This episode unpacks a multi-cohort study showing that tumors evolving toward neutral genome-level selection (dN/dS ≈ 1) during or after therapy are associated with treatment resistance and poorer outcomes.
0:00Welcome to Base by Base, the papercast that brings genomics to you wherever you are. Thanks for listening, and don't forget to follow and rate us in your podcast app. So you are looking at a scan, and, you know, the tumor is shrinking.
0:12Everyone in the room breathes this huge sigh of relief, the chemotherapy, or the targeted radiation seems to be working. But beneath the surface, at a microscopic cellular level, there is a massive, agonizing, real world clinical problem.
0:27Doctors often don't actually know if that tumor is truly dying or if it is just, you know, silently adapting to the drugs until it's completely immune. Right. It really becomes this incredibly high stakes waiting game.
0:39I mean, historically, the medical field has relied almost entirely on those physical imaging scans. Or, uh, we look for very specific previously identified genetic mutations to see if the tumor is building resistance.
0:50But the thing is tumors are incredibly dynamic. They operate as these constantly evolving ecosystems. Yeah, and trying to predict the behavior of an entire ecosystem by tracking a single gene is, well, it's like trying to understand the strategy of an entire army by looking at one soldier through a tiny keyhole.
1:05That is exactly it. It's just too narrow of a view. So what if the secret to predicting treatment failure isn't found in looking for like one specific super mutation or a notorious resistance gene? What if instead we could read the evolutionary stress levels of the entire tumor genome all at once?
1:23That shift in perspective changes everything. We move away from cataloguing individual genetic parts and start evaluating the evolutionary state of the whole biological system. Today we celebrate the work of the teams at the NIH, Moffat Cancer Center and Stony Book University, who have advanced our understanding of how tumors evolved to resist therapy.
1:44What really happens when a tumor outsmarts chemotherapy. And how could reading the whole genome's evolutionary state fundamentally change how we track treatment success for you or your loved ones? We are taking a deep dive into the evolutionary biology of cancer.
1:57It's going to be a really fascinating deep dive, and to grasp the magnitude of what these teams discovered, We have to establish what tumor evolution actually looks like first. Right, set the stage. Exactly.
2:08Tumors are entirely shaped by selective pressures, much like species in the wild. In a natural, untreated state, a developing tumor faces all these physiological barriers. Like just trying to survive in the body.
2:20Yeah, it has to survive the patient's immune system. For one, it has to secure its own blood supply to get nutrients. And if it's going to spread, well, it has to figure out how to detach, survive in the bloodstream, and then colonize completely new distant tissues.
2:34It's essentially running a biological obstacle course just to stay alive. And that is just its natural environment. Then we introduce clinical barriers, which means therapy. Chemotherapy, radiation, targeted small molecule drugs.
2:48The big guns. Right. These are massive, sudden, unnatural selective pressures. And the complication for researchers and oncologists is that distinguishing between the genetic changes caused by natural tumor progression, and the changes induced by the therapy itself is, well, it's incredibly difficult.
3:05Because it all just looks like genetic chaos. Right. pretty much. And the main driver of that difficulty is something called intra-tumor heterogeneity. You know, I always used to picture a tumor as just a big lump of identical malfunctioning cells, a Xerox machine that just got stuck.
3:19But that's not the case at all. Far from it. A single tumor is made up of a highly genetically diverse population of subclones. It is a varied community. Wow. So when you apply chemotherapy, you are applying intense pressure to this diverse community.
3:35Some cells die instantly, some are severely damaged, and some just happen to have a genetic work that lets them survive. Trying to track how the overall community adapts to that pressure is incredibly noisy.
3:48Let me try to visualize this. Think of a tumor like, um, like a roofless startup company. Early on, it needs specific driver innovations to survive and get off the ground. It needs to break the rules of normal cellular growth.
4:01Right. those core disruptive ideas. Exactly. But as it rapidly expands and copies its DNA 1000000s of times over, it accumulates passenger inefficiencies. These are mutations that don't actually help it grow, and, you know, might even drain its resources.
4:15They're just along for the ride. Right. It's like a startup hiring way too many useless middle managers just because it's expanding so fast. We need a way to see if our medical treatments are actually bankrupting this startup, or just forcing it to fire the middle managers, streamline its operations and become leaner and meaner.
4:33I love that. We can push that startup analogy even further. What this research team did was figure out how to look at the tumor's fundamental financial ledgers. It's DNA. Exactly. It's DNA. They looked at it to see exactly where and how it was cutting costs or investing in new survival strategies.
4:49To see how this biological startup handles the stress of therapy versus natural growth. Researchers needed a specific metric. They needed a way to measure the evolutionary pressure across the entire genome simultaneously.
5:02So if tracking a single mutation is like looking through a keyhole, how on earth do you zoom out? Like, how do you actually measure the stress level of an entire genome across 1000000s of different cells?
5:13Well, the foundation of this lies in whole XM sequencing or WES. The XM is the protein coding region of the genome. Just the parts that make proteins. Right. It's a small percentage of our total DNA, but it's the part that actually builds the machinery of the cell.
5:26The researchers gather this sequencing data. from highly diverse cohorts of cancer patients. But to track evolution, you know, a single snapshot is useless. Because you can't see the motion. Exactly. So they specifically filtered for quantifiable patients, meaning patients who had at least two distinct samples taken over time.
5:46Oh, so you need a before and after picture to see if the tumor strategy is actually changing. Precisely. So an early primary tumor compared to a late primary tumor, or a primary tumor compared to a metastasis in the lung or liver.
5:58And most importantly for the therapy question. A sample taken before chemotherapy compared to a sample taken after the tumor relapses. Yes. Now, to measure the evolutionary state of the tumor genome between those 2 snapshots, they utilize 2 universal parameters.
6:13The 1st is a relatively simple concept, which they call in. Okay, and what does that stand for? It stands for the total mutational burden, basically, and is the sheer number of non-synonymous mutations.
6:24Non-synonymous, meaning they actually do something. Right. Those are mutations that alter the amino acid sequence and actually change the resulting protein. For the purposes of this analysis and acts as a proxy for the tumor's biological age.
6:37Because, I assume, the longer the tumor has been growing and dividing, the more protein changing mutations it naturally accumulates over time. Exactly. It's just a raw account. But the 2nd parameter is where the real analytical power lies.
6:51It is called a DNDS ratio. D-N-D-S. Okay break that down for me. This measures the strength of evolutionary selection at the protein level. It compares the rate of non-synonymous mutations, the ones we just talked about that change proteins, to synonymous mutations.
7:06Synonymous, meaning the DNA sequence changes, but the final protein stays exactly the same. Yes, because the genetic code has redundancy built into it. Right. You can change a letter in the DNA spelling, but the cell still produces the exact same amino acid.
7:20That redundancy is key, because these synonymous mutations don't change the protein, they don't help the tumor, but they also don't hurt it. They are completely invisible to natural selection. Wait, hold on.
7:31If the silent mutations don't do anything to the tumor's fitness, why even count them? Aren't they just genetic background noise? That background noise is actually the control group built into the tumor itself.
7:42Oh, wow. Really? Yeah, because they are invisible to selection, synonymous mutations accumulate at a steady random rate, purely due to the mechanics of DNA replication errors. They act as a neutral ticking clock.
7:55Oh, I see. So by comparing the rate of protein changing mutations to this steady background hum of silent mutations, we can see if the tumor is actively shaping its genome. Okay, let me make sure I'm following the math on this DNDS ratio.
8:09If the ratio is greater than one, that means the tumor has way more protein changing mutations than we would expect from just random chance. Exactly. Because the tumor is actively keeping those specific changes.
8:19They provide a survival advantage. This is called positive selection. The tumor is holding onto those driver innovations. And if the ratio is less than one, then the silent random noise outnumbers the functional changes.
8:31So the tumor must be actively throwing things out. Right, it is under negative or purifying selection. The mutations that are occurring are harmful to the tumor's survival in its current environment. So any cell that gets those mutations just dies off.
8:46So going back to the analogy, the tumor is firing those inefficient middle managers to stay viable. Yes. And then there was the 3rd scenario where the ratio is exactly one. Okay, what happens at one? That is neutral evolution.
8:58The rate of functional changes perfectly matches the rate of random background noise. The tumor is just coasting. Coasting. Like it's not trying to adapt anymore. Exactly. It is found a biological sweet spot where it isn't desperately acquiring new genetic advantages, nor is it frantically purging disadvantages.
9:17The brilliant step in this deep dive is applying that ratio at the entire genome level for individual patients. We aren't just checking if one gene is under positive selection. We are taking the evolutionary temperature of the entire cancer.
9:29Armed with this universal stress test metric. The researchers established a baseline by looking at untreated cancers. They analyze cohorts of untreated colorectal, esophageal and lung cancers. They wanted to see the natural progression of the disease without any drugs interfering.
9:45So what does a tumorous financial ledger look like when it's left completely alone? The findings were incredibly striking. Despite massive genetic variations between different patients. The DNDS value within an individual patient remains highly stable and strictly linear during natural primary progression.
10:03Meaning, if a patient's tumor starts out under heavy positive selection like, it's grabbing lots of new mutations to survive. It stays on that exact same trajectory as it grows. The pressure doesn't randomly spike or drop while it's sitting in the primary organ.
10:17Correct. The evolutionary trajectory remains remarkably stable. Now, when the tumor metastasizes and moves to a new organ, it faces a new environment. So you do see some cancer specific signatures of relaxed or positive selection, as it adapts to, say, the liver or the bone.
10:32That makes sense. But largely the baseline rule is stability. Okay, so if the baseline is this linear, predictable track, what knocks it off course? What happens when we look at the data from tumors fighting for their lives against chemotherapy?
10:46This is where the paradigm of cancer treatment might fundamentally shift. The researchers looked at diverse cohorts of treated cancers. Pediatric acute lymphoblastic leukemia, chronic lymphocytic leukemia, ER positive breast cancer, bladder cancer, and glioblastoma.
11:03That's a huge variety of cancer. It is. These patients underwent heavy therapy and unfortunately eventually developed resistance, and the data revealed a universal pattern. What was? When tumors resist treatment.
11:14Their DDS distribution shifts heavily toward neutrality. The ratio moves toward one. Regardless of whether it is a liquid blood cancer or a solid brain tumor, When it figures out how to beat the chemotherapy, its entire evolutionary state shifts into this coasting neutral zone.
11:29Yes. And to prove this was an anomaly, the team utilized a massive original data set of 624 multiple myeloma patients, which gave them 780 distinct sequencing samples. Multiple myeloma. That's a blood cancer, right?
11:43Yes, it's a cancer of the plasma cells and the bone marrow. And it is uniquely suited for the study because it has a well-documented pre-malignant face. You can track patients for years before they even need treatment.
11:56Oh, that longitudinal timeline is everything. You get to see the tumor's behavior when it's just a benign quirk in the bone marrow all the way through to aggressive relapse disease. Exactly. The multiple myeloma data set proved the exact same phenomenon on an unprecedented scale.
12:11Pre-treatment during the benign and early active phases, the evolutionary path was linear and stable. But post treatment, resistance fundamentally broke that correlation. The intense pressure of the therapy forced a shift to neutrality.
12:26Yes, exactly. Let me try to synthesize this. The tumor goes through a fascinating biphasic evolution. In the early days, it is grabbing useful driver mutations to get established. That's positive selection as it rapidly expands.
12:41It starts acquiring all those useless passenger mutations, which drain its energy so it has to purge them. That's negative selection. But when it hits that neutral zone, when the DNDS ratio settles right around one, it is like a fighter jet, hitting its optimal cruising altitude.
12:55The aerodynamic thrust of the drivers is maximized, the drag of the passengers is minimized, and the tumor is operating at its absolute peak fitness. I think the cruising altitude analogy is perfect. And in that neutral zone, it has achieved an evolutionary balance where its current genetic configuration is perfectly optimized to ignore the chemotherapy.
13:17It's untouchable. Pretty much. It doesn't need to mutate further to survive the drug because the genome it currently has is already resistant. If that neutral zone represents the tumor's peak fitness, the implications for a patient currently undergoing treatment are massive, how does an oncologist actually use this information?
13:34The clinical reality found in this study is grim. but absolutely vital for the future of oncology. When they perform survival analysis on the glioblastoma and multiple myeloma data, they found that a posttreatment shift toward neutral evolution is significantly associated with rapid relapse in a much worse prognosis.
13:52Because hitting neutrality means the tumor has successfully adopted, it's just brushing off the chemotherapy like it's nothing. Yes. On the flip side, patients whose tumors escaped the neutral regime, meaning their DNDS ratio, stayed far away from one after receiving therapy, had significantly better survival times.
14:10Because their tumors were still under severe evolutionary stress, the drugs were still actively hurting the cancer. Exactly. This elevates the DNDS ratio from a cool biological observation to a universal evolution guided marker for clinical decision making.
14:26Let's play out a scenario here to really grasp the impact. Imagine a patient comes in with glioblastoma. Month one, they get surgery and start a heavy chemo regimen. Month 3, they get an MRI and the scan is completely clear.
14:39The tumor looks stable or even invisible. But if the oncology team sequences the microscopic tumor DNA is circulating in the blood, and they calculate that the DNDS ratio is creeping from, say, 0.5 up to 0.9.
14:51That tells the oncologist that the treatment is actively failing, even though the physical MRI scam looks perfect. That is wild. It is. The tumor genome is steadily approaching neutrality. It is adapting to peak fitness right under our noses.
15:05long before the physical symptoms of relapse manifest. So, you could catch the resistance months before the tumor physically grows back. An oncologist could look at that neutral ratio and say, okay, the tumor has adapted to drug A, we need to switch to drug B immediately to knock it back off balance.
15:21Yes, you are getting ahead of the physical disease by tracking the evolutionary mechanics. Conversely, if a tumor starts near neutrality, but shifts far away from a ratio of one after you administer a new drug, you know instantly that the therapy is working.
15:36You have disrupted its optimal cruising altitude. Now, in any complex biological system, we always have to look for the blind spots, what are the limitations of relying on this math? If I'm designing a clinical trial around this tomorrow?
15:48What pitfalls do I need to watch out for? That's a great question. Researchers and bioinformaticians must be careful of specific mutational biases. For example, some cancers are classified as hypermutators.
16:00Hypermutator sounds aggressive. It is. Due to defects in their DNA repair mechanisms, they accumulate mutations at a staggering, chaotic rate. In these specific environments, you might see a rapid saturation of silent synonymous mutations.
16:15Ah, so the background noise gets incredibly loud. Exactly. That saturation could mathematically skew the denominator of our DNDS ratio, making it look like the tumor is approaching neutrality when it actually isn't.
16:27Oncologists would need to adjust the algorithms to account for these specific hypermutators' signatures to avoid false readings. Okay, there's one detail in the data regarding how the tumor actually survives in this neutral zone that I want to push back on.
16:41Sure, what is it? The study notes that even as the tumor adapts to the therapy, the actual number of genetically distinct subclones inside the tumor stays relatively small and stable. We are talking maybe 2 or 3 major clones.
16:52If the underlying genetics aren't going crazy and creating dozens of new variations, how on earth is the tumor shape shifting to become fitter and resist the drugs? That observation points directly to the next major frontier of cancer biology.
17:05The stability in the number of genetic clones implies that the tumor is not solely relying on altering its underlying DNA sequence to survive the chemotherapy. Then what is it doing? It strongly suggests the tumor is heavily utilizing epigenetic modifications.
17:20Epigenetics. So, the tumor isn't mutating the DNA code itself. It's modifying how those genes are turned on or off. Yes, think of it as hardware versus software. The underlying DNA sequence is the hardware.
17:33The epigenetic chemical tags attached to the DNA are the software. I see. The tuber maintains the exact same hardware those 2 or 3 genetic clones, but when the keyotherapy hits, it quickly runs a different software program.
17:45It changes its gene expression to survive the toxic environment without ever rewriting its core genetic code. It's a non-genetic diversification. The tumor just flips a few biological switches to maintain its peak fitness cruising altitude.
17:58It is a remarkable survival mechanism. And it means that in the future, just tracking the genetic mutations won't be enough. We will need to measure the epigenetic landscape simultaneously to get a truly complete picture of the tumor's evolutionary state.
18:13That is just incredible. To bring all of this complex biology together into a clear take home message for you. Tumor evolution is a biphasic process that universally shifts toward neutral evolution when resisting therapy, marking the point where the cancer reaches its peak fitness state.
18:29Tracking this genome wide selection pressure through the DNDS ratio offers a powerful universal new metric to predict if a cancer treatment is failing at a molecular level long before clinical symptoms even appear.
18:42It allows the medical field to read the tumor's evolutionary playbook in real time. This metric strips away the noise of individual mutations, and gives clinicians a totally new lens through which to view treatment efficacy across almost all cancel types.
18:55Which leaves us with a massive, thought-provoking question for the future of medicine. What does this mean for how we design the timing and sequence of cancer therapies? Could we one day use this evolutionary metric to intentionally sequence our drugs rotating therapies not based on calendar days, but based on the genome stress levels, to intentionally keep tumors off balance and forever locked out of their neutral comfort zone?
19:19This episode was based on an open access article under the CCBY 4.0 license. You can find a direct link to the paper and the license in our episode description. If you enjoyed this, follow or subscribe in your podcast app and leave a 5 star rating.
19:33If you'd like to support our work, use the donation link in the description. Now, stay with us for an original track created especially for this episode and inspired by the article you've just heard about.
19:42Thanks for listening and join us next time as we explore more science base by base.
