Nous examinons théories et faits neuroscientifiques (anatomie, histologie, biologie, physiologie du système nerveux central ou SNC) du début du siècle dernier à partir de l’emploi psychiatrique du nucléinate (forme d’ADN). Nous nous centrerons sur les causes des troubles mentaux et les marqueurs d’efficacité des traitements au travers de l’anatomopathologie, de l’analyse du liquide céphalorachidien (LCR) et des tentatives de reproduction de symptômes psychiatriques par injection de substances supposées toxiques. Le nucléinate augmente le nombre des leucocytes et entraîne de la fièvre, favorisant théoriquement la phagocytose. La médecine l’emploie contre les toxi-infections. En 1907, Lépine l’utilise en psychiatrie. Ses hypothèses neuroscientifiques (encéphalite tuberculeuse subaiguë, faiblesse métabolique neuronale et des oxydations) sont peu étayées, les résultats thérapeutiques modestes. En 1909, pour Donáth, le nucléinate éliminerait les métabolites toxiques apparaissant lors de la paralysie générale (PG) via oxydation, fièvre et leucocytose. Ses résultats sont optimistes. Plus empirique, Fischer propose le nucléinate dans la PG avec un argumentaire plus immunitaire que neuroscientifique. Avant-guerre, le test de Wassermann dans le LCR donne l’espoir d’un marqueur de rémission, associé à la lymphocytorachie et à la protéinorachie. Cependant, Jolowicz n’en observe pas la normalisation en combinant nucléinate et antisyphilitique. En 1918, Cazamian espère augmenter le taux d’antisyphilitique dans le LCR en combinant nucléinate en sous-cutané et Enésol en intraveineux. C’est un échec. Dans la démence précoce (DP), les théories d’intoxications externes ou internes, souvent bactériennes, étaient centrales. En 1911, Itten expérimente le nucléinate dans la schizophrénie (SCZ), espérant stimuler les oxydations cérébrales ou fixer des toxines, sans résultats probants. En 1913, Donáth évoque dans la DP, en sus d’auto-intoxications hypothétiques, une fragilisation préalable du SNC, non prouvée. Ses résultats cliniques dans la DP/SCZ sont médiocres. Spéculations et biologie prenaient le pas sur les faits neuroscientifiques, phénomène accentué par le filtre du nucléinate. Dans la PG, les neurosciences permirent d’améliorer les connaissances pathogénétiques et thérapeutiques. Dans la DP/SCZ, elles ne validèrent pas les modèles autotoxiques. Des hypothèses proches (infections, microbiote, neuro-inflammation, stress oxydatif) sont encore considérées dans la SCZ actuellement. Les neurosciences manquaient d’outils d’exploration in vivo et de modèles animaux. La méthodologie statistique était balbutiante.

We examine some neuroscientific theories and facts at the beginning of the last century, in light of the therapeutic use of nucleinate (a form of DNA) in psychiatry. Methods for exploring the central nervous system (CNS) in vivo were limited, however the knowledge about its anatomy, histology, chemistry and physiology were well developped. Infections and autointoxication were highly considered as causes of general paralysis (GP) and dementia præcox (DP). We considered as neuroscientific: theories and facts about anatomomapathology; exploration and measurement of the cerebrospinal fluid (CSF); attempts to reproduce the psychiatric symptoms by injection of suspected substances, in GP and DP.

About 20% of patients admitted to asylums suffered from GP. DP accounted for 5 to 15% of admissions. Treatments were varied but ineffective and often toxic.

Nucleinate, used in medicine since 1892, increased the number of leucocytes (leucotherapy) and caused fever, giving hope of promoting phagocytosis and counteracting toxi-infections. In 1907, Lépine tried it in asylum patients with chronic conditions. His neuroscientific etiologic hypotheses (subacute tuberculous encephalitis, weakness of neuronal metabolism and oxidation) were poorly supported. The therapeutic results were very modest.

In 1909, Donáth believed that toxic metabolites appeared during GP, nucleinate being supposed to eliminate them through oxidation and hypermetabolism associated with fever and leukocytosis. His results were presented optimistically. Fischer's approach was mainly empirical, drawing on artificial fevers (Wagner-Jauregg). He cited a theory relating GP with a trouble of lecithin metabolism. He used a control group. Before the war, the Wassermann's test in CSF offered hope for a marker of remission, associated with lymphocytosis and proteinorachy. In 1913, Jolowicz failed to find a normalization of the Wassermann's test in CSF by combining nucleinate and an antisyphilitic drug. In 1918, Cazamian used neuroscientific arguments about the meningeal permeability to increase arsenic (component of antisyphilitic drug) levels in the CSF. Assuming that systemic inflammation caused by subcutaneous injection of nucleinate would increase this permeability, he combined it with intravenous injection of an antisyphilitic drug (Enesol). It was a failure. The blood-brain barrier was not yet known.

Bacterial and internal toxins (auto-intoxication), metabolic products of colonic bacteria, were invoked. In 1911, Itten experimented nucleinate in schizophrenia (SCZ), hoping to stimulate metabolism (oxidations) or to bind toxins, but his clinical results were poor. In 1913, Donáth was interested in counteracting autointoxication with nucleinate. He considered the possibility of a prior non-specific weakening of the CNS. His results were also deceptive. In the 1920s, independently of nucleinate, Baruk worked on experimental catatonia with various substances including bulbocapnine and a coliform toxin, reopening the field of experimental neurobiology.

Speculative intuitions and general biology took precedence over neuroscientific facts, with the nucleinate filter accentuating this trend, because it was primarily an immunological treatment. In GP, the neurosciences helped to improve pathogenetic hypotheses and treatments. Regarding DP/SCZ, autotoxic models lacked histological and neurophysiological evidence. Some hypotheses similar to those reviewed here are always under investigation (infections, microbiota, neuroinflammation, oxidative stress, neurodegeneration). At this time, neuroscience lacked tools, especially in vivo exploration, animal models, and statistical methodology.