Allergen cross reactions: a problem greater than ever thought?


  • Edited by: Reto Crameri

Pascal Pfiffner, University Institute of Immunology, Sahli Haus 2, Inselspital, 3010 Bern, Switzerland.
Tel.: +41 31 632 22 89
Fax: +41 31 632 35 00


To cite this article: Pfiffner P, Truffer R, Matsson P, Rasi C, Mari A, Stadler BM. Allergen cross reactions: a problem greater than ever thought? Allergy 2010; 65: 1536–1544.


Background:  Cross reactions are an often observed phenomenon in patients with allergy. Sensitization against some allergens may cause reactions against other seemingly unrelated allergens. Today, cross reactions are being investigated on a per-case basis, analyzing blood serum specific IgE (sIgE) levels and clinical features of patients suffering from cross reactions. In this study, we evaluated the level of sIgE compared to patients’ total IgE assuming epitope specificity is a consequence of sequence similarity.

Methods:  Our objective was to evaluate our recently published model of molecular sequence similarities underlying cross reactivity using serum-derived data from IgE determinations of standard laboratory tests.

We calculated the probabilities of protein cross reactivity based on conserved sequence motifs and compared these in silico predictions to a database consisting of 5362 sera with sIgE determinations.

Results:  Cumulating sIgE values of a patient resulted in a median of 25–30% total IgE. Comparing motif cross reactivity predictions to sIgE levels showed that on average three times fewer motifs than extracts were recognized in a given serum (correlation coefficient: 0.967). Extracts belonging to the same motif group co-reacted in a high percentage of sera (up to 80% for some motifs).

Conclusions:  Cumulated sIgE levels are exaggerated because of a high level of observed cross reactions. Thus, not only bioinformatic prediction of allergenic motifs, but also serological routine testing of allergic patients implies that the immune system may recognize only a small number of allergenic structures.

Determination of specific IgE in patient sera is a valuable test for allergologists (1). The number of potential allergens is steadily increasing (2) and suppliers of allergy tests are providing ever longer lists of allergenic preparations to be used for in vitro assays. In most instances, allergens are still relatively crude extracts of organisms or parts thereof (3). Recently, allergen diagnosis has improved by the use of highly purified natural or recombinant allergens and protein microarrays (3–8). This may improve allergy diagnostics in the future.

Cross reactions are allergic reactions against other allergens without prior sensitization. They have been extensively studied and a handful of well-defined cross reactivity syndromes are clinically highly important, e.g., the pollen-food syndromes (9). Cross reactions between recombinant allergens are also documented (3, 7, 10). Thus, the immune system might recognize common structures, allowing to predict allergic reactions that have not been tested physically but were derived by similarity.

We have previously shown that a bioinformatic approach is capable to define a much lower number of potentially allergenic structures, termed motifs, than the number of known protein sequences of allergens (11). These motifs represent a scaled profile over a window of 50 amino acids, derived from all currently known allergen protein sequences. They serve as an identifier for evolutionary conserved protein domains. Consequently, if protein sequences match a given motif, these proteins are predicted to fold into the same protein domain and therefore exhibit similar surface structures. We showed that this method of cross reactivity prediction is superior to the FAO/WHO rule, which states that a protein is allergenic if it has either an identity of at least six continuous amino acids or more than 35% sequence similarity over a window of 80 amino acid residues. Especially in view of false positive matches (67.3% of all Swiss-Prot proteins were predicted to be allergenic by the FAO/WHO rule), the motif-based approach performed much better (2.6% predicted to be allergenic) (11).

Thus, the question remains whether the in silico prediction of allergenicity may be confirmed by wet lab data. For this purpose, we have analyzed 5362 sera corresponding to 203 283 specific IgE determinations. We could demonstrate that the degree of cross reaction was greater than ever thought.

Materials and methods

Serum samples

Data on 5456 serum samples were obtained by testing for IgE using Phadia’s ImmunoCAP® (former UniCAP®, Phadia AB, Uppsala, Sweden) systems. These are sandwich immunoassay systems where serum IgE antibodies react with anti-IgE covalently coupled to the system in case of total IgE determination or with solid-phase bound allergen extracts to determine specific IgE. Bound antibodies are detected and quantified using enzyme-labeled anti-IgE-antibodies and fluorescence detection.

Tests were performed in the years from 1988 to 2006 in 17 different countries in different laboratories. Raw, anonymized IgE data (no age, sex, and other demographic and clinical information) were collected as quality assurance; therefore, no selection criteria were applied. Test results were collected in a clinical setting; most sera are presumably from patients with atopy.

All IgE levels are expressed in kilo units of antigen per liter serum (kUA/l). Specific IgE levels >0.35 kUA/l (Class I and higher) were regarded as a positive test result, levels >100 kUA/l were capped at 100 kUA/l, which affected 1578 values.

Included in the database were serum levels for 99 allergens as well as the total IgE level. According to the manufacturer, the 99 allergen extracts used to determine the specific IgE values are the 99 most tested allergens among a list of more than 700 allergens available in Phadia’s catalog. Table 1 lists the extracts and groups them into major subsets.

Table 1.   The allergen extracts used to determine specific IgE levels. Identifier for extracts is Allergome’s accession number and Phadia’s product code. The number of known proteins (excluding isoforms, within brackets the total number of proteins including isoforms) is estimated based on available sequence data. The number of distinct motifs present within the extract is derived from contained proteins
GroupExtractAllergome accession numberPhadia product code# known proteins in extract# motifs associated with extract
EpidermalsCat epithelium and dander1819e16 (8)3
Dog dander1756e55 (7)4
Guinea pig epithelium1765e62 (2)0
Horse dander1813e35 (6)3
Mouse epithelium; serum proteins and urine proteins1881e882 (2)2
Rat epithelium; serum and urine proteins1957e872 (2)2
Foods of animal originBeef2019f275 (5)3
Blue mussel1413f371 (1)1
Egg white1832f16 (8)4
Fish (cod)1831f31 (2)1
Milk1747f27 (11)6
Scallop2012f3383 (3)2
Foods of plant originAlmond1948f204 (5)2
Apple1871f494 (71)4
Banana1882f921 (1)1
Barley2040f67 (13)4
Brazil nut1738f182 (2)2
Buckwheat1816f114 (9)3
Carrot1799f312 (8)2
Celery1721f854 (5)2
Cherry1946f2424 (7)3
Coconut3559f361 (1)0
Garlic1706f471 (2)1
Hazel nut2028f177 (17)5
Kiwi1697f848 (21)3
Maize; Corn2092f84 (5)3
Onion1704f481 (1)1
Orange1774f333 (6)2
Pea1931f122 (4)1
Peach1949f953 (6)2
Peanut1723f1313 (36)9
Potato1977f356 (19)4
Rice2058f92 (2)1
Rye2076f51 (2)0
Sesame seed1971f107 (7)3
Soya bean1834f1413 (32)7
Strawberry2251f443 (14)3
Tomato1870f255 (12)5
Wheat1993f413 (29)8
White bean1923f1500
Yeast1960f452 (2)1
Grass pollensBermuda grass1796g29 (22)6
Cocksfoot1798g35 (16)2
Common reed1927g700
Johnson grass1979g101 (1)0
Sweet vernal grass1718g11 (2)0
Timothy1924g610 (36)8
InsectsCockroach; German1742i611 (115)10
MicroorganismsAlternaria alternata (tenuis)1708m612 (16)10
Aspergillus fumigatus1730m324 (27)14
Aureobasidium pullulans2197m1200
Botrytis cinerea1630m700
Candida albicans1757m53 (3)2
Cladosporium herbarum (Hormodendrum)1775m212 (13)7
Epicoccum purpurascens1810m142 (2)1
Fusarium moniliforme1554m900
Helminthosporium halodes1125m800
Mucor racemosus2291m400
Penicillium notatum1912m15 (8)2
Phoma betae2303m1300
Rhizopus nigricans1622m1100
Stemphylium botryosum2637m101 (1)1
MitesDermatophagoides pteronyssinus1803d116 (32)12
Dermatophagoides farinae1801d220 (72)14
Tree pollensAmerican beech2249t500
Common silver birch1741t36 (67)6
Japanese cedar1784t175 (31)6
Maple leaf sycamore; London plane1932t113 (4)3
Mountain juniper1851t63 (4)3
Oak1955t71 (5)1
Olive1888t911 (97)6
White ash2253t1500
White pine2312t1600
VenomsCommon wasp (Yellow jacket)2008i34 (5)4
Honey bee1722i19 (13)8
Weed pollensCocklebur2401w1300
Common ragweed1710w19 (17)5
Dandelion6146w81 (1)1
Goosefoot; Lamb’s quarters1768w103 (3)3
Marguerite; Ox-eye daisy1567w700
Mugwort1728w66 (25)7
Plantain (English); Ribwort1933w900
Saltwort (prickly); Russian thistle1961w112 (3)1
Scale; Lenscale2193w1500
Sheep sorrel2353w1800
Wall pellitory (Parietaria officinalis)1906w191 (8)0
Wall pellitory (Parietaria judaica)1904w214 (9)3
Western ragweed1711w21 (2)1

Sera had to be tested for total IgE, against at least 10 different allergens and yield at least one positive specific IgE test result to be allowed for the final database. With a total of 203 283 specific IgE tests, 5362 sera met our criteria and were used for the analysis.

Databases and software

We created a MySQL database to hold the serum data (MySQL 5.0, obtained from Allergen protein sequences were extracted from the Allergome database ( as of January 2009). MEME 3.5.7 (12) (obtained from and pftools 2.3.4 (13) (obtained from were used for the iterative allergen motif discovery. Perl 5.8.8 (, PHP 5.2+ (, and R 2.8 (14) ( scripts were created to extract the desired statistical calculations.

Allergen motifs

We performed the iterative allergen motif discovery according to Stadler and Stadler. (11) using 2189 protein sequences from Allergome. These sequences are known to encode allergenic proteins (2). We identified 97 motifs with a residue length of 50 amino acids. Three hundred and four of the sequences used for identification did not match an allergen motif, 96 thereof were shorter than 50 amino acids in length and therefore could not match a motif, and 26 additional proteins were known to only encode protein fragments.

To identify the motifs present in each allergen extract, the proteins used in the motif identification process had to be linked to extracts in which they occur. This linking was achieved by first matching the allergen extract to the corresponding allergen source within Allergome and then assigning the proteins to the extract as defined by Allergome (15).

Our database now made it possible to computationally determine which motifs occur in which extract(s), and therefore which extracts are likely to cross react due to their structural similarity.


The serum database

We used the data of specific and total IgE determinations comprising a total of 5362 sera. Figure 1A shows the distribution of IgE levels in this serum collection. As expected for a serum collection of allergic individuals, total IgE levels peaked above the threshold value of 100 kUA/l (16), namely between 200 and 400 kUA/l. For our study purpose, it was not necessary to define a lower cut-off of total IgE levels.

Figure 1.

 Grouping the sera by their total IgE content resulted in the distribution shown in (A). The amount of total IgE had no influence on the number of tests being performed (B); however, it affected the outcome of the tests (C). The lines in B and C indicate the median, the error bars represent the 25th and 75th percentile, respectively.

Most sera were from the US (2264) and Sweden (2119), from Western European countries (635) and Russia (230) while the residual sera were either obtained from Japan, Southern Africa or Canada (81) or its origin was unspecified (33).

All sera were tested against a subset of a panel consisting of 99 allergens, as described in Table 1. Among these sera, 1471 sera have been tested against 90 or more of the allergens and 3448 sera against 10–30 allergens.

We used the traditional cutoff of 0.35 kUA/l (Class I or higher) to test positively against a given allergen extract. This yielded 99 276 positive values representing IgE from classes I–IV from a total of 203 283 allergen-specific IgE determinations.

Despite the similar number of tests that were performed (Fig. 1B), Fig. 1C shows that the percentage of positive specific IgE tests steadily increased with increasing total IgE, resulting in more than 90% positive tests for sera with very high total IgE (>3200 kUA/l).

Cumulated specific IgE

Next, we cumulated all individual specific IgE levels within the 5362 sera and plotted them against the total serum IgE (Fig. 2A). The data show that in 88.95% of the sera, total serum IgE exceeds the cumulated specific IgE. Figure 2B shows the percentage of cumulated specific IgE if sera were grouped according to total IgE levels. Of all data, 90.2% lay within a range between >0 and 1600 kUA/l and in this range, specific IgE was between 25% and 30% of total IgE.

Figure 2.

 Relation of cumulated specific IgE to total IgE. The scatterplot (A) plots the cumulated specific IgE per serum to the total IgE. The dashed line in A indicates total IgE = cumulated specific IgE. For 90% of the sera, the cumulated specific IgE represented about 25–35% of the total IgE (B). The line in B represents the median, the error bars the 25th and 75th percentile, respectively. The histogram (C) shows the percentage of sera for which at least one extract value has been capped at 100 kUA/l.

Interestingly, the residual 10% of sera with more than 1600 kUA/l total IgE showed a decreasing percentage of specific IgE. Thus, even though high total IgE sera at a greater number result in positive tests (Fig. 1C), their specific cumulated IgE fraction was lower (Fig. 2B). However, sIgE determinations were capped at 100 kUA/l and we found that with increasing total IgE sera were more likely to contain sIgE values affected by this capping (Fig. 2C), which would underestimate the percentage of cumulated specific IgE in these sera.

Cross reactivity by motifs

Based on the most recent sequence database, we found 64 of the 99 tested allergens to contain proteins containing our defined motifs. Some extracts contain more than one motif resulting in 86 motifs to be associated with the tested allergen extracts. In 13 extracts, we found only one motif (Table 1). On the other hand, extract d2 (House Dust Mite, Dermatophagoides farinae) contained 14 motifs, m3 (Aspergillus fumigatus) also 14, d1 (House Dust Mite, Dermatophagoides pteronyssinus) 12, i6 (German Cockroach) 10, and so on. Excluding those extracts without a known motif, we found a median of three motifs per extract.

Figure 3A shows that also the number of recognized motifs steadily increased with higher IgE levels. On the other hand, Fig. 3B shows that there are approximately three times less motifs recognized than extracts from different allergenic sources (correlation coefficient: 0.967).

Figure 3.

 The number of recognized motifs within a serum increases with the total IgE content (A). In A, the line represents the median, the error bars the 25th and 75th percentile, respectively. Comparing the number of recognized motifs to the number of positive extracts shows that the number of recognized motifs is about three times lower than the number of positive extracts (B). In B, the line represents the mean while the error bars represent the standard deviation.

The question remained whether this relation between motifs and extracts is linear and is directly due to cross reactions. Our next assumption was simple: If different extracts from different sources theoretically contain the same motif, one would expect a cross reaction. For this purpose, we analyzed all motifs that occurred in three or more different extracts. Table 2 shows all motifs that fulfill this criterion. For example, it is seen that motif 1 (corresponding to the group of the Bet v 1 allergen) occurs in 25 extracts of our allergen panel.

Table 2.   Range of cross reactivity by allergen motif. The percentage of sera reacting with another extract containing the same motif. For an illustration of the degree of cross reaction, see Fig. 4
Motif idNumber of extracts containing motifMost frequent cross reactionLeast frequent cross reaction
% (extract)Number of positive sera% (extract)Number of positive sera
 12591.3 (f33)73960.0 (g6)3132
 21384.0 (t17)88764.1 (g6)3132
 71394.5 (f33)73972.1 (w1)2102
 41092.7 (f85)15169.6% (t3)2916
 8973.1 (f27)27245.2 (d1)2586
22882.2 (f45)65652.7 (g6)3132
28877.5 (f27)27239.3 (e5)3044
60877.2 (m3)118462.4 (t3)2916
11793.7 (f12)77774.8 (f13)1975
10691.6 (f18)77778.0 (f13)1975
19679.6 (f242)3957.0 (f49)1179
20679.6 (f27)27250.0 (e5)3044
32558.9 (i1)60447.0 (m6)1526
33588.3 (f6)105183.3 (f17)1588
52594.3 (e87)15353.6 (e5)3044
 3494.9 (f37)35858.9 (d1)2586
18484.8 (f35)90768.1 (t9)1667
23483.4 (i6)110266.0 (f13)1975
30484.6 (f92)23063.5 (f4)1472
37473.8 (m2)120147.4 (f20)858
45478.8 (m3)118463.7 (f4)1472
54492.4 (f11)91978.4 (f13)1975
57494.5 (w11)104370.1 (g6)3132
63461.3 (m2)120154.4 (i3)641
13398.3 (g2)141592.8 (g3)1950
15386.7 (i6)110274.6 (d1)2586
62383.5 (m2)120144.6 (d2)1718
65369.9 (m14)66853.0 (i1)604

Figure 4 depicts two examples from the list in Table 2. We have chosen motif 4 occurring in 10 different extracts and motif 8, occurring in nine different extracts, as examples. Motif 4 was chosen because it is an example for relatively high cross reaction, while motif 8 is an example for a ‘low’ cross reactive motif. Our results, depicted in a spider form, show how closely related the different allergen motif containing extracts actually are. The graphical depiction of the relationship between the motif defined cross reactions also allows to create a ranking. For example, motif 4, extract f85 (Celery), shows cross reaction at an average of 92.7% with all other allergen extracts in the group, while the lowest extract (t3, Common silver birch) cross reacts at 69.6%. For illustration, we have chosen this high and low cross reactive level from Table 2 that shows at the same time the absolute number of sera falling within this group.

Figure 4.

 Degree of cross reaction. Spider plots for the extracts containing motif 4 or motif 8 as listed in Table 2. The number of sera positive against two extracts is expressed as a fraction of the number of sera positive against other extracts containing the same motif. The dashed line represents the mean fraction for each extract.

As cross reactions seemed to be very frequent, we analyzed how many sera recognized extracts without cross reaction. Three hundred and eighty-five sera (7.2%) were positive against one extract only. However, 12 single positive sera could not be considered as there are yet no protein sequences containing a motif within the positive tested extract [2x guinea pig epithelium (e6); 2x elm (t8); 2x wall pellitory (w19); 1x rye (f5); 1x oat (f7); 1x shrimp (f24); 1x chicken (f83); 1x mucor racemosus (m4); 1x walnut (t10)].

To truly exclude cross reactions, for each of the motifs occurring in the positive extract, the single positive sera would have to be tested against at least one other extract theoretically containing the motif. Otherwise, one might argue no cross reactions were observed because not all other extracts containing the potentially cross reactive motifs had been tested.

Only in 24 of the residual 373 sera, we observed a ‘true’ single positive reaction as all motifs theoretically contained within the positive extract were analyzed. In seven of the 24 sera, the specific IgE level against the single positive extract was only barely positive (≤0.5 kUA/l), but in general, the amount of specific IgE was not significantly different from the amount of specific IgE against these extracts in other sera (data not shown).

Additionally, we found that with an increasing number of extracts being tested, the fraction of single positive sera decreased. While 10.1% of the sera tested against 20 or less extracts were positive against one extract only (324 out of 3213 sera), this number dropped to 1.6% for sera tested against 90 or more extracts (15 out of 956 sera).


We recently described a bioinformatic approach to predict allergenicity (11). Motifs in size of 50 amino acids were generated using bioinformatic algorithms using the most comprehensive database of allergen sequences, Allergome (2). We are presently at approximately 100 theoretical structures (motifs); whether such a general denomination of an allergen may truly predict allergen cross reactivity may be disputed.

Thus, we have used a large database of specific IgE determinations. These data seem to be normally distributed around an elevated IgE level as expected. The literature does not provide any assessments on how much of the specific IgE is actually directed against allergens causing symptoms. Our database was tested against a panel of 99 allergens and 1471 sera were tested against 90 or more of the allergens. During the rather long period over which our data were collected, the quality of allergen extracts has been changed by the manufacturer. Nevertheless, our analysis showed no statistically significant specific IgE level changes over time, neither in general nor grouped by country (data not shown).

Because it has been shown that ImmunoCAP sIgE determinations are accurate on a mass unit level (17), IgE levels of different extracts can be compared. As total and specific IgE determinations are based on a different measuring principle, one might argue their respective levels are not comparable, representing a systematic error. As quantification of human IgG and IgM has been transitioned from international units to absolute (g/l) values almost 30 years ago (18), this may also apply to IgE. The scientific literature has been converting 1 unit to 2.4 ng for IgE determinations performed on ImmunoCAP systems, and recent work comparing a new total IgE quantification method has shown that the absolute results are comparable to ImmunoCAP results (19). Thus, we assumed that total IgE determinations are accurate on a mass unit level and therefore may be compared to sIgE levels.

One may argue that the allergens of our panel were the true sensitizers for the generation of IgE (20, 21). We speculated that by using such a great number of allergens, most of the specific IgE might represent the total IgE level. Interestingly, this simple hypothesis delivered a value of 20–30% specific IgE in patient sera. While this percentage dropped in sera with total IgE higher than 1600 kUA/l, it is likely that we underestimate the amount of cumulated sIgE in these sera because specific IgE levels have been capped at 100 kUA/l, especially in these high total IgE sera, as Fig. 2C shows.

Another interesting finding is the percentage of positive test results related to total IgE levels. For the 169 sera with very high total IgE levels (higher than 3200 kUA/l), on average 85.8% of specific IgE tests yielded positive results. This suggested a broader reactivity against allergens in relation to higher total serum IgE.

The question remains whether the residual IgE, not to be attributed to one of these frequent allergens, is directed toward unknown allergens or represents IgE that cannot be assessed by using commercial allergen preparations. Of course, it may also be speculated that ‘non specific’ IgE may be a result of somatic mutation; in other words, IgE without a known specificity.

Here, we postulate that the estimation of 20–30% specific IgE may be exaggerated as a great proportion of this IgE is cross reacting. By comparing the specific IgE determinations against the various allergen extracts with the theoretically postulated motif, one may speculate that the accumulation of specific IgE using different extracts results in an overestimation. Indeed, as Fig. 3B shows, there seems to be an almost strict linear correlation between the number of recognized extracts and motifs. However, there seems to be about three times fewer motifs recognized than allergen extracts. Based on this result, one would have to assume that the actual percentage of specific IgE in patient sera is actually below 10% of the total IgE.

Figure 3B also addresses the question whether co-sensitization might be mistaken as cross reactivity. We observed a direct linear relationship of 3 : 1 between the recognition of allergens containing defined motifs and the number of motifs that were recognized. Co-sensitization would occur independently of structural similarities between proteins, resulting in a ratio closer to 1 : 1 as each recognized extract might be recognized by a different epitope. This is also exemplified by our finding for Bermuda grass (g2), an allergen that is not naturally occurring in Scandinavia. 61.6% of sera derived from Sweden were positive against Bermuda grass, arguably because of cross reaction rather than co-sensitization.

As motifs can easily be associated with known allergen sequences, we have grouped the cross reactions according to the theoretical presence of the motif. This approach resulted in an astonishing high frequency of cross reactions. It was also possible to generate a hierarchy which may or may not be of clinical relevance. Within one motif, it was always possible to rank cross reactions by their mean percentage of cross reactions with other allergen extracts containing the same motif. This, of course, may be a bias as the allergen extract with the highest degree of cross reaction may be an allergen extract where the epitopes that can be associated to motifs are well preserved, well presented or most prominently present in a given industrial allergen preparation coupled to a solid phase (22). However, our data strongly suggest that patients and doctors should realize that sensitization and reaction to a given allergen extract may not be the sensitizing organism or part thereof, like birch pollen, but that it is a molecular entity occurring in different organisms.

We were also interested whether there were patient sera without cross reactions. In our 5362 sera, we found 385 that were positive against one extract only. Twenty-four of these sera had been tested against at least one extract for every motif present in the positive extract, yet showed no cross reactions. Twelve sera were positive against an extract containing no currently known motif, and in the remaining 349 sera, a sensitization against one of the motifs that had not been tested would be possible. One might hypothesize that these patients developed an IgE response against an epitope not shared among proteins bearing the same motif, or even simpler that not all cross reactive motifs have been found yet. Furthermore, the fact that with an increasing number of extracts tested, the percentage of single positive sera decreased from approx. 10% for sera tested against 20 or less extracts to 1.6% for those tested against at least 90 extracts suggests that the true number of single positive sera must be low.

For allergy diagnosis, the way out of this uncertainty whether a patient reacts to a certain part of a plant or a common allergenic structure may be diagnostic procedures using recombinant allergens. Thus, in the future, we intend to analyze whether a small number of recombinant proteins covering most of the known motifs would suffice to diagnose other sensitizations by predicting the allergen via the motif.


This work was supported by grant no 8803.1 from the Commission of Technology and Innovation (CTI), Switzerland.