Discussion
This multinational study indicates that, based on fish PV- and extract-sIgE detection for 263 fish-allergic patients, up to 90% may tolerate cartilaginous fish. In addition, up to 21% of fish-allergic patients may tolerate some bony fish species. Multiplex in vitroIgE quantification, using purified allergens and extracts from various fish species, may be utilized to identify patient candidates and fish species for FCs, with the ultimate goal to confirm the clinical tolerance of some species.
Fish allergens from different species share limited similarities. While some IgE epitopes of parvalbumins are highly conserved, species-specific epitopes have also been identified.18,19 Additionally, not all species have equal allergenic potential.20 We previously showed the low reactivity to cartilaginous fish among bony-fish allergic patients due to the evolutionary distance between beta- and alpha-PVs from bony and cartilaginous fish, respectively.4 In addition, a significant proportion of fish-allergic patients may tolerate certain bony fish.5 Besides PVs, other fish allergens have been identified (e.g. tropomyosin, lactate dehydrogenase, glucose-6-phosphate isomerase, creatine kinase) but their cross-reactivity potential is unknown.21,22
In addition to species and allergen diversity, eating habits differ across the world. Fish is subject to different cooking methods, which may impact allergen stability and IgE reactivity.23,24Beta-PVs demonstrated higher heat stability than enolase and aldolase.3 The variety of species, allergens and processing methods present a challenge for diagnosis. Certain region-specific important species may be absent from common diagnostic tests, and routine approaches often focus on limited number of species.25 High-risk patients with fish allergy are recommended strict avoidance of all fish, which is often unnecessary.25 Next-generation diagnostic approaches are hence necessary and multiplex molecular allergy diagnosis emerges as a promising tool.26,27
A research version of the ALEX2 multiplex assay (Macro Array Diagnostics), with values of sIgE strongly correlated to those obtained by ImmunoCAP-ISAC (Thermo Fischer)28 was used here to quantify total, fish extract- and PV-sIgE in fish-allergic patients’ sera. The ALEX2 platform was previously successfully utilized in quantifying serum IgE of patients with atopic dermatitis, dust mite allergy and nut allergy.29-31 We first investigated the IgE levels to PVs from 10 fish species. The species were selected based on published phylogenetic analyses, covering species relevant for consumption, and those responsible for monosensitizations.3,32 The median sIgE values were the highest for tuna and mackerel PVs (>10 kU/L), followed by herring, carp, salmon and swordfish (> 8kU/L), ocean perch (6 kU/L), cod (5 kU/L), sole (2.5 kU/L) and ray (<0.3 kU/L) (Figure 1A ). High IgE levels to tuna and swordfish PVs were unexpected, considering previous reports on their lower allergenicity due to low PV abundance in their dark muscles.33,34 Our data indicate the presence of shared IgE epitopes with other beta-PVs. However, the clinical relevance of these findings, and their clinical cross-reactivity require future investigations. Significantly lower IgE for cod PV than for several other beta PVs demonstrated that cod PV, a commonly used diagnostic marker, may not be sufficient for accurate fish allergy diagnosis. In this study, 22% of the patients were sensitized to ray PV (Figure 1B ), showing that tolerance of cartilaginous fish may be a possibility for many but not all bony-fish sensitized individuals.
Fish consumption differs across geographic regions in terms of quantity and processing.8 According to the Food and Agriculture Organization, China was the top fish consumer in 2017 worldwide based on total supply. Among the countries from our study, the highest per capita fish consumption was estimated for Norway (51.4 kg) followed by Spain (42.5 kg) and China (38.8 kg).35 IgE to PVs from different species may reflect exposure and indicate the species most relevant for diagnosis. Interestingly, we observed the highest IgE to all PVs for cohorts from Norway, China and Spain, and each tested country displayed weaker IgE to ray and sole PVs than to others (Figure 1D ). However, IgE levels to most commonly used PVs in diagnostics differed depending on the country. Cod PV-sIgE was overall low and the highest for patients from Spain and Norway, while tuna PV-IgE was high (in comparison to other PVs) in all cohorts. Currently we cannot reach conclusions about the clinical relevance of IgE binding to tuna PV.34 Nevertheless, tuna PV appears as highly cross-reactive protein to other beta PVs and its potential use in diagnostic assays should be investigated in the future. Although the country-specific distribution of IgE to different PVs may in part be a result not only from IgE specific to single PVs but also cross-reactive IgE, our data indicate important PVs for fish-allergy diagnosis for each country.
The diversity of PV-sIgE levels was analyzed next (Figure 2 ). Shannon’s diversity did not depend on patient’s symptom severity, age or gender, suggesting that IgE profiles to different PVs are largely independent of these factors, and each patient should be subjected to a wide-range diagnosis with multiple fish species.
In our study, up to 41% of patients were negative to some of the bony fish PVs and up to 97% to ray PV (Figure 3A ). Previous studies indicated that up to 30% of patients may tolerate some fish species.3 This variable reactivity to different PVs emphasized the importance of including allergens from diverse fish families into diagnostic approaches. Furthermore, based on probability calculations (Figure 3B ), we demonstrated that tuna PV may be utilized to predict whether patients may be negative to other PVs when negative to this protein.
Additional allergens such as enolase, aldolase or collagen may be important for some patients. Our previous study on 101 fish-allergic patients showed sensitization to fish collagen by 21%.36 Another study reported the varying prevalence of sensitization to enolase and aldolase depending on fish species, being as high as 70% for cod enolase.5 Some allergens lost their IgE-binding capacity upon heat treatment.37In contrast, HEs may contain gelatin, still able to bind IgE.38,39 In this study, lower IgE levels to extracts were observed compared to PVs (Figure 4 ), indicating that for patients with predominant sensitization to PV, the dilution factor of the parvalbumin in whole extracts coated on allergen-detection systems may yield a negative test result. In contrast, to determine the sensitization to other known and yet unknown allergens, whole extracts may be required. This was emphasized by negative result for PVs but positive results for extracts for some patients. Additionally, for most of the species, correlation between IgE levels to PVs and extracts was weak, indicating that all three components (PV, RE and HE) are required for diagnosis.
Finally, we investigated patients with no IgE to PV, RE and HE from the same fish, as potential candidates for the tolerance-confirming FCs. As many as 90% of the patients were detected as candidates for future FCs with ray and up to 21% with specific bony fish species (Figure 5A ). Negativity to ray was confirmed for 6 of 8 tested Chinese patients by PPT. For the PPT, the Dipturus innominatus ray, available in China, was used. Although Dipturus innominatus and Raja clavata (used in the ALEX2 assay) belong to the same fish family (Rajidae)40 and high IgE cross-reactivity is expected, the possibility of unshared epitopes on some allergens cannot be excluded, possibly explaining the 2 positive PPT outcomes. Larger studies are required to confirm negative in vitro IgE data using functional assays or FC, and to calculate the predictive values of the multiplex in vitro diagnosis. When IgE quantification for many species cannot be performed, our calculations of probability to be negative to some fish if known to be negative to another (Figure 5B ) may be used as a guideline for selecting the panel for in vitro testing. A recently published prospective clinical study showed development of fish tolerance with age.41 Our data may therefore also be used in the context of determining the species for testing tolerance development over time. A limitation of our study is the absence of FCs for tolerance confirmations. Schulkes et al. indicated a frequent serological, but limited clinical cross-reactivity between fish species.42 Although a further investigation of the clinical significance of our data is required, we provide a clear direction for future studies which should carefully choose the species for specific patients to explore tolerance.
Together, the presented data demonstrate the need for fish extracts in both raw and heated form, and the PVs from several evolutionary distant species for next-generation fish allergy diagnosis, which will enable identification of potentially safe-to-consume species for individual patients. Ultimately, combining the knowledge about the important species and allergens with novel bioinformatic approaches will permit the design of region-specific diagnostic arrays, which will significantly improve safety and wellbeing of fish-allergic individuals.