ARTICLE IN PRESS

Variation of the group 5 grass pollen allergen content of airborne pollen in relation to geographic location and time in sea son

The HIALINE working group: Jeroen Buters, PhD,a,b Marje Prank, MSc,c Mikhail Sofiev, PhD,c Gudrun Pusch, Dipllng,a,b

Roberto Albertini, PhD,d lsabella Annesi-Maesano, MD, OSe, PhD,e,f Celia Antunes, PhD,9 ·h Heidrun Behrendt, MD,a,b

Uwe Berger, MBA,; Rui Brandao, PhD,9 t Sevcan Celenk, PhD} Carmen Galan, PhD,k Lukasz Grewling, PhD)

Bogdan Jackowiak, PhD) Roy Kennedy, PhD,m Auli Rantio-Lehtimiiki, PhD," Gerald Reese, Ph0,0 lngrida Sauliene, PhD,P

Matt Smith, PhD} Michel Thibaudon, Pharm,q Bernhard Weber, PhD, 0 and Lorenzo Cecchi, MD' Munich and Reinbek,

Germany, Helsinki and Turku, Finland, Parma a11d Florence, Jtaly, París and Brussieu, Fra!ICe, Évora and Coimbra, Portugal, Vienna, Austria,

Gorukle-Bursa, Turkey, Cordoba, Spain, Pozna11, Poland, Worcester, United Kingdom, and Siauliai, Lithuania

Background: Allergies to grass polleo are the number one cau..:;e of outdoor hay fever. The human immune system reacts with symptoms to allergen from polleo. Objective: We investigated the natural variability in release of the major group S allergen from grass polleo across Europe. Methods: Airborne polleo and allergens were simultaneou.<;ly collected daily with a volumetric spore trap and a high-volume cascade impactor at 10 sites across Europe ft1r 3 consecutive years. Group S allergen levels were determined with a Phi p S­specific ELISA in 2 fractions of ambient air: particulate matter

From ~LAUM Ccntcrof Allergy& linvironment, Helrnholt:r.Center Munich!Hx:hnischc Universitat München, Mcmbcr of thc German Ccnter for Lung Rcscarch (DZL), Munich; "thc Kühnc Foundation, Christinc Kühnc--Ccnter for Allergy Rescarch and l.iducation (CK-CARI.i), Munich; "the Finnish Mctcorological lnstitutc, lirik Palmenin Aukio 1, Helsinki; "the Laboratory of Allergology, Dcpartrncnt of Cli nical and Experimental Medicine, University of Parma, U.O. Medica! lrnrnunology, University Ho&pital of Parma; "the EPAR Dcpartrnent, INSERM, UMR_SI 136, lnstitut l'icrrc Louis d'l.ipidérniologic e t de Santé Publique, France; ' the EPAR Dcpartrnenl, Sorbonnc Universités, UPMC Univ Paris 06, UM R_S 1 136, lnstitul l'icrrc Louis

d'Epidémiologie c t de Santé Publique, Paris; "ICAAM- Institute of Mcditcrrancan Crop and Environmcntal Scicnccs, Univcrsity of Évora~ hthc Ccntcr for Ncuroscicncc

and Ccll Biolo)(y (CNC), University of Coimbra; ;thc Medica! Univcrsity of Vienna, Dcpartment of Ot(>-Rhino-Laryngology, Rcsearch Unit Acrobiology and Pollen lnfonnalion, Vienna~ jthe Acrobiology Laboratory. Biology OcpartmcnL, Scicncc and Art.< Faculty, Uludag Univcrsity, Gorukle- Bursa; 'thc Dcpartmcnt of Botany, Ecology and Plan! Physiology, University of Córdoba, lntcmational Campus of Exccllcncc on Agrifood (cciA3), Córdoba; 1the Laboratory of Acropalynology, Facully of Biology, Adam Mickicwic-z Univcrsily, Poznaó; mthc Nalional Pollcn and Acrobiology Rcsc.arch Unit, Univcrsily of Worccstcr, nthc Univcrsily of Turku,

CERU'I: Acrobiology Unit, Turku; "A IIergophanna GmbH & Co KG, Rcinbck; "thc Department of Environmcntal Rcscarch, Siauliai Univcrsity, Siauliai; " RNSA (Réscau Nalional de Survcillancc Aérobiologiquc), Bru.~sicu; and fthc lntcrdcpartrnental Centre of Bioclirnatology, Univcrsity of Horcncc.

tDcccased. ·n,is publication ariscs from HIALJNE, a project that ha.< rcccived funding from the

Euro~n Union in thc framcwork of thc Hcalth Programmc. Oisclosurc of pOLcntial conOict of intcrcst: J. Butcrs ha..:; rcccivOO or ha.~ granL'; (X!nding

from Foundation C K-CARE. S. Cclenk's institution ha.< rcccivcd fw1ding from the Scicntiflc and Tcchnological Rcsearch Council ofTurkcy (T UB ITAK 109S032). B. Weber is cmploycd by Allergophanna GmbH & Co. KG. Thc rcst of the authors declare thatthcy ha ve no rclcvant conOicL~ of intcrcst.

Reccivcd for publication January 22, 2014; reviscd January 7, 2015; acceptcd for publi­cation January 12, 2015.

Corrcsponding author: Jerocn Buters, PhD, ZAUM- Ccnter of Allergy & Environrncnl, HclmhoiLZ Ccntcr MunicWICchnischc Universillit Münchcn, Bicdcrstcincrstra..~sc

29, 80802 Munich, Gennany. E-mail: butcrs(qltum.dc. 0091 -6749/$36.(1() © 2()15 American Academy of Allergy, Aslluna & lmmunology bttp://dx.doi.org/1 O. JO 1 6/j.jaci.20 15.01.049

of greater than 10 fJ.m in diameter and particulate matter greater than 2.S J.tm and less than 10 fJ.m in diameter. Mediator release by ambient air was determined in FceRI-humanized basophils. The origin of polleo was modeled and condensed to polleo potency maps. Results: On average, grass polleo released 2.3 pg of Phi p S per polleo. Allergen release per polleo (potency) varied substantially, ranging from less than 1 to 9 pg of Phi p S per polleo (S% to 9S% percentile). The main variation was locally day to day. Average potency maps across Europe varied between years. Mediator release from basophilic granulocytes correlated better with allergen levels per cubic meter (? = 0.80, P < .()()}) than with polleo grains per cubic meter (? = 0.61, P < .()()}). In addition, polleo released ditTerent amounts of allergen in the non- pollen­bearing fraction of ambient air, depending on humidity. Conclusion: Across Europe, the same amount of polleo released substantially ditl'erent amount<; of group S grass polleo allergen. This variation in allergen release is in addition to variations in polleo count'l. Molecular aerobiology (ie, determining allergen in ambient air) might be a valuable addition to polleo counting. (J Allergy Clin lmmunol 201S;•••:•••--.)

Key words: Allergen, grass, Phi p 5, Europe, PM/0, PM2.5, exposure, allergy, basophil, modeling, SJLAM, HJAUNE

Grass pollen is the major cause of allergic sensitizations to outdoor allergens all o ver Euro pe 1 and ranks second to house dust mite a~ a cause of respiratory allergy in most countries.2-5 The role of grass poli en in respiratory allergy was recently confirmed in a large epidemiologic study from the European Community Respiratory Health Study.6

Grass pollen originates from many species of the Poaceae family and release severa] proteins and glycoproteins,7 sorne of which are identified as allergens. These allergens are classified in 1 O groups on the basis of their high intragroup cross-reactivity (see World Health Organization!Intemational Union of Immunological Societies Allergen Nomenclature Sub-Committee at http://www.allergen.org/; Allergome at http://www.allergome.org/). Phleum pratense (timothy grass) is a major source of grass pollen in temperate areas, and because of the cross-reactivity mentioned above, its extracts and single allergen components are widely used for the in vitro and in vivo diagnosis of grass pollen allergy and specific immunotherapy.8

Although sensitizations against Phi p 1, 2, 4, 5, 6, 7, 11, 12, and

2 BUTERS ET AL

Abbreviations used

ffiALINE: European Un ion Project Health Tmpacts of Airbome Allergen Tnformation Network

1 0>PM>2.5: Particulate matter Jarger !han 2.5 11-m and Jess !han JO 11-m in diameter

PM>lO: Particulate matter Jarger !han JO 11-m in diameter

13 are described, Phi p 1, 2, 5, and 6 are markers of grass poli en sensitization because they are present exclusively in grass pollen.8•

9 Phi p 1, 4, and 5 show the highest rates of sensitization (>50%) in patients with grass pollen-induced allergy. 10

•11

However, recent findings have shown that the prevalence of sensitization and clinical importance of Phi p 4 might be overesti mated. 12

•13 Other than for the clinical relevance, Phi p 5

is important for its high allergenic activity and good characterization. 14 For these reasons, the variant Phi p 5.0102 (Phi p 5a) was adopted as reference material for the standardiza­tion of extracts for immunotherapy and skin prick tests, as indicated by the Development of Certified Reference Materials for Allergenic Products and Validation of Methods for Their Quantification (CREATE) project's findings. 15 Different species of the Poaceae family express homologues of Phi p 516

•17 that

show similar immunologic responses. 18

Tt is not possible to distinguish grass polleo grains from different grass species by using light microscopy, but the identity of allergens and the amount of allergen released from polleo varíes between species.8•

17 Natural variation of allergen from biological sources was reported for birch polleo, 19 olive pollen,20

horses,21 and cats.22 Allergen release also varíes during the lifespan of pollen?3

Monitoring allergens in ambient air enables the detection of non-pollen-bound allergens, as encountered during special weather conditions, such as thunderstorms? ·24 In addition, the determination of thresholds for allergic symptoms based on pollen counts has produced variable results25

•26 and might be

improved by adding airbome allergen monitoring?7

In the framework of the European Union Project Health Tmpacts of Airborne Allergen Tnformation Network (HIALINE; www.hialine.eu), we monitored daily amounts of grass polleo and their major group 5 allergens simultaneously across Europe. We showed that levels of Phi p 5 released from grass polleo was variable and emphasized the necessity to consider molecular aerobiology in addition to counting polleo for the assessment of exposure to airborne allergens, especially in clinical trials and epidemiologic studies.

METHODS Pollen in ambient air

Airbome pollen concentmtions were collected al JO sites in Europe (Table T and see Table El and Fig El in this article's Online Reposi tory al www.jacionline.org) by using Hi rst-type vol umetric spore traps, accordi ng 10 !he requirements of the European Aeroallergen Network28 Quality of counts was controlled, as described previously.19

Allergen in ambient air Air was simultaneously sampled al each site wi th ChemVol (Butraco, Son,

The Netherlands) high-volume (800 Umin) cascade impactors equipped with dry polyurelhane impacting substmtes (that were prewashed with 0.1 mol/L

J ALLERGY CUN IMMUNOL

- 2015

ammoni um bicarbonate and water) using the stages for particulate m alter greater !han 1 O 11-m (PM> 1 O) and particu late m alter greater !han 2.5 11-m but Jess !han 1 O 11-m in di ame ter (1 0>PM>2.5), as described previousty? 9 Allergen was extracted in 0.1 mol/L ammoni um bicarbonate contai ning 0.1 % BSA in an end-over-end rotator al 100 rpm for 4 hours, Jyophilized, and redissolved in one tenlh of Lhe original volume in PBS, according 10 standard procedures.23

,30 Phi p 5 allergen was determined by means of ELTSA with the mAbs lDl l and BO 1 (AIIergopharma, Reinbek, Germany)31 calibmted against recombinant Phi p 5.0102.15 The assay recognizes group 5 allergens of Pooideae (Phi p 5.0102 and 5.0201 ); has no cross-reactivity wi th groups 1, 2, 3, 4, 6, and 13; and has a limit of detection of 1.0 ng of Phi p 5/mL, equali ng about2 poli en grains/m3 (1 mL of extrae! =256m3 of air). Response factors for the different group 5 allergens vary less than 2-fold (data not shown; Allergopharma, Reinbek, Germany). The assay recognizes C-terminal group 5 fragments,32 and ELTSA correlated with mass spectromelric quantification (data not shown, Allergopharma). Quality control of ELTSA data was performed accordi ng to !he ru les developed in ffiALINE. 19 Values Jess !han the detection limit were reponed a~ zero.

Basophil degranulation assay An immune cell response independent ofELTSA was determined based on

induction of mediator release by ambient air extmcts from an Fe~;RJ­

humanized ral basophi l cell fine (at-, ¡:\-, and -y-chain).33 The cells were passively sensitized wi th human serum of a palien! with gmss pollen-induced symptoms, a positive SPT response, anda RAST score of greater than 3. The patient wa~ sensitized to Phi p 1, 213, 4, Sb, 6, and 11 and Cyn d 1 but not Phi p 7 and 12. Levels of ¡:1-hexosaminidase a~ a proxy for histamine relea~e was measured, as descri bed previously.19 Each environmental sample was diluted sequentially, and only data from the linearpanofthe degranulation curve were reported.

Statistics Poli en potency wa' determined seveml ways: (1) as the slope of the linear

regression curve of poli en per cubic meter versus allergen per cubic meter with the intercept forced through zero and !he strenglh of the relation expressed by the coefficient of correlation (?); (2) as the mode of the histogmm of daily potencies; (3) as the median potency calculated for the whole data set and for subranges of daily poli en counts; and (4) a~ the 5th to 95th percentile mnge. The analysis excluded pollen counts of Jess than JO gminslml_ 19

•34

The stalion- and year-wise variabil ity of potency was assessed from histograms for the individual si tes. Differences were analyzed by using !he Student t test, unless stated otherwise.35 A P value of Jess than .05 was considered statistically significan!. Val ues beyond 3 SDs of !he mean were considered outliers.

The explanalion factors for the potency variations included meteorological variables that were either monitored al the Che m Vol locations or obtained from the nearest meteorological station.

Modeling The System for Integmted Modell ing of Atmospheric Composi tion

(STLAM)36•37 was used to eval u ate the Howering period and poli en atmo­

spheric transport and also 10 compute the footprints of the observations, as described previously. 19TheSTLAM model was run with a ti mestep of 15 mi­nutes, and the footpri nts were calculated 60 hours backward in ti me for each daily observation al each si te.

Dai ly observed potency was mapped to !he origin of pollen through !he footprints. The lowest values that make up 1% of !he footprint integral were cut off 10 reHect the limited tmnsport distance of gmss pollen. The potency value was attributed 10 all areas covered by the footprint that Howered atlhe Lime of !he footpri nt passage. The allributed potency values were then avemged overtheseason, finally resu lti ng in potency maps that show the mean potency of pollen relea~ed from the area during a specific year.

All simulations used the same configumtion that included 8 vertical Jayers up to approximately 6 km above the ground. The horizontal grid cell si1-e was

J ALLERGY CUN IMMUNOL

VOLUME - . NUMBER •

BUTERS ET AL 3

TABLE l. Average yearly pollen count (pollen index) and allergen release per pollen for the stations in the d ifferent countries

2009-2011, average :!:

1999-2008• 2009 2010 2011 SO (SO %)

Polleo iodex, Polleo iodex, Poteocy, Phi Polleo iodex, Poteocy, Phi Polleo iodex, Poteocy, Phi Poteocy, Phi :!: Graios/m3/y :!: Graios/m3/y p 5/graio (pg) :!: Graios/m3/ y p 5/graio (pg) :!: Graios/m3/y p 5/graio (pg) p 5/graio (pg)

Austria Vieooa 3,024 ::!: 606t NA NA 2,604 3.546 2,411 2.94 1 3.244 ::!: 0.428 (13) Fio laod Turku 740 ::!: 267 654 1.592 738 1.992 782 1.535 1.706 ::!: 0.249 (15) Fraoce Varíes¡ 9,671 ::!: 3,808 10,420 2.405 7,130 2.403 6,708 2.937 2.582 ::!: 0.308 (12) Germany Munich 2,033::!: 604 2,041 3.042 1,821 2.127 2,184 2. 133 2.434 ::!: 0.527 (22) Italy Parma 7,521 ::!: 2,973 4,135 1.508 4,362 2.871 5,409 0.82 1 1.733 ::!: 1.043 (60) Polaod Poznao 4,697 ::!: 1 ,472§ 6,078 2.703 7,898 3.398 4,482 1.996 2.699 ::!: O. 701 (26) Portugal Evora 6,617 ::!: 8,83611 5,725 2.144 17,113 1.499 22,815 1.545 1.729 ::!: 0.360 (21) Spaio Cordoba 5,609 ::!: 2,777 4,014 2.268 8,693 4.149 5,888 2.423 2.947 ::!: 1.044 (35) Turkey Bursa 1,242 ::!: 490, 1,881 NA 2,892 3.1 26 3,812 NA 3.126 United Worcester 6,715::!: 2,223 4,885 5.906 5,141 3.570 3,531 4.178 4.551 ::!: 1.212 (27)

Kiogdom Average 2.696 ::!: 1.395 2.868 ::!: 0.840 2.279 ::!: 0.989

potency::!: SD SD (%) 52 29 43

Potency (relcasc of allcrgcn pcr pollcn) was dctcrmincd by thc slopc of thc linear regrcssion (scc thc Mcthods section). Mcans :!: SDs are prcscntcd. Exact location oftlte stations is givcn in Tablc E l.

NA, Not applicablc. *A cornplcLc prcc.cding dccadc was not ava.ilablc for all SL.ations, in which ca.;;;c thc maxirnum pcriod availablc wa.'i takcn. tVicnna 1999-2008: 2 sitc.';: (A) 1999-2002 ATWIAK; ( B) 2003-2008 A:IWII.iN. ·n ,c trap was moved from A:IW IAK (A KH- AIIgcmcincs Krankcnhaus) to A:IWIEN (ZAMG­Zcntralanstalt ttir Mctcomlogic und Gcodynamik).

t ln 1999-2008, thc station was in Lyon; in 2010, it wa• in Saint-Genis 1' Argcnticrc; and in 2011, it wa• in Brussicu. §From 2005-2008.

IIFrom 2001-2008. 1Fmrn 1999-2000, 2003-2005, and 2008; no data from 2001 -2002 and 2006-2007.

25 km. Input meteorological data were taken from the operational archive of the European Centre for Medium-Range Weather Forecast.

RESULTS Pollen in ambient air

Airbome grass polleo is ubiquitous in Europe, but large variations in polleo counts exist between countries. A map of the stations is presented in Fig El . The Finnish station in Turku (seaside) showed the Iowest pollen counts, whereas the Portuguese station in Evora (rural site) recorded the highest counts (Table n. It should be noted that the years studied for each station were within the normal range of the preceding years for all stations (season timing and total polleo counts were within 1 SD from the long-term mean values, data not shown). The stations were distributed over Europe to obtain the most representative picture for the continent.

With 3 Hirst-type pollen traps operated simultaneously within 5 m of each other at about 9 m in height for a 3-week period during the grass polleo season (0-890 pollen grains!m\ we determined the variability between the traps to be about 5 polleo grains/m3

plus 20% (data not shown).

Allergen In addition to vanattons in airborne polleo counts, pollen

differed in allergen release per pollen (potency, Fig 1). Determined by me.:'lns of linear regression (see the Methods section), the average group 5 allergen re lease per polleo was 2.3 pg of Phi p 5 per polleo(? = 0.497, P < .001, n = 1629; see Fig E2 in this article's Online Repository at www. jacionline.org). This number agrees well with the histogram

mode (2-2.5 pg of Phi p 5 per pollen, see the histogram in Fig E3 in this article's Online Repository at www.jacionline. org). Median potency is somewhat greater, 2.6 pg of Phi p 5 per polleo, because of the asymmetric tails of the histogram. The polleo potencies for each station are given in Table 1, and histograms are shown in Fig E4 in this article's Online Repository at www.jacionline.org. Potency distribution can have quite a wide spread that is different for different regions. Interannual variability of the histograms at the same station is smaller than spatial differences. Taking theextremes in pollen counts (Fin1aod vs Portugal), allergen re lease per pollen was not different (not significan!). Potency values over Europe were correlated with polleo counts (Fig 2).

We detected many days when polleo did not release any Phi p 5 at all. These pollen grains without group 5 allergens were especially noticeable at the beginning of the grass pollen seasons. This also explains why pollen potency distribution is not normally distributed (see Fig E3). Using the 5th to 95th percentile, polleo potency ranged between less than 1 and 9 pg ofPhl p 5 per polleo (excluding low polleo counts [ie, <10 polleo grains/m3]).

Mediator release from basophils For one station (Munich, Germany), we determined the daily

mediator release by basophils passively sensitized to grass polleo allergens. In Munich no overlapping polleo season was measured, except for nonallergenic Pinus species (pine) and Picea species (spruce) pollen. The degranulation was standardized as if 1 m3

of air was delivered to the cells, which could result in calculated degranulation rates of greater than 100% (Fig 3). The serum used showed sensitization against severa! major and minor grass polleo allergens (Phi p 1, 2/3, 4, 5b, 6, and 11). Al though these specific

4 BUTERS ET AL

Austria

Finland

France

2009 2010

""" "' n $ " ' o•v n . Mio. -

.. .•

J ALLERGY CUN IMMUNOL

- 2015

2011 . -. .. 1· :1 i ~~

.tl~J¡, .... z.lML .. '" 1

M . . ,_, u1t. ~ ~u. nA ,._,, .. , u .

~·

"l~:: . .. t~ ~ = i" . , .. 1 t. : : 1 . :: o ~.

1&4. u. oU. ».&. •U. V6. u.t. -.. .•

Germany .. , . . ' ' • ' '" 1 1" ·~ : .• 1 a: •: ~ .

• • "' ·SO

r:: :··· 1"·~·-1 . . t.. .. :•

ltaly

Poland

Portugal

Spain

Turkey

UK

~~LLJJ . ~1 •. J } •· . . . IUJIU.U. .. JH. IU . lU. U , u.t. t,7, <IU. -

!~[¡]Il :1 a:. • l(lf 1 . . . . .

1:7A I U, JU U . 1JA V. a.1. ,_., 11 .. -

... ..- _._,_:. U t1-t. tU oU :116. o)t. Jli.l U -:r t 1 1~ • ;: JiU~

~=L_L·=1 ~: ··l . -~ . .

IU Al- ll • . ~o. IIU. ~!. 111. JU, U . lt.J, -

-

H 11.$, IU, I'"A al IU, ;ll,7, U lJA -

. ' t.\ hio. Ailo. lii)Ai .,, 01,1 >11. Ul. -

-,:ll ~1 r: /WJ;A - -~ ,, ~,

1UliO.S- OI4J14 0UlS.11.t. - --H -

t 1

::1 ~ ' ...,J. td., 1! .... J • ...l ª 1

. ) UA l!t.<l 9). U.l. .... ;Nt. .. 1 0$1 ...

!~:=:¡]· :1 ~ .: -1 .. ~ . .

11.1 "" IU, u , )U, . , . U A li'A IV. IU. -

U. IU :1M. IH DA 1V IV. U. ti:A -FIG 1. Palien flight fram all statians in 2009 until 2011 cancamitant w ith palien allergen release patency. Gray vertical bars represen! palien, and purple fines represen! daily allergen released per palien (daily palien patency). UK, United Kingdam.

J ALLERGY CUN IMMUNOL

VOLUME - . NUMBER •

1 ~8.00

64.00

32.00

16.00

- 8.00 r:: .!!! •1.00 o Q.

....... 2.00

"' Q. 1.00

:E Q. o.so .. Q.

- 0.2S . ~ r:: 0. 13 J!l o 0.06 Q.

0.03

0.02 ID 20 40

Dm

-- 2%

. ... 16% -•·median

-- 84% - - 1)8% - 2.s.td

- btd

80 160 Pollen/ m3

320 640 1280

RG 2. Experimental variability o f grass pollen potency versus its estimated uncertainty. So/id fines, Potency uncertainty range calcu lated from assumed errors of pollen counts and ELISA. Dashed fines, Observed percentiles; thick dots indicate the centers of bins used for percentiles estimation. The observed potency vari ability was much larger than can be explained by experimental error. Pollen counts of less than 10 pollen grainslm3 were excluded.

TgEs to severa! majar or minor allergens couJd have mediated basophil degranulation, mediator release from basophilic granulocytes correlated well wi th Phi p 5 Jevels per cubic meter (? = 0.80, P < .001) and with polleo Jevels per cubic meter (? = 0.61, P < .001 ). Two outliers (probably because of the large dilution needed to achieve the linear part of the degranulation curve) were identified and indicated in Fig 3.

Potency maps Potency maps compiled for each year (Fig 4, left column)

showed a substantial variability over Europe. Sampling stations were Jocated between 2 and 22 m (see Table El ). Fewer grass pollens were sampled at higher monitor Jocations because of the substantial sedi mentation velocity of grass polleo. However, numeric simulations showed that the separation of different sizes along the height is negl igible ( <20% in the most extreme cases) at the used heights (data not shown). Thus the differences in sampling height in our study do not explain the differences in polleo potency.

Severa! regions, such as Franceand Germany, appeared to have quite high potency of the poli en released, reaching values greater than 5 pg ofPhl p 5 per polleo. The Jowest potency in all years was registered in Eastern Europe in 20 JI at less than 2 pg ofPhl p 5 per poli en. However, most regions appeared to vary from year to year.

A crude estímate of reliability of these maps can be obtained from the SD of potency (Fig 4, right column). The high potency estimated in France is usually accompanied by a large spread that is similar but less than the values themselves. The stripes visible on the maps are individual footprints corresponding to substantially different observed potencies. Their overlap results in high uncertainty of the final pattern. In other regions, however, the uncertainty was lower and independent of the actual potency value.

Relation of potency and pollen counts The average polleo potency stayed almost constan!

throughout the observed polleo count range (Fig 2, thick black

BUTERS ET AL 5

A

ME ~i

/ 7 •

y= 0.87x + 8.7 /~ r'=0.80 1 1500

1000 e <E---outliers

~ 400

"' ()) -¡¡; 0:: • • ~ 300 • e .Q o; :;

• • 200

e ~

"' ()) o

• 100

100 200 300 400 600 700

Concentration Phi p 5 lpg/m3]

50 100 150 200

Concentration palien 1Grainstm3]

RG 3. Degranulation of FceRI-humanized rat basophils by extracts of ambient air. A and B, Degranulation caused by allergen of PM>10 (Fig 3, A) or degranulation caused by pollen (Fig 3, 8). Two extremes indicated in the graph were excluded in both calculations of correlation (see the Methods section).

dashed Jine denotes its median). However, s light decreasing trends are noticeable for low and high counts. To the contrary, the spread clearly increases toward low counts. One can es ti mate what fraction ofthis variability is dueto observational errors and what is to be attributed to natural potency variabi Ji ty. As stated above, the experimental error for poli en counts is about 5 poli en grainslm3 plus 20% of the observed value, and the error of ELTSA is approximately 20%. Because potency distribution is non-Gaussian, robust estimates of the 1 and 2 SD ranges will be the differences between the potency values at 16% to 84% and 2% to 98%, respectively. As one can see, polleo varied more in potency than can be explained by the experimental errors.

Humidity Allergen was collected in 2 fractions of ambient air. Within

all ofEurope, more than 89% :t 11% ofthe allergen was collected as PM> 1 O. The geometric diameter of grass polleo is 20 to 40 J.Lm,38 exemplifying that allergen in ambient air coincides with polleo. However, on sorne days, allergen in 10>PM>2.5,

6 BUTERS ET AL

A

Potency 2010

B

e

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4 .5 5.0 6.0 7.5 10.0

Standard deviation 2009

Standard deviation 2010

Standard deviation 2011

J ALLERGY CUN IMMUNOL

- 2015

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 6.0 7.5 10.0

FIG 4. Map of Europe showing the potency of palien to release a llergen in 2009 (A), 2010 (B), and 2011 (C), depending on the palien ori gin (source). Source attribution is needed because ripening of allergen occurs at the source. Concomitantly, the SD in the corresponding years is g iven. Pollen potency is not fixed to certain regions and changes between years.

~ "' o

.!::. E ::1. "! N

" ~ 0..

" E ::1.

o ....

40 y = 0.2261 X + 1.200

r2 = 0.8153

30

I 20

10

0+-----~~----~------~----~ 20 40 60

Humidity (%) 80 100

FIG 5. Ph i p 5 as 10>PM>2.5 versus total Phi p 5 in ambient air (total, ratio in percentages) for a ll stations for all years, depending on local relative humidity, showing a llergen in the non- pollen-containing fraction of ambient air. Zero group 5 allergen values and values of less than 10 palien grains/m3 were excluded. Means :!: SDs are presented (P< .001, n = 1252).

where nonnally few polleo is detected, reached 20% of total airbome allergen. The allergen content of this fraction of air can contain smaller particles, which correlated with humidity (P < .001, Fig 5).

DISCUSSION Polleo is the major outdoor cause for allergic rhinitis, and more

than 350 poli en traps o ver Europe monitor poli en ftight dai ly.39 The immune system reacts with symptoms to allergens,40 which are carried and released by polleo. The natural variability across Europe of how much allergen is carried by grass polleo was unknown.

This study shows that, on average, grass polleo in Europe rele.:1sed about 2.0 to 2.5 pg of group 5 allergens, the major grass polleo allergen, per polleo grain. This average value has a large variation because the amount of group 5 allergen released by the same amount of polleo ranged from less than 1 pg (detection limit) to 9 pg per polleo (5th to 95th percentile) and varied 7-fold between countries and years (P <.O 1 ).

Although polleo counts are indispensable for agriculture, phenology, climate change, and health, airborne allergen adds a new dimension to understanding allergic rhinitis and asthma. Tndeed, depending on humidity, up to 20% of the allergen could be detected in the fraction of air containing smaller particles than polleo that could penetrate deeper into the lungs.

Polleo in ambient air correlated with allergic symptoms,26•40

also for grass pollen.41 However, the correlation is not clear because symptoms lag behind exposure and later in the season higher polleo counts do not evoke more symptoms. The human immune system reacts to allergens with symptoms, and allergen

J ALLERGY CUN IMMUNOL

VOLUME - . NUMBER •

release of poli en is not constant (Fig 1 and Table 1). In addition, more events with high-potency polleo were observed when poli en counts were low (Fig 2). This could explain why even low levels of2 to 9 polleo grains!m3 can result in allergic symptoms.42 Al so, at high humidity, more allergen was detected in the fraction of air containing smaller particles that can penetrate deeper into the lungs43 (Fig 5; ie, depending on humidity, a part of the allergen could penetrate deeper into the airways, approximately double at 100% relative humidity [rainy days) vs dry days with 30% relative humidity; P < .001 ). As a consequence, poli en concentration predicted allergen exposure but only with an ,2

correlation of 0.497 (P < .001, see Fig E2).

Allergen in ambient air For birch poli en, a 1 0-fold natural variation in allergen release

per polleo was published; for olive polleo, this was 12-fold. 19•20

For the sake of comparing the allergen release variability for grass with those results, we also followed the approach of those studies and compared the mean potency ofthe top 10% ofvalues with the lowest 10%. For grass, this additionally required excluding poli en without group 5 allergens and polleo with potencies of greater than 20 pg Phi p 5 per polleo. Also excluded were days with polleo counts of less than 10 polleo grains/m3 . Then the variability in potency for grass polleo was estimated to be 17-fold. With all the ambiguity of such a comparison, the range of variations among the considered taxa appeared to be the largest for grass.

A weakness of this study is that no other grass poli en allergens, particularly group 1 allergens, were determined. Grass polleo exposure involves a succession of polleo from different grass species.44 Grass polleo of different species release similar, cross­reacting group 5 allergens that are not identical (www.allergome. org, accessed August 20 13). Thus Dactylis glomerata poli en releases Dac g 5, whereas P pratense releases Phi p 5. The response factor of each Phi p 5 analogue is different, 18 as reported for other allergens, such as Phi p 1 . 17 We calibrated all samples against Phl p 5.0102, analogue to calibrating protein concentra­tion against BSA,45 because Phi p 5.0102 is the pharmacopeia standard for Phleum species preparations used in specific immunotherapy? 1 For group 5 allergens, the difference in response factor between the early ftowering major species D glomerata and the late blooming species P pratense is low, maximally about 2-fold, with Phi p 5 having the lower response factor. Differences in response factor are not sufficient to explain the observed difference in poli en potency (range, <1-9 pg ofPhl p 5 per polleo [5th to 95th percentile)). 17

•18

The ftowering succession of grass species includes, among other minor species, Alopecurus pratensis, Poa pratensis, Lolium perenne, D glomerata, and P pratense.44 A pratensis (beginning of the season) and Phragmites species (end of season) do not release Phi p 5 (and data not shown),8•

18 and indeed, at the beginning of the season, several stations detected polleo without group 5 allergens (Fig 1 and see Figs E3 and E4). For instance, the lower potency of polleo in Munich at the beginning of the grass polleo season in 201 1 (Fig 1) was not dueto a lower response factor because D glomerata ftowers before P pratense and has a higher response factor to Phl p 5.0102 but must be dueto a lower release of total group 5 allergens per poli en.

Analogous to having polleo without group 5 allergens at the beginning ofthe grass poli en season, at theend ofthe grass poli en

BUTERS ET AL 7

season in Finland, where Phragmites species (common reed) are more abundant, also polleo with no Phi p 5 is detected. In other countries the contribution to polleo counts by this species is limited. 18

'46

Mediator release In the basophil degranulation bioassay J3-hexosaminidase as a

proxy for histamine release by the particle equivalent of 1 m3 of ambient air correlated well with allergen levels per cubic meter (? = 0.80, P < .001) and poli en levels per cubic meter (? = 0.61, P < .001, Fig 3). Poli en releases more allergy­modulating factors than allergens.47 The degranulation assay also agrees with our hypothesis that polleo releases different amounts of Phi p 5, as reported previously for other polleo species. 19

•20 However, the basophil assay is not a substitute for

allergy symptoms. It should also be taken into account that the reliability of the group 5 ELTSA as a predictor of biological activity depends on the variability of the group 1/group 5 ratio. This highly significant correlation presumably indicates that the group 1/group 5 ratio in our air samples, which were all from the Munich area, is relatively constant and that the contribution of other allergens is small. We cannot exclude that polleo collected in Munich might have a different andlor more stable group 1/group 5 ratio than polleo collected in other parts ofEurope.

Potency variations in time and space Polleo grains varied in the amount of allergen they release.

However, most data stem from single samples and might not be representative because of interday, intercountry, and interyear variability of the natural sources. 17

•18

•48 In addition, allergen

levels throughout the literature were quantified with different ELTSAs, yielding variable results. Therefore we used the Development of Certified Reference Materials for Allergenic Products and Validation of Methods for Their Quantification proposed methods and reference materials, 15

•49 eliminating this

variability. Considering the potency maps and their SDs, one should

keep in mind the peculiarity of grass polleo: in different regions and at different times, it is released by different species, which are likely to have different potencies. This is corroborated by substantial variability between the monthly potency maps for each season (data not shown). In Fig 4 this temporal variability contributes to the overall seasonal SD (Fig 4, right column).

At the source, variability in poli en potency can be the result of2 ripening processes occurring simultaneously. First, pollens increase their allergen content rapidly on ripening inside the anthers just before pollination? 3 Second, when anthers have ripened sufficiently, they react to actual weather conditions, releasing pollen.50 These parallel ripening processes mean that the amount of allergen in the released polleo depends on the actual weather before and during the ftowering season, resulting in weather-dependent variations in potency at the monitoring si te.

Pollens varied more in their potency than can be explained by experimental error (Fig 2). Potency has an uncertainty stemming from both uncertainty in polleo count and uncertainty in Phi p 5 determinations. For polleo counts, the inaccuracy is poorly

8 BUTERS ET AL

known,51 and we estimated itas 5 polleo grains/m3 plus 20% of the observed value, which is similar to the inaccuracy for birch polleo counts. 19 For ELISA in this study, the inaccuracy was 20.4%, which is similar, as generally reported for ELISAs. 19

.52

Using both uncertainties, we showed that the observed potency variability was several-fold higher than what could be explained by the experimental error (Fig 2). Because we determined the amount of allergen in both fractions, PM> 1 O and 1 0>PM>2.5 of ambient air, 2 explanation are possible. Either pollens with low polleo counts release more allergen or more allergen was already present as 10>PM>2.5, which could be free allergen.

Humidity At higher relative humidity of ambient air during pollination,

more allergen was detected as 10>PM>2.5 (Fig 5). Polleo grains were detected in this fraction of air, despite their large size, because of incomplete separation by means of cascade impaction? 9

•53 At higher humidity, either more pollen was

collected as 1 0>PM>2.5 because of poli en grains changing thei r aerodynamic diameters or more allergen is present in the form of smaller particles. In contras! to other observations, which described extreme weather conditions, such as thunderstorms, when these effects appeared,24

•54

•55 we could show that by less

extreme weather conditions, such as higher humidity, more allergen, either as changed polleo or as smaller particles, was available in ambient air that could penetrate deeper into the lung, evoking more severe symptoms. As grass pollen swells and increases its aerodynamic diameter on exposure to higher humidity (A. LeMoal, unpublished observation), we think the presence of more small Phi p 5-containing particles at higher humidity is the most likely explanation. This hypothesis was also suggested by others41 and even more closely related to the results of Schappi et al48 after light rainfall condition.

The results indicate that the polleo potency is not constan! and, in addition, could be explained by a free fraction of allergen in ambient air that could depend on humidity and is more easily detected when less polleo is in the air (Figs 2 and 5).

Conclusion Pollens across Europe varied in their natural capacity to release

group 5 allergens. This is in addition to the natural variation in pollen counts. In a biological test system mediator release correlated better with ambient allergen concentrations than with pollen counts. No clear geographic pattern of poli en potency was detected. Although pollen is an excellent proxy for exposure, more can be learned from the actual monitoring of allergens in ambient air. Indeed, at higher humidity, more allergen seems to be present on smaller particles that could penetrare deeper into the airways.

The technical assistance of Danijel Kupresanin, Tngrid Weichenmeier, Christi ne Huber, Amelie Le Moai-Le Berre, Elsa Caeiro, Raq uel Ferro, Manuela Ugolotti, Malgorzata Nowak, Agata Szymanska, Lukasz Kostecki, and Satu Saaranen is greatly appreciated. We thank Roberl Gebauer for the construction of the database and Dr S. Vieths for donating the humani1-ed basophils. The excellent managemenl assistance of Annina Sorgner was pivota! for the project.

Key messages

J ALLERGY CUN IMMUNOL

- 2015

• The same amount of poDen can release variable amounts of allergen rangi.ng from less thao 1 to 9 pg of Phi p S per poDen (Stb to 9Stb percentile).

• This dift'erence varies per location in Europe and during tbe poDen season.

• At higher hum.idity, tbe amount of group S allergen not associated witb poDen increases.

REFERENCES l. Burbach GJ, Heinzcrling LM, Edenharter U, Bachcrt C. Bindslev-Jcnsen C,

Bonini S, el al. GA(2)LEN sldn test study 11: clinical rclevance of inhalan! allergcn sensitizations in Europc. Allergy 2009;64: 1507-15.

2. Sunyer J, Jarvis O, Pekkanen J, Chinn S, Janson C, Leynacrt B, et al. Gcographic variations in the effect of atopy on asthma in the Europcan Community Respiratory Health Study. J Allergy Clin lmmunol 2004;114:1033-9.

3. Jarvis O, Lucqnska C. Chinn S, Pous J, Sunyer J, Janson C. et al. Change in prcvalence of lgE sensitization and mean total lgE with age and cohort J Allergy Clin lmmunol 2005; 116:675-82.

4. Scala E, Ale.'\.~andri C, Bernardi ML, !'errara R, Pala.u.o P, Pomponi O, et al. Cross-scctional survey on immunoglobulin E rcactivity in 23,077 subjccts using an allergenic molccule-based micn)3rray detection systern. Clin Exp Allergy 2010;40:911-21.

5. Haflcnberger M, Laussmann O, Ell(.-rt U, Kalcklosch M, Langen U, Schlaud M, et al. 1 Pn.-valencc of sensitisation to acraoallergcns and food allergcns: rcsults of !he Gcrman Health lnterview and Examination Survey for AduiL~ (OEGSI)). Bundesge.~undhcitsblau Ge.~undheiL~forschung GesundhciL'\.~chutz 2013;56: 687-97.

6. Canova C, Heinrich J, Anto JM, Leynaert B, Smith M, Kucnzli N, et al. Thc inOuence of sensitisation to pollens and mould.~ on seasonal variations in asthma auacks. Eur Respir J 2013;42:935-45.

7. Abou Chakra OR, Sutra JP, Dcmey Thoma.~ E., Vinh J, Lacn)ix G, Poncet P, et al. Protoomic analysis of major and minor allergens fn)m isolated pollen cytopla.~mic granule.~. J fu)loome Res 20 12; 11: 1208-16.

8. Gangl K. Niederberger V, Valenta R. Multiple gr«-'\.~ mixes a.~ opposed to single gra.'\.~S for allergen immunotherapy in allergic rhiniti~. Clin Exp Allergy 2013; 43: 1202-16.

9. Anders.~n K, Lidholm J. Characteristics and immunobiology of gra.~s pollen allergcns. lnt Arch Allergy lmmunol 2003;130:87-107.

10. Tripodi S, l'rcdiani T, Lucarclli S, Macri 1', Pingitorc G, Di Rien.t.O Businco A, et al. Molecular profiles of lgE to Phleum pratense in childrcn with grass pollen allergy: implications for spccific immunotherapy. J Allergy Clin lmmunol 2012; 129:834-9.

11. Moverarc R, Kosunen TU, Haahtela T Change in thc pauem of lgE rcactivity to timothy gro~.'\.~ and birch pollen allergens over a 20-year pcriod. J lnve.~tig Allergol Clin lmmunol 2006;16:274-8.

12. WestriL~hnig K, Horak 1', Swoboda 1, Balic N, Spitzauer S, Kuodi M, et al. Differcnt allergcnic activity of gra.~s pollen allergens rcvealed by ~kin te.~ting. Eur J Clin lnve.~t 2008;38:260-7.

13. Zafrcd O, Nandy A, Pump L. Kahlert H, Keller W. Crystal structurc and immunologic characteri,.ation of !he major gra.'\.~ pollen allcrgen Phi p 4. J Allergy Clin lmmunol 2013;132:696-703.

14. l·licker S, Vrtala S, Steinbcrger P, Vangelista L, Bufe A, Petersen A, et al. A human monoclonal lgE antibody defines a highly allergenic fragment of thc major timothy gra.~s pollen allergcn, Phi p 5: molecular, immunological, and structural characteri.Gation of !he epitopc-containing domain. J lmmunol 2000; 165:3849-59.

15. Chapman MO, l'crrcira 1', Villalba M, Cn)mwell O, Bryan O, Bccker WM, et al. "lltc Europcan Union CREATE projcct: a model for international standardi.Gation of allergy diagnostics and vaccines. J Allergy Clin lmmunol 2008;122:882-9.

16. Bullimorc A, Swan N, Alawodc W, Skinner M. A protoomic style approach to characteri.t.e a gra_'\.~ mix product rcveals potential immunoth(.-rapcutic bcnefit. World Allergy Organ J 2011 ;4: 140-6.

17. Duffort O, Quintana J, lpscn H, BarbL-r O, Polo E Antigenic similarity arnong group 1 allergens fn)m gra.'\.~es and quantitation ELISA using monoclonal antibodies to Phi p l. lnt Arch All(.-rgy lmmunol 2008; 145:283-90.

18. Rarnircz J, Obispo TM, Ouffort O, Carpizo JA, Charnorro MJ, BarbL-r O, et al. Group 5 determination in Pooideac gro~.<s pollen extracts by monoclonal

J ALLERGY CUN IMMUNOL

VOLUME - . NUMBER •

antibody-bascd EL!SA. Correlalion wilh biologic aclivily. Allergy 1997;52: 806- 13.

19. BulersJTM, Thibaudon M, Smilh M, Kenncdy R, Rantio-Lchlimaaki A, Albcrtini

R, el al. Relca<e of Be! v 1 from birch pollen from 5 l.iuropcan counlries. ResulL• from !he H!AL!Nl.i Sludy. Almos l.inviron 2012;55:496-505.

20. Galan C, Anlunes C, Brandao R, Torres C, Garcia-Mo,_o H, Caeiro l.i, el al.

Airbome olivc pollen eounL• are no! representative of ex¡x'>Sure 10 !he major olivc allergen Ole e l. Allergy 2013;68:809- 12.

21. Millehner W. Allergy againsl horres. Are curly horse.• an allemative for horsc-allergic riders'l Allergo J 2013;22:244-5 1.

22. Bienboire-l'rosini C, Coui A, Lafonl-Lecuelle C, VL-rvloel D, Ronin C, Pageal P.

lmmunological ditTcrenee.• in !he global relea.e of !he major cal allL-rgen !'el d 1 are inRuenccd by scx and behaviour. Ve! J 2012;193:162-7.

23. Bulers JT, Weichenmeier 1, Ochs S, Pusch G, Kreyling W, Boerc AJ, el al. The

allergen Be! v 1 in fractions of arnbienl air deviále.• from bireh pollen counls. Allergy 201 0;65:850-8.

24. D' Amalo G, Cccchi L, Liccardi G. 'lltunderslorm-relalcd a.U1ma: no! only gr.tss pollen and spores. J Allergy Clin lmmunol 2008;121:537-8.

25. de Weger L., Dahl A, Bcrgmann K, Bclmonle J, Ranli<>-Lchlimaaki A, lltibaudon M, el al. lmpacl of pollen. In: Bergman CK, Sofiev M, cditors. Allergenie pollen. New York: SpringL-r; 2012.

26. Caillaud DM, Martín S, Scgala C, Be.'\ancenol JP, Clol B, Thibaudon M. Nonlincar shorHerm ctTecls of airbome Poaccac levels on hay fever symploms. J Allergy Clin lmmunol 2012;130:812-4.

27. Cccchi L. From p<)llen counllo pollen ¡x'>lency: !he molecular era of aerobiology.

l.iur Re.•pir J 20 13;42:898-900. 28. l.iuropcan Aerosol Sociely. Mínimum rcquiremenL• 10 managc acrobiological

moniloring slátions includcd in a national nelwork involved in !he l.iAN. lnlemalional Aerobiology Newsleuer 2011 ;72: l .

29. Demokrilou P, Uuplá 1; Ferguson S, Koulrakis P. Dcvelopmenl and laboratory

characlerizalion of a prololypc coarsc particle conccnlralor for inhalalion loxicological sludie.•. J Aerosol Sci 2002;33: 1111-23.

30. Bulers JT, Ka.ehe A, Weichenmeicr 1, Schober W, Klaus S , Traidl-HotTmann C, el al. Ycar-1<>-year variation in relca.e of Be! v 1 allergen from bireh pollen: evidence for geographical ditTerences bclween Wcsl and Soulh Germany. lnl Areh Allergy lmmunol 2008;145:122-30.

3 1. Yie!hs S, Barbcr D, Chapman M, Costanzo A, Daa• A, l·íebig H, el al. l.islablishmenl of recombinanl major allergens Be! v 1 and Phi p 5a a• Ph. l.iur. refcrencc standard.< and validation of l.iUSA melhods for lheir mca<uremenL ResulL• from fca<ibilily sludies. Pharmeur Bio Sci Notes 2012;2012:118-34.

32. van Oort l.i, Dieker MC, de HCL-r PG, Pcllre G, Aalbersc RC, van Ree R. The major gra'l.• p<)llen group 5 allergen from Dactylis gUJmerata and ils C-!L-rminal

splil producl bolh behave a< dimers: implications for allergen standardi.t.ation. lnl Areh Allergy lmmunol 2005;136:113-22.

33. Yogel L, Lunkopf D, Halahel L., Hauslein D, Vielhs S. Dcvelopmenl of a funclional in viln) a'l.'\ay a< a novel lOo! for !he slándardization of allergen exlracL• in !he human syslem. AllL-rgy 2005;60: 1021-8.

34. LA>rcnz RJ. Biomctric. GrundbcgriiTc U.cr Biornctric. 2nc.l cd. Stuugan: Gus:t.av

l'ischer Yerlag; 1989. 35. Livingston EH. Who wa< S!udenl and why do wc carc so much aboul his Hest'l

J Surg Res 2004; 118:58-65. 36. Prank M, Chapman DS, Bullock JM, llélmonle Soler J, Berger U, D-ahl A, cl al.

An opcrational model for forcea.Ung ragwecd p<)llen relca.e and dispcrsion in l.iuropc. Agríe l'or Me!eorol 2013;182- 183:43-53.

BUTERS ET AL 9

37. Sofiev M, Siljarno P, Ranlá H, Linkosalo T, Jaeger S, Ra•musscn A, el al. A numerical model of bireh pollen emission and dispersion in !he atmosphere. Dcscription of !he emission module. lnl J Biomcleorol 2013;57:45-58.

38. Brown HM, lrving KR. llte size and weighl of common allergenic pollens. An investigation of lheir number pcr micn)grarn and size distribulion. Acta Allergol 1973;28:132-7.

39. Berger U, Kara~~s K, Jaeger S, VoukanL•is D, Sofiev M, Brand! O, cl al. Personalizcd pollen-relalcd symplom-forcea<l informalion scrvice.• for allergic rhinitis patienL• in l.iuropc. AllL-rgy 2013;68:963-5.

40. Janlunen J, Saarinen K, Rantio-Lchlimaaki A. Allergy symploms in relalion 10 alder and bireh pollen concentrations in l'inland. Aerobio logia 20 12;28: 169-76.

4 1. Brilo 1', Mur Gimeno P, Carnes J, l'emandcz-Calda• l.i, Lara P, Alonso A, el al.

Gra<s pollen, acroallergens, and clínica! symploms in Ciudad Real, Spain. J lnve.•lig Allergol Clin lmmunol 2010;20:295-302.

42. DcllaValle Cl', Triche EW, Lcaderer BP, Bell ML. l.iffecL• of arnbienl pollen concenlr.ttions on frequency and scverily of a.rnma sympl<lms arnong a•U•malic children. Epidcmiology 2012;23:55-63.

43. lnlL-rnalional Commission on Radiological Pn'>leclion. Human respiralory !rae! model for radiological prolection, !CRP publicalion 66. New York: Pergarnon

Prcs<, Elsevier Science; 1994. 44. l'renguelli U, P'.t.•salacqua G, Bonini S, l'iocchi A, lncorvaia C, Marcucci F, el al.

Bridging allergologic and bo!ánical knowlcdge in sca<onal allergy: a role for phenology. Ann AllL-rgy AsU1ma !mm uno! 20 10; 105:223-7.

45. Pclerson GL. A simplification of !he prolein a'l.<ay melhod of Lowry cl al. which is more gener.Uly applicable. Anal Biochern Im;83:346-56.

46. Lcon-Ruiz E, Al=~ P, Dominguez-Vilches l.i, Galan C. S!udy of Poaccae phenology in a Mcditerranean climale. Which species conlribule mosl!O airbome

pollen counL<? Aerobiologia 2011 ;27:37-50. 47. Gilles S , l'ekcle A, Zhang X, Beck 1, Blume C, Ring J , cl al. Pollen me!abolome

analysis revcals adenosine a• a major regulalor of dendritic cell-primcd T(H) ccll responses. J Allergy Clin lmmunol 2011;127:454-{íl.

48. Schiippi UF, 'laylor PE, P'.Un MC, Cameron PA, Den! AW, Slaff lA, cl al.

Concenlrations of major gra.'l.• group 5 allergens in pollen grains and almospheric particles: implications for hay fever and allergic a<U1ma sufferers scns ilized 10 gra<s p<)llen allergens. Clin l.ixp Allergy 1999;29:633-41.

49. van Ree R, Chapman MD, l'errcir.t ¡.; Vielhs S, Bryan D, Cromwell O, el al. lltc CREATE projCCl: dcvclopmenl of certif.cd referencc malcrials for allergenic pmd­ucL< and validalion of mclhod.• for lhcir quantifiCálion. Allergy 2008;63:310-26.

50. 'laylor PE., l'lagan RC, Valenlá R, Glovsky MM. Release of allergens a•

respirable aL-rosols: a link belween gr.t.•s pollcn and a•lhma. J Allcrgy Clin lmmunol 2002;109:51-6.

5 1. Bergmann KC, Simolcil A, Wagener S , Mücke HG, WLTChan M, Zberbicr T. YL-rteilung von Pollen und Feinstaub in einern Sliidlischen Ballungsraum arn Bcispiel der gro6s!adlllérlin. AllL-rgo J 2013;22:471 -5.

52. King EM, l'ilep S, Smilh B, Plálls-Mills T, Harnillon RG, Schmechel D, cl al. A mulli-cenler ring tria! of allergen analysis using Ruorescenl mulliplex array

technology. J lmmunol Melhod.< 2013;387:89-95. 53. Püucring H, Lcnarcic J. DirccLivc 2008/50/I:!C uf Lhc l!un>¡xan Parliamcnt anc.l of

!he council. OOicial Joumal of !he Europcan Union 2008; 152:1-44.

54. 'laylor Pl.i, Jons.•on H. Thunderstorm a•thma. Curr Allergy Asll1ma Rep 2004;4: 409- 13.

55. Ycnables KM, Alliu U, Collier CG, Emberlin J, Ureig JB, Hardaker PJ , el al.

Thunderslorm-rela!ed a<lhma !he L-pidemic of 24/25 June 1994. Clin l.ixp Allergy 1997;27:725-36.

9.e1 BUTERS ET AL

RG E1. Map o f station locations used in HIALINE.

J ALLERGY CUN IMMUNOL

- 2015

J ALLERGY CUN IMMUNOL

VOLUME - . NUMBER •

3000

1 Q.

:E Q.

2000 ~ e

" ~ j!

< 1000

o

Potency pollen across Europe with lin. regression

200 400 600 800 1000 1200 1400 1600 1800

Grass pollen (grains/m3)

• Poacea vs Sum Phi p 5 --All Europe HIALINE Regr

RG E2. Pollen flight from all stations in 2009 until 201 1 concomitant w ith airborne allergen concentration. Pollen values o f less than 10 pollen grains/m3 were deleted for this correlation . The slope of the Pearson linear reg ression curve is g iven and represents allergen release per pollen (potency, 2.3 pg Phi p 5/pollen, ,.2= 0.495, P< .001, n = 1629).

BUTERS ET AL 9.e2

9.e3 BUTERS ET AL

0.12

0.1

-c o.os o ~

l > 0.06 u e .. :;:¡ r:r .. .z 0.04

0.02

o o "' ó

o

"1 ": "' ":' "' <;' Ñ m "' "' "1 ó ,..¡ "'

~ "' ": "' ~ "' ":' "' "' ..; "' "' "' "' m ... "' "' "' ..;

Potency (pg Phi p S 1 pollen)

1 1 1 • 1 • "' ":' "' ~ "' o ,.: có .,; ...

"' "' ,... ,.: 00 có "' "' .,;

o .... 1\

J ALLERGY CUN IMMUNOL

- 2015

FIG E3. Histogram of grass poli en potency across Euro pe. The high frequency o f low-potency po llen was dominated by pollen releasing no Phi p 5.

J ALLERGY CUN IMMUNOL

VOLUME - . NUMBER •

so -

40 -

!: 30

~ ~ .,. ! 20

10

o 1 1 1

o 0 · 1 1 ·2

so

40

~30

~ !!: .,. ! :10

10

J] ~ o •

o 0·1 1· 2

50

40

lJo 1

i:' e ~ .,. ~ 20

J -

o 11 1 1 o 0 · 1 1 · 2

BUTERS ET AL 9.e4

so --1 Austria 2010 1 Finland 2009

1 Austria 2011 1 Finland 2010

40 1 Finland 2011

* 30 1 j 1

~ .,. ~ 20 -

1-- 10 t - 1

t .• Jt 11L ~ • 1 1 _n 1 1 1 o 1 ._,_ J 2· 3 3 · 4 •·S 5·6 • · 7 7·8 8·9 9 · 10 """'

1

o 0· 1 1· 2 2·3 3·4 4 · S S·& 6·7 7 ·8 8 · 9 9· 10 More

Potency )pg Phi p S 1 pollen) Potency (pg Phi p S 1 poli en)

50

1 France 2009 1 Germany 2009 1 France 2010 1 Germany 2010 1 France 2011 40 1 Germany 2011

~30

~ !!: .,. ! 20 -

l ht. Ji hl 10 1-- n 1 • • o 1 ~llli J 1 1 L•

2 · 3 3 · • 4 · 5 5 · 6 6 · 7 7 · 8 8 · 9 9 ·W More

1

o 0· 1 1 · 2 2 · 3 3 -4 • ·S S · ~ 6 ·7 ... ... 9· 10 ..... Potcncy (pg Phi p S 1 pollcn) Potency )pg Phi p S 1 pollen)

50

l ltaly 2009 1 Poland 2009 l ltaly 2010 1 Poland 2010 l ltaly 2011 40 1 Poland 2011

~ 30 i:' e ~ .,. ~ 20

J 1

10

~ h l •.. 1 L,. o 111 h 1 M ~ • . ..

2·3 3· 4 4 · S S · & 6 · 7 7 · 8 8 · 9 9 ·10 "'~· o 0· 1 1 ·2 2 · 3 3 · 4 4 · 5 S· & • · 7 7 · 8 ... 9 · 10 More

Potency (pg Phi p S 1 pollen) Potency (pg Phi p S 1 pollen)

FIG E4. Histograms of g rass poli en potency for each observation site and year.

9.e5 BUTERS ET AL

so

40

~ 30

f ~ O'

~ 20

10

0 · 1 1 · 2 2· 3 3 · 4 4 · S S 4 , . 1

Poten<Y (pe Phi p S 1 pollen)

so

40

~30 ?: ~ O'

~ 20

10

0 · 1 1 ·2 2 · 3 3 · 4 4 · 5 5 · 6 6 · 1

Poten<Y (p¡ Phi p S 1 pollen)

so 1 • Portugal 2009 • Portugal2010 • Portugal 2011

·o I ~ 30

f ~ O' ~ 20

10

1· 8 8 · 9 9 · 10 MQt~

50

• Turkey 2010

40

~30 ?: e .. ~ O' ! 20

10

1· 8 8 · 9 9 · 10 -· FIG E4. (Continued).

Ool 1 · 2

0 · 1 1· 2

J ALLERGY CUN IMMUNOL

- 2015

• Spain 2009 • Spain 2010 • Spain 2011

~ · $ 3· 4 4 · S S · & 6 · 1 1 · 8 8 · 9 9 · 10 Mor~

Poten<Y (pe Phi p S 1 pollen)

• UK 2009 • UK 2010 • UK 2011

2 · 3 3·4 4·5 5·6 6·7 7·8

Potency (pg Phi p S 1 poli en)

J ALLERGY CUN IMMUNOL

VOLUME - . NUMBER •

TABLE E1 . Stat ion descript ions

Austria Vi en na

Fin land Turku

France Lyon 2009-2010*

Lyon 20JOt

Germany Munich

Ita! y Parma

Poland Poznan

Portugal Evom

Spain Cordoba

Turkey Bursa

Uni ted Ki ngdom Worcester

location

48. 14'56.0" North, 16•21 '22.0" East

60"27' 18.34" North, 22• 17'07.49"' East

45•42' 40.3" North, 4"29'34.6" East

45.45'09.5" North, 4"32' 17.9" East

48.09' 52.3" North, 11. 35'35.4 .. East

44•48' 16.2" North, 10"1 8'56.9" East

52"28'01.8 .. North, 16. 55'27.3" East

38"34'07.2" North, 7•54'68.9" West

37. 53'0.5" North, 4•46'45.0" West

40. 13'21.2" North, 28"51 '48.2" East

52. 19' 69.0" North, 2"24'20.0" West

ChcrnVol and HirsHypc pollcn trap wcrc localcd al lhc samc localion al hcighL< wilhin 5 m or cacto olhcr. *Saint-Gcnis 1• ArgcnLicrc. tBru.<sicu (8 krn apart).

BUTERS ET AL 9.e6

Above ground (m) Above sea level (m)

9.0 209

15 40

2 478

2 448

1.8 520

18 52

18 93

12 300

22 123

9 169

JO 43