Read the Full Video Transcript

Andrea Miyahira: Hi, everyone. I'm Andrea Miyahira at the Prostate Cancer Foundation. I'm excited to be joined by Dr. George Butler of University College, London. He also has an adjunct position at Johns Hopkins. He will discuss his new paper, "No Evidence for Peto's Paradox in Terrestrial Vertebrates" that was published in PNAS. Dr. Butler, thanks for joining us.

George Butler: Hey, thank you for the invitation. It's great to see you. Today, I'm going to be talking about a recent publication, "No Evidence for Peto's Paradox in Terrestrial Vertebrates." So I'm going to give a little overview first about what is Peto's paradox because this might be new to some people, then how we tested Peto's paradox, and then some of the results that we found to do with getting big fast and how that potentially relates to mechanisms of cancer defense.

So what is Peto's paradox? So Peto's paradox was an observation that was originally formulated in 1977 by Sir Richard Peto. And Sir Richard said, well, if we think about the difference between a mouse and a human, a human is many times larger than a mouse and has many more cells than a mouse. So we might expect that if we take a really basic view of cancer, that you'd expect the human would have a higher prevalence of cancer compared to a mouse.

And we can extrapolate this same idea out to think about larger species as well, such as an elephant, or a lion, or a giraffe, for instance. So if you were to make this plot on a graph, you'd expect to see cancer prevalence, shown here on the y, to be positively associated with body size. But rather weirdly, when Sir Richard started to look at this, he found that, that observation didn't appear to be true. And since then, other groups have also started to probe this observation and found that, actually, it appeared as if there was no association between cancer prevalence and body size, hence the paradox.

So part of the challenge up until now has been, how do you test these ideas at scale? We're talking about multiple species. We need to get an idea of cancer prevalence, which is very challenging. Unfortunately, earlier this year, there was a paper that was released in Cancer Discovery that published a very large, broad, pan-species cancer prevalence data set. So it spans approximately 16,000 individual necropsies across 292 species. This is a really big data set and a massive effort from the guys.

It covers amphibians, birds, mammals, and reptiles. And it records the prevalence of neoplasia and malignancy. And in this context, neoplasia being the prevalence of benign and malignant tumors, and malignancy being the prevalence of just malignant tumors. And that categorization was performed by board-certified pathologists. So using this data, we wanted to go back and start to ask these fundamental questions that Sir Richard posed nearly 40 years ago.

So we use what's known as a phylogenetic generalized mixed model. Which sounds fancy, it's not, really. The idea here is that you account for the fact that you've got shared ancestry between individual species, and then you can do sort of normal regression analysis. So at the top here, I'm showing you a plot of neoplasia on the y with snout-vent length on the x for amphibians and reptiles. And then in panel B here, neoplasia, again, on the y and body mass on the x for mammals and birds.

And we separated amphibians and reptiles from birds and mammals because of their different growth pattern. So amphibians and reptiles will grow throughout their entire life, whereas birds and mammals will reach a given size and then they won't get any bigger. And we found that when we looked at neoplasia, we found a positive association between neoplasia and snout-vent length and neoplasia and body mass.

And this has been shown before by some other groups. But what was really interesting is when we then looked at malignancy, we also found a positive association between malignancy and snout-vent length and malignancy and body mass. This is really key because this contradicts many studies that have come previously to say, actually, there's no evidence for Peto's paradox in terrestrial vertebrates. Meaning that larger species or longer species are actually getting more malignancy compared to their smaller counterparts.

So then we wanted to step back and ask ourselves, well, how is this happening? We know that we've got really large species, such as an elephant, that roam the Earth. How have they been able to get so big, but yet they've got this increased risk of cancer? Well, getting big is not a homogeneous process. Certain species will get big much faster than others.

So for instance, an elephant, we know, very recently has increased dramatically in size. So we wanted to be able to take this into account. So we accounted for the fact that some species will have a faster rate of body size evolution compared to others. And when we built this into our model, we found that mammals and birds actually have a negative association between neoplasia prevalence and what we refer to as pathwise rate, which is basically the speed at which these species are getting bigger.

So this means that bigger species are still getting more neoplasia compared to their smaller counterparts. But the species that have got big fast have been able to pick up some sort of adaptive mechanisms along the way. And what's really interesting is when we then switch gears and look at malignancy prevalence, we find the same thing is also true. Meaning that while bigger species of birds and mammals are still getting more cancer compared to their smaller counterparts, the species that have got big quickly are getting less cancer.

Which is an interesting thing to think about as we look at these scales constantly between staving off the threat of cancer but being big for other reasons that might be beneficial, for instance, being able to access food or to avoid predation. And if this sounds a bit heady, another way to think about this is—so we know that the bigger species are getting more cancer. What's the effect these rates of evolution are having?

Well, the thick black line here is the amount of cancer that you would predict for a given species of a certain size, excluding the rapid rates of evolution. And then if you account for those rapid rates of evolution, you get this dotted line here. Meaning that an elephant actually has a cancer prevalence akin to a species the size of a tiger. And importantly, this effect has a much greater impact on larger compared to smaller species.

So in summary, bigger species have a higher prevalence of cancer compared to smaller species, refuting Peto's paradox. But interestingly, we find that faster rates of body size evolution are associated with a reduction in cancer prevalence in birds and mammals. So what's really interesting about this is, well, what are the molecular mechanisms that these species have evolved to get so big without incurring this cancer burden? And that's the kind of stuff we're hoping to look at in the future.

Andrea Miyahira: Thank you so much, Dr. Butler, for sharing this. So how does lifespan play into this paradox?

George Butler: Yes, that's a great question. So typically, bigger species often live longer compared to their smaller counterparts. So we started to be able to test how lifespan was intertwined. We found that there was no association between malignancy prevalence and lifespan. And that's a little bit counterintuitive. There are a few caveats there about data access.

But I think it starts to ask an interesting question about the different levels at which evolution is acting. So you can imagine that we know that longer lifespan is associated with cancer within a given species, but across species, the same trend might not be true. But it's an interesting thing to ask in the future when more data becomes available.

Andrea Miyahira: Thank you. And why do you think that you and your team were able to find a different result compared to the previous studies?

George Butler: Yeah, no, so I think there's a couple of reasons there. At a practical level, it's the ability to combine these really large, powerful data sets with modern phylogenetic statistical methods. And marrying those two things together enables you to question fundamental ideas.

But at a broader level, I think a point to highlight here is that we've really relied on having a multidisciplinary team. So we've got medical oncologist, cancer biologist, evolutionary biologist, me as a kind of computational statistician. And I think that's important because it highlights everyone's different strengths and enables you to ask these challenging questions.

Andrea Miyahira: OK, thank you. And where do humans fall into the paradox?

George Butler: So yeah, so humans are difficult because we have access to modern medicine, which makes it a little bit more complicated. But we did take the kind of average size of a human—I think it was about 180 pounds—and then used that to estimate what cancer prevalence you'd expect. And interestingly, it came out for a species about the size of a bat, which is obviously much, much smaller than us. So hopefully, that's good news. But like I say, it would be interesting as we move forward to translate some of these ideas.

Andrea Miyahira: Thanks. And can we use these same methods and approaches to better understand prostate cancer in humans?

George Butler: Yeah, no, definitely. So whilst we focused on cancer prevalence across species, you could imagine looking at the same thing to understand tumor evolution at a cellular level. Some of the work that we're doing at UCL using research autopsies is to understand how cancer spread within an individual patient, how it's metastasized to those different distant sites, and to be able to understand those evolutionary dynamics in the same way that we have here at a species level.

Andrea Miyahira: OK, well, thank you so much for sharing this with us.

George Butler: Thank you.