Discussion
In this study, we delineated the allergen profiles of salmon and grass carp using sera from two Asian populations. With the purified allergens from both fish species, our study illustrates parvalbumin and collagen as important biomarkers for fish allergy diagnosis. Our results also indicate a wider allergen spectrum including sensitization to heat labile allergens, as well as stronger sensitization to salmon than grass carp among the Japanese fish allergic subjects where grass carp is unavailable. Such differences highlights that the relevant fish species as diagnostic marker and allergen sensitization are distinct in different populations due to diverse dietary exposures and habits.
Ruethers et al. reported recently 10 allergens from catfish and 6 allergens from salmon13. Specifically, tropomyosin and TPI identified as fish allergens in their study are known shrimp allergens23. Despite positive shrimp-sIgE by ImmunoCAP in six HK subjects, none showed reaction to fish tropomyosin or TPI in our western blot. Our subjects also showed no detectable IgE binding to allergens including creatine kinase, pyruvate kinase, L-lactate DH and G-6-PI that were previously reported as catfish allergens. This might partly indicate that these are only catfish-specific allergens and/or the reactions are population-specific. Apart from the well-defined fish allergens (i.e. parvalbumin, enolase, aldolase and collagen), we consistently identified GAPDH as a putative fish allergen of both salmon and grass carp12. Such findings supported GAPDH to be an important fish allergen.
Our present report further focuses on studying the more well-defined fish allergens in terms of their sensitization rates. In agreement with most studies that parvalbumin accounted for 70-95% of allergic reactions9,24, parvalbumin is the major allergen in grass carp and salmon with 74% sensitization rate determined by western blotting and ELISA. Collagen is the second commonest fish allergen with 37.5% of subjects being sensitized. The reported rates of IgE binding to collagen vary in different populations. For instance, Kalic et al. analyzed samples from 101 fish allergic patients in Australia and showed that 21% of subjects displayed IgE binding to collagen from tuna, barramundi and/or salmon14. Kuehn et al. reported 19.3% sensitization rate to gelatin of cod, salmon and/or tuna in 62 fish allergic patients16. With only 21 patients, Shimojo et al. reported a higher rate of 52.4% sensitization to collagen than to parvalbumin25 while Kobayashi et al. reported also 50% of sensitization rate to mackerel collagen26. IgE binding to collagen was also more frequent in patients with more severe fish allergic symptoms than those with mild-moderate responses while independent reports have shown anaphylaxis provoked by gelatin/collagen15,27. On the other hand, Kalic et al. reported that 8/21 of collagen-IgE positive patients showed IgE binding to collagen only but not parvalbumin14. Comparatively, seven of our 37 collagen-IgE positive subjects demonstrated moderate IgE binding to collagen but absence of parvalbumin-specific IgE. Taken together, collagen is a prevalent allergen in Asian populations. Fish allergic patients can have IgE binding to collagen only, supporting that testing for collagen should be included in the diagnostic workup for fish allergy.
To our knowledge, only one study with 62 fish allergic subjects reported the sensitization rates of 62.9% and 50.0% to enolase and aldolase, respectively16. Another smaller study by Ruethers et al. showed that half of their 16 Australian patients showed weak to moderate IgE reactivity to enolase in commercial tuna and salmon SPT extracts28. In our cohort, IgE binding to aldolase was detected in 38.5% of subjects while that to enolase was present in 17.8% of subjects. We did not anticipate a major difference in the reactivity of natural and recombinant aldolase and enolase as the reason for lower sensitization as the recombinant enzymes were confirmed to be reactive consistently by ELISA (data not shown) and western blots. Such discrepancy might, on the other hand, be explained by differences in the ethnicity and age of the studied populations: this study only recruited children below 18 years old while many subjects studied by Kuehn et al. were young adults (mean age 28.7, range 13-62 years). Among seven subjects without parvalbumin-specific IgE in our cohort, only one Japanese subject showed IgE binding to salmon and grass carp aldolase. Testing of enolase and aldolase thus do not substantially improve the diagnostic yield for fish allergy in Asians while detecting IgE against parvalbumin and/or collagen might be sufficient to identify patients with fish allergy.
This study found minimal binding of sera from HK subjects to allergens other than parvalbumin by western blotting. A possible explanation was the use of different serum samples for ELISA and western blot, different sensitivity of the assays, intrinsically low abundance of these allergens in fish extracts and the insoluble nature of collagen in aqueous solution29. Nevertheless, both ELISA and western blot assays robustly indicated a wider spectrum of allergen sensitization in Japanese than HK subjects. Japanese commonly consume raw fishes, but this is a less likely reason for this observation given the age-range of our recruited patients; the consumption of raw fish might not be a major attribute to the sensitization to a wider spectrum of allergen including the heat labile allergens. Our present study showed, for the first time, differences in fish protein and allergen profiles in relation to cooking methods at the same core temperature. Despite higher cooking temperature with frying/baking compared with steaming/boiling, frying/baking retained more heat liable allergens in salmon meat. Steaming and boiling to make fish congee and soup are the preferred methods of fish cooking in Hong Kong while baking and frying of fish are more popular in Japan. For instance, 73.8% of the interviewed Japanese workers consumed fish by broiling (baking) comparing to only 14.1% for raw fish and 10% by stewing30. Although we are aware that our present study is limited by the lack of direct investigation into the dietary habits of our studied population, our results support the importance of dietary exposures, consumption and cooking preferences in determining fish allergen sensitization in particular when the two populations did not differ in terms of symptom severity. This factor should be considered during the selection of allergens for allergy testing and the interpretation of CRD results.
In the context of fish exposure, Japanese patients should not have eaten grass carp due to the unavailability of this freshwater fish. Thus, IgE binding to grass carp allergens in Japanese patients should be a result of cross-reactivity that can be partly reflected from the stronger positive correlation between grass carp and salmon sIgE levels than HK subjects whom could be independently sensitized to the two fish species. However, we observed that more Japanese patients reacted to grass carp parvalbumin with higher OD values in ELISA, although salmon-sIgE level was higher than grass carp-sIgE and dietary exposure to grass carp is lacking in this population. The higher salmon-sIgE level might thus be explained by IgE binding to other salmon allergens, particularly collagen, when considering the lower parvalbumin content in crude salmon extract, lower IgE affinity of salmon parvalbumin in ELISA, and the more diverse salmon allergen binding pattern in western blots. On the other hand, we also showed that one-fifth of subjects reacted only to grass carp parvalbumin, which indicated that grass carp parvalbumin has higher allergenicity than salmon parvalbumin. However, we are yet to know the clinical reactivity to grass carp among the Japanese subjects and so the question of improving assay sensitivity and specificity using fish allergen (parvalbumin) with higher allergenicity remains open. The varying allergenic potential of fish parvalbumins could be explained by the specific sensitization of patients from their clinical history8. With the lack of grass carp in the local Japanese diet, our results suggested the presence of higher diversity, distinct and/or species-specific IgE-binding epitopes for grass carp parvalbumin31. It is of note that IgE sensitization based on parvalbumin did not correlate well with patients’ clinical allergy status18, and we postulate that the identification of parvalbumin epitopes may help to resolve the inconsistency in cross-reactivity of different fish species.
In summary, this study extends our understanding on the allergen profile of two important fish species that parvalbumin and collagen are important components for fish allergy diagnosis. We also illustrated that the fish species extracts relevant for diagnosis and allergen sensitization pattern differ among populations. Overall, this study provides additional information for the implementation of relevant extracts and components for precise fish allergy diagnosis.