The healing of tympanic membrane (TM) perforations is a complex wound healing process including inflammation, migration of keratinocytes and tissue remodelling. Most TM perforations in human heal spontaneously, however some perforations become chronic, and the reason to why is still largely unknown. In cutaneous wound healing plasminogen (plg) has been shown to play an important role. Plg is converted into the protease plasmin regulated by two plasminogen activators (PA), urokinase type PA (uPA) and tissue-type PA (tPA). The aim of the present thesis was to evaluate the role of plg in healing of TM perforations, both in vivo and in vitro. The main objectives were to determine the healing capacity of the TM, the involvement of keratinocytes, fibrin(ogen) and inflammatory cells in the healing process. The studies were performed in plg deficient and uPA deficient mice, with littermate wild type (wt) mice as controls It was shown that myringotomies of the TMs in plg deficient mice still remained open 143 days following a perforation. The wound area was characterized by an abundant recruitment and accumulation of inflammatory cells; mainly macrophages and neutrophils, an arrested keratinocyte migration and a fibrin deposition covering the surface of the TM. The TM perforations in the wt mice all healed within 11 days. Interestingly, the myringotomies of the plg deficient mice could be closed by reconstitution with systemic injections of plg, whereas injections of PBS had no affect on the healing. To characterize mechanisms involved in the development of persistent TM perforations in plg deficient mice after a myringotomy the early inflammatory response during the first 48 hours was studied. The recruitment and accumulation of inflammatory cells in the perforated TMs was found to be similar between the plg deficient and the wt mice. Myringotomized TMs in uPA deficient mice healed similar to perforations of wt controls. Neither did the keratinocyte migration nor the occurrence of inflammatory cells differ between these genotypes. In the in vitro experiments TMs from plg deficient and wt mice, were dissected out, perforated and cultured in absence or surplus of plg. A decrease in perforation size was seen in all groups regardless of genotype or amount of plg in the medium. In conclusion, the present studies show: • Plg is essential for the healing of TM perforations in mice. • The altered healing process after a myringotomy in plg deficient mice involves a disturbed keratinocyte migration, a massive deposition of fibrin and an abundant accumulation of inflammatory cells in the wound area. • Plasminogen deficiency does not alter the early inflammatory response, following a myringotomy. • Deficiency of uPA does not influence the healing of TM perforations. • During in vitro conditions healing of TM perforations is initiated irrespectively of genotype of the explant (plg deficient or wt) or supply of plg. The increased knowledge of the involvement of plg in the healing of TM perforations may open therapeutical possibilities in the treatment of chronic TM perforations in humans.

L'étude systématique de l'hémostase post-thrombolytique a été peu étudiée. Chez 80 malades thrombolysés consécutifs, une étude prospective a comporté l'étude – aux heures 0, 2 et 24 – des facteurs de l'hémostase suivants: fibrinogène, plasminogène, PDF (produits de dégradation de la fibrine et du fibrinogène), D-dimères, alpha2-antiplasmine et facteur XIII, ainsi que l'hématocrite et la numération plaquettaire. Des calculs statistiques approfondis ont exploré les corrélations des variations des facteurs hémostatiques entre eux et avec 37 paramètres cliniques et étiologiques. Processus moléculaires post-thrombolytiques. Le rt-PA induit deux processus, indépendants statistiquement à la 2ème heure: d'une part une élévation des PDF et des D-dimères; d'autre part, une baisse du fibrinogène, corrélée à une baisse du plasminogène (r=0,48, p=0.01), de l'alpha2-antiplasmine (r=0.48, p =0.004) et du facteur XIII (r=0.44, p=0.01). La baisse du plasminogène est corrélée significativement avec celle de l'alpha2-antiplasmine (r=0.77, p<0.001), et du facteur XIII (r=0.47, p=0.02). La mise en jeu de facteurs anti-fibrinolytiques, qui n'avait jamais été décrite précédemment, peut jouer un rôle dans une limitation de la fibrinolyse et dans la rethrombose. Des corrélations sont notées entre la baisse précoce du plasminogène et l'étiologie cardioembolique (p=0.04), et un mauvais pronostic final (p=0.03), possiblement en rapport la thrombolyse intense de gros caillots. Les hématomes intra-cérébraux parenchymateux (HP) sont liés significativement à la baisse du fibrinogène (p=0.01) et à l'augmentation des PDF (p=0.01). Une baisse du fibrinogène au-dessous de 2g/L multiplie la probabilité de HP précoce par un facteur 12,82. Ainsi est confirmé le modèle d'une “coagulopathie précoce avec dégradation du fibrinogène”», prédictive de l'hématome, proposé par l'équipe lyonnaise de thrombolyse en 2004 / A systematic study of post-thrombolytic haemostasis has rarely been performed. In 80 consecutive patients, we have prospectively studied at hours 0, 2 and 24 the following parameters: fibrinogen, plasminogen, alpha2-antiplasmin, factor XIII, fibrin(ogen) Degradation Products (FDP), D-dimers, haematocrit and platelet count. Comprehensive statistical studies calculated correlations of the haemostatic values betwen themselves and with 38 etiological and clinical parameters. Molecular dynamics. Two changes between h0 and h2 were statistically independent: an increase in FDP and D-Dimers; a decrease in fibrinogen, plasminogen, alpha2-antiplasmin and factor XIII. At h2, the decrease in fibrinogen was significantly correlated with that of plasminogen (0.48, p = 0.01), alpha2-antiplasmin (0.48, p = 0.004), and factor XIII (0.44, p = 0.01). The decrease in plasminogen was significantly correlated with those of antifibrinolytic components, alpha2-antiplasmin (r=0.77, p<0.001) and factor XIII (0.47, p=0.02). To our knowledge, such an activation of antifibrinolytic components had not hitherto been mentioned. The h2 decrease of plasminogen was correlated with cardioembolic etiology (p=0.04) and final poor oucome (p=0.03), a fact possibly due to intense thrombolysis of large clots. Patients having early parenchymal hematomas (PH) showed h2 haemostasis disturbances: high FDP (p=0.01), and low fibrinogen (p=0.01). The decrease in fibrinogen less than 2g/L multiplies the odds of early PH by a factor 12.82. Thus, we confirm the model of an “early fibrinogen degradation coagulopathy” predictive of hematomas, which had been coined by the Lyon thrombolysis team in 2004