How Frogs Freeze in Winter: The Strange Biochemistry of Vertebrate Cryopreservation

A wood frog (Rana sylvatica) overwintering in the leaf litter of a Quebec forest is, by most reasonable measurements, dead. Its heart does not beat. Its blood does not flow. Its body is cold enough that ice has nucleated in the spaces between its cells and crystallized through about two-thirds of its body water. If you picked one up, it would be hard, cold, and brittle. There is no metabolic activity that conventional methods can detect. In every meaningful biological sense the animal has stopped functioning.

In the spring, over the course of a few hours, the frog thaws. Ice melts in the body cavity. The heart starts beating, weakly at first, then normally. Circulation resumes. Within a day the frog is hopping around the forest floor looking for food and mates, with no apparent damage from having spent four months as a block of ice. This sequence is one of the more remarkable physiological feats in vertebrate biology, and the biochemistry that makes it possible was not properly worked out until the 1980s.

The freezing problem

The basic problem with freezing a vertebrate is that ice is destructive in three different ways. Ice crystals that form inside cells puncture cell membranes. Ice crystals that form outside cells draw water out of the cells by osmosis as the extracellular fluid concentrates, dehydrating the cells until their proteins denature. And the volume change of water freezing to ice is about 9 percent, which produces mechanical stresses that rupture tissue if it happens in confined spaces like blood vessels.

Most vertebrates die when they freeze because all three damage modes occur uncontrollably. The freeze starts wherever ice happens to nucleate, propagates through the tissue with no regard for what it disrupts, dehydrates cells past the point of recovery, and produces mechanical damage that is not reversible by thawing. A frozen human would have ice crystals through every tissue, denatured proteins, and ruptured membranes; thawing recovers none of it.

The wood frog solves all three problems with a coordinated biochemical strategy that controls where the ice forms, prevents it from forming inside cells, prevents cellular dehydration, and limits the mechanical damage.

The glucose surge

The first and most striking element of the frog's strategy is a massive glucose surge that begins within minutes of ice nucleation on the skin. Blood glucose in a normal frog is around 5 mmol/L, comparable to a human. As soon as ice begins to form on the surface, the frog's liver dumps glucose into the bloodstream and the level rises within hours to 200 to 500 mmol/L — fifty to a hundred times the normal concentration, and well above what would kill a human within minutes through osmotic damage.

The glucose serves as an intracellular cryoprotectant. It diffuses into cells, where it raises the osmolarity to match the increasingly concentrated extracellular fluid as ice forms outside the cells. Without the glucose, water would leave the cells by osmosis until they were dehydrated past recovery; with the glucose, cells maintain a viable hydration level even as the extracellular space freezes around them. The mechanism is similar to how trehalose protects fungi and tardigrades during desiccation: a small soluble molecule that occupies the place water would have occupied.

The fact that the frog tolerates 200 mmol/L glucose is itself a biochemical puzzle. At those concentrations, the glucose would normally produce extensive glycation of proteins — the same chemistry that causes complications in diabetic humans. The frog appears to have evolved protein structures that resist glycation, possibly through modifications to lysine residues, and the high glucose state is metabolized back to glycogen within hours of thawing without lasting damage.

The ice-nucleating proteins

The second element of the strategy is controlling where the ice forms. Ice crystals that form randomly in tissue are destructive; ice crystals that form in predictable extracellular locations can be controlled.

The wood frog produces ice-nucleating proteins that are secreted into the extracellular space and into the blood plasma. These proteins provide a structured surface on which ice can begin to crystallize at temperatures just below 0°C, which is much warmer than the supercooling temperatures at which random nucleation would occur. By controlling where nucleation happens, the frog ensures that ice forms in the spaces between cells and in the blood vessels (where it is mechanically containable) rather than inside cells (where it would be lethal).

The combination of ice-nucleating proteins (controlling where) and high glucose (preventing cellular damage where ice does not form) means the frog freezes in a controlled, predictable pattern: extracellular ice, intact cells, viable on thawing.

The metabolic shutdown

The third element is the controlled shutdown of metabolism as the frog freezes. Heart rate slows, then stops; breathing stops; metabolic rate drops to undetectable levels. The frog is not in a low-metabolism state like hibernation — it is in a no-metabolism state, with ATP production and consumption both at zero.

The viability question is therefore not "how does the frog maintain itself during freezing" but "how does the frog survive having no metabolism for months." The answer is that the cellular machinery is preserved by the cryoprotectants and the controlled freezing, and the metabolism simply resumes when conditions allow. The frog is not running in a low gear during winter; it is parked. The remarkable thing is that it starts again.

The startup is choreographed. Heart contraction resumes before circulation is fully restored; the heart pumps against ice plugs in the blood vessels until they melt. Liver glycogen production resumes within hours of thawing, pulling the high glucose levels back to baseline. Neural activity returns within minutes of thawing, with the frog responsive to touch and capable of coordinated movement.

The other freeze-tolerant species

The wood frog is the most-studied example but not the only one. Several other amphibians in the genera Rana, Hyla, and Pseudacris have freeze tolerance, generally based on the same glucose or glycerol strategies. Some terrestrial insects (especially in the genera Eurosta and Pterostichus) have analogous strategies based on glycerol and antifreeze proteins. Certain turtle hatchlings (Chrysemys picta, the painted turtle) overwintering in shallow nests have a limited freeze tolerance.

What is notably absent from the freeze-tolerant species list is mammals and birds. Endothermic vertebrates do not freeze and recover; they hibernate at low but positive metabolic rates. The reasons are partly that endotherms cannot afford to lose metabolic capacity for extended periods (their organs are heavily oxygen-dependent), partly that freeze tolerance seems to require evolutionary trajectories that endotherm physiology has not historically passed through.

The cryobiology research interest is partly basic biology and partly applied: organ preservation for transplant is a multi-billion-dollar problem, and the mechanism the wood frog uses to preserve cell viability through freezing is directly relevant. The clinical translation has been slow because mammalian tissue is much more sensitive than amphibian tissue, but the biochemistry of cryoprotectants in modern organ preservation owes a substantial debt to wood frog research.

The deeper observation

The vertebrate biology textbook of fifty years ago would have described the wood frog as impossible. Vertebrates were not supposed to freeze and recover; the textbook freezing-injury mechanisms were thought to be insurmountable. The discovery and characterization of the wood frog's strategy in the 1980s, primarily by Kenneth Storey's group at Carleton University in Ottawa, required revising the textbook account of what vertebrate cells could tolerate.

The pattern recurs across biology: phenomena that are universally true under the conditions where they were studied turn out to have exceptions when the conditions change. Cells dehydrated below 30 percent of normal water content are dead, except in tardigrades. Vertebrate brain tissue without oxygen for more than four minutes is irrecoverable, except in certain turtles that overwinter underwater for months. Vertebrates that freeze are dead, except in wood frogs. The universe of biological possibilities is consistently larger than the schoolroom version of biology suggests, and the cases that fall outside it are some of the most informative places to look.