A novel image analysis approach to characterise the effects of dietary components on intestinal morphology and immune system in Atlantic salmon

Abstract

The intestinal tract of salmonids provides a dynamic interface that not only mediates nutrient uptake but also functions as the first line of defence against ingested pathogens. Exposure of the immune system to beneficial microorganisms and different dietary immunostimulants via the intestine has been shown to prime the immune system and help in the development of immune competence. Furthermore, the morphology and function of teleostean intestines are known to respond to feed components and to ingested and resident bacterial communities. Histological appraisal is still generally considered to be the gold standard for sensitive assessment of the effects of such dietary modulation. The aim of the present study was to improve understanding of salmonid intestinal function, structure and dynamics and to use the knowledge gained to develop a model for analysis, which would allow intestinal health to be assessed with respect to different intestinal communities and feed components. Virtual histology, the process of assessing digital images of histological slides, is gaining momentum as an approach to supplement traditional histological evaluation methodologies and at the same time, image analysis of digitised histological sections provides a practical means for quantifiable assessment of structural and functional changes in tissues, being both objective and reproducible. This project focused on the development of a rapid, practical analytical methodology based on advanced image analysis, that was able to measure and characterise a range of features of the intestinal histology of Atlantic salmon in a quantitative manner. In the first research chapter, the development of a novel histological assessment system based upon advanced image analysis was described, this being developed with the help of a soybean feed model known to induce enteropathy in Atlantic salmon. This tool targeted the evaluation of the extent of morphological changes occurring in the distal intestine of Atlantic salmon following dietary modulation. The final analytical methodology arrived at, could be conducted with minimal user-interaction, allowing rapid and objective assessment of 12 continuous variables per histological frame analysed. The processing time required for each histological frame was roughly 20-25 min, which greatly improved the efficiency of conducting such a quantitative assessment with respect to the time taken for a subjective semi-quantitative alternative approach. Significant agreement between the fully automated and the manual morphometric image segmentation was achieved, however, the strength of this quantitative approach was enhanced by the employment of interactive procedures, which enabled the operator / observer to rectify preceding automated segmentation steps, and account for the specimen’s variations. Results indicated that image analysis provided a viable alternative to a pathologist’s manual scoring, being more practical and time-efficient. In the second research chapter, feeding Atlantic salmon a high inclusion level of unrefined SBM (25 %) produced an inflammatory response in the distal intestine as previously described by other authors. The model feed trial successfully generated differentiable states, although these were not, for the most part, systemically differentiable through the majority of standard immunological procedures used, being only detectable morphologically. Quantitation of morphometric parameters associated with histological sections using the newly developed image analysis tool successfully allowed identification of major morphological changes. Image analysis was thus shown to provide a powerful tool for describing the histomorphological structure of Atlantic salmon distal intestine. In turn, the semi-automated image analysis methods were able to distinguish normal intestinal mucosa from those affected by enteritis. While individual parameters were less discriminatory, use of multivariate techniques allowed better discrimination of states and is likely to prove the most productive approach in further studies. Work described in the third research chapter sought to validate the semi-automated image analysis system to establish that it was measuring the parameters it was purported to be measuring, and to provide reassurance that it could reliably measure pre-determined features. This study, using the same sections for semi-quantitative and quantitative analyses, demonstrated that the quantitative indices performed well when compared to analogous semi-quantitative descriptive parameters of assessment for enteritis prognosis. The excellent reproducibility and accuracy performance levels indicated that the image analysis system was a useful and reliable morphometric method for the quantification of SB-induced enteritis in salmon. Other characteristics such as rapidity, simplicity and adaptability favour this method for image analysis, and are particularly useful where less experienced interpreters are performing the analysis. The work described in the fourth research chapter characterised changes in the morphology of the intestinal epithelial cells occurring as a result of dietary modulation and aspects of inflammatory infiltration, using a selected panel of enzyme and IHC markers. To accomplish this, image analysis techniques were used to evaluate and systematically optimise a quantitative immunolabelling assessment protocol. Digital computer-assisted quantification of labelling for cell proliferation and regeneration; programmed cell death or apoptosis; EGCs and t-cell like infiltrates; mobilisation of stress-related protein regenerative processes and facilitation of nutrient uptake and ion transport provided encouraging results. Through the description of the intestinal cellular responses at a molecular level, such IHC expression profiling further characterised the inflammatory reaction generated by the enteropathic diet. In addition, a number of potential diagnostic parameters were described for fish intestinal health e.g. the relative levels of antigenicity and the spatial distribution of antigens in tissues. Work described in the final research chapter focused on detailed characterisation of intestinal MCs / EGCs in order to try to elucidate their functional role in the intestinal immune responses. Through an understanding of their distribution, composition and ultrastructure, the intention was to better characterise these cells and their functional properties. The general morphology, histochemical characteristics and tissue distribution of these cells were explored in detail using histochemical, IHC and immunogold staining / labelling, visualised using light, confocal and TEM microscopy. Despite these extensive investigations, their physiological function and the content of their granules still remain somewhat obscure, although a role as immunodulatory cells reacting to various exogeneous signals through a finely regulated process and comparable to that causing the degranulation of mammalian MCs is suggested. The histochemical staining properties demonstrated for salmonid MCs / EGCs seem to resemble those of mammalian mucosal mast cells, with both acidophilic and basophilic components in their granules, and a granule content containing neuromodulator / neurotransmitter-peptides such as serotonin, met-enkephalin and substance-p. Consequently, distinguishable bio-chromogenic markers have been identified that are of utility in generating a discriminatory profile for image analysis of such cells.