Clinicai and Experimental Allergy, 1995, Volume 25, pages 234-239

Seasonal appearance of grass pollen allergen in natural,pauci-micronic aerosol of various size fractions.Relationship with airborne grass pollen concentration

F. TH. M. SPIEKSMA, B. H. NIKKELS and J. H. DIJKMAN

Department of Pneumology, University Hospital, Leiden, The Netherlands

Summary

In a study during the 1993 grass pollen season at Leiden, the relationship betweenatmospheric pollen allergen carried by five size fractions of pauci-micronic (fewmicrons) particles and the grass pollen count was investigated. Sampling was carriedout on dry days, and atmospheric pollen allergen in the particle fractions was assessedby a RAST-inhibition assay while grass pollen quantities were measured with avolumetric pollen trap. It appears that the atmospheric presence of grass pollenallergen in all size fractions is restricted mainly to the period of presence of grasspollen grains. Before and after the grass pollen season atmospheric grass pollenallergen quantities are generally very low. It is concluded that a routinely performedgrass pollen count is a reliable measurement for the estimation of the amount ofatmospheric grass pollen allergen, also in the pauci-micronic particle fraction.

Clinical and Experimental Allergy, Vol. 25, pp. 234-239. Submitted 7 March 1994;revised 7 June 1994; accepted 17 June 1994.

Introduction

The airborne presence of pollen allergens outside thepollen grains has been demonstrated for several pollentypes, such as ragweed {Ambrosia) [1-3], birch (Betula)[4], oak (Quercus) [5], Japanese cedar {Cryptomeria) [6],and grasses {Poaceae) [7-11]. In some of these studies,using size fractionating sampling devices, the allergenicactivity was shown to be carried by particles measuring afew microns or less, i.e. of pauci- or even sub-micronicsizes [1-6,10,11].

The nature and origin of these smaller micronic pollen-allergen carrying particles is not yet known. It has beensuggested [11] that osmotic bursts of humidified rye-grasspollen grains creates a natural aerosol of starch granulesof pauci-micronic size, carrying a specific allergen Loliumperenne {Lol p V), which is responsible for the rise inairborne grass pollen allergen following episodes of rain[7,8]. This finding implies a more or less direct relation-ship between the airborne presence of pollen grains andallergen carrying micronic particles. On the other hand,

Correspondence: Dr F. Th. M. Spieksma, Laboratory of Aerobiology.Department of Pneumology, C3-P, University Hospital, P.O. Box 9600,NL-2300 RC Leiden, The Netherlands.

for ragweed [3] and for oak [5] the presence of pollenallergen outside, particularly after the pollen season hasbeen reported, incriminating non-pollen plant parts aspossible origin of the airborne allergenic activity, just ashas been described for Parietaria and cocksfoot [12].Also, the possibility of transfer from pollen grains toenvironmental (pauci-)micronic particles, by an extractionprocess under humid conditions, has been put forward[13]. This study does not aim at an answer to thesequestions about the origin of the small particles, as itwas limited to dry weather conditions.

Worldwide, in many places, airborne quantitiesof allergenic pollen grains are monitored, to collectknowledge about levels of allergen to which pollenallergic patients are exposed to by inhalation. Becausein Europe, grass pollen is the major cause of pollmosis, itis of great clinical interest to determine the airbornepresence of grass pollen allergen in relation to theatmospheric quantities of grass pollen grains. Morespecifically, it is important to know if grass pollenallergen carried by particles other than grasspollen grains is present in the outdoor air outside thegrass pollen season, in significant amounts capable ofprovoking hay fever symptoms.

234

Seasonal appearance of grass pollen allergen 235

Methods

The methods used for measuring airborne quantities ofpollen grains and for assessing the presence of airbornegrass poilen allergen are similar to those described inextenso earlier [10]. Therefore, only a summary of thesetechniques is presented here, together with the minorchanges.

Particle sampling

Particle sampling was carried out on the roof of theLeiden University Hospital building which is five storeyshigh, in the west of the city of Leiden, in a semi-urbanenvironment, with private houses with gardens anduniversity facilities with parks, and with some waste-land.

As usual, airborne pollen-grain concentrations,expressed as the mean number of pollen grains per m'̂over periods of 24 h. were measured continuously withthe volumetric Burkard pollen trap (Rickmansworth,UK). To collect the airborne particles to be studied fortheir allergen carrying properties, a high volume airsampler (Ecotech Model 2000, Blackburn, Victoria,Australia) was used. This instrument has the samesampling characteristics as the one used in the earlierobservation [10]. To be able to detect relatively smallamounts of allergen, the running time was extended fromapproximately 8 to approximately 12 h compared withthe earlier study. Sampling with the high volume samplerwas carried out on 18 occasions, covering the grasspollen season from beginning to end, on days with dryweather to prevent the intake of rain or fog droplets andthus the wetting of the impaction filters.

A five-stage cascade impactor (Andersen InstrumentsInc., Atlanta, GA, USA) was mounted on top of the highvolume sampler. Also as described earlier, with the flowrate of the air sampler set on 1 L^m-'min"', the sizefractions containing 80% of the mass were: Stage I:>iO^m: stage II: 4-9-10/im; stage III: 2-7-4-9/im;stage IV: 1 •3-2-7/^m; and stage V: 0-6-1 3 ^m. For thisstudy the back-up filter underneath the five impactionstages was left out of consideration.

The impaction strips were cut, and placed in thecascade impactor, again as described earlier [10], fromsheets of glass fibre with hydrophilic properties (GF 50;Schleicher & Schuell. Dassel, Germany), necessary forthe extraction, incubation and elution of the allergen inthe allergen measuring procedure.

Measurement of allergen

The glass-fibre impaction strips need to be hydrophilic to

allow moistening, and thus extraction of the allergenfrom the impacted dust particles, and release of theallergen into the strip. After drying, the allergen in thestrip was incubated with specific anti-grass-pollen IgEcontaining pooled serum during a descending elutionprocedure, exactly as described earlier [10]. Detection ofgrass pollen allergenic activity was established by block-ing of anti-grass pollen IgE antibodies by the allergen inthe strip, and subsequent measuring of remaining specificIgE reactivity in the eluate by RAST ('RAST-inhibi-tion'), using Dactylis pollen allergen discs (G3; KabiPharmacia Diagnostics AB. Uppsala. Sweden). Levelsof specific anti-grass-pollen IgE antibodies in the serumpool without prior application to, and descending elutionfrom the strip (minimum inhibition), non-specific bindingof these IgE antibodies to clean, un-exposed strips (blankinhibition) and baseline, non-specific inhibition by serumwithout anti-grass pollen IgE antibodies were routinelyperformed, and used for the calculation of the percentageRAST-inhibition as a measure of the (relative) amountof aliergenic activity on the strips. Minor changes in theearlier described incubation and elution technique,which were caused by the use of a ditferent type ofimpaction glass-fibre substrate, were the application onthe strip of 50 /il (instead of 100 /J1) serum, the collectionof 1 0 ml (instead of 0 5 ml) eluate from the strip, and thelonger elution time of 6 h (instead of 60 min). To enhanceand standardize elution time and conditions, this set upof the laboratory equipment was covered by vapour tightglass containers to keep air humidity around the strips ata constantly high level.

Results

Pollen quantities

Tn the Netherlands, the 1993 grass pollen season, asassessed with the pollen trap at the University HospitalLeiden, was characterized by an early start from mid-May, and by an extraordinarily high and rather earlypeak period during the first 11 days of June (Fig. la). Thesecond half of June showed relatively low airborne grasspollen concentrations, whereas in July, August, andSeptember the levels were decreasing and low, as isusual for the west European grass pollen season.

Pollen grains were also collected by the high volumeair sampler which was located at 2 m distance from thepollen trap. The airborne grass pollen concentrations(numbers of grains per m"^) during the operation of thehigh volume sampler were derived from the trap mea-surements, and served for the calculation of the quantityof pollen grains collected by the high volume sampler.Sampling was performed on 18 occasions between the

236 F. Th. M. Spieksma, B. H. Nikkels and J. H. Dijkman

200 000

100 000

0

(f)<

April

300

250

200

150

100

50

Moy June July August September

0

(c)

1 — IApril May June July August September

Fig. 1. Seasonal appearance of airbornegrass pollen and grass pollen allergen, atLeiden, in 1993. (a). Daily airborne grasspollen concentration, (b). Calculatedquantities ofgrass-pollen grains collectedby the high volume sampler, on 18occasions, before, during and after thegrass pollen season, (c). Sums of RAST-inhibition percentages as measure forrelative amounts of grass pollen allergenin five size fractions (I V) of atmosphericdust, presented as stacked bars. n. StageI. • . Stage n, ^ . Stage III. ^. Stage IV.a. Stage V.

end of April and of September, with intervals of 3-7 daysduring the main grass pollen season, and of 12-20 daysafter the main season. The sampling times and durations,and the calculated quantities of collected pollen grainsare given in the left part of Table 1. For this calculation,that time area of the pollen-sampler slide is countedwhich corresponds with the operation time of the highvolume air sampler, and then the pollen count is multi-plied by the volume of air (m )̂ sampled. These quantitiesof grass pollen grains are also shown in Fig. l(b). It is

obvious that these quantities, collected by the highvolume sampler during a part of the day, do notalways have the same ratio to the average airbornegrass-pollen concentrations during 24 h. But generallythere appears to be a good relationship (r ^ +0-89)between the amounts of grass pollen grains collected bythe pollen trap and those collected by the high volumesampler. Due to their aerodynamic size, grass pollengrains will be accumulated selectively in the top stage 1of the cascade impactor [10].

Seasonal appearance of grass pollen allergen

Airborne allergen

The relative amounts of grass pollen allergen in theattnospheric dust material collected in the five impac-tion stages of the cascade impactor, measured as RAST-inhibition percentages, are given in the right half of TableI, and shown in Fig. lc. To be able to show the results ofthe assessment of pollen-grain quantities and relativeallergen amounts on an equal time scale, the sum ofthe RAST-inhibition percentages of the five stagesis presented in stacked bars, so that the theoreticalmaximum values of 100 may be exceeded. From theresults shown in Fig. I it becomes clear thatthe amount of atmospheric grass-pollen allergen is verysimilar to the airborne grass-pollen concentration. Thisappears to be true not only for the bigger particle dustfraction of stage I which contains the pollen grains, butthere is also a strong correlation between the airbornepollen concentration and the amount of allergen in thedust of the pollen free pauci-micronic size fractionscollected in stages II to V of the cascade impactor. Thecoefficients of correlation between the quantities of

collected pollen grains and the amount of allergen inthe five dust fractions are given at the bottom of Table I.Figure 2 shows two examples of these correlations, forstage I, containing the pollen grains; and for stage IVwith small, pauci-micronic particles of 1-3-2-7 ^m. Forthe other stages the results are very similar. It is also clearfrom Fig. 1 that before and after the main grass pollenseason, in April, and in August and September,atmospheric grass pollen allergen is not completelyabsent on days with dry weather conditions.

Discussion

The results of the observations on the airborne presenceof grass pollen allergenic activity presented here, confirmand corroborate the earher findings with only twosamplings [10]. One of the questions raised in that earlierdiscussion was whether a relationship exists betweenthe airborne allergen levels and the airborne pollenconcentration.

The aim of this study was to investigate the seasonal

Table I. Dates and durations of high volume air samplings, calculated quantities of collected grass pollen grains, and RAST-inhibition percentages as measure for relative amounts of grass pollen allergen in five size fractions of atmospheric dust, with thecoefficients of correlation (r) between the quantities of pollen grains and RAST-inhibition percentages

Sampling

12

45fl

789

10II12131415161718

Correlation

Date 1993

28/29 April11 May17 May25 May29 May

1 June5 June9 June

17 June22 June29 June7 July8 July

22 July2 August

19 August28 August29 Sept.

coefficient (r):Level of significance (/*);

Duration ofsampling(h: min)

18:2011:3010:457:10

11:4011:3010:2011:4511:1012:1512:4511:5511:4012:2511:4011:4511:0511:55

Calculated quantityof collected

pollen grains

6600148509405044 5508085079200

292050207 9008085033000

1534501650033000247501980099003 3001650

I>10

to192522434553474832361116312055

13

+0-78<0001

RAST-inhibition percentages in dustcascade impactor-stagcs

II49-10

92243374839595753284027162124

16

13

+0 81<000I

III2-7-49

716234046374656362041134

14604n

+079<0 001

IV13-2 7

6108

4424334748281333111894038

+0 79<0001

in five

V0-6-13

52115442543484728113776000

3020

+0 69<0002

238 F. Th. M. Spieksma, B. H. Nikkels and J. H. Dijkman

50

40

30

20

:? 10

tion

(

(a)

-

-

•- • •

•

• •

1

•

• -••

•

1 1 1 1

60 r

501-

4 0 -

20

10

b)

50 100 150

xiOOO pg

200 250 300

Fig. 2. Two examples of the relationship between RAST-inhibition percentages as measure for relative amounts ofgrass pollen allergen in atmospheric dust collected on stage I(particle sizes > 10 ^m) and on stage IV (particle sizes 1 3 -2-7 ̂ m) of a cascade impactor, and different quantities ofcollected grass pollen grains (p.g.), on 18 occasions before,during and after the grass pollen season, (a) Stage I, r = +0-78.(b) Stage IV, r = +0 79.

course of the atmospheric grass pollen allergen amountin relation to the airborne grass pollen concentration, inorder to judge the value of the routinely performed,so-called pollen count as a measure for atmosphericpollen-allergen amounts. It is not surprising thai such arelationship {r = +0-78) exists for the dust fractioncollected in stage I of the cascade impaction, where thepollen grains are accumulated (Fig. 2). But it is note-worthy that these same high levels of relationship alsoexist between quantities of pollen grains and allergenamounts in the pollen-free, pauci-micronic fractionscollected in stages II to V. For a biological phenom-

enon, the relationship between quantities of airbornepollen grains and atmospheric pollen allergen amountsin all size fractions is remarkably high, and the similarityin the seasonal courses (Fig. 1) is also rather great. This isillustrated, for instance, by the simultaneous drop in botlithe airborne pollen concentration and the atmosphericallergen amount on 22 June (sampling nr. 10).

Shortly before, and long after the main grass pollenseason, during sampling nr. 1 on the night of 28-29April, and more evidently during sampling nrs 16, 17 and18 in August and September, atmospheric grass pollenallergen is not completely absent, but the level is ratherlow in all size fractions (except in stage V, after Iheseason). These low levels of natural allergen exposure areprobably too low to provoke symptoms of pollinosis.The conclusion is that, within an acceptable range ofbiological variation, the routinely performed assessmentof airborne grass pollen concentration is a reliablemeasurement of the amount of atmospheric grasspollen allergen. This also holds for the smaller size,pauci-micronic dust fraction, which is thought to becapable of penetrating into the lower regions of therespiratory tract [10].

This conclusion is not in agreement with those ofFernandez-Caldas et al. [5] for atmospheric oak-pollenallergen, and of Rantio-Lehtimaki et al. [4] for birch-pollen allergen. Possibly, the difference in the conclusionis caused by the difference in the length of the periodsconsidered, and with the amount of variations in boththe pollen counts and allergen assessment which isaccepted as not essential. It is true that both for thegrasses and the birches the relationship between atmos-pheric pollen concentrations and allergen amountsduring the pollen season is not complete on everyindividual day, and big aberrations are sometimesobserved. It is also true that both for the grasses andthe oaks there are pollen-allergen carrying particles in theatmosphere after the pollen season is over. But the mainconclusion is that high atmospheric pollen allergen levelsin all the size fractions are observed only within thepollen season itself, and predominantly on days with ahigh pollen count.

In the observations presented here, no special attemptwas made to collect atmospheric dust immediately afteran episode of rain, under which condition other authors[7,8,11] found high levels of allergen under low pollen-concentration conditions. So, this potential controversycontinues. The question of the nature and origin of thepauci-micronic allergen carrying particles, particulariyduring dry weather conditions, cannot be answered bythe observations presented here. The three main hypoth-eses are: (i) allergen carriage by starch granules releasedin the atmosphere after osmotic rupture of pollen grains

Seasonal appearance of grass pollen allergen 239

[11]; (ii) airborne presence of non-pollen plant partscarrying small amounts of pollen allergen [3,12]; (iii)transfer of allergen from pollen grains to natural dustby some extraction and dispersion process [13]. Theturbulent air in the lower layers of the atmosphere isalways loaded with particles, mostly of pauci-micronicand sub-micronic sizes. Apparently., a portion of theseparticles is carrying pollen allergen particularly duringthe main pollen season. In the period after the mainpollen season., when airborne pollen concentrationsbecome low, this portion of allergen carrying pauci-micronic particles also becomes low, but remainspresent for some time. The fact that the allergen isiipparently distributed over particles of several sizefractions makes the third hypothesis of transfer tonatural dust more likely than atmospheric carriage bymore specific non-pollen plant parts or starch granules,which have a narrower range of pauci-micronic sizes.

Acknowledgements

The gift by Kabi Pharmacia Nederland bv of reagentsfor the RAST-inhibition procedure is gratefullyacknowledged. The authors thank Dr P. S. Hiemstrafor critical comments, Mr P. van Noort for providing thedrawings., and Mrs C. C Zuiderduin for the preparationof the manuscript.

References

1 Solomon WR, Burge HA, Muilenberg ML. Allergencarriage by atmospheric aerosol. I. Ragweed pollendeterminants in smaller micronic fractions. J Allergy CiinImmunol 1983; 72:443-7.

2 Habcnicht HA, Burge HA, Muilenberg ML, Solomon WR.Allergen carriage by atmospheric aerosol. II. Ragweed

pollen determinants in submicronic atmospheric fractions.J Allergy Clin Immunol 1984; 74:64-7.

3 Agarwal MK, Swanson MC, Reed CE, Yunginger JW.Airborne ragweed allergens: association with variousparticle sizes and short ragweed plant parts. J AllergyClin Immunol 1984; 74:687 93.

4 Rantio-Lehtimaki A. Viander M, Koivikko A. Airbornebirch pollen antigens in different particle sizes. Clin ExpAllergy 1994; 24:23 8.

5 Fernandez-Caldas E, Swanson MC. Pravda J. Welsh P,Yunginger JW. Reed CE. Immunochemical demonstrationof red oak pollen aeroallergens outside the oak pollinationseason. Grana 1989; 28:205-9.

6 Takahashi Y. Sakaguchi M, Inouye S. Miyazawa H,Imaoka K. Katagiri S. Existence of exinc-iVec airborneallergen particles of Japanese cedar (Cryptomeria japonica)pollen. Allergy 1991; 46:588-93.

7 Stewart GA, Holt PG. Submicronic airborne allergens.MedJ Aust 1985; 143:426-7.

8 Schumacher MJ, Griffith RD. O'Rourke MK. Recognitionof pollen and other particulale aeroantigens by immunoblotmicroscopy. J Allergy Clin Immunol 1988; 82:608 16.

9 Jensen J, Poulsen LK. Mygind K, Weeke ER. Wecke B.Immunochemical estimations of allergenic activities fromoutdoor aero-allergens, collected by a high-volume airsampler. Allergy 1989; 44:52 9.

10 Spieksma FThM, Kramps JA, van der Linden AC et al.Evidence of grass-pollen allergenic activity in the smallermicronic atmospheric aerosol fraction. Clin Exp Allergy1990; 20:273-80.

11 Suphioglu C. Singh MB, Taylor Ph et al. Mechanism ofgrass-pollen-induced asthma. Lancet 1992; 339:569-72.

12 D'Amato G. De Palma R, Verga A et al. Antigcnic activityof nonpollen parts (leaves and stems) of allergenic plants(Parietaria judaica and Dactylis glomerata). Ann Allergy1991;67:42I 4.

13 Behrendt H, Becker WM, Friedrichs KH. Darsow U,Tomingas R. Interaction between aero-allergens and air-borne particulate matter. Int Arch Allergy Immunol 1992;99:425-8.