of how we might target, but potentially because there are no currently developed drugs or any clear way of developing drugs, the one way we could do it is basically the same way as we did in the study. We used a genetic approach called RNAi inhibition and so that RNAi, an ambition in our mouse model, decreased metastasis by over 99%. So we know, you know, in the real world the drug using the same approach may not be as effective, but we know it possibly could be effective. So, one of the things we’re gonna work on is developing an RNAi inhibition type of drug to accomplish the same thing. Now, as far as time, I mean sort of the cost and the time to do clinical development is a bit daunting. But, I think now that we have the targets and we’re comfortable about moving forward with sort of the top two or three and actually if we adopt an RNAi approach, I think potentially we could be in the clinic within 3 to 4 years, and of course that would be the very first phase of clinical development. So, it’s a phased program of clinical development and approval. So, phase one of the study, which we might start in three to five years, would be it just a safety study and then followed by phase two, which is looking more at the activity in efficacy, and then phase three typically compares it directly to whatever is used at the current standard of care to see if it’s better. So, typically that can you know the whole thing from beginning to end might take a minimum of eight years and a maximum of up to fifteen years. So, we’re looking at a few years ahead but I think we’ve got a pretty good basis from which to start that.

Leigh: Lastly, Ryan and I were interested in hearing John’s thoughts on the promise of developing genetic therapies that could halt the progression of cancers, whether it’s through the use of CRISPR for gene editing or some other method.

Lewis: We’ve spent many years in research sort of trying to figure out what is the sequence of our genome, what does it mean, what are all the proteins in the body, what are all the metabolites circulating through the body. But, our ability to sort of intervene and make changes to something that, you know, obviously the body is extremely complicated but exquisitely programmed. So, once we could decode that programming, the question then becomes what are the tools we can use to basically perturb that programming in a way that would help health. So, I think technologies like CRISPR and other genetic technologies are potentially very exciting, but we’re still in the very early days of CRISPR. So, despite all the excitement and the hype, we’re not gonna have a successful CRISPR drug in the next couple of years. It’s going to take, you know, several years to get to the point where we can reliably and safely edit genes, but that day is coming. So, I’m really both excited and terrified by genetic technologies that could use something like CRISPR to propagate changes to the entire body, these have already been demonstrated. So, those we know they’re possible and I think they’re really gonna transform medicine. So, once we can specifically edit any gene, we want in a safe and reproducible way and most importantly propagate that change throughout the body. I think we’re going to be in a completely new era.

Leigh: That was John Lewis, discussing the article “Quantitative in vivo whole genome motility screen reveals novel therapeutic targets to block cancer metastasis,” which he published with Konstantin Stoletov and thirteen other researchers on June 14, 2018 in the journal Nature Communications. You’ll find a link to their paper on https://www.parsingscience.org/e33, along with bonus content and other material discussed during the episode.

Watkins: Since we launched it in August, Parsing Science’s weekly newsletter on the latest developments in science continues to gain new subscribers every day! You can sign up at: https://www.parsingscience.org/newsletter/, or if you’d like to check out our first seven issues, go to: https://www.parsingscience.org/newsarchive/. That’s one word, “newsarchive.”

Leigh: Next time on Parsing Science, we’ll continue to explore the latest research into cancers with Mike Feigin from Cold Spring Harbor Laboratory in New York. He’ll talk with us about his discovery of mutations in part of the genome that most people have so far tended to ignore, but which regulate the expression of genes that drive the formation of cancer tumors in the pancreas.

Michael Feigin: We think that these mutations are causing a decrease in gene expression, and we know from the shRNA gene studies that a decrease in expression of these genes is, you know, really important for the ability of these cells to grow. And so, we think that’s how this process is working is that the decrease in gene expression of these genes is critically important for a growth of the cells.

Leigh: We hope that you’ll join us again.