Cancer is one of the most difficult diseases to treat because cancer cells are remarkably adaptable. They can reprogram metabolism, remodel their surroundings, evade cellular regulation, and withstand therapeutic stress. Although conventional anticancer treatments can be effective, their limited specificity can also damage healthy tissues. The fundamental challenge, therefore, is not simply to eliminate cancer cells, but to distinguish them from healthy cells and intervene with greater precision.

When the challenge is not simply to kill, but to distinguish.

We found a useful analogy in the idea of a healthy society. A stable society depends on cooperation, communication, and collective wellbeing. When disruptive influences emerge, effective intervention requires identifying the source of the problem and responding proportionately rather than indiscriminately. Similarly, the human body is an extraordinarily organized biological society. Its cells communicate, cooperate, specialize, and maintain tissue homeostasis. Cancer disrupts this balance when malignant cells escape regulatory control, proliferate uncontrollably, remodel their microenvironment, and resist mechanisms that would normally eliminate abnormal cells.

Question of “What kills cancer?” to “What can sense cancer?”

We asked whether a molecule could sense its intracellular environment, respond to a biological signal, and dynamically construct its therapeutic architecture where needed. Inspired by nature’s ability to transform molecular information into functional structures – such as proteins fold into functional structures, lipids organize into membranes, and nucleic acids assemble into sophisticated architectures, we sought to translate this principle into a synthetic supramolecular system. Our answer was G-BFmocY, a tyrosine-based molecular platform containing a phenylboronic-acid functionality responsive to hydrogen peroxide (H₂O₂), a biologically relevant reactive oxygen species. Rather than designing G-BFmocY simply as a cytotoxic agent, we developed its molecular components to integrate several functions, including cellular recognition, chemical responsiveness, self-assembly, and biological activity. This design allowed the molecule to respond to its chemical environment and potentially generate therapeutic structures in-situ.

A molecule that evolves with its environment

In aqueous conditions, G-BFmocY formed a well-defined single nanofibril network with fibres ~15 nm wide. Upon exposure to H₂O₂, the ROS-responsive boronic-acid functionality underwent 1,6-elimination and hydrolysis. The resulting molecular components reorganized through non-covalent interactions, transforming the original network into a distinct double-network architecture containing fibres >20 nm and larger nanofibres of ~120 nm. Rather than simply undergoing chemical degradation, G-BFmocY therefore used the chemical signal as an instruction for structural evolution. This finding revealed that a relatively simple molecular building block could  transform into double-network supramolecular architecture in response to its chemical environment. It also raised an intriguing question: could the intracellular environment itself provide the chemical instructions required to drive such molecular transformation and therapeutic function?

From the vial to the living cell

Moving from controlled aqueous conditions to living cells was one of the most challenging stages of the project. Unlike a vial, a cell presents a complex and dynamic environment containing proteins, membranes, metabolites, ions, enzymes, and organelles. Thus, a molecule that behaves predictably in solution may respond differently within a biological system. This led us to a key question: where inside the cell would our molecular transformation occur?

We focused on mitochondria, the cell’s “powerhouse” organelle and important regulators of intracellular ROS. Because cancer cells often display altered redox balance and elevated oxidative stress, mitochondria offered both a chemical trigger and a potential therapeutic target. Our strategy was therefore to exploit the cancer cell’s oxidative environment to activate the molecular system and direct its activity toward mitochondrial machinery. Rather than causing indiscriminate cellular damage, this approach would exploit a biochemical vulnerability associated with cancer cells and selectively disrupt the machinery that supports their survival.

Watching chemistry become biology

Experiments with normal HaCaT cells and cancerous HeLa cells provided encouraging evidence that our molecular design could produce a meaningful cellular response. Confocal microscopy and MitoTracker analysis confirmed mitochondrial localization of G-BFmocY in cancer cells. Within this environment, the ROS-responsive system underwent its programmed transformation, generating a double-network supramolecular architecture. This molecular reorganization was associated with mitochondrial disruption and cancer-cell death, while normal cells retained substantially greater structural integrity under the conditions investigated.

This was a particularly rewarding stage of the project, as we moved from a molecular concept developed at the chemistry bench to observing its behavior within living cells. Importantly, the cellular environment actively participated in constructing the therapeutic structure rather than merely serving as its destination. Unlike conventional drug delivery, our approach demonstrates the potential for therapeutic structure and function to emerge in situ in response to an endogenous biological signal.

A molecular “surgical strike”

In our work, “surgical strike” is a metaphor for chemical precision: selectively activating therapeutic function by exploiting biochemical differences between cancerous and healthy cells. The cancer-associated redox environment provides a potential cue for triggering supramolecular transformation and mitochondrial activity, with the aim of minimizing unnecessary effects on healthy cells. An equally important lesson was that molecular simplicity can generate functional complexity. Our findings suggest that relatively small, single-amino-acid-derived building blocks can undergo sophisticated structural evolution through non-covalent interactions. This approach opens opportunities to design simpler yet intelligent supramolecular systems whose structures and functions dynamically adapt to their biological environment.

What comes next?

These findings open exciting questions for the next stage of our research. Can the system recognize other cancer-associated signals, improve selectivity across cancer types, or dynamically evolve within living cells? A key goal is to move beyond cellular models toward increasingly complex biological systems and, ultimately, a "bench-to-bedside" strategy. We welcome collaborators from complementary disciplines to help address these challenges and accelerate translation. More broadly, this work represents a shift from conventional drug delivery toward molecular evolution: from passive materials to responsive systems and from conventional targeting to exploiting cellular chemical identity. Our long-term vision is to develop intelligent therapeutic materials that sense their surroundings, respond to appropriate biological signals, and assemble their functions precisely where needed. G-BFmocY represents an important step toward realizing this vision and advancing responsive supramolecular therapeutics.

A greater scientific outlook and related results can be found in our recent publication in Communication Materials (https://doi.org/10.1038/s43246-026-01408-2)