Brucella species infection in North Sea seal and cetacean populations

albumin to 1:750 and 1:1500.Overnight incubation at 4°C was the best condition for the different primary antibodies employed.Swine anti-rabbit immunoglobulins 1:20 for use as a secondary antibody and PAP complex 1:100 were both made up in phosphate buffered saline with I per cent bovine serum albumin and incubat- ed for 30 minutes at room temperature.To detect the immune complexes, the sections were incubated in a solution of 60 mg of 3,3-diamino-benzidine tetrahydrochloride, 100 ,ul hydrogen per- oxide and 100 ml of phosphate buffered saline for eight minutes and counterstained with Mayer's haematoxylin.Ganglia from all horses with either mal seco or EGS showed numerous affected neurons with their cytoplasm positively stained for anti-whole horse serum (Fig 1).No differences between immunostaining with anti-whole horse serum preincubated with bovine serum albumin or horse serum albumin were detected.Fewer neuronal perikarya stained positively for anti-horse IgG.The highest dilutions of anti-whole horse serum and anti-horse IgG causing the strongest staining with the lowest background were 1:1500 and 1:750, respectively.PAP staining of affected neurons was not observed when normal serum replaced the anti-whole horse serum or the anti-horse IgG.Ganglia from the horse destroyed due to volvulus showed few neurons positively stained for both primary antibodies.Neurons in similar sections from normal horses failed to immunostain with either primary antibody (Fig 2).The degenerative changes of the nerve cell body observed in the haematoxylin and eosin-stained ganglia sections from the affected horses consisted mainly of chromatolysis and vacuolation in the damaged neuron perikarya, neuronophagia, intercellular and intracytoplasmic eosinophilic bodies and pyknotic and eccentric nuclei.As they are considered pathognomonic lesions (Pollin and Griffiths 1992, Griffiths and others 1993) they confirmed a clini- cal diagnosis of mal seco or EGS.This study presents the first evidence of intraneuronal serum proteins in cases of mal seco.These results have not been reported before and are consistent with those described for EGS (Griffiths and others 1994b).The putative neurotoxin suspected of being the aetiological agent in EGS may reach neurons by retrograde axonal transport (Griffiths and others 1994a).If this is true for mal seco, a cell- mediated and/or a humoral immune response, either to the unknown aetiological agent or to some component of the nervous system, could explain the presence of serum-derived molecules in the nerve cell body as was considered for EGS (Griffiths and others 1994a).It is also possible that, as in other species (Olsson 1984), the blood vessels of autonomic ganglia of horses have no blood- nerve barrier.If this were the case, the serum proteins could be non-specifically absorbed by the affected neurons.Why this only occurs in damaged neurons, what stimulates it to occur and what the significance is of these serum proteins in the neuronal perikarya remain to be elucidated.