For decades, we’ve understood cancer as a disease of broken genes, a series of typos in the DNA code. But a new wave of research is revealing that the story is far more complex and elegant. It’s not just about the code, but the physical shape it takes. Guiding us through this paradigm shift is Ivan Kairatov, a biopharma expert whose work at the intersection of technology and R&D provides a unique lens on this emerging field. He joins us to discuss a groundbreaking new study on how the three-dimensional architecture of the genome can either suppress or unleash malignancies like lymphoma. Our conversation will explore this novel concept of “architectural tumor suppression,” delving into how the collapse of specific DNA loops silences our body’s natural defenses. We will also touch upon the sophisticated AI tools that are making these invisible structures visible and discuss how this new understanding could revolutionize everything from cancer diagnostics to treatment, moving us from merely fixing genetic errors to potentially repairing the very blueprint of the cell.
Your research introduces “architectural tumor suppression.” Could you walk us through the step-by-step process of how losing one copy of SMC3 or CTCF leads to the silencing of tumor suppressor genes? Please share a specific anecdote or metric from your research that first illustrated this phenomenon.
Absolutely. Think of proteins like SMC3 and CTCF as the engineers responsible for maintaining the critical infrastructure within the cell’s nucleus. Their job is to fold the immense strand of DNA into precise loops, connecting gene “switches,” called enhancers, to the genes they control. When a cell loses one of its two copies of the gene for SMC3 or CTCF—a state we call haploinsufficiency—you essentially cut your engineering team in half. The immediate result isn’t a total collapse of the city; the major highways might still be intact. Instead, you see a subtle, insidious decay of the local roads. These are the short-range enhancer-promoter loops. The first time we truly saw this in action was with a critical tumor suppressor gene called Tet2. We had lymphoma cells where Tet2 was completely silent, but when we sequenced it, the gene was perfect—no mutations. It was a real mystery. Then we looked at the Hi-C maps, the architectural data. We saw that the physical loop connecting Tet2 to its power switch had vanished. The lights were off not because the bulb was broken, but because someone had cut the wire.
You used AI to analyze massive datasets from Hi-C maps and RNA sequencing. What specific patterns in enhancer-promoter loops did the AI reveal that were otherwise invisible? Can you explain how these computational tools helped connect that architectural erosion to the B-cell “decision bottleneck” you described?
The sheer scale of this data is staggering; a single Hi-C map contains billions of potential contact points. A human simply cannot see the meaningful patterns in that noise. This is where AI truly transformed our understanding. We fed it the architectural maps from both healthy cells and those with SMC3 loss. The AI began to recognize a distinct pattern of erosion. It wasn’t random; it specifically flagged the decay of short-range loops—the very ones responsible for fine-tuning gene expression. It then performed the crucial next step: it integrated that architectural data with our single-cell RNA sequencing results. This is where the magic happened. The AI drew a direct line from cause to effect, showing us that the B-cells with the most degraded loops were the very same cells getting stuck in a developmental limbo. They couldn’t complete their maturation into plasma cells. This “decision bottleneck,” which creates a pool of vulnerable, pre-malignant cells, wasn’t just correlated with architectural loss; the AI showed it was a direct downstream consequence.
The study links lower SMC3 expression in DLBCL patients to poorer outcomes. Beyond using this as a biomarker, how do you envision a therapy that could “repair architecture”? Could you outline a hypothetical first step or strategy for developing a treatment that restores these critical DNA loops?
This is the most exciting frontier. Using SMC3 levels as a biomarker is a powerful first step, allowing us to stratify patients who might be at higher risk. But the ultimate goal is to intervene. A hypothetical first-generation “architectural therapy” probably wouldn’t involve rebuilding loops from scratch, as that’s an immense bioengineering challenge. A more practical first step would be to develop a drug that acts as a “molecular glue” or a stabilizer for the remaining SMC3 protein complexes. If we can make the 50% of the protein that’s still there work more efficiently or hold onto the DNA longer, we might be able to preserve those critical tumor suppressor loops. Another strategy would be to bypass the broken loop entirely. We could use epigenetic modifier drugs to hyper-activate the silenced gene’s promoter, essentially forcing the “on” switch, even if its primary connection has been severed.
You described this architectural disruption as “losing the blueprint for a building.” Using this analogy, how does this paradigm shift change the day-to-day priorities for cancer researchers? What new experimental questions or screening methods does this open up for labs studying lymphoma and other malignancies?
This analogy truly gets to the heart of the shift. For years, cancer research has been overwhelmingly focused on finding and interpreting the “typos” in the blueprint—the genetic mutations. Our day-to-day work involved sequencing, sequencing, and more sequencing. Now, we understand we also have to check if the blueprint is folded correctly. This completely changes our experimental priorities. In my lab, for instance, we’re no longer just running genetic screens. We’re developing “architectural screens.” Instead of asking, “Does this drug kill cells with a KRAS mutation?”, we’re now asking, “Does this drug restore the Kmt2d enhancer-promoter loop in our lymphoma cell lines?” This opens up a whole new class of potential drug targets that aren’t the genes themselves, but the structural machinery that organizes them. It forces us to think in three dimensions.
The finding that haploinsufficiency erodes specific short-range loops, rather than causing total genomic collapse, is a crucial detail. Was there a specific “aha” moment or a key piece of data that revealed this nuance? Please describe how your team pinpointed this subtle but critical effect.
There absolutely was an “aha” moment, and it came from a result that was initially quite disappointing. Based on the central role of these proteins, our first hypothesis was that losing half of them would cause catastrophic, widespread architectural chaos. We expected the Hi-C maps to look like a tangled mess. When the first results came back, the large-scale structures—the massive compartments that organize whole chromosomes—looked surprisingly normal. We were stumped. The breakthrough came when a graduate student decided to computationally “zoom in” on the contact maps, focusing on high-resolution, short-range interactions. Suddenly, there it was. It wasn’t a bomb that had gone off; it was more like a series of precise, surgical snips. The damage was highly localized to these small, regulatory loops, while the broader scaffolding held firm. That was the moment we realized this wasn’t about total collapse. It was a subtle, targeted erosion of the most critical connections, which is a far more insidious and therapeutically interesting problem.
What is your forecast for this area of research?
I believe that within the next five to ten years, 3D genomic profiling will become a standard component of cancer diagnostics, sitting right alongside DNA sequencing. We will no longer just ask what genes are mutated, but also which crucial regulatory loops are broken. This will give us a much richer, more functional picture of what’s driving a patient’s tumor. On the therapeutic side, I predict that the first “architectural-repair” drugs, perhaps stabilizers for proteins like SMC3 or targeted epigenetic activators, will be entering early-phase clinical trials. We’re moving from a one-dimensional view of cancer to a three-dimensional one, and that added dimension holds the key to a whole new generation of smarter, more effective treatments.
