Skin mucus N-glycolylneuraminic acid relates to mucous cell parameters in Atlantic salmon

Grigory V. Merkin^1*, Anette Sauve², Mearge A. Okubamichael¹, Trond-Inge Kvernevik³, Karin Pittman^1,4**

¹QuantiDoc AS, 5006 Bergen, Norway

²Vitas AS, 0349 Oslo, Norway

³Caligen Holding AS, 6700 Måløy, Norway

⁴University of Bergen, 5006 Bergen, Norway.

---

Abstract

Sialic acids (Sias) in fish skin mucus play a crucial role in mediating interactions with the surrounding environment, serving as key components of the protective mucus that coats the fish. Sias contribute to the physical and chemical barriers that protect fish from pathogens and support mutualistic relationships with bacteria. In Atlantic salmon (Salmo salar), N-Glycolylneuraminic sialic acid (Neu5Gc) concentration in skin mucus and skin mucous cell (MC) parameters such as size, abundance, and volumetric density were monitored during the transition from freshwater to seawater across five sampling dates. A significant positive regression was observed between Neu5Gc concentration (free Neu5Gc and total Neu5Gc, encompassing both free and bound forms) and MC parameters in four of the five sampling points, suggesting a relationship between mucous barrier characteristics and Neu5Gc concentration in the mucus. Moreover, the transition to seawater resulted in a significant decrease in both MC parameters and Neu5Gc concentration (p < 0.05). Together, these findings support the interpretation that MCs represent a source of Neu5Gc in the skin mucus and that both Neu5Gc levels and MC characteristics appear to change during the transition to seawater.

---

Introduction

External mucosal surfaces acts as immunological barriers^1,2. The fish mucus contains a wide range of antimicrobial molecules and immune factors that contribute to host defense^{1, 3}. Beyond its immunological functions, skin mucus also plays key roles in other physiological processes ⁴, including maintaining osmotic balance⁵. The mucosa can be affected by smoltification⁶ and seawater transfer (SWT)⁷ a key transition in the normal life cycle of Atlantic salmon (Salmo salar).

Fish skin secretes a mucous layer primarily consisting of mucins, which are synthesized by mucous cells (MCs) ⁸. Sialic acids (Sias)⁹, are constituents of mucins and play a crucial role in maintaining the integrity of the mucosal barrier in vertebrates ^{10, 11}. Sias also join various physiological and pathological processes, functioning as key ligands for cell-adhesion molecules ¹². They serve as interaction points with various bacteria in the mucosa ^{13, 14, 15}, which has adaptations that support colonization by a diverse community of commensal microbes ^{1, 16}. In vertebrates, Sias function as recognition targets for various pathogens, such as viruses, bacteria, and protozoa, either enhancing innate immune protection or facilitating pathogen invasion ¹².

In fish, as in most vertebrates, the conversion of N-Acetylneuraminic acid (Neu5Ac) to N-Glycolylneuraminic acid (Neu5Gc) occurs through hydroxylation of the N-acetyl group ¹². Although Neu5Gc is less common than Neu5Ac (18% vs. 58%) in the Atlantic salmon skin mucus ¹⁷, the skin mucins carry a high amount of terminal Neu5Gc, resulting in approximately 70% of the glycans in glycoconjugates being acidic ¹⁸. The presence of skin Neu5Gc as terminal constituents in the mucins highlights its potential role in coordinated communication at the tissue level between fish and their aquatic environment, including interactions with microbiota. Moreover, it has been proposed that Sias, including Neu5Gc, may function as a decoy for certain pathogens ¹².

The stress in Atlantic salmon induces a remodelling of mucin O-glycosylation in the skin mucus, characterized by an increased proportion of large O-glycan structures that affect sialylation pattern of skin mucin O-glycans ¹⁹. Although changes in mucin O-glycosylation (including sialylation) have been documented in salmonids under infection- or inflammation-related conditions ^{20, 21}, the effects of salinity transitions on the detailed Sias composition in skin mucus remain largely uncharacterised. While a correlation between the abundance of skin MCs and Neu5Ac concentration in fish skin mucus has been demonstrated ²², we now report a similar relationship between MC parameters and Neu5Gc.

Materials and Methods

The experiment was conducted on Atlantic salmon (Stofnfiskur Iceland) in a flow-through system in ILAB (Bergen, Norway) from 4.08.2020 (Day 1) to 18.11.2020 (Day 107). The fish were monitored from the parr to post-smolt stages (Fig.1A), transitioning from freshwater to seawater (Appendix 1). The fish were kept in freshwater until 14 September. The fish were then gradually adjusted to seawater by being held for two days at 20 parts per thousand (ppt) salinity before being exposed to full-strength seawater on 16 September (Appendix 1). The fish (87.27±76.06 g) were sampled (six tanks per sampling date and five fish per tank) in freshwater (11.08.2020, Day 8) and after salinity was increased to 34 ppt (22.09.2020, Day 50 and 20.10.2020, Day 78) (Appendix 1). On Day 79 (Fig.1A), the fish were divided into two challenge groups, each exposed to a different pathogen on the same day: one group to salmon lice (Lepeophtheirus salmonis) and the other to infectious salmon anemia virus (ISAv) (Appendix 2), to simulate the challenges that farmed fish face from pathogens after their transfer to sea cages. The group exposed to salmon lice was sampled on 05.11.2020 (Day 94), and the group exposed to ISAv was sampled on 18.11.2020 (Day 107) (Fig.1A). All the fish (n=150) were sampled for skin mucus (Appendix 3) and for skin biopsies (Appendix 4).

Mucosal mapping technology (Appendix 5), which was developed in academia ^{23, 24}, was applied to measure MC area (μm²), the volumetric density of MCs in the epithelium (in %), and MC defense activity (the abundance of MCs in the tissue) ^{25, 26, 27}.

Total (bound and free) and free Neu5Gc were quantified in salmon skin mucus samples ²⁸ (n=150, Appendix 1) using LC-HRMS (Appendix 3) at a commercial laboratory (Vitas AS, Oslo, Norway). Total Neu5Gc includes both free (unbound) and bound forms. Statistical data analysis (Appendix 6) was performed in R ²⁹.

Results

As compared with the first sampling in fresh water (Day 8, Appendix 1), six days after adjustment water salinity to 34 ppt (Day 50, Appendix 1), the overall change in the skin mucosa was a decrease in MC size, abundance, and volumetric density (Fig. 1B, p < 0.05; Appendix 7). This change corresponded to a significant decrease in both total and free Neu5Gc concentrations (5 and 15 times respectively, Table 1; p < 0.05, Appendix 8). Further exposure of the salmon to seawater (Day 78) did not result in significant changes in MC parameters (Fig. 1B; p > 0.05, Appendix 7) or in either total or free Neu5Gc concentrations (Table 1; p > 0.05, Appendix 8).

At fifteen days post-exposure to salmon lice (Day 94), there was a significant decrease in MC size only (Fig. 1B; p < 0.05; Appendix 7) and a significant increase in free Neu5Gc levels (Table 1; p < 0.05, Appendix 8). By contrast to the parasite, twenty-eight days after the ISAv virus was induced, a significant increase in both MC size and abundance was observed (Day 107, Fig. 1B, p < 0.05; Appendix 7), but no significant changes in the levels of free and total Neu5Gc (Table 1, p > 0.05; Appendix 8).

At the start of the experiment (Day 8), only the positive regression models for MC size vs. total Neu5Gc and MC volumetric density vs. free Neu5Gc were significant (Table 1). By Day 50, only the positive regressions for MC density vs. total Neu5Gc and defense activity vs. both total and free Neu5Gc were significant. At Day 78, all positive regression models assessing the relationship between Neu5Gc and the MC parameters were statistically significant (Table 1). However, at Day 94, the regression between MC size and total Neu5Gc concentration was significantly negative (Fig. 1C, Table 1), while no significant relationship was observed with free Neu5Gc. By the end of the experiment (Day 107), all positive regression models evaluating the association between Neu5Gc and MC parameters were again statistically significant (Table 1).

Moreover, a descriptive comparison of the regression outputs showed that the proportion of variance explained (adjusted R²) was more than twofold higher at Day 78 and at the end of the experiment than at the initial freshwater sampling (Table 1). In general, the positive regression between mucosal parameters and total Neu5Gc concentration was more significant than the regressions observed for free Neu5Gc (Table 1).

Discussion

In this study, a reduction in MC parameters after adjustment of water salinity to seawater (Day 50) supports previous findings ⁶ and corresponds with the statistically significant decrease in Neu5Gc concentrations in response to salinity adjustment, which, to our knowledge, is reported here for the first time in Atlantic salmon. Thus, both the decrease in MC parameters and the reduction in Neu5Gc concentration appear to reflect physiological responses to changes in salinity and/or smoltification, including reduced mucus production, changes in mucus composition. Because the salinity in the tanks was adjusted gradually and without handling the fish, the observed reductions in MC parameters and Neu5Gc concentrations are not attributable to handling stress.

This finding aligns with previous observations in salmonids, which demonstrate that fish mucus undergoes compositional changes in response to shifts in salinity ³⁰. A number of genes for mucosal proteins were down and up-regulated, indicating a change in mucus composition during the transition to seawater ³¹. In line with this, the change from brackish water to seawater led to a reduction in the expression of several genes involved in immune defence in Atlantic salmon ³².

Possibly due to previous salmon lice exposure, in addition to some decrease in skin MC size, only on Day 94 was a significant negative regression between MC size (Fig. 1C, Table 1) and total Neu5Gc concentration. The overall disruption of the skin mucosal barrier by salmon lice leads to changes in mucus composition ³³ and could potentially result in alterations to Neu5Gc profile in the skin mucus, although we are not able to exclude the effects of handling two weeks prior to sampling (Appendix 2), or to seawater adaptation.

After ISAv exposure, at the end of the experiment (Day 107), there was some increase in MC abundance and size (Fig 1B). Although ISAv attaches specifically to 4-O-acetylated Sias on salmon epithelial surfaces ³⁴, there is currently no information on whether ISAv can specifically attach to Neu5Gc or directly affect Neu5Gc metabolism patterns. Moreover, according to the Quantidoc database, the skin MC values for all sampling dates in seawater, including post-ISAv sampling, fell within the most common range (mean ± 0.5 standard deviations) for salmon (green zone, Fig. 1B) ²⁶, suggesting that the effect of exposure to the virus was relatively minor compared to the initial impact of the salinity increase to 34 ppt. Although, it remains challenging to differentiate the effects attributable to seawater adaptation from those associated with ISAv exposure, these increase in MC parameters after the fish have adapted to the seawater environment at the end of the experiment (Day 107) (Fig 1B) aligns well with a previous study showing compensatory increase in MC abundance approximately one to two months after the transfer to seawater ³¹.

Summarizing the measurements of mucosal cellular characteristics and Neu5Gc (Table 1), we observed a consistent trend of positive regression between MC size (Fig. 1C, Table 1), volumetric density (Fig. 1E, Table 1), and abundance (Fig. 1G, Table 1) and the total Neu5Gc concentration in skin mucus across all sampling dates, except following salmon lice exposure (Day 94). Lower adjusted R² values (Table 1), reflecting a smaller proportion of explained variance, were observed in freshwater (Day 8) and following adjustment of water salinity to seawater (Day 50). These patterns may be associated with physiological changes related to smoltification and early seawater adaptation. In contrast, higher adjusted R² values (Table 1) after prolonged seawater acclimatization (Day 78) and at the end of the experiment (Day 108) suggest that the association between Neu5Gc concentration and MC parameters became stronger following adaptation to marine conditions.

In general, the positive regression between mucosal parameters and total Neu5Gc concentration was more significant than the correlations observed for free Neu5Gc (Fig. 1, Table 1). Total Neu5Gc includes both free (unbound) and bound forms, where bound Neu5Gc is integrated into glycoproteins like mucins ^{12, 17}, and free Neu5Gc exists as a monosaccharide not attached to proteins or lipids. In vertebrates, certain bacteria take up free Sias from their hosts, which may play a crucial role in supporting the growth and colonization of commensal bacteria on mucosal surfaces ³⁵. Free Neu5Gc in the mucus might be more readily utilized by microbial activity or physically removed through natural mucus shedding, thus potentially leading to less consistent associations with mucosal parameters in the current study. In contrast, bound Neu5Gc, due to its incorporation into stable structures like glycoproteins, might reflect more persistent and functionally relevant changes in the mucus composition.

Concluding Remarks

Skin MCs likely represent a source of Neu5Gc in skin mucus, similar to Neu5Ac ²². However, variability in Neu5Gc levels, MC dynamics, and the degree of association between these parameters across sampling dates in the current study may have been influenced by physiological changes related to smoltification and seawater adaptation, and probably also by host responses to salmon lice exposure. Variability in mucus turnover and mucin hydration may introduce additional variance in measured Neu5Gc concentrations that is independent of MC parameters. In future studies, parallel quantification of total carbohydrates or glycans would help normalize Neu5Gc levels and distinguish between dilution effects related to mucus hydration and true compositional changes. Moreover, limitations associated with the mucus sampling method ²⁸ may have influenced the observed association between MC parameters and Neu5Gc concentrations (Appendix 1).

Because salinity was increased without handling, the reductions in MC parameters and Neu5Gc concentrations observed after salinity adjustment cannot be attributed to handling stress in this study. Moreover, mucus viscosity is higher in seawater than in freshwater, a salinity-dependent change that is thought to reflect reduced mucin hydration in seawater rather than major differences in glycoprotein composition ³⁰. If decreased hydration in seawater resulted in a simple concentration of mucus components, an increase in Neu5Gc concentration would be expected. However, the opposite pattern was observed in the present study, indicating that the reduction in Neu5Gc cannot be explained by changes in mucus hydration, but is likely associated with the decrease in the size and abundance of MCs. These findings support the hypothesis that MCs contribute to Neu5Gc content in fish mucus.

The results represent an important step toward establishing a conceptual model in which, increases in MC size, abundance, and volumetric density are associated with enhanced mucin synthesis and secretion, leading to elevated Neu5Gc levels in skin surface mucus.

Figure Legends

Figure 1: Overview of the experimental set-up, mucosal parameters, and glycolylneuraminic acid (Neu5Gc) concentration in Atlantic salmon (Salmo salar). The fish were sampled in freshwater (Day 8, FW) and after the salinity was increased to 34 ppt (Day 50, SW, and Day 78, SW). Post-acclimatization to seawater, fish were divided into two groups for exposure to salmon lice (Lepeophtheirus salmonis) and infectious salmon anemia virus (ISAv), respectively. The fish group exposed to lice was sampled on 05.11.2020 (Day 94), and the group exposed to ISAv was sampled on 18.11.2020 (Day 107).

Fig. 1A. Experimental set-up (Atlantic salmon sampling dates).

Fig. 1B. Mucosal bivariate reference intervals: Comparison of group arithmetic means (represented by X in the figure) of skin mucous cell (MC) parameters—MC area (MCA, μm²) and MC defense activity (the abundance of MCs in the tissue)—in salmon (n = 150) relative to the salmon skin database of Quantidoc (N = 2158). The individual dots represent individual values from the database. The graph highlights the most common zone (38%, in green), a potential transition zone from homeostasis (38%-68%, in yellow), a transition from homeostasis area (68%-95%, in red), and a zone indicating homeostatic disturbance (in white) ²⁶ (Appendix 9).

Fig. 1 C, D, E, F, G, H: Regression between MC parameters (MCA (μm²), MCD (*100=%), MC defense activity (the abundance of MCs in the tissue) and N-Glycolylneuraminic acid (Neu5Gc) concentration (μg/g mucus) in salmon skin mucus. All variables were natural log–transformed (ln) prior to analysis.

Fig. C: Regression between MCA and total Neu5Gc (n=150).

Fig. D: Regression between MCA and free Neu5Gc (n=150).

Fig. E: Regression between MCD (*100=%) and total Neu5Gc (n=150).

Fig. F: Regression between MCD (*100=%) and free Neu5Gc (n=150).

Fig. G: Regression between MC defense activity and total Neu5Gc (n=150).

Fig. H: Regression between MC defense activity and free Neu5Gc (n=150).

Table 1: Total and free N-Glycolylneuraminic acid (Neu5Gc) concentrations in skin mucus in Atlantic salmon (n=150) and their regression with skin mucous cell (MC) parameters (MC area (MCA), the volumetric density of MCs in the epithelium (MCD) and the abundance of MCs in the tissue (defense activity) (n=150).

Day Treatment	Neu5Gc(µg/g mucus)	MCA (μm²) vs Neu5Gc (µg/g mucus)(p-value, R²adj †)	MCD (*100=%) vs Neu5Gc (µg/g mucus)(p-value, R²adj †)	Defense activity vs. Neu5Gc (µg/g mucus)(p-value, R²adj †)
Day 8:FW	Total: 468.0±70.2aFree: 34.0±8.84a	Total: p=0.029R²adj=0.129‡,Free: p= 0.352	Total: p=0.107 R²adj=0.058‡,Free: p=0.041R²adj=0.111‡	Total: p=0.737,Free: p=0.054R²adj = 0.095‡
Day 50:SW	Total: 89.3±19.8bFree: 2.24±0.336b	Total: p=0.441,Free: p=0.848	Total: p=0.003, R²adj=0.257‡,Free: p=0.059 R²adj=0.091‡	Total: p=0.001 R²adj=0.284‡,Free: p=0.031 R²adj=0.126‡
Day 78:SW	Total: 76.9±19.2bFree: 1.95±0.238b	Total: p=0.0002 R²adj=0.372,Free: p= 0.0007R²adj= 0.320‡	Total: p=5.13e-06, R²adj= 0.513‡,Free: p=9.57e-06 R²adj=0.492‡	Total: p=0.0004 R²adj=0.347‡,Free: p=0.0003R²adj= 0.355‡
Day 94:SW, lice	Total: 119.0±27.3bFree: 4.54±0.678c	Total: p=0.018R²adj=0.155§,Free: p=0.325	Total: p=0.858,Free: p=0.654	Total: p=0.145,Free: p=0.834
Day 107:SW, ISAv	Total: 110.0±48.8bFree: 2.15±0.377b	Total: p=0.0001 R²adj=0.389‡,Free: p=0.008R²adj=0.214‡	Total: p=4.31e-07R²adj=0.590‡,Free: p=0.0002 R²adj=0.403‡	Total: p=1.64e-06 R²adj=0.550‡,Free: p=0.0001 R²adj=0.420‡

The statistical differences (p<0.05) for Neu5Gc concentrations at different sampling points (Appendix 8) are indicated using Latin letters.

Acknowledgment

The authors thank ILAB personnel, including Linda Andersen and Steffen Blindheim, for conducting the fish trial. We would like to thank Professor Yanran Cao (NTNU, Norway) for her guidance on the mucus collection protocol. Aquatics AS financially supported this study, and QuantiDoc AS supported the preparation of this manuscript.

Ethical Considerations: The care and use of experimental animals complied with the European Union Guidelines (Directive 2010/63/EU) and the Norwegian Animal Welfare Act, as approved by the Norwegian Food Safety Authority under FOTS ID 24246. All key personnel involved in the trial held FELASA C certification (Experiment DIM No. 2868, ILAB).

Conflict of Interest Statement:

Dr. Grigory V. Merkin – Data Analyst (QuantiDoc AS, 5006 Bergen, Norway)

Anette Sauve - Head of Pharma Services at Vitas AS (0349 Oslo, Norway)

Mearge A. Okubamichael – Data Analyst (QuantiDoc AS, 5006 Bergen, Norway)

Trond-Inge Kvernevik – Aquatics AS Board member (6155 Ørsta, Norway)

Dr. Karin Pittman – Professor at the University of Bergen (5006 Bergen, Norway) and QuantiDoc AS Board member (5006 Bergen, Norway)

References

1. Salinas, I., Ding, Y., Fernandez-Montero, A. & Sunyer, J.O. Mucosal Immunity in Fish. In: Buchmann, K. & Secombes, C.J. (eds). Principles of Fish Immunology. Springer, 2022.
2. Guardiola, F.A., Cuesta, A. & Esteban, M.Á. Chapter 8 - Mucosal immunology in fish. In: Monzón, I.F. & Fernandes, J.M.O. (eds). Cellular and Molecular Approaches in Fish Biology. Academic Press, 2022, pp 251-284.
3. Díaz-Puertas, R., Adamek, M., Mallavia, R. & Falco, A. Fish Skin Mucus Extracts: An Underexplored Source of Antimicrobial Agents. Mar Drugs 21 (2023).
4. Reverter, M., Tapissier-Bontemps, N., Lecchini, D., Banaigs, B. & Sasal, P. Biological and Ecological Roles of External Fish Mucus: A Review. Fishes 3, 41 (2018).
5. Ordóñez-Grande, B., Guerreiro, P.M., Sanahuja, I., Fernández-Alacid, L. & Ibarz, A. Environmental Salinity Modifies Mucus Exudation and Energy Use in European Sea Bass Juveniles. Animals (Basel) 11 (2021).
6. O'Byrne-Ring, N., Dowling, K., Cotter, D., Whelan, K. & MacEvilly, U. Changes in mucus cell numbers in the epidermis of the Atlantic salmon at the onset of smoltification. Journal of Fish Biology 63, 1625-1630 (2003).
7. Merkin, G.V., Barlaup, B.T. & Pittman, K. Farmed Atlantic salmon (Salmo salar) skin mucosa during seawater transfer compared with wild juveniles. Aquaculture International 33, 637 (2025).
8. Mittal, A.K., Ueda, T., Fujimori, O. & Yamada, K. Histochemical analysis of glycoproteins in the unicellar glands in the epidermis of an Indian freshwater fish Mastacembelus pancalus (Hamilton). Histochem J 26, 666-677 (1994).
9. Münster-Kühnel, A.K. & Hinderlich, S. Sialic acid biosynthesis in vertebrates. Sialobiology: Structure, Biosynthesis and Function 1, 76-114 (2013).
10. Ghosh, S. Sialic acid and biology of life: An introduction. Sialic acids and sialoglycoconjugates in the biology of life, health and disease, 1 (2020).
11. Ma, X. et al. Sialylation in the gut: From mucosal protection to disease pathogenesis. Carbohydrate Polymers, 122471 (2024).
12. Burzyńska, P., Sobala, Ł.F., Mikołajczyk, K., Jodłowska, M. & Jaśkiewicz, E. Sialic acids as receptors for pathogens. Biomolecules 11, 831 (2021).
13. Jennings, M.P., Day, C.J. & Atack, J.M. How bacteria utilize sialic acid during interactions with the host: Snip, snatch, dispatch, match and attach. Microbiology 168 (2022).
14. Padra, J.T. et al. Aeromonas salmonicida growth in response to atlantic salmon mucins differs between epithelial sites, is governed by sialylated and N-acetylhexosamine-containing O-glycans, and is affected by Ca2+. Infection and immunity 85, 10.1128/iai.00189-00117 (2017).
15. Yao, Y. et al. Mucus sialylation determines intestinal host-commensal homeostasis. Cell 185, 1172-1188 e1128 (2022).
16. Gomez, D., Sunyer, J.O. & Salinas, I. The mucosal immune system of fish: The evolution of tolerating commensals while fighting pathogens. Fish Shellfish Immun 35, 1729-1739 (2013).
17. Jin, C. et al. Atlantic salmon carries a range of novel O-glycan structures differentially localized on skin and intestinal mucins. Journal of proteome research 14, 3239-3251 (2015).
18. Benktander, J., Sundh, H., Sharba, S., Teneberg, S. & Linden, S.K. Aeromonas salmonicida binds alpha2-6 linked sialic acid, which is absent among the glycosphingolipid repertoires from skin, gill, stomach, pyloric caecum, and intestine. Virulence 13, 1741-1751 (2022).
19. Benktander, J. et al. Stress Impairs Skin Barrier Function and Induces α2-3 Linked N-Acetylneuraminic Acid and Core 1 O-Glycans on Skin Mucins in Atlantic Salmon, Salmo salar. International Journal of Molecular Sciences 22, 1488 (2021).
20. Benktander, J. et al. Gill mucus and gill mucin O-glycosylation in healthy and amebic gill disease-affected Atlantic salmon. Microorganisms 8, 1871 (2020).
21. Venkatakrishnan, V. et al. Exploring the Arctic Charr Intestinal Glycome: Evidence of Increased N-Glycolylneuraminic Acid Levels and Changed Host–Pathogen Interactions in Response to Inflammation. Journal of Proteome Research 18, 1760-1773 (2019).
22. Pickering, A.D. The distribution of mucous cells in the epidermis of the brown trout Salmo trutta (L.) and the char Salvelinus alpinus (L.). Journal of Fish Biology 6, 111-118 (1974).
23. Pittman, K. et al. Body site matters: an evaluation and application of a novel histological methodology on the quantification of mucous cells in the skin of Atlantic salmon, Salmo salar L. J Fish Dis 36, 115-127 (2013).
24. Pittman, K. et al. Novel method for quantifying salmonid mucous cells. J Fish Dis 34, 931-936 (2011).
25. Merkin, G.V., Girons, A., Lazado, C.C. & Pittman, K. Mucosal responses in the skin and gill filaments of Atlantic salmon (Salmo salar) infected by Neoparamoeba perurans. Comparative Immunology Reports, 200181 (2024).
26. Merkin, G.V., Girons, A., Okubamichael, M.A. & Pittman, K. Mucosal epithelial homeostasis: Reference intervals for skin, gill lamellae and filament for Atlantic salmon and other fish species. J Fish Dis 48, e14023 (2025).
27. Merkin, G.V., Girons, A. & Pittman, K. Variations in mucous cell parameters in the gills and skin of farmed Atlantic salmon (Salmo salar) from freshwater flow-through and recirculating aquaculture systems compared to the gills of wild Atlantic salmon juveniles. Aquaculture International 33, 214 (2025).
28. Fanjara, E. et al. Sampling time for different matrices in stress assessment of farmed Atlantic salmon post-smolt. J Steroid Biochem Mol Biol 242, 106542 (2024).
29. Crawley, M.J. The R book. John Wiley & Sons 2012.
30. Roberts, S.D. & Powell, M.D. The viscosity and glycoprotein biochemistry of salmonid mucus varies with species, salinity and the presence of amoebic gill disease. J Comp Physiol B 175, 1-11 (2005).
31. Karlsen, C. et al. Atlantic salmon skin barrier functions gradually enhance after seawater transfer. Sci Rep 8, 9510 (2018).
32. van Muilekom, D.R. et al. Salinity change evokes stress and immune responses in Atlantic salmon with microalgae showing limited potential for dietary mitigation. Front Physiol 15, 1338858 (2024).
33. Provan, F. et al. Proteomic analysis of epidermal mucus from sea lice-infected Atlantic salmon, Salmo salar L. J Fish Dis 36, 311-321 (2013).
34. Fosse, J.H. et al. The infectious salmon anemia virus esterase prunes erythrocyte surfaces in infected Atlantic salmon and exposes terminal sialic acids to lectin recognition. Front Immunol 14, 1158077 (2023).
35. McDonald Nathan, D., Lubin, J.-B., Chowdhury, N. & Boyd, E.F. Host-Derived Sialic Acids Are an Important Nutrient Source Required for Optimal Bacterial Fitness In Vivo. mBio 7, 10.1128/mbio.02237-02215 (2016).

Supplementary Information

Appendix 1

The study, water parameters, light manipulation, sampling and the weight of the analysed Atlantic salmon (Salmo salar) in the trial

This study reports results, some of which were previously mentioned only briefly within a broader set of anonymized trials in an earlier publication ¹.

Smoltification had been induced using light manipulation, starting with a 12:12 h light regime followed by continuous light (24 h) from 18 August onward.

Table 1: Adjustment of water parameters in the trial.

Date/Parameter	Salinity of water (promille)	Temperature of water (°C)
04.08.2020	Fresh water	12
31.08.2020	Fresh water	14
14.09.2020	20	12
16.09.2020	34	14
20.10.2020	34	12

Table 2: The weight of Atlantic salmon sampled for skin MC parameters and skin mucus analysis (n=150).

Sampling date	Fish weight (x̅±2SD (g))
11.08.2020	43.9±6.04
22.09.2020	59.9±12.9
20.10.2020	93.4±48.5
05.11.2020	103.±33.4
18.11.2020	136.±65.0

Appendix 2

Fish exposure to salmon lice (Lepeophtheirus salmonis) and salmon anemia virus (ISAv)

Fish exposure to salmon lice (Lepeophtheirus salmonis) and salmon anemia virus (ISAv) took place at ILAB on October 21, 2020. In brief, 20 fish from each tank were weighed, tagged, and infected with salmon lice (30 copepodites per fish). The salmon lice infection followed the standard ILAB method, which involves draining the tank to half its volume, adding the lice, and then refilling the tank. The remaining fish (28-30 per tank) were infected with ISAv (Infectious Salmon Anemia Virus) at a dose of 2x10² TCID50/L in 150 liters of aerated water for 1 hour (TCID50 stands for Tissue Culture Infective Dose 50%). The virus was purchased from the Norwegian Veterinary Institute (Oslo).

Appendix 3

Mucus sampling and quantification of free and total Neu5Gc in mucus samples via HPLC-MS analysis

All mucus sampling was conducted in an ILAB laboratory facility (Bergen) at ambient room temperature under standardized laboratory conditions. For each sampling event, fish were handled individually, and skin mucus was collected immediately after removal from the water using a consistent and predefined protocol, minimizing the time between air exposure and mucus sampling. The same personnel performed all sampling using identical materials and procedures across all sampling dates. The water temperature varied modestly between sampling dates (12–14 °C, Table 1).

Skin mucus was gently scraped from the lateral surface of the fish using a cell scraper and transferred into 15 mL centrifuge tubes ². The head, fins, and ventral surface were avoided to minimize contamination from blood, scales, or other body fluids. Care was taken to apply minimal pressure to avoid disruption of the underlying epidermis. The skin samples for histological processing ^3,4 were taken from another fish side.

Aliquots of mucus were accurately weighed and precipitated with 5% trichloroacetic acid for measurement of free Neu5Gc or hydrolysed by adding aliquots of 0.1M sulphuric acid and heating at 80°C for 60 minutes for measurements of total Neu5Gc. Supernatants from both preparations were analysed using an Agilent 1260 HPLC (Agilent Technologies, Palo Alto, USA) coupled to an Agilent 6230 TOF high resolution mass spectrometer operated in negative mode scanning from 50-400 u using a Jetstream ESI ion source. Separation was performed as reversed phase liquid chromatography using a Supelcosil (Merck KGaA, Darmstadt, Germany) ABZ+ 150 mm x 4.6 mm x 3.0 μm column and quantification was performed against an external eight-point standard curve.

Appendix 4

Summary of a standardized protocol for specific body sites: Verification of mucosal barriers (mucosal mapping, now trademarked as Veribarr™) – Sampling of gills and skin (adopted from the previous publication ^3-6)

1. Complete the project form and use a pencil to pre-label histocassettes with fish numbers and tissue types. Apply a single continuous numbering sequence for all fish. Clearly mark containers (BiopSafe®, Mermaid Medical, Denmark) with the project title and date.
2. Ensure that the salmon either very freshly dead or terminally anesthetized, as waiting time causes tissue degradation. Avoid handling the sampling area to prevent wiping off mucous cells. Transport the fish carefully to the sampling table.
3. Wear protective gloves. Cut out the desired sample using a sharp scalpel or scissors. For the SKIN: With a scalpel, make a rectangle of about 1-2 cm x 1-2.5 cm just below the dorsal fin, aligning with the front edge of the fin. Gently lift one side and remove the skin piece and remove the skin rectangle along with a bit of the underlying muscle.
4. Place each sample into a separate histocassette, handling only the ends of the tissue. Then, place the histocassette in a Biopsafe container.

Appendix 5

Summary of histological processing for mucosal mapping (now trademarked as Veribarr™, adopted from the previous publication ^3-6)

Histological samples were first rinsed with phosphate-buffered saline (PBS) and then decalcified using a formic acid solution (250 mL formic acid, 19.8 g NaOH in 1 L water) from Riedel-de Haën, for 24 hours at room temperature. After decalcification and a further PBS rinse, the samples were dehydrated in ethanol and embedded in paraffin (Histowax; Leica). The paraffin-embedded tissues were then sectioned into tangential slices, running almost parallel to the non-planar epidermal stratifications. Sections were stained with Periodic Acid Schiff (PAS) and Alcian Blue, in accordance with the protocols by Pittman et al. (2011, 2013). The stained sections were then mounted onto microscope slides using a suitable mounting medium, and finally, the slides were scanned with a high-resolution digital slide scanner to create digital images.

Mucosal mapping technology, which was developed in academia ^3,4, was applied to examine MC dynamics in skin biopsies using tangential sections, approximately 90° from the perspective of traditional histology, to measure MC area (μm²), the volumetric density of MCs in the epithelium (in %), MC defense activity, defined here as the abundance of MCs within the tissue rather than a direct measure of immune function ^5-7. This technology is now trademarked as Veribarr™.

Appendix 6

Statistical Analysis

A linear mixed-effects model (LME) using the lme() function (nlme package, R) was fitted to examine the effects of sampling dates on mucous cell (MC) parameters. Sampling dates were used as a fixed effect, and Tank was included as a random effect to account for variability across tanks. Diagnostic plots were generated to check model assumptions: a residuals vs. fitted values plot assessed homoscedasticity and linearity, and a Q-Q plot checked the normality of residuals. Tukey's post-hoc test (emmeans package, R) was used to check for significant difference in MC parameters across the sampling dates. The same strategy was applied to examine the effects of sampling dates on free and total Neu5Gc concentrations at different sampling points.

To evaluate the relationship between MC parameters and the concentration of Neu5Gc (free and bound forms), a linear regression model (lm() function, R) was applied, with a log transformation used to normalize the dataset. All statistical analyses were conducted in R, following Crawley et al. (2012).

Appendix 7

Significant differences in skin MC parameters across the sampling dates

A linear mixed-effects model (LME) was fitted to examine the effects of sampling dates on mucous MC parameters. Tukey's post-hoc test (Appendix 6) was used to check for significant differences in skin MC parameters across the sampling dates:

The effects of sampling dates on MCA (mucous cell area (size)):

Comparisons (Contrast)	p-value
(11/08/2020) - (22/09/2020)	<0.0001
(22/09/2020) - (20/10/2020)	0.9940
(20/10/2020) - (05/11/2020)	0.0003
(20/10/2020) - (18/11/2020)	0.0005

The effects of sampling dates on MCD (volumetric density of MCs in the epithelium):

Comparisons (Contrast)	p-value
(11/08/2020) - (22/09/2020)	<0.0001
(22/09/2020) - (20/10/2020)	0.5252
(20/10/2020) - (05/11/2020)	0.4873
(20/10/2020) - (18/11/2020)	0.0001

The effects of sampling dates on Defense activity:

Comparisons (Contrast)	p-value
(11/08/2020) - (22/09/2020)	<0.0001
(22/09/2020) - (20/10/2020)	0.1747
(20/10/2020) - (05/11/2020)	0.9876
(20/10/2020) - (18/11/2020)	0.0006

Appendix 8

Significant differences in total and free Neu5Gc concentrations in the skin mucus across the sampling dates

Total Neu5Gc

Table 1: Total Neu5Gc concentrations in the skin mucus (n=150) across sampling dates.

Date	Concentration (μg/g mucus) Mean±SE
11/08/20	468.0±35.1
22/09/20	89.3±9.89
20/10/20	76.9±9.62
05/11/20	119.0±13.6
18/11/20	110.0±24.4

Figure 1: Natural log–transformed (ln) total N-Glycolylneuraminic acid (Neu5Gc) concentration (μg/g mucus) in skin mucus of Atlantic salmon (n=150) across sampling dates.

A linear mixed-effects model (LME) was fitted to examine the effects of sampling dates on log-transformed total N-Glycolylneuraminic acid (Neu5Gc) concentration. Tukey's post-hoc test (Appendix 6) was used to assess significant differences in the log-transformed total Neu5Gc concentration (μg/g mucus) in skin mucus across the sampling dates:

Comparisons (Contrast)	p-value
(11/08/2020) - (22/09/2020)	<0.0001
(22/09/2020) - (20/10/2020)	0.7619
(20/10/2020) - (05/11/2020)	0.0921
(20/10/2020) - (18/11/2020)	0.8059

Free Neu5Gc

Table 2: Free Neu5Gc concentrations in the skin mucus (n=150) across sampling dates.

Date	Concentration (μg/g mucus) Mean±SE
11/08/20	34.0±4.42
22/09/20	2.24±0.168
20/10/20	1.95±0.119
05/11/20	4.54±0.339
18/11/20	2.15±0.189

Figure 2: Natural log–transformed (ln) free N-Glycolylneuraminic acid (Neu5Gc) concentration (μg/g mucus) in skin mucus of Atlantic salmon (n=150) across sampling dates.

A linear mixed-effects model (LME) was fitted to examine the effects of sampling dates on the log-transformed concentration of free N-Glycolylneuraminic acid (Neu5Gc). Tukey's post-hoc test (Appendix 6) was used to assess significant differences in the log-transformed concentration of free Neu5Gc (μg/g mucus) in skin mucus across the sampling dates:

Comparisons (Contrast)	p-value
(11/08/2020) - (22/09/2020)	<0.0001
(22/09/2020) - (20/10/2020)	0.8629
(20/10/2020) - (05/11/2020)	<0.0001
(20/10/2020) - (18/11/2020)	0.9921

Appendix 9

Homeostasis, normality, reference values and mucosal epithelial health in aquaculture

A vast body of scientific literature exists on the definitions of health ⁸. Homeostasis, a cornerstone concept across various biological disciplines often used in discussions about health, signifies the organism's capacity to autonomously maintain a stable internal environment ⁹. Multiple physiological processes fit the notion of homeostasis, including the functioning of mucosal epithelia, which responds with changes in MC sizes and abundance to environmental stimuli. Epithelial tissue homeostasis is achieved by balancing cell divisions with cell growth, death, and extrusion (in intestinal epithelia) ¹⁰.

The observation that animals’ biological parameters typically fluctuate within a narrow range, with significant deviations signaling illness or potentially fatal conditions, has evolved into the concept of reference values ¹¹. These ranges of biological parameters serve as benchmarks for identifying healthy states in humans and animals despite ongoing debates about the theoretical basis of reference values. If there is a significant change in biological characteristics beyond its normal range, mechanisms to maintain homeostasis are activated.

In the medical field, pathological conditions are labelled as abnormal, while healthier states are considered normal. Clinical textbooks define “normal” for various health metrics as the statistical mean encompassed by a specific “normal variation” range. Medical authors often emphasize that the so-called normal values do not represent the average individual but rather the average individual in good health ^{8, 12}.

Reference values, describing the natural fluctuations observed in healthy populations, have gained universal acceptance as indispensable tools in both medicine and veterinary science, significantly enhancing the clinical decision-making process, where typically about 120 “reference individuals” can be used to generate a reference interval ¹¹. A reference sample group consists of an adequate number of measurements selected to represent a well-defined reference population. Reference limits, derived from this group, pinpoint values within the distribution of reference value (a reference distribution) and serve crucial descriptive purposes. The Reference Interval (RI) is the span between two reference limits and typically encompasses the central 95% of the distribution, thus excluding the extreme 5% of measurements from reference individuals ¹¹. The values that reflect large, sudden, or long-term deviations from the homeostatic range (beyond 95% of the distribution) that body mechanisms cannot correct naturally are termed homeostatic ‘imbalance’ ¹².

This initiative serves as a reference tool to safeguard barrier health in farmed fish ⁵.

Appendix 10

Discussion of methodological limitations related to mucus collection and changes in skin mucus chemistry during seawater transfer (SWT)

The effect of the water temperature

Although water temperature varied modestly between sampling dates (12–14 °C, Table 1), these differences were small and are unlikely to have caused systematic dehydration effects during the short sampling window. Sampling procedures were consistent across individuals within each sampling event; therefore, any minor effects of temperature or handling time would apply equally within groups and are unlikely to explain the observed between-group differences or interindividual variation.

The effect of sampling sites

Mucous cell abundance, as well as mucus composition, varies depending on the location on the fish surface. Although the sampling sites for both the skin tissue samples (see sampling protocol) and the skin mucus were well defined and consistently applied across all sampling events, the area used for mucus collection was larger than that used for skin tissue sampling. Consequently, the analyzed mucus may have represented a broader epithelial region than the histologically evaluated tissue samples. Theoretically, this spatial difference may have influenced the strength of the correlations between Neu5Gc concentration and mucous cell (MC) parameters compared to sampling from identical regions. Nevertheless, the standardized sampling procedure ensured consistency across fish and sampling dates, supporting the reliability of the observed relationships.

The effects of a method for harvesting mucus

There are several methods for harvesting skin mucus, which can broadly be classified into scraping, absorption, and bagging ¹³. The scraping method employs different instruments, including cell scrapers, spatulas, or glass slides, to collect mucus from the surface of the fish ¹³. This is the most commonly used method for collecting fish skin mucus samples ^{2, 13}. The sampling procedure is relatively simple and efficient. However, the scraping process exerts slight pressure on the skin surface, which may stimulate additional mucus secretion and cause dislodgement of scales ¹³.

In the present project, sampling was performed by trained personnel under specialist supervision using a standardized protocol ². Although minor contamination of mucus samples with fragments of mucous cells (MCs) cannot be completely excluded, we consider the potential impact of such contamination on the measured Neu5Gc concentrations to be negligible.

Appendix 11

Additional information

Neu5Gc synthesis and CMAH activity in fish: regulation and salinity effects (brief comment)

The enzyme CMP-N-acetylneuraminic acid hydroxylase (CMAH) is responsible for the synthesis of Neu5Gc. The CMAH gene is thought to be present in most vertebrates, including Atlantic salmon, but has been inactivated in several lineages, including humans and some fish species such as Atlantic cod (Gadus morhua) ¹⁴.

Salinity, via smoltification, may plausibly influence the Neu5Gc pathway indirectly by altering mucous cell dynamics and overall mucin biosynthesis and glycosylation; however, direct evidence that CMAH itself is regulated by salinity in salmon skin remains limited.

A short discussion linking the observed Neu5Gc dynamics to known changes in skin microbiota during seawater transfer.

The observed changes in Neu5Gc levels during and after seawater transfer may also be interpreted in the context of known shifts in the skin-associated microbiota during seawater transfer. The transition from freshwater to seawater is accompanied by pronounced restructuring of the salmon skin microbiome, likely driven by changes in salinity, ionic composition, and environmental microbial exposure ¹⁵. In vertebrates, sialylated mucins are known to influence microbial adhesion, colonization, and nutrient availability at mucosal surfaces ¹⁶. Consequently, alterations in mucous cell capacity to biosynthesis of sialic acids and hence Neu5Gc abundance in mucus may contribute to the reorganization of microbial communities on the skin during seawater transfer, reflecting adaptive modulation of the mucosal interface in response to changing microbial pressures.

A short justification explaining why Neu5Gc was prioritized over Neu5Ac

The three most well-characterized sialic acids in fish mucus are Neu5Ac, Neu5Gc, and Kdn. Although Neu5Gc is less abundant than Neu5Ac (18% vs. 58%) in the Atlantic salmon skin mucus, the skin mucins have a high amount of terminal Neu5Gc ¹⁷. The presence of skin Neu5Gc as terminal constituents in the mucins ¹⁷ highlights its potential role in mediating interactions at the tissue–environment interface, including communication between the host and the skin-associated microbiota. Terminal sialic acids on mucins are known to influence microbial adhesion, colonization, and nutrient utilization, suggesting that Neu5Gc may contribute to shaping microbial community structure on the salmon skin surface.

Given that Neu5Gc potentially serves as an important interaction point with various bacteria, future studies on this Sia should include fish challenges with bacterial infections and measures of mucosal protection in the vital skin barrier.

Appendix 12

Statement of significance

To our knowledge, this study is the first to establish a correlation between mucous cell (MC) parameters in the skin and the concentration of N-Glycolylneuraminic acid (Neu5Gc) in the skin mucus of Atlantic salmon (Salmo salar). The study suggests that MCs are the primary source of Neu5Gc. The observed changes in both MC parameters and Neu5Gc concentration during transitions from freshwater to seawater provide insights into how fish adapt their mucosal defenses to environmental challenges. These results are important for future research on mucosal immunity and have applications in aquaculture by potentially improving the management of salmon mucosal health during smoltification.

References

1. Merkin, G.V., Barlaup, B.T. & Pittman, K. Farmed Atlantic salmon (Salmo salar) skin mucosa during seawater transfer compared with wild juveniles. Aquaculture International 33, 637 (2025). https://doi.org/10.1007/s10499-025-02268-2
2. Fanjara, E. et al. Sampling time for different matrices in stress assessment of farmed Atlantic salmon post-smolt. J Steroid Biochem Mol Biol 242, 106542 (2024). https://doi.org/10.1016/j.jsbmb.2024.106542
3. Pittman, K. et al. Novel method for quantifying salmonid mucous cells. J Fish Dis 34, 931-936 (2011). https://doi.org/10.1111/j.1365-2761.2011.01308.x
4. Pittman, K. et al. Body site matters: an evaluation and application of a novel histological methodology on the quantification of mucous cells in the skin of Atlantic salmon, Salmo salar L. J Fish Dis 36, 115-127 (2013). https://doi.org/10.1111/jfd.12002
5. Merkin, G.V., Girons, A., Okubamichael, M.A. & Pittman, K. Mucosal epithelial homeostasis: Reference intervals for skin, gill lamellae and filament for Atlantic salmon and other fish species. J Fish Dis 48, e14023 (2025). https://doi.org/10.1111/jfd.14023
6. Merkin, G.V., Girons, A., Lazado, C.C. & Pittman, K. Mucosal responses in the skin and gill filaments of Atlantic salmon (Salmo salar) infected by Neoparamoeba perurans. Comparative Immunology Reports, 200181 (2024). https://doi.org/10.1016/j.cirep.2024.200181
7. Merkin, G.V., Girons, A. & Pittman, K. Variations in mucous cell parameters in the gills and skin of farmed Atlantic salmon (Salmo salar) from freshwater flow-through and recirculating aquaculture systems compared to the gills of wild Atlantic salmon juveniles. Aquaculture International 33, 214 (2025). https://doi.org/10.1007/s10499-025-01894-0
8. Boorse, C. Health as a theoretical concept. Philosophy of science 44, 542-573 (1977).
9. Cannon, W.B. Organization for physiological homeostasis. Physiological reviews 9, 399-431 (1929). https://doi.org/10.1152/physrev.1929.9.3.399
10. Guillot, C. & Lecuit, T. Mechanics of epithelial tissue homeostasis and morphogenesis. Science 340, 1185-1189 (2013). https://doi.org/10.1126/science.123524
11. Geffre, A. et al. Reference values: a review. Veterinary clinical pathology 38, 288-298 (2009). https://doi.org/10.1111/j.1939-165X.2009.00179.x
12. Clancy, J. & McVicar, A. Physiology and anatomy for nurses and healthcare practitioners: A homeostatic approach. CRC Press, 2017.
13. Lee, K.M. et al. Comparisons of Methods for Mucus Sampling and Mucin Semi-Quantification on Barramundi (Lates calcarifer) and Atlantic Salmon (Salmo salar) Epithelial Sites. Mar Biotechnol (NY) 27, 140 (2025). https://doi.org/10.1007/s10126-025-10515-z
14. Peri, S., Kulkarni, A., Feyertag, F., Berninsone, P.M. & Alvarez-Ponce, D. Phylogenetic Distribution of CMP-Neu5Ac Hydroxylase (CMAH), the Enzyme Synthetizing the Proinflammatory Human Xenoantigen Neu5Gc. Genome Biol Evol 10, 207-219 (2018). https://doi.org/10.1093/gbe/evx251
15. Lokesh, J. & Kiron, V. Transition from freshwater to seawater reshapes the skin-associated microbiota of Atlantic salmon. Sci Rep 6, 19707 (2016). https://doi.org/10.1038/srep19707
16. Linden, S.K., Sutton, P., Karlsson, N.G., Korolik, V. & McGuckin, M.A. Mucins in the mucosal barrier to infection. Mucosal Immunol 1, 183-197 (2008). https://doi.org/10.1038/mi.2008.5
17. Jin, C. et al. Atlantic salmon carries a range of novel O-glycan structures differentially localized on skin and intestinal mucins. Journal of proteome research 14, 3239-3251 (2015). https://doi.org/10.1021/acs.jproteome.5b00232