The formalin test does not probe inflammatory pain but excitotoxicity in rodent skin.

The most widely used formalin test to screen antinociceptive drug candidates is still apostrophized as targeting inflammatory pain, in spite of strong opposing evidence published. In our rat skin-nerve preparation ex vivo, recording from all classes of sensory single-fibers (n = 32), 30 units were transiently excited by formaldehyde concentrations 1-100 mM applied to receptive fields (RFs) for 3 min, C and Aδ-fibers being more sensitive (1-30 mM) than Aβ-fibers. From 30 mM on, ~1% of the concentration usually injected in vivo, all RFs were defunctionalized and conduction in an isolated sciatic nerve preparation was irreversibly blocked. Thus, formaldehyde, generated a state of 'anesthesia dolorosa' in the RFs in so far as after a quiescent interphase all fibers with unmyelinated terminals developed a second phase of vigorous discharge activity which correlated well in time course and magnitude with published pain-related behaviors. Sural nerve filament recordings in vivo confirmed that higher formalin concentrations (> 42 mM) have to be injected to the skin to induce this second phase of discharge. Patch-clamp and calcium-imaging confirmed TRPA1 as the primary transducer of formaldehyde (10 mM) effects on mouse sensory neurons. However, stimulated CGRP release from isolated skin of TRPA1 and TRPA1 mice showed a convergence of the saturating concentration-response curves at 100 mM formaldehyde, which did not occur with nerve and trachea preparations. Finally, skin-nerve recordings from C and Aδ-fibers of TRPA1 mice revealed a massive reduction in formaldehyde (30 mM)-evoked discharge. However, the remaining activity was still biphasic, thus confirming additional unspecific excitotoxic actions of the fixative that diffuses along still excitable axons as previously published. The multiplicity of formaldehyde's actions requires extensive discussion and literature review, leading to a fundamental reevaluation of the formalin test.