cryonics-posts post 1 draft-post

Why are humans (not) solid?

Epistemic status: I spent a day on Portland on 2026-04-23 talking to the Nectome team (Aurelia, Charlie, Jasmine) about brain preservation, and this post is me chasing down one throwaway remark Aurelia made that wouldn't leave my head. The comparative-biochemistry part is well-sourced. The last section — "maybe cell death is a partial version of the same chemistry" — is speculation I am taking seriously but not asserting. Footnotes are in the companion research notes.

The question

Aurelia said something like this, and I'm paraphrasing:

An egg is easy to turn solid. You heat it and it sets. That's because it's basically a sack of protein, and heat denatures and cross-links proteins. A human body is much harder. But aldehyde fixation is this little thing — one class of chemical reaction — that causes the whole system to turn solid at ultrastructural resolution. And I wonder if a lot of cell death is already mediated by related chemistry, just uncontrolled.

I'd like to take each piece of that seriously.

Part 1: eggs and cooking

Start with the literal claim. Is an egg mostly protein?

A whole raw egg is about 76% water, 13% protein, 10% lipid by mass [USDA/PMC composition data]. That doesn't sound "basically protein" — it's mostly water. But the whole-egg number hides the structure. Egg white — the part that goes opaque-white-solid when you fry an egg — is 88% water, 11% protein, 0.2% fat. The yolk is a lipid-dense emulsion and behaves differently. When people say "an egg sets," they mean mostly the white.

What makes egg white set is that ovalbumin (~54% of the white's protein) and a handful of other egg-white proteins denature at 62–65 °C, exposing hydrophobic residues and free cysteines. The free cysteines then find each other and form intermolecular disulfide bonds via sulfhydryl-disulfide exchange (primary mechanism papers), and those S-S bonds, along with hydrophobic patches sticking together, aggregate the unfolded proteins into a space-filling three-dimensional gel network. The network traps the water. The whole thing becomes a solid.

Two things make this easy:

  1. Homogeneity. Egg white is basically one chemical system: a soluble protein pool in water. The setting reaction is self-similar everywhere in the fluid.
  2. Endogenous cross-linkers. The proteins bring their own cross-link chemistry with them. You don't need to add anything — you just need to give them energy. Heat breaks non-covalent folds; the cysteines take it from there.

So the cartoon "egg = sack of protein → heat sets it" is simplified but not wrong. The load-bearing detail is that egg white is a dilute protein solution where the proteins auto-polymerize through disulfide chemistry when unfolded.

Part 2: why humans don't work the same way

Now compare:

An adult human is ~58% water (men) / ~48% (women), ~20% protein, ~15-25% lipid [body-composition review]. On a percent-protein basis, a lean human is actually more proteinaceous than a whole egg. So "too little protein" isn't the issue.

The issues are heterogeneity and architecture.

Heterogeneity. The human body is not one chemical system. It's simultaneously: - Intracellular proteins (cytoskeleton, enzymes, organelles) - Extracellular matrix (collagen, elastin, glycosaminoglycans — collagen alone is ~25–30% of all protein mass in the body) - Lipid membranes (phospholipid bilayers, cholesterol, sphingolipids) - Myelin (up to ~80% lipid by dry weight in the central nervous system — O'Brien & Sampson 1965) - Aqueous extracellular fluid (~20% of brain volume is extracellular space) - Mineralized bone matrix (up to 70% mineral plus type I collagen) - Whole-blood fluids with cells and electrolytes

Each component has its own characteristic chemistry of going solid. Heat denatures intracellular enzymes at ~50–70 °C. Collagen melts to gelatin around 65 °C (which is not setting — it's the opposite). Phospholipid bilayers are fluid and don't directly cross-link on heat. Cholesterol crystallizes at different conditions than phospholipids.

Architecture. Egg white is an emulsion with a simple goal: become a gel. Human tissue has geometry you want to preserve, not scramble: the precise topology of synapses, the layered structure of myelin, the positions of individual receptors in post-synaptic densities. Heat-based thermal aggregation is blind — it doesn't care what ends up bonded to what. Even if it produced a solid, the solid would be wrong in ways that matter.

If you boil a piece of brain, you get mush. If you boil an egg white you get a coherent gel. The difference is whether the chemistry operates over a uniform substrate or over a mosaic whose structure is the thing you care about.

Part 3: aldehyde fixation — the "one reaction" that solves most of the problem

Glutaraldehyde is OHC–(CH₂)₃–CHO. Five carbons, two aldehyde ends, relatively small, very reactive.

The core reaction is simple. An aldehyde carbon is electrophilic. A protein side chain with a free amine — typically the ε-amino group of lysine, which is by far the most reactive nucleophile accessible on the surface of a folded protein — attacks the aldehyde, forming a carbinolamine, which dehydrates to a Schiff base (C=N), which can be attacked by a second amine to give a methylene bridge of the form Protein–NH–CH₂–NH–Protein [JBC review on formaldehyde crosslinking]. Lysine is the dominant reactive group; cysteine thiols, arginine guanidines, and histidine imidazoles also react. Glutaraldehyde's two aldehyde ends let it bridge two proteins at a distance, which formaldehyde (monomer, one aldehyde) does only rarely.

A few striking facts:

So what does aldehyde fixation actually do, mechanically? It turns the protein fraction of a tissue into a covalently cross-linked gel, at the scale of the native protein lattice. The lipids are still there, still bilayer, but their protein contacts are pinned. The water is still there. Small metabolites are still there. Everything is stuck in place relative to the protein scaffold, which is now so densely cross-linked that nothing can drift or rearrange.

This is enormously important for cryopreservation because of a boring logistical reason: vitrification takes hours. You have to gradually replace the water in the tissue with a high concentration of cryoprotectant (McIntyre & Fahy's ASC uses 65% w/v ethylene glycol, whose glass transition is ~-131 °C). Without fixation, cells swell and shrink as osmotic gradients shift, mitochondria release stuff, membranes blebb. With fixation, the cells can't move. You can take as long as you want. You're perfusing cryoprotectant through a chemically frozen tissue that just happens to still be a room-temperature gel.

This is Nectome / 21st Century Medicine's ASC protocol in a nutshell: perfuse glutaraldehyde, ramp ethylene glycol over several hours, cool to −135 °C. It won the Brain Preservation Foundation's Small Mammal Prize in 2016 and Large Mammal Prize (pig) in 2018. The 2018 evaluators' call was that processes were traceable and synapses were crisp on FIB-SEM throughout the brain.

So Aurelia's "a little thing that causes it to all turn solid" is a compression, but it's a fair one. The protein fraction turns solid via aldehyde cross-linking; the water and lipid fractions turn solid via the subsequent vitrification. The protein-gel step is what makes the vitrification step possible without scrambling the ultrastructure.

It's worth saying what aldehyde fixation doesn't do. It shrinks the brain by about 30% in volume and collapses extracellular space from ~20% to ~5% (unless you tune osmolarity carefully, as Mikula, Denk, and Pallotto's groups have demonstrated). Lipid bilayers remain fluid within the gel. Small metabolites can still diffuse until vitrification locks them in. Organelle geometries are shifted, sometimes meaningfully. This is a real freeze-frame at one level (the connectome) and a shifted-and-smeared snapshot at another (exact geometry, active-zone nanostructure, precise membrane topology).

Part 4: is there something to the "cell death is a partial ASC" idea?

This is the speculative part. I want to split Aurelia's third claim into a strong version and a weak version, because only one of them survives scrutiny.

Strong version: "Most cell death is aldehyde-mediated cross-linking chemistry."

This isn't right. Most cell death is driven by other mechanisms. Apoptosis is a caspase cascade — proteolytic, not cross-linking. Necrosis is ionic dysregulation and membrane rupture. Pyroptosis and necroptosis are their own programmed pathways. Ischemic death is primarily energy failure and Ca²⁺ overload triggering mitochondrial permeability transition. Aldehyde damage is downstream of these processes in many cases but is rarely the first cause.

Weak version: "Endogenous aldehyde chemistry is a significant, convergent damage modality in aging and in some specific cell-death pathways, and it uses the same reaction types as glutaraldehyde fixation."

This is solidly supported by the literature.

The short list of endogenous reactive aldehydes:

There is an entire theory of aging — the cross-linking theory of aging, proposed in 1942, and the Maillard theory of aging from the 1980s — that posits accumulated non-enzymatic protein cross-links as a primary driver of age-related tissue stiffening, loss of function, and disease. The evidence on causality vs correlation is mixed, but AGEs are unambiguously associated with diabetes, cardiovascular disease, kidney disease, osteoporosis, Alzheimer's, and skin aging.

And there is a specific mode of cell death — ferroptosis — whose proximate lethality is pore formation in membranes by lipid-derived aldehydes, not lipid peroxides themselves. Cells have evolved ALDH3A1 and ALDH7A1 to detoxify reactive aldehydes before they form these adducts, and loss of ALDH activity makes cells more vulnerable to ferroptotic death.

So here's a more defensible version of Aurelia's framing:

Cells have evolved a dedicated enzyme family whose job is to prevent exactly the kind of chemistry glutaraldehyde does — reactive aldehydes covalently attaching to lysine, cysteine, and histidine residues on random proteins. When this defense fails, tissues stiffen, proteins misfold, membranes perforate, and cells die by ferroptosis or accumulate damage that looks like accelerated aging. ASC is doing, on purpose and uniformly, a strict-superset version of what happens naturally when this defense fails.

That's not mechanistic identity. It's chemical homology. The uncontrolled version uses mono-functional aldehydes at micromolar concentrations appearing sporadically over years; the controlled version uses a dialdehyde at tens-of-millimolar, perfused uniformly, over minutes. You can fairly say ASC is "doing the chemistry of aging on purpose at lethal speed to capture the structure before anything else can happen." That frames it well rhetorically, and it's chemically grounded — but the inference that "cells die by aldehydes ⇒ aldehyde fixation is a controlled death" has a quantitative gap that's hard to close from current literature.

Part 5: so what

I care about this question for two reasons.

One is that it's satisfying to have the actual chemistry. "Glutaraldehyde cross-links proteins" is a phrase that gets used as a black box even by people who work in this field. The reaction is real — Schiff base formation, methylene bridges, lysine ε-amines — and the fact that it works on tissue as well as it does has a structural reason: protein density and lysine density in cytosol are high enough that a mesh of cross-links forms throughout the tissue on perfusion, and the mesh is mechanically rigid enough to immobilize everything it contains.

Two is that the framing matters for public communication of brain preservation. Aldehyde fixation sounds scary — you're chemically embalming someone's brain — and the usual defense ("it preserves ultrastructure") is an argument from connectomics expertise that laypeople don't have purchase on. But the "it's doing the same chemistry your body is doing slowly and badly" frame is a useful intuition pump. You're already accumulating lysine-lysine cross-links from methylglyoxal. You're already accumulating HNE-protein adducts. That's part of aging. Glutaraldehyde does more of it, faster, more uniformly, and for a specific purpose: freezing the information content of the brain before the natural version of the same chemistry (plus everything else) degrades it further.

The honest caveat: this is a helpful metaphor with real chemical overlap, not a proof that ASC preserves "what matters." That's a separate empirical question about what information needs to be in a brain for a person to be recoverable from it, and the answer to that one is still open.

Sources in research/ folder; full URL list in sources/source-urls.md. Confidence tiers and direct quotes are in the research notes. Major unknowns: quantitative comparison of cross-link density in ASC-fixed vs aged brain (couldn't find a direct paper); whether glutaraldehyde perfusion preempts vs propagates endogenous cell-death pathways during the fixation process (also couldn't find). Both are interesting follow-ups.

ai gen