Why has nobody died from a hallmark of aging?

I recently re-read a paper from Michael Levin’s lab arguing that aging results from the loss of goal directedness in morphogenesis. The core thesis states that organisms are evolved to develop successfully but lack goals for maintaining anatomical stability after the developmental period. In short, evolution gave us a blueprint to build our bodies, but no manual for maintenance afterwards. Importantly, they argue that damage merely accelerates aging, but does not fundamentally cause the decline.

It’s still not clear to me that loss of anatomical form is the key driver of aging. Their argument relies on cellular automata models, with no comparative analysis of long-lived species, no aging clocks, no fancy multi-omic analysis. In any case, new theories of aging come out every week. Why should this one matter?

Yet sometimes the most useful theories come not from exhaustive data, but from asking better questions. Whether or not Levin’s mechanism is correct, his ideas led me to a simple question: Why has nobody died from a hallmark of aging?

If you have spent any time researching aging you have probably come across a list of aging hallmarks: telomere attrition, mitochondrial dysfunction, genomic instability, etc.. Yet there is a paradox worth pondering. Nobody dies from telomere attrition. They die when their heart stops coordinating contraction. Nobody dies from mitochondrial dysfunction. They die when their lungs can’t exchange gas. The hallmarks describe what’s broken inside cells, but death happens when tissues can’t coordinate.

To understand this, consider what a cell actually knows about the world around it.

A macrophage in inflamed tissue has its own transcriptional profile, metabolic state, and activation status. But it has no direct knowledge of the infection happening around it. Everything the macrophage knows comes through its membrane: cytokine receptors, chemokines, ECM adhesions. This boundary is all it has. The macrophage infers “infection” from these signals alone, since it has no clue what reality looks like even a few micrometres away from itself. The cell must constantly ask: given what I’m sensing at my boundary, what environment am I in, how should I respond?

This matters because organisms are hierarchies of such boundaries. Cells interface with extracellular matrices. Tissues interface through paracrine signalling and mechanical coupling. Organs interface via systemic circulation and neural coordination.

The key insight is that to keep the organism running, we do not need every cell’s interior to be perfect. We need the inference process at each boundary to remain coherent.

Let’s consider a few thought experiments:

1. Replace a cell with a black box containing a silicon chip. If that black box presents identical surface markers, responds identically to stimuli, and exchanges metabolites at the same rate as a biological cell, the organism cannot distinguish it from a native cell. The internal implementation is irrelevant. DNA or silicon, doesn’t matter.

2. Next, think of an entire tissue where every cell has internal damage, yet collectively they maintain normal boundary behavior. Correct surface markers, proper response dynamics, normal metabolic exchange. From the organism’s perspective, nothing is wrong. The damage is real but functionally invisible because the coordination machinery remains intact.

Current aging paradigms conflate two distinct problems: molecular damage inside the cells and degraded inference between them. These are related but not identical. Hallmarks matter only insofar as they corrupt boundary behavior.

My goal is not to dismiss the hallmarks or deny connection between intracellular damage and organismal decline. Damage can indeed accumulate, break the boundaries, and lead to system degradation. But the framework’s value lies in identifying what actually fails when organisms age, which is not the presence of damage itself, but the breakdown of coordination (which damage can cause, though damage may not be the only cause).

If coordination at boundaries is what actually matters, here are predictions we can make:

1. Internal damage should be neither necessary nor sufficient for aging

   1. In humans, mutated POLE/POLD1 could lead to ~7x higher somatic mutation rates yet no overt premature aging. The damage is present, but the boundary behavior compensates.

   2. Naked mole rats have higher oxidative damage compared to mice, yet have longer lifespan. The boundary condition (immune system, ECM, etc..) remains intact despite internal damage.

2. Environmental context should override internal damage

   1. If we place old cells in a young signalling environment, they should behave young, and vice versa. Sun et al. show that both young and old MSCs cultured on young ECM show significantly higher telomerase activity compared with those on old ECM or plastic. They also found correlations with improved proliferative capacity and osteogenic potential. Heterochronic parabiosis experiments are also a great testament to the importance of environmental context.

3. Small changes to boundary signals should have outsized systemic effects

   1. Injecting only 0.5 - 1 * 10^6 senescent cells into a mouse which consists of billions of young cells is enough to move systemic physiology into an old/frail state. An extremely small population can corrupt the signalling environment at the boundary for everyone else.

   2. Lei et al. used FOXO3 engineered, senescence resistant human mesenchymal progenitor cells to modulate everything from senescence, proteostasis, inflammation to neurodegeneration. All this is achieved by altering what signals the surrounding tissue receives, not its inner composition.

4. Boundary breakdown should be detectable before phenotypic aging

   1. Here, admittedly, the evidence is sparse. There is simply a lack of time course experiments that would be able to measure this boundary. For instance, Levin’s group has already shown that bioelectric states predict and control anatomical outcomes, showing that boundary level information precedes structural changes. However, whether this directly translates to aging trajectories remains to be tested.

How exactly does boundary inference break down during aging? The honest answer is we don’t know. We haven’t been asking this question. What I can say is that the question itself is different from asking what accumulates inside cells. Instead of finding and fixing every type of molecular damage, what if we could restore coordination just by repairing communication environments? How does signal fidelity degrade across tissue boundaries? Do cells lose the ability to interpret signals correctly, or does the environment become too noisy? When coordination fails at one hierarchical level, does it cascade to others? These are all measurable phenomena that we can explore.

While this work is still in it’s infancy, I believe this is one of the most exciting avenues where aging research could be moving. At the end of the day, we are much better at solving computational inference problems than fixing damage on the nanoscale.

You don’t need to accept Levin’s framework entirely or discard hallmarks completely to find value here. You just need to sit with one question: Why has nobody died from a hallmark of aging?