HMGs and antimirobial peptides in intestinal mucus
On fish mucin molecules, relatively few studies have been conducted, however, the data available indicate that fish mucus is similar in composition to the mucins secreted by mammalian goblet cells from epithelial tissues (Fletcher & Grant 1968; Asakawa 1974; Wold & Selset 1977; Lumsden & Ferguson 1994; Neuhaus et al. 2007b). For C. carpio, mucus glycoproteins have so far been described under physiological circumstances and after bacterial challenge for C. carpio (Neuhaus et al. 2007a,b; Schroers et al. 2009; Van der Marel et al. 2010). For mammals, it is well known that different parts of the gastrointestinal tract have different mucus thicknesses (Szentkuti & Lorenz 1995; Atuma et al. 2001) and express different mucus genes (Gendler & Spicer 1995; Van Klinken et al. 1995). For carp, so far two mucin genes have been found of which only one, Muc2, was expressed in the intestinal tract (Van der Marel et al. 2012). As only one Muc gene has been identified in the intestinal tact, it is not known whether different Muc gene profiles exist along the gut axis. However, morphological and functional differences have been previously found along the gut axis. Whether these differences between different parts of intestinal tract also include differences in mucus HMGs has to our knowledge not been extensively examined so far.
In this study, mucus glycoproteins within the goblet cells appeared to have a similar charge due to their glycosylation, as differences between the applied stainings could not be found. However, HPLC differences in the monosaccharide content of the mucus HMGs could be observed for both the first and second segments. Based on its molecular weight, the unidentified sugar in this study is probably ribose (Marko-Varga 1987). Ribose and 2-deoxyribose were the predominant sugars in rainbow trout intestinal mucus. For the first segment, the size area of the ‘ribose-peak’ was the highest. Fucose that is thought to be especially important for the viscoelasticity of the mucus, may vary strongly in intestinal mucus or anal gland mucus of mammals (Tsukise et al. 2000). The fucose content of C. carpio intestinal mucus was dependent on the origin of the mucus (first or second segment). For non-secreted mucus in the second segment, fucose was the most predominant sugar, as was also observed for C. carpio skin mucus (unpublished data). Non-secreted mucus is thought to be present in a native form, whereas secreted mucus might be altered by environmental influences such as bacterial degradation. In the first segment, NeuNAc/gal was predominant. The sugar galNAc could not be detected with the method used in the second segment, which was also the case for C. carpio skin mucus (unpublished data). In mammals, the monosaccharide profile of mucins is largely genetically determined (c.f. Freitas et al. 2005), however, can be modulated in response to bacterial colonization (Freitas et al. 2005). Research on germ-free animals indicated that the glycoproteins do not only play a role in binding of adverse molecules such as hormones, enzymes or microbial toxins, but also serve as nutrients for the commensal microbial flora. In germ-free mice, for instance, the expression of NeuNac was increased after oral application of bacteria from the conventional microflora, while the expression of GlycNac was decreased (Freitas et al. 2005). A high expression of Gal1,3Gal at the portal to the zebrafish intestine was considered to have a function in the transient recruitment of gut mutualists. Aeromonads members from the zebrafish gut flora preferentially bound to this sugar when expressed on cells. (Cheesman & Guillemin 2007). This might explain the high expression of GalNac in the first segment (recruitment of gut flora) and its decreased expression in the second segment. In general, on the monosaccharide level, C. carpio skin mucus resembled the mucus of the second segment more than the mucus of the first segment.
Besides sugar analysis, protein content was examined by size-exclusion chromatography. In all examined samples, a biphasic profile could be observed. Non-secreted and secreted mucus samples from the different intestinal segments contained almost similar protein profiles: their residues are between 120–180 kDa and 12–70 kDa. The proteins found between 120–180 kDa most likely represent mucins which adhere to the epithelium, as they are similar to the adherent mucins (AMs) from previous reports (Enss et al. 1996). The proteins found between 12 and 70 kDa probably consist of mucin constituents mixed with luminal proteins, the so-called luminal mucins (LMs) (Enss et al. 1996). Smaller molecules (LMs) probably represent the soluble luminal mucus. The AMs form a framework in which smaller molecules might be entrapped.
Luminal mucin might be essential for protection of the intestinal surface, as higher bacterial adhesion to smaller mucus molecules has been observed (Schroers et al. 2008). Mucin degradation by pathogenic bacteria has been suggested earlier for C. carpio (Van der Marel et al. 2008; Schroers et al. 2009) and might largely take place at the carbohydrate side chains of the mucins. In humans, it is known that bacteria with extracellular glycosidases may contribute to the damage of intestinal mucins (Jonas et al. 1977). In clinically healthy C. carpio, bacterial degradation of the protein core of mucus HMGs appears to be limited because secreted and non-secreted mucus have comparable-sized mucins indicated by their similar protein profiles.
However, between different intestinal segments differences in carbohydrate staining could be observed, indicating differences in pH of the mucus. Differences between the first and second segment in the staining intensity of the glycoproteins in the goblet cells support the role of mucus in the defence against pathogens. This fits well in the immunological role, as previously suggested for the second segment (Rombout et al. 1989). The mucus from the intestinal bulb might have also a different function in defence against bacteria as differences in carbohydrate staining to that of first and second intestinal segment could be seen.
A possible different role for mucus in the defence against bacteria is underlined by the different staining for BD2. Staining for BD2 was more pronounced for tissue samples of the first than for the second segment. It is probable that under physiological circumstances, the release of BD2 in the first segment is high enough, so that BD2 molecules can also be functional in the second segment. Staining for BD3 and lysozyme was low and similar between the first and second segment. Therefore, under unchallenged conditions, the humoral substances BD3 and Lys appear to be less important in C. carpio.
Besides the LMs, also AMs were found. Functionally, large molecules (AMs) are regarded to form the mucus layer, which adheres to the epithelium. The main function of these proteins appears to be some kind of stabilizer or carrier for the mucus. In the present study, size areas of the major protein residue between 120 and 180 kD were variable in height, but were a single peak with little variation in retention time. Retention volume of the first peak was similar for samples from the first and second intestinal segment (peak at approximately 150 kDa) but was slightly but significantly larger (approximately 140 kDa) for mucus of the intestinal bulb. Size shifts of around 10 kDa are often caused by a difference in the glycosylation pattern of a protein. The smaller size of the molecules might play a role in food transportation, as smaller molecules are considered to be washed away easier. The intestinal bulb has a food storage function, with the mucus enveloping the food for easier transport. Furthermore, a large part of the digestion has already taken place in the intestinal bulb.
Mucus alterations upon SBM
Soybean meal is used in fish feed, as it is a cheap source of protein. The use of soybean in fish feed, however, is sometimes causing problems, as SBM-containing diets are known to induce an inflammatory response in the hindgut of certain fish species. Contrary to previous observations made with S. salar, C. carpio start to recover or adapt to the SBM feed from the fourth week after the SBM feeding (Uran et al. 2008).
After SBM feeding, the second gut segment of C. carpio did not show external disease symptoms as found previously for carp by Uran et al. (2008). A clear loss of supranuclear vacualistion could not be found. However, a clear increase in goblet cell number in the AB2.5 staining by week three, as found by Uran et al. (2008), could also be observed. This indicates that the carp in this study suffered from an acute but mild enteritis process. In carp, the enteritis process is accompanied by damages to the intestinal epithelium (Uran et al. 2008). In the present study, a higher number of bacteria could be isolated from internal organs after SBM feeding. This might be explained by damage to the intestinal epithelium. The increased bacteria number in the internal organs indicates that the mucosal barrier as first line of defence was compromised, supporting the theory that the intestinal barrier was affected. For humans, increased gut permeability has been described following enteritis. If changes after enteritis persist, a chronic inflammation can develop (Dunlop et al. 2006). Between the control and week 3 after 20 SBM feeding, no difference could be found in bacterial numbers in intestinal organs, indicating that the intestinal barrier function recovers over time.
All samples from the SBM experiment showed a biphasic protein profile with a major protein peak at approximately 150 kDa. The size of the large mucins did not change markedly. However, switching to the 20 SBM diet led to a significant increase in peak height of 150 kDa proteins in non-secreted mucus of the first segment, whereas the amount of 150 kDa in secreted mucus remained relatively stable. The obseved increase in peak height became less over time. Mucus has a high turnover rate, as mucus traps pathogens that can be removed from the body through a constant flow of mucus. The increase in the non-secreted mucus indicates that the SBM diet induces an increased mucus synthesis. The amount of newly synthesized mucus might increase the mucus flow, which would explain why the amount of secreted mucus remains stable. An increased mucus flow stimulates the removal of pathogens from the entire subsequent intestine, which might hereby prevent pathogens from entering the soybean-damaged intestinal epithelium. Changes in mucin levels during enteritis have been described for chronic IBD in humans.
In IBD, genetic mutations in mucin genes, changes in sulphation, degree of glycosylation, mucin mRNA, protein levels and degradation of mucins have been described. Changes of immunological or bacterial factors during an initial or ongoing inflammation can influence mucin production, which could have further adverse effects on mucosal–bacterial interactions, hereby sustaining the chronic character of the inflammation (Einerhand et al. 2002). In ulcerative colitis (a form of IBD), mucin protein levels (Einerhand et al. 2002) as well as the expression of MUC2 were reduced (Van der Sluis et al. 2008). MUC2 is the structural component of the colonic mucus layer in humans (Van der Sluis et al. 2008). Reduced Muc2 levels were also observed in interleukin (IL)-10 knockout mice that develop colitis (Van der Sluis et al. 2008). The expression of mucin 2 proteins and the anti-inflammatory cytokine IL-10 by type-2 T-helper cells and activated macrophages, which induces its expression, seems to be essential in the control of enteritis in mice. The increase in mucus glycoproteins as seen in the first intestinal segment in carp under SBM diet decreased over time, which indicates that mucus composition returns to the state before switching to the SBM diet. An initial up-regulation of IL-10 at week 1, followed by a down-regulation to slightly below or to the level before the enteritis induction was observed previously for C. carpio (Uran et al. 2008).
In the present study, a 20 SBM diet induced changes in the mucus, indicating that the mucosal layer showed altered mucus secretion upon 20 SBM feeding. In the first segment, the amount of sulphated glycoconjugates decreased, suggesting that similar to IBD reports (Einerhand et al. 2002), SBM feeding also leads to changes in mucus sulphation in C. carpio. In addition, mucus from SBM-fed C. carpio showed a different glycosylation pattern as was also reported for IBD (Einerhand et al. 2002). The different glycosylation pattern is indicated by the significantly increased amount of neutral glycoconjugates in the first intestinal segment and the significantly decreased amount of acid and neutral glycoconjugates in the second intestinal segment.
Besides the changes in mucus HMs, also a difference in staining for BD3 could be observed. One week after switching to the SBM diet, the amount of goblet cells that were stained for the antimicrobial peptide BD3 was significantly reduced. This might indicate that BD3 was released to the intestinal lumen to fight pathogens, as BD3 in humans has a strong bactericidal activity (Maisetta et al. 2003). No goblet cells could be stained at week 1 after the diet switch, this might indicate that BD3 was released more rapidly than it could be produced. After the initial depletion of the goblet cells, synthesis and secretion of BD3 seemed to return to their initial equilibrium, as on week 2 or 3, no differences compared with the control could be observed anymore.