Download as PDF here

Development of in vitro models for
hazard evaluation of dust from
the indoor environment.
Correlation to organic parameters and
building related symptoms.
Ph.D.-Thesis by
Leila Allermann
National Institute of Occupational Health
Arbejdsmiljøinstituttet

Preface
This Ph.D.-project was carried out at the National Institute of Occupational Health (AMI) in
collaboration with the “The Copenhagen school study” with Harald Meyer from the
Department of Occupational and Environmental Medicine (AMK), Bispebjerg hospital as
co-ordinator. The parties involved in this project are AMK, AMI, Danish Building Research
Institute (SBI) and ALK Abelló. The epidemiological part of the project and gathering of
data from the other parties was managed by AMK. AMI performed the microbiological
investigations and development of in vitro methods for evaluation of the inflammatory
potential of dust from the indoor environment. The technical investigation of the buildings
was made by SBI, and ALK Abelló tested the dust for content of different allergens.
“The Copenhagen school study” has fostered two Ph.D. projects. Harald Meyer from AMK
has on the basis of the epidemiological part and data from the above parties written his thesis
entitled: “Copenhagen school study. A study of indoor air quality” (Meyer 2000). The
second theses, the one that you are about to read, therefore contain many references to the
thesis of Harald Meyer. Both projects are matriculated at the Faculty of Health Sciences,
University of Copenhagen.
Regarding guidance and technical assistance in connection to sampling of dust I would like
to thank environmental consultant Jan Bach Nielsen and senior researcher Ole Valbjørn from
SBI, and head of department Thomas Schneider from the department of indoor climate,
AMI. Laboratory technician Esben Kjær Sørensen AMI has been a very appreciated help in
the process of sampling the dust from the 20 schools. In the laboratory I would like to give a
special thank to laboratory technician Anne Karin Jensen for good technical assistance and
for helping me with stimulation of the cells. In total about 350 in vitro experiments were
performed. Also technicians Birgitte Korsholm and Lourdes Petersen I thank for technical
guidance in the laboratories. Microbiologist Helle Würtz, laboratory technician Mirella
Simkus and laboratory technician Charlotte Lohmann Kristensen has performed the
microbiological analysis and are also to be thanked for guidance in interpretation of the
results. Statistician Erik Holst, AMI I thank for good assistance and patience in the statistical
analysis.
1

A special thank I owe to Ph.D. student Harald Meyer, AMK for good and effective
collaboration between our two projects.
I thank my supervisor from AMI, research director Otto Melchior Poulsen for giving me the
opportunity for performing my Ph.D. project here on AMI and for always supporting me and
leading me in the right direction. Professor Finn Gyntelberg my supervisor from AMK I also
thank for excellent support and guidance.
Specials thank to my fellow students on AMI, Lone Borg, Susanne Clausen, Marianne
Dybdahl, Søren Thor Larsen and Andrea Wilks for cosy and inspiring moments, at and out
side AMI.
Grants from The Danish Research Academy (Forskerakadamiet) and The Danish Working
Environment Fund (Arbejdsmiljøfondet) supported this Ph.D.-study.
Preliminary results of the present study have been published in the following papers:
I. Hansen, L.A., Poulsen, O.M. and Nexø, B.A. (1997) Inflammatory Potential of Organic
Dust Components and Chemicals Measured by IL-8 Secretion From Human Epithelial
Cell Line A549 In Vitro. Ann Agric Environ Med 4, 27.
II. Hansen, L.A., Poulsen, O.M. and Würtz, H. (1999) Endtoxin potency in the A549 lung
epithelial bioassay and the limulus amebocyte lysate assay. J Immunol Methods 226,
49.
III. Allermann, L. and Poulsen, O. M. (1999) Correlation between the In Vitro inflammatory
potential and microbial parameters in dust from indoor environments. Indoor Air 99 2,
719-724. Edinbourgh, Scotland, Indoor Air 99. (Conference Proceeding).
IV. Meyer, H. W., Allermann, L., Nielsen, J. B., Hansen, M. Ø., Gravesen, S., Nielsen, P.
A., Skov, P., and Gyntelberg, F. Building conditions and building-related symptoms in
the Copenhagen school study. Indoor Air 99 2, 298-299. 1999. Edinbourgh, Scotland,
Indoor Air 99. (Conference Proceeding).
Reprints are enclosed at the end of this thesis. In addition, we have published preliminary
studies and use of in vitro bioassays for evaluation of the inflammatory potential of dust in:
Allermann, L. and Poulsen, O. M. (2000). Inflammatory potential of dust samples from
waste handling facilities measured as IL-8 secretion from lung epithelial cells in vitro.
Ann Occup Hyg, 44, 259-269.
2

Hansen, L. A., Nexø, B. A., Borg, L., and Poulsen, O. M. (1997). Inflammatory potential of
airbornr microorganisms. Phase1: Development of a measurement method (UK title).
Luftbårne mikroorganismers evne til at udløse betændelse. Fase 1: Udvikling af en
målemetode. (DK title). Arbejdsmiljøfondet, Copenhagen. 1-36 ISBN: 87-7359-838-0
(Report in Danish, with English summary).
At last I would like to thank my husband and my family for their support.
Leila Allermann, March 2001
3

Content
Abbreviations............................................................................................................................................ 7
Summary ................................................................................................................................................... 8
Dansk sammendrag................................................................................................................................ 12
1. Objectives............................................................................................................................................ 16
2. Background......................................................................................................................................... 19
2.1. Building related symptoms related to indoor air quality. ............................................................. 19
2.2. Contaminants in the indoor environment...................................................................................... 22
2.2.1. Microorganisms and their indoor air relevant constituents.................................................... 23
2.2.2. Volatile organic compounds (VOCs) .................................................................................... 27
2.2.3. Dust........................................................................................................................................ 27
2.2.4. Measurements of inflammatory reactions of dust.................................................................. 31
2.3. In vitro models and cytokines........................................................................................................ 32
2.3.1. Cytokines............................................................................................................................... 33
2.3.2. Interleukin 8........................................................................................................................... 34
2.3.3. Airway inflammation............................................................................................................. 36
3. Materials and methods....................................................................................................................... 38
3.1. The school investigation................................................................................................................ 38
3.2. Sampling and handling of dust...................................................................................................... 39
3.2.1. Floor dust............................................................................................................................... 39
3.2.2. Surface dust ........................................................................................................................... 41
3.2.3. Ventilation shafts................................................................................................................... 41
3.2.4. Sieving and fractionation of the dust ..................................................................................... 42
3.3. Test compounds............................................................................................................................. 44
3.3.1. Lipopolysaccharides (LPS).................................................................................................... 44
3.3.2. Glucans .................................................................................................................................. 44
3.3.3. Pure compounds .................................................................................................................... 45
3.3.4. Cytokines............................................................................................................................... 45
3.3.5. Surfactants ............................................................................................................................. 45
3.3.6. Dust........................................................................................................................................ 45
3.4. A549 bioassay ............................................................................................................................... 46
3.5. Monocyte assays (U937 and THP-1 bioassay) ............................................................................. 47
3.6. ELISA ............................................................................................................................................ 48
3.7. Microbiological parameters ......................................................................................................... 49
3.8. Allergens ....................................................................................................................................... 50
3.9. Organic content of the dust samples ............................................................................................. 50
3.10. Statistical models and methods................................................................................................... 50
3.11. Quality control and method evaluation....................................................................................... 51
4

3.11.1. Bioassays ............................................................................................................................. 51
3.11.2. Microbiological parameters ................................................................................................. 52
4. Development of in vitro methods ...................................................................................................... 53
4.1. Results from the A549 bioassay .................................................................................................... 56
4.1.1. Growth experiment ................................................................................................................ 56
4.1.2. Quality control....................................................................................................................... 56
4.1.3. Endotoxins............................................................................................................................. 56
4.1.4. Glucans .................................................................................................................................. 58
4.1.5. Chemical compounds............................................................................................................. 59
4.1.6. Surfactants ............................................................................................................................. 61
4.2. Results from the U937 bioassay.................................................................................................... 62
4.2.1. Growth experiment ................................................................................................................ 62
4.2.2. Quality control....................................................................................................................... 62
4.2.3. Endotoxins............................................................................................................................. 64
4.2.4. Glucans .................................................................................................................................. 64
4.2.5. Chemical compounds............................................................................................................. 64
4.2.6. Surfactants ............................................................................................................................. 65
4.3. Results from the THP-1 bioassay.................................................................................................. 65
4.3.1. Growth experiment ................................................................................................................ 65
4.3.2. Quality control....................................................................................................................... 66
4.3.3. Endotoxins............................................................................................................................. 67
4.3.4. Glucans .................................................................................................................................. 68
4.3.5. Chemical compounds............................................................................................................. 68
4.3.6. Surfactants ............................................................................................................................. 69
4.4. Discussion of development of in vitro methods............................................................................. 69
4.4.1. Growth experiment ................................................................................................................ 69
4.4.2. Quality control....................................................................................................................... 70
4.4.3. Pure test substances ............................................................................................................... 76
5. Dust samples from schools tested in the in vitro methods............................................................... 80
5.1. Results from dust collection .......................................................................................................... 80
5.2. Results from dust tested in the in vitro assays............................................................................... 81
5.2.1. Dust tested in the A549 bioassay........................................................................................... 81
5.2.2. Dust tested in the U937 bioassay........................................................................................... 84
5.2.3. Dust tested in the THP-1 bioassay......................................................................................... 85
5.3. Discussion of test of dust samples in the in vitro methods............................................................ 87
5.4. Evaluation of the A549 bioassay................................................................................................... 91
5.4.1. Estimation of cut-off values – An example ........................................................................... 91
6. Correlation of results from in vitro methods with parameters in the dust.................................... 94
6.1. Correlation to content of organic fraction.................................................................................... 94
6.1.1. A549 bioassay........................................................................................................................ 94
6.1.2. U937 bioassay........................................................................................................................ 96
6.1.3. THP-1 bioassay...................................................................................................................... 96
6.2. Correlation to microbiological parameters..................................................................................96
6.2.1. A549 bioassay........................................................................................................................ 97
6.2.2. U937 bioassay........................................................................................................................ 98
5

6.2.3. THP-1 bioassay...................................................................................................................... 98
6.3. Correlation to content of allergens............................................................................................... 98
6.3.1. A549 bioassay........................................................................................................................ 99
6.3.2. U937 bioassay........................................................................................................................ 99
6.3.3. THP-1 bioassay...................................................................................................................... 99
6.4. Discussion of correlation of in vitro results to parameters in the dust......................................... 99
7. Correlation of results from the in vitro methods with epidemiological data............................... 106
7.1. Correlation of results to parameters of the technical investigation and selected
epidemiological data.......................................................................................................................... 106
7.2. Discussion of correlations to parameters of the technical investigation and selected
epidemiological data.......................................................................................................................... 108
8. Conclusion......................................................................................................................................... 112
9. References ......................................................................................................................................... 115
10. Appendix for protocols and raw data (in Danish)....................................................................... 127
11. Papers ....................................................................................................................................................
6

Abbreviations
AMI Arbejdsmiljøinstituttet (National
Institute of Occupational Health)
AMK Arbejds- og miljømedicinsk
klinik (Department of
Occupational and Environmental
Medicine)
ATCC American Type Culture
Collection
BAL Broncho alveolar lavage
BRS Building related symptoms
BRI Building related illness
CNS Central nervous system
DG 18 dichloran glycerol 18
ELISA Enzyme linked immunosorbent
assay
GM-CSF Granulocyt macrophage colony
stimulating factor
HVS-3 Dust sampler for sampling on
floors
ICAM Intracellular adhesion molecule
(Ig superfamily)
IL Interleukin
LAL assay Limulus amebocyte lysate assay
LAS Linear alkylbenzene sulfonates
7
IFN-γ Interferon gamma
LOD Limit of detection
LPS Lipopolysaccharide
MCP Monocyte chemotatic protein
MNC Mononuclear cells
MMA Methyl metacrylate
MVOC VOCs derived from
microorganisms
ODTS Organic dust toxic syndrome
PF Potency factor
SBI Statens
Byggeforskningsinstitut
(Danish Building Research
Institute)
SBS Sick building syndrome
SDBS Sodium dodecyl benzene
sulfonate (Dodecylbenzene
sulfonic acid)
SDS Sodium dodecyl sulfonate
TGF Tumor growth factor
TNF Tumor necrosis factor
TVOC Total amount of VOCs
VOC Volatile organic compound
WHO World Health Organisation

Summary
Development of in vitro models for hazard evaluation of dust from the indoor
environment. Correlation to organic parameters and building related symptoms.
Background
Building related symptoms (BRS) are characterised as non-specific complaints such as
irritations of the mucus membrane in the eyes, nose and throat, headache, dizziness, fatigue
and lethargy, lack of concentration, cough, and dryness of the skin. Chemical, biological,
physical, and psychosocial factors have been related to the BRS, and symptoms often have a
multifactorial cause. Also gender, job satisfaction, job category, job function, work stress,
and degree of crowding are important factors related to BRS. Illnesses as hay fever and
asthma are specific clinical conditions caused by specific exposures from the indoor
environment, also called building related illnesses (BRI). Such specific illnesses must be
distinguished from the BRS.
Contaminants in the indoor environment can be derived from different sources. A special
focus has been placed on biologic contaminants such as bacteria, fungi and components from
these organisms e.g. endotoxins and glucans. Also dust mites, hair and dander from animals
can contribute to this pool of contaminants. Organic dust is characterised as unspecific
samples of both living and dead organic material of vegetable, animal and microbial origin.
The composition of dust varies according to the environment, even within a building the
composition can vary. It is therefore difficult to point out one or just a few chemical factors
or components of microbial origin that may cause or ad to the multitude of symptoms in the
indoor environment.
Inflammation is common to many of the symptoms and illnesses related to bad indoor
climate. Inflammation could therefore be considered an overall parameter that integrates the
effect of the total exposure load from an indoor environment. Interleukin-8 (IL-8) belonging
to the proinflammatory cytokines together with the interleukins IL-1, IL-6 and tumor
necrosis factor (TNF) has chemotactic and activating properties especially on neutrophils.
The epithelial cells and the macrophages constitute the first line of defence against
8

xenobiotics in the lung. When inflammatory agents, such as endotoxins are inhaled a cascade
of proinflammatory cytokines are secreted, which initiate the inflammatory process.
The purpose of this Ph.D.-project is to develop and use in vitro methods for measuring the
inflammatory potential of organic dust from the indoor environments in public schools.
Furthermore to correlate the obtained inflammatory potentials to biological constituents
found in the dust, and to BRS and BRI found via questionnaires given to workers and
students at 20 schools in Copenhagen.
Materials and methods
The school investigation was performed as an epidemiologic cross sectional study
investigating both the buildings of 75 schools, and their employees and students from the 8 th
grade and up. 20 schools were selected: 10 schools with the lowest prevalence of BRS
(“good”) and 10 schools with the highest prevalence of BRS (“bad”). From these 20 schools
dust samples from the floor, the horizontal surfaces and the ventilation shafts were collected.
Floor dust was sampled with the HVS-3 sampler connected to a normal “Miele” vacuum
cleaner. Collection of surface dust and dust from exhaust ducts was made by a portable
vacuum cleaner with a special nozzle. The dust samples were sieved on a 300 µm filter and
stored at
–20 °C.
Three in vitro bioassays were developed and characterised with biological and chemical
agents. Four different lipopolysaccharides (LPS) from Escherichia coli O55:B5, Klebsiella
pneumoniae, Pseudomonas aeruginosa and Salmonella enteritidis. Glucans from bakers
yeast, Saccharomyces cerevisiae and a Gram-negative bacteria Alcaligenes faecalis.
Chemicals with allergenic or irritative effect such as nickel sulphate, methyl metacrylate, and
formaldehyde. Four surfactants: dodecylbenzene sulfonic acid, sodium dodecyl sulphate,
coconut oil and genapol X-80. Cytokines as tumor necrosis factor-α (TNF-α) and granulocyt
macrophage colony stimulating factor (GM-CSF).
The cells (human lung epithelial cells A549, and human monocytic cells U937 or THP-1)
were grown in 24 well plates and incubated with the test substance for 24 hours at 36.5°C
9

and 5% CO2. The media were harvested and the concentration of IL-8 measured by ELISA.
As positive controls LPS (E. coli) and TNF-α were used.
The concentration of endotoxin in the dust sample was measured by the standardised
Limulus Amebocyte Lysate (LAL) assay. Viable bacteria and fungi were cultivated for 7
days at 25 °C. The content of mite, cat and dog allergen was measured by ELISA techniques.
The content of organic matter in the dust samples was measured by incineration.
The potency factor (PF) was calculated from the slope of the initial linear part of the dose
response curves, i.e. the released IL-8 versus the concentration of dust. To reduce day-to-day
variation in the bioassays the PF is corrected against the value of a positive control.
Results and discussion
Three in vitro methods with the cell lines A549, U937 and THP-1 were developed and
evaluated with the biological and chemical agents described above. Method evaluation was
performed on all three bioassays and revealed that the methods were in control especially
when correcting against one of the positive controls. The bioassays with the mononuclear
cells (U937 and THP-1) were much more sensitive than the epithelial cell line (A549), when
pure chemicals were applied.
A total of 158 dust samples were collected from the 20 schools. A significant positive
correlation was found between the PF of all the dust samples tested in the A549 bioassay and
the content of organic matter in the sample. PF of the total dust from mechanical ventilation
ducts was found to be higher than PF from naturally ventilated ducts. A positive correlation
between the mean PF of the floor samples and the PF from the surface dust samples tested in
the A549 bioassay was observed, indicating that these samples reflect each other. Tested in
this assay the PF for dust samples from a “good” school were significantly lower than the PF
of a “bad” school.
Correlations to biological parameters in the dust and to symptoms of employees and students
were only made for the A549 bioassay. The PF of surface dust and the concentration of
endotoxin in the sample were significantly positively correlated. However, since the
measured endotoxin concentrations were below the limit of detection of the bioassays the
10

correlation may indicate that endotoxin could be a marker for other biological exposures. A
significant difference between the content of the single microorganism in the three types of
dust samples (the surface dust, floor dust and dust from the exhaust ducts) was found.
Significant correlation was found between the content of viable bacteria in the floor dust and
the PF of floor dust and between all tested microorganisms and the PF of dust from exhaust
ducts. A correlation was found between mite, cat and dog allergen in floor dust samples and
the PF, and between cat and dog allergens in dust from the exhaust ducts and the PF. The
allergen contamination was generally considered as low, and below the limit (2000 ng/g) for
mediating an IgE response. Correcting the PF against the organic fraction did generally not
result in better correlations with the microbiological data or with allergens, indicating that
none of the tested parameters are dominating factors in the organic fraction of the dust, and
the potency may be multifactorial in origin.
Buildings with flat roof had a significant higher PF than buildings with a pitched roof. And
less volume per person gave a higher PF of the dust. Positive associations between the PF of
floor dust and symptoms of itching of the eyes, index of positive symptoms of at least two
out of five symptoms of the mucous membrane and the skin, and to headache were found.
The PF of floor dust did not correlate with asthma or hay fever.
Using cut-off values CO1 = 2.0 and CO2 = 4.5 ng IL-8/mg dust for the PF of floor dust
tested in the A549 bioassay, 78% of the schools were ”good” in the area below CO1 and
22% were ”bad”. In the area above CO2 9% of the schools were ”gode” and 91% were
”bad”. 54% of the samples from “good” schools were also found below CO1 and 43% of the
samples from “bad” schools were found above CO2. Symptoms like headache and itching
eyes were found more often in schools above the CO2 than below. The “good” schools were
significantly found below the CO2 and the “bad” schools above.
Conclusion
The A549 bioassay seems to be able to differentiate clearly between the “good” and the
“bad” schools. Therefore this bioassay could be a useful screening tool in the evaluation of
indoor air problems. However, a study of the predictive value of the bioassay will be
valuable to estimate the real potential of the assay. The THP-1 bioassay needs to be
evaluated further.
11

Dansk sammendrag
Udvikling af cellemetoder til at bedømme skadelige egenskaber af støv fra indeklimaet,
og sammenhæng med organiske parametre og bygnings relaterede symptomer.
Baggrund
Bygnings relaterede symptomer (BRS) er karakteriseret som uspecifikke klager så som
irritationer i slimhinder i øjne, næse og hals, hovedpine, svimmelhed, træthed og sløvhed,
koncentrationsbesvær, hoste og tør hud. BRS menes at være multifaktoriel i sin
årsagssammenhæng. Kemiske, biologiske, fysiske og psykosocial faktorer er alle associeret
med BRS. Også køn, job kategori og funktion, tilfredshed med arbejdet, arbejdspres og antal
personer pr. m 2 er vigtige faktorer med indflydelse på BRS. Bygnings relaterede sygdomme
(building related ilnesses, BRI) er specifikke sygdomme, f. eks. høfeber og astma, som kan
skyldes specifikke eksponeringer fra indeklimaet. Det er vigtigt at disse specifikke
sygdomme skelnes fra BRS.
Der er forskellige kilder til forurening af indeklimaet. Speciel fokus har ligget på de
biologiske faktorer som bakterier og svampe, samt komponenter herfra som endotoxiner og
glucaner. Støvmider, hår og skel fra dyr kan også bidrage til puljen af forureningskilder.
Organisk støv er karakteriseret som en uspecifik samling af både levende og dødt materiale
stammende fra planter, dyr og mikroorganismer. Sammensætningen af disse faktorer i støvet
kan variere afhængig af omgivelserne, selv inden for samme bygning kan der være forskel.
Derfor kan det være vanskeligt at pege på en, eller blot få kemiske faktorer eller mikrobielle
komponenter som årsag til denne vifte af symptomer, der kan forårsages af påvirkninger i
indeklimaet.
Betændelsesreaktionen, også kaldet inflammation, observeres ofte i forbindelse med de
symptomer som findes hos personer med BRS. Man kan derfor betragte inflammation som
en general parameter, der integrerer effekten af den totale eksponering fra indeklimaet.
Interleukin-8 (IL-8) hører til gruppen af proinflammatoriske cytokiner sammen med IL-1,
IL-6 og tumor necrose faktor (TNF). IL-8 er et kemotaktisk og aktiverende protein specielt
over for neutrofile celler. Epitelceller og makrofager er en del af kroppens første
forsvarssystem mod fremmed-stoffer og indtrængende organismer. Hvis en inflammatorisk
12

agent, som f.eks. endotoxin, indåndes vil en kaskade af proinflammatoriske cytokiner blive
secerneret og igangsætte den inflammatoriske proces.
Formålet med dette Ph.D.-projekt er: 1) At udvikle og anvende in vitro metoder på udvalgte
cellelinier til måling af det inflammatoriske potentiale af støv fra indeklimaet. 2) At korrelere
de fundne inflammatoriske potentialer fra indeklimastøv med biologiske faktorer i støvet,
samt til BRS og BRI fundet via spørgeskemaer givet til ansatte og elever på 20 skoler i
København.
Materialer og metoder
”Skoleundersøgelsen i København” er en epidemiologisk tværsnitsundersøgelse, som
inkluderer ansatte og elever fra 8. klasse og opefter, på 75 skoler. Herudfra udvælges 20
skoler: 10 skoler med de laveste forekomster af BRS (”gode”) og 10 med de højeste
(”dårlige”). Fra disse 20 skoler er der indsamlet støvprøver fra gulv, vandrette overflader og
fra udsugnings-kanaler.
Gulvstøvet blev opsamlet med en opsamler kaldet HVS-3 tilsluttet en almindelig ”Miele”
støvsuger. Støv fra vandrette eller næsten vandrette overflader (overfladestøv) og støv fra
udsugningskanaler blev opsamlet med en bærbar støvsuger med et specielt mundstykke.
Støvet blev sigtet på et 300 µm filter og opbevaret ved -20°C.
3 in vitro bioassays blev udviklet og karakteriseret med forskellige biologiske og kemiske
agens. 4 forskellige lipopolysakkarider (LPS) fra Escherichia coli O55:B5, Klebsiella
pneumoniae, Pseudomonas aeruginosa and Salmonella enteritidis. Glucaner fra bagergær,
Saccharomyces cerevisiae og fra den Gram negative bakterie Alcaligenes faecalis. Rene
kemiske stoffer med allergene eller irritative effekter var nikkel sulfat, metyl metacrylat og
formaldehyd. 4 surfaktanter: Dodecylbenzen sulfonsyre (SDBS), natrium dodecyl sulfat
(SDS), kokosnødeolie og genapol X-80. Cytokiner som tumor nekrose factor-α (TNF-α),
granulocyt makrofag koloni stimulerende faktor (GM-CSF) blev også testet.
Cellelinierne (epitelcellelinien A549 og monocytcellelinierne U937 og THP-1), blev dyrket i
24 brønds bakker. Efter 24 timers inkubation med teststofferne ved 36,6°C og 5% CO2
13

høstes mediet og IL-8 koncentrationen måles med en kommerciel ELISA. LPS (E. coli) og
TNF-α anvendes som positive kontroller.
Endotoxin koncentrationen i støvet blev målt med et standardiseret limulus amebocyt lysat
(LAL) kit. Levende bakterier og svampe blev kultiveret i 7 dage ved 25 °C. Støvets indhold
af mide, katte og hunde allergener blev målt med ELISA teknikker. Den organiske fraktion i
støvet blev bestemt ved glødning.
Støvets potens også kaldet potensfaktoren (PF) udtrykkes som den initiale hældning af dosis
respons kurverne, dvs. det frigivne IL-8 versus støv koncentrationen. For at reducere dag til
dag variationen i bioassay'ene, korrigeres PF i forhold til værdien af en af de positive
kontroller.
Resultater og diskussion
3 in vitro metoder med cellelinierne A549, U937 og THP-1 blev udviklet og evalueret med
de forskellige biologiske og kemiske teststoffer beskrevet ovenfor. Metodeevaluering blev
udført for alle 3 bioassays, og alle metoderne var i kontrol specielt efter korrektion med
positiv kontrol. Metoderne på monocyt celler (U937 og THP-1) var væsentlig mere sensitive
end metoden på epitelcellelinien (A549) når rene stoffer var anvendt.
Totalt blev der indsamlet 158 støvprøver fra de 20 skoler. Der blev fundet en statistisk
signifikant korrelation mellem PF af alle støvprøverne testet i A549 bioassay'et og indholdet
af organisk indhold i prøven. PF af den totale støvprøve fra udsugningskanaler med
mekanisk ventilation var højere end PF af støvprøver fra kanaler med naturlig ventilation.
Der fandtes en positiv korrelation mellem PF af gulvprøverne og PF fra overfladeprøverne
testet i A549 bioassay'et, hvilket tyder på at disse prøver afspejler hinanden. Ligeledes var
PF fra støvprøverne testet i A549 bioassay fra de ”gode” skoler statistisk signifikant lavere
end PF fra ”dårlige” skoler.
Korrelationer til målte biologiske faktorer blev kun udført for PF fra A549 bioassay. Kun PF
af overfladestøv og koncentrationen af endotoxin var svagt korrelerede. Da mængden af
endotoxin i støvprøverne var under detektionsgrænsen for cellemetoderne kan dette indikere
at endotoxin evt. er en markør for andre biologiske eksponeringer. Der blev fundet en
14

signifikant forskel mellem indholdet af de enkelte mikroorganismer i de 3 typer støvprøver
(overflade støv, gulv støv og støv fra udsugningskanaler). En statistisk signifikant
korrelation blev fundet mellem indholdet af levende bakterier i gulvstøvet og PF af
gulvstøvet og mellem alle de målte mikroorganismer og PF fra støvet fra udsugningskanaler.
Støvprøvernes indhold af mide, katte og hunde allergen var generelt lav, lavere end grænsen
for udvikling af et IgE respons (2000 ng/g). Korrektion af PF med den organiske fraktion
gav generelt ikke bedre korrelationer med de mikrobiologiske parametre eller med
allergener, hvilket tyder på at ingen af de testede parametre er dominerende faktorer i støvet.
Potensen kan betragtes som værende multifaktoriel i sin årsag.
Bygninger med flade tage havde statistisk signifikant højere PF end bygninger med tage der
havde hældning. Mindre volumen per person gav et højere PF fra støvet. Positive
associationer blev fundet mellem PF af gulvstøv og symptomer som øjenkløe, index for
mindst 2 tilstedeværende symptomer ud af 5 fra slimhinderne og huden, og til hovedpine. PF
af gulvstøv korrelerede ikke med astma eller høfeber.
Ved anvendelse af skæringspunkterne (cut-off values) CO1=2,0 og CO2=4,5 ng IL-8/mg
støv på PF fra gulvprøverne testet i A549 bioassay, var der i området under CO1, 78%
”gode” og 22% ”dårlige” skolermens der i området over CO2 var 9% ”gode” og 91%
”dårlige” skoler. 54% af prøverne fra de ”gode” skoler blev også fundet under CO1 og 43%
af prøverne fra de ”dårlige” skoler over CO2. Flere tilfælde af hovedpine og kløe i øjnene
blev fundet i skoler over CO2 end under. Der blev fundet signifikant flere “gode” skoler
under CO2 og signifikant flere “dårlige” skoler over.
Konklusion
A549 bioassay'et ser ud til at kunne differentiere mellem “gode” og “dårlige” skoler. Dette
bioassay kan derfor være et værdifuldt screenings redskab i evalueringen af indeklimaet. Et
studie af metodens prædiktive værdi er dog vigtig for vurdering af metodens reelle
potentiale. THP-1 bioassay'et skal evalueres yderligere.
15

1. Objectives
Severe pulmonary effects after exposure to high concentrations of organic dust have been
known for many years. In 1555 the Danish Bishop Olaus Magnus described his findings of
the threshers’ disease. The threshers beat the grain with flails during the wintertime, and
must take care and choose a time with suitable winds to sweep away the grain dust, so that it
will not damage the vital organs of the threshers (Magnus 1555). He describes further that
the dust is so fine that it will almost unnoticeably penetrate into the mouth and accumulate in
the throat. This should be dealt with by drinking fresh ale or the thresher may newer again or
only for a short period eats what he has threshed (Chan-Yeung, Clark et al. 1994). Severe
pulmonary diseases caused by high level exposure to organic dust are toxic pneumonitis
(Organic Dust Toxic Syndrome = ODTS), bronchitis and byssinosis (Sigsgaard, Abell et al.
1994; Poulsen, Breum et al. 1995a; Poulsen, Breum et al. 1995b ). However, the exposure
concentrations in the indoor climate are much lower, and unspecific symptoms like slight
irritations of the eyes and the moucus membranes of the throat as well as diseases like hay
feber and asthma have been associated with the indoor climate (Redlich, Sparer et al. 1997).
To day complaints of the indoor environment are followed by physical, biological, chemical
and psychosocial investigations, to clarify the cause of symptoms (Lahtinen, Huuhtanen et
al. 1998; Hodgson 1991). This includes technical investigations of the building, parameters
of comfort, measurements of microorganisms, microbiological constituents, volatile organic
compounds (VOC) etc. (Lahtinen, Huuhtanen et al. 1998; Hodgson 1991). The building
related symptoms (BRS) are believed to be multifactorial in origin, and it has not been
possible to identify one or a few single causal factors (Brooks 1994).
Exposure to organic dust has the inflammatory reaction as a central part of the development
of symptoms. The underlaing hypothesis of the present study is that in the indoor
environment the inflammatory potential of dust can be used as a marker of the overall
chemical and biological load contributing to the development of symptoms of the mucous
membranes in the eyes, nose and throat. Using a screening method for evaluation of the
inflammatory potential of dust from the indoor environment may therfore exclude or verify,
if the observed symptoms are caused by factors in the dust, or whether other aspects of the
16

working environment needs to be investigated. Such a pre-screening of dust from the indoor
environment could in cases of e.g. physical or psychosocial causes save the expenses of
chemical and biological analysis. In the case of findings of dust with high inflammatory
potential the causative agents may be soughed and eliminated from the environment.
The purpose of this Ph.D.-project was to develop and use in vitro methods for measuring the
inflammatory potential of dust from schools in Copenhagen. Furthermore to correlate the
obtained inflammatory potentials to allergens and microbiological constituents in the dust,
and to correlate the inflammatory potential of dust to building related symptoms (BRS) and
building related illnesses (BRI) found via questionnaires given to employees and students at
the schools.
In the first in vitro method a lung epithelial cell line is used as a model of the epithelial
lining of the lungs. The lung epithelial cells are not directly part of the immune system but
the epithelial lining is involved in the first line of defence against intruding agents or
microorganisms (Devalia and Davies 1993). Induced inflammation of the airways is likely to
be influenced by modulation of epithelial synthesis and release of mediators (Devalia and
Bayram 1997). The A549 cell line is known to secrete IL-8, which is involved in the early
inflammatory response. The macrophages are involved in the inflammatory process, removal
of foreign materials and development of immunity towards specific allergens. In vitro assays
using macrophages will provide information about the effect of dust in a central part of the
inflammatory response. No macrophage cell lines were available at the time of the study.
Hence two monocytic cell lines (U937 and THP-1) were used as a model of the tissue
macrophages of the lungs.
In vitro methods is a gross extrapolation made in the laboratory and the observed responses
cannot directly be compared to complex biological systems. The hypothesis of this study is
built on the assumption that the ability of a test substance to stimulate a cytokine release
from isolated cells in vitro correlate to observed symptoms from humans exposed to the
same test substance. This hypothesis is very difficult to verify, but indirectly we try to do so
by the test of substances with known irritative properties and substances as endotoxin and
dust known to cause symptoms in humans after exposure. Evaluation of the inflammatory
potential based on the secretion of cytokines from cells grown in culture could therefore be
17

used as a screening tool, differentiating buildings with a bad indoor climate from buildings
with a good. This differentiation, however, is only valid with the respective factors that are
in the dust.
18

2. Background
2.1. Building related symptoms related to indoor air quality.
The replacement of older naturally ventilated buildings with more energy efficient and
airtight buildings has increased the number of indoor air problems since the middle of the
1960’s (Redlich, Sparer et al. 1997; Andersen and Lundquist 1966). In 1983 the World
Health Organisation (WHO) defined the concept of the sick building syndrome (SBS) (WHO
1983). Many articles still uses the term SBS, but because of disagreement of definition and
use of the term we today see an opinion going towards the phrase building related symptoms
(BRS). In the rest of this thesis the more descriptive BRS will be used. Non specific
complaints such as mucous membrane irritations in the eyes, nose and throat, headache,
dizziness, fatigue and lethargy, lack of concentration, cough, and dryness of the skin have all
been related to the BRS (Gyntelberg, Suadicani et al. 1994; Redlich, Sparer et al. 1997; Skov
1992). Table 1 gives a short overview of the symptoms.
Table 1. Common symptoms of the “Sick” building syndrome (SBS).
Common symptoms in SBS
Mucus-membrane irritation Eye irritation
Nasal irritation
Throat irritation
Cough
Neurological effects Headaches
Fatigue
Dizziness
lethargy
Lack of concentration
Respiratory symptoms Shortness of breath
Cough
Wheeze
Chest tightness
Gastro-intestinal symptoms Nausea
Skin symptoms Rash
Dryness
Itching
Chemosensory changes Enhanced or abnormal odour perception
Visual disturbances
Modified from (Redlich, Sparer et al. 1997)
Both the type and the severity of symptoms vary from person to person, even within a
building. The BRS are often described for occupational environments in non-industrial
19

uildings as offices, schools etc, but reports from people’s homes are also seen. The BRS
start gradually, within days or months after the beginning of work in a “sick” building or
after changes are made in a building (Skov 1992). The symptoms usually disappear after
work, in the weekends or during vacation (Wolkoff and Kjærgaard 1992). Normally a
building is characterised as “sick” if more than 20% of the users complain of these nonspecific
symptoms (Wolkoff and Kjærgaard 1992).
Chemical (Klein 1991), biological (Seltzer 1995), physical (Jaakkola, Heinonen et al. 1989),
and psychosocial (Skov, Valbjørn et al. 1989a) factors have been related to the BRS, and
symptoms often have a multifactorial cause (Brooks 1994). Also gender, job satisfaction, job
category, job function, work stress and degree of crowding are important factors related to
BRS (Salvaggio 1994; Brooks 1994; Skov, Valbjørn et al. 1989a; Bachmann and Myers
1995).
Building related illnesses are specific clinical conditions caused by specific exposures from
the indoor environment. Examples are hay fever and asthma, caused by exposure to moulds,
spores, bacteria or other allergenic substances (Redlich, Sparer et al. 1997; Brooks 1994).
Also food additives and perfumes in the products used today, and evaporations from
furnitures or paints as well as the cleaning standard of the room and smoking may contribute
to the development of these allergic diseases (Ott and Roberts 1998). Symptoms related to
allergic diseases must be distinguished from the BRS (Redlich, Sparer et al. 1997). However,
workers suffering from allergic diseases more often complain of mucosal irritation and
aggravation of their illness after a stay in a “sick” building (Skov 1992). The susceptibility of
the worker also has an influence on the development of symptoms. Factors affecting a
person’s susceptibility are genetic background (including gender), age, atopy, airway
hyperreactivity, and pre-existing asthma etc. (Brooks 1994).
A case report from Sweden describes how a non-smoking school teacher over a six year
period developed a chronic allergic alveolitis related to the indoor environment of the school,
where indoor air quality problems had existed for several years (Thörn, Lewné et al. 1996).
However, since allergic alveolitis is caused by wery high concentration of e.g. microbiel
spores this case may reflect an extreme situation, not normaly found in the indoor climate.
20

In Denmark the working population spend about 90% of their time indoors, at home, at work
or during transport (Wolkoff and Kjærgaard 1992; Seltzer 1995). This percentage is even
larger for old and sick persons. In the indoor environment we have a higher concentration of
pollutants than in the outdoor environment (Ott and Roberts 1998). This could partly be
caused by the large amount of consumer products emitting chemical compounds to the
indoor air in the homes or compounds and particles being tracked in on people’s shoes and
clothes. Inside a building emission products may accumulate in the dust, carpets and
furnitures (Ott and Roberts 1998). Also microbial parameters are important factors for our
well being when staying indoors. Bacteria, fungi and their constituents are found in the
indoor environment. They can be growing inside the building e.g. in mouldy buildings or
come from the outdoor environment (Seltzer 1995). In a study by Waegemaaekers et al.
(1989) a correlation between dampness, high average spore count and higher prevalence of
respiratory symptoms were found in homes (Waegemaekers, Van Wageningen et al. 1989).
Mechanical ventilation e.g. with recirculation of air, and apparatus for control of temperature
and humidity of the building, together with the use of synthetic materials, number of people
working in the room, carpets etc. are related to the outbreaks of BRS (Redlich, Sparer et al.
1997). Inadequate ventilation is often picked out as the cause of bad indoor environment
causing BRS. Hedge et al. (1989) reported a higher concentration of formaldehyde, volatile
organic compounds (VOC) and respirable particles in air-conditioned offices (Hedge,
Sterling et al. 1989). They also reported that symptoms like sleepiness, nasal irritation,
difficulties to concentrate, and cold/flu-like symptoms were significant more prevalent in airconditioned
offices compared to naturally ventilated offices (Hedge, Sterling et al. 1989).
Another study showed that the age of the building was important, with older buildings
having a better indoor environment (Skov 1992). This could be because older buildings are
not as airtight as newer buildings and the older buildings have natural ventilation, no
humidifier system and a pitched roof (larger cubic capacity or fewer water damages). These
factors ensure good ventilation with air from outside the building, and no accumulation of
chemical and biological contaminants compared with recirculation of the air within a
building.
21

2.2. Contaminants in the indoor environment
Sources of contaminants in the indoor environment can derive from emissions from e.g.
paints, carpets, flooring, insulation materials, adhesives, cleaning agents, and office
machines (Redlich, Sparer et al. 1997). For instance, fibres of sound absorbing mineral
fibreboards may be the cause of reported irritations of the mucosa and the skin (Thriene,
Sobottka et al. 1996). Biologic contaminants are generally derived from living organisms
such as bacteria, fungi (mould and yeast), protozoa, insects, and arachnids (e.g. dust mites)
or components from these organisms (Seltzer 1995; Bischoff 1989). Hair and dander from
pets and animals also contribute to the pool of biologic contaminants (Seltzer 1995) (figure
1).
Geographical factors
E.g. altitude,
climate,
Residential area
E.g. River,
industry,
Algae
Insects
Pollen
Wild animals
Outdoor
Environment
Outdoor
allergens
Sick buildings
Indoor
Pollution
Inflammation /
Hyperreactivity
Unspecific symptoms
and disease
Figure 1. Sources of indoor air pollution. (Modified from Munir and Björkstén 1992)
22
Carpets
Furniture
Indoor
Environment
Bedding material
Mites
Indoor
Allergens
Cleaning
Heating
Ventialtion
Moulds
Cockroach
Pets
Rodents

Emissions from the human body and pollution from the outdoor environment can also be
concentrated in the indoor environment (Ott and Roberts 1998). Many of these constituents
in the biologic fraction of the indoor environment have allergenic properties (Becher,
Hongslo et al. 1996).
2.2.1. Microorganisms and their indoor air relevant constituents
Building materials or carpets damaged by water or high humidity in the building may
provide a basis for the growth of microorganisms, especially bacteria and fungi. Some
studies have found a correlation between the concentration of Gram negative bacteria in the
indoor dust and BRS (Gyntelberg, Suadicani et al. 1994; Teeuw, Vandenbroucke-Grauls et
al. 1994).
Gram-negative bacteria have a cell wall with a multilayer structure, composed of an inner
and outer membrane. In the outer membrane of Gram-negative bacteria a lipopolysaccharide
(LPS) termed endotoxin is incorporated (figure 2A). The structure of the endotoxin molecule
determines its position in the membrane. The lipid-A part of the toxin is embedded in the
bacterial membrane with its long fatty acid chains. The polysaccharide chain protruding
from the bacteria is divided in an inner core, an outer core and an o-specific chain (figure
2B) (Rietschel and Brade 1992). The o-specific chain is a 20 to 40 units long chain,
depending on the species and growth conditions. Endotoxin is liberated during growth and
death of the bacteria and the toxic effects of endotoxin are related to the lipid-A component,
where the polysaccharide chain determines the virulence (Rietschel, Kirikae et al. 1993). The
physical structure of lipid A is essential for some of the biological activity of endotoxin. The
purified LPS, which are used in research, contain a low content of protein. Endotoxins or
purified LPS are measured by the endotoxin specific limulus amebocyte lysate (LAL) test,
originally developed from the clotting enzymes in the blood of the horseshoe crab Limulus
polyphemus (Levin 1985). In the chromogenic method, the natural LAL coagulogen, which
is a peptidase, is coupled to a synthetic substrate (5-amino-acid-polypeptide, 5-Pep bound to
p-nitroaniline, pNA). LPS activates the coagulase that liberates free pNA upon cleavage of
the substrate. Liberated pNA is yellow in solution and can be measured
spectrophotometrically (Jacobs 1997). This chromogenic test can be performed as either an
endpoint test or a kinetic test.
23

A B
Figure 2. A) The structure and placement of endotoxin in the membrane of Gram negative bacteria. B) The Ospecific
chain with repeating sugar units in the polysaccharide part of endotoxin is protruding from the cell to
the environment whereas the core part and the Lipid A is embedded into the membrane (Modified from
Rietshel and Brade 1992).
Initially, the pyrogenicity of endotoxin was of particular concern in pharmacology and
clinical medicine. The pyrogenicity of endotoxin was documented to be dependent upon the
strain of Gram-negative bacteria (Nowotny 1990; Dutkiewicz, Skorska et al. 1988; Pearson,
Weary et al. 1982; Weary, Donohue et al. 1980). Teeuw et al. (1994) found a six-fold
increase in airborne endotoxin concentration in “sick” buildings compared to the
concentration found in “healthy” buildings (Teeuw, Vandenbroucke-Grauls et al. 1994).
Endotoxin has also proinflammatory and adjuvant-like properties (Rylander 1997a), which
may have a serious impact on health. Exposure to endotoxins has been reported in relation to
various symptoms and diseases, such as toxic pneumonitis, airway inflammation, bronchitis,
allergic asthma and systemic effects such as fever, pain in muscle and joints, and excessive
fatigue (Rylander 1997a).
24
Lipid Polysaccharide
O-specific chain
Outer core
Inner core
Lipid A
n
Repeating unit
Heptose
Ethanolamine
Kdo
Phosphate
Glucosamine
Fatty acid

There has been little focus on the Gram-positive bacteria in relation to occupational
respiratory symptoms and diseases. Only the actinomycetes, characterised by a filamentous
mycelium, have been associated with diseases such as rhinitis, asthma and allergic alveolitis
(Flannigan 1992; Lacey 1997). In the indoor environment the concentration of actinomycetes
is rather low compared to the concentration of viable bacteria and fungi (Rautiala, Reponen
et al. 1996; Nevalainen and Jantunen 1988). However in some studies, the presence of
actinomycetes was associated to complaints of odour or mould growth in homes, schools,
offices and day-care centres (Nevalainen, Kotimaa et al. 1990). The Gram positive bacterial
membrane consists of a single layer composed of 90% peptidoglycan, a polysaccharide
composed of N-acetylglucosamine and N-acetylmuramic acid, and amino acids. These
constituents are assembled in layers, forming the glycan tetra peptide (figure 3) (Brock,
Madigan et al. 1994a).
HOOC
NH
N-Acetylglucosamine N-Acetylmuramic acid
O
H
CH OH
2
O H O
OH O
H
H
NH
C=O
O
H
HC CH
CH 3
HC
3
C=O
NH
CH C
Figure 3. Structure of the glycan tetrapeptide showing one of the repeating units of peptidoglycan cell wall
structure. The Gram positive membrane consists of 90% peptidoglycan, and Gram negative bacteria consists of
10 %. (From Brock and Madigan, 1994).
25
CH OH
2
=
3
O
H
NH
C=O
CH 3
NH
O =
C CH CH CH COOH
NH
C CH CH CH CH C
H
2
(1 ,4 )
2 2 2
HC
3
2 2
=
NH
(1 ,4 )
O
CH COOH
O
(1 ,4 )
H
Lysozyme-sensitive bond
L-Alanine
D-Glutamic acid
Meso-diaminopimelic acid
L-Alanine

In damp and moist buildings the dominating filamentous fungi are generally the mould
genera Penicillium, Cladosporium and Aspergillus (Flannigan 1992). In an investigation of
the indoor climate in a day-care centre the dominating fungi were Penicillium, Cladosporium
and Aspergillus. However, only the Aspergillus was associated with the reports of BRS (Li,
HSU et al. 1997). Fungi are often found on filters in air conditioning equipment (Schata,
Jorde et al. 1989). Schata et al (1989) investigated 150 patients with allergic disease, whose
symptoms mainly occurred in air-conditioned rooms. The patients were challenged with
extracts of fungi from the air conditioning equipment. In 135 patients, a positive allergic
reaction towards the tested fungi was induced (Schata, Jorde et al. 1989).
CH OH
2
H
H
O
HO
OH H
H
H OH
O
CH OH 2
H
O
H
H
HO H
CH2 H
H
O
HO
O
H
H
CH OH 2
H O
H
O H
HO H
H OH H OH H OH
Figure 4. β, 1-3 Glucan with a 1-6 side chain. The polyglucose chain may be straight, with side chains or form
helices, mainly triple helices. (Modified from Rylander and Goto eds. 1991).
A major constituent of the cell wall of fungi is glucan, a polyglucose compound with
glucopyranosyl rings attached in chains and sometimes with side chains in α or β positions
with 1-3, 1-4 or 1-6 linkages (figure 4) (Fogelmark, Goto et al. 1991). The β 1-3 D-glucan
has been associated with a significant immunobiological activity (Rylander, Williams et al.
1993). β 1-3 D-glucan has the potential to induce production of tumor necrosis factor (TNF)
in macrophages through reaction with specific glucan receptors (Goto, Yuasa et al. 1994).
Rylander et al (1992) found a significant correlation between symptoms such as dry cough
and itching of the skin and the concentration of airborne β 1-3 glucan measured by the
glucan specific LAL assay. They also found a significant correlation between airborne
endotoxin measured by the LAL assay and skin rashes (Rylander, Persson et al. 1992). These
results suggest that components of both fungi and bacteria may contribute to the symptoms
26
O
n

BRS found in “sick” buildings. Other components of the fungus cell wall are mannan,
cellulose, pullulan and polymers of N-acetyl-glucosamine (Rylander, Williams et al. 1993).
Fungi and some bacteria (actinomycetes) produce spores, which in comparison to the
vegetative stadium of the microorganism are more resistant to physical and chemical
treatment (Brock, Madigan et al. 1994b; Brock, Madigan et al. 1994c). Fungi and
actinomycetes form spores after asexual growth in favourable conditions. Fungi also form
spores after sexual reproduction (Brock, Madigan et al. 1994c). Some of these spores contain
constituents with allergenic (Schata, Jorde et al. 1989) and cytotoxic properties (Smith,
Anderson et al. 1992 ), some also produce toxins that may contribute to the BRS.
2.2.2. Volatile organic compounds (VOCs)
The primary source of VOCs in the indoor air, are emissions from building materials,
furnitures and carpets, combustion fumes, cleaning compounds, paints or stains (Koren,
Graham et al. 1992), and microorganisms. VOCs may be distinguished according to origin.
VOC from microorganisms are termed MVOC, and the total amount of VOCs is termed
TVOC. VOCs have a boiling point between 50-260°C, defined by the WHO (WHO 1989a).
Approximately 300 different VOCs have been identified from the indoor environment
(WHO 1989a). Some of the different VOCs bind to dust, and exposure could be via
inhalation of dust. Several studies indicate an association of exposure to VOCs and
occurrence of asthma like symptoms or BRS (Norbäck, Bjornsson et al. 1995; Norbäck,
Torgen et al. 1990).
2.2.3. Dust
Organic dust is characterised as unspecific samples of both living and dead organic material
of vegetable, animal and microbial origin (Chan-Yeung, Clark et al. 1994). Airborne dusts
will sediment on surfaces, from where it can be whirled up. Surface dust can be broken down
to smaller pieces or lump together in larger aggregates. From the sedimented dust,
containing living and dead material, evaporation of gasses can occur, or gasses from the air
can adsorb to the dust. This is described in figure 5.
27

VOC
Aggregation
Airborne dust
Sedimentation Whirling
Sedimented dust
Figure 5. Schematic overview of the turnover of dust in the indoor environment. Airborne dust may, e.g.
because of aggregation into larger particles, sediment on surfaces such as the floor, tables, shelfs etc. Different
kinds of volatile organic compounds (VOCs) can adhere to or be liberated from the particles. Aggregated and
sedimented dust may break down in to smaller pieces, which may lead to the particles being whirled up into the
air again. Also binding and evaporation of substances to and from the sedimented dust can occur.
In a cross-sectional study involving 205 workers from 6 primary schools, the observed
correlation between presence of wall-to-wall carpets and BRS may be explained by the
accumulated dust and dirt in the carpets (Norbäck and Torgén 1989). In a study by the
Danish indoor climate group, a correlation was found between macromolecular dust and
symptoms like headache, and dizziness (Gyntelberg, Suadicani et al. 1994). In the
Copenhagen Town Hall study a correlation was found between macromolecular organic dust
from the floors and BRS symptoms (Skov, Valbjørn et al. 1989b). These studies indicate a
connection between dust from the indoor environment and BRS.
Many different opinions exist on what to measure in the indoor environment, and in general,
different studies elucidate correlations between different environmental parameters and BRS.
Gyntelberg et al. (1994), found a significant correlation between the prevalence of Gramnegative
bacteria in the indoor dust and symptoms such as fatigue, heavy-headedness,
headache, dizziness and lack of ability to concentrate, and symptoms from the mucus
membrane of the upper respiratory tract (Gyntelberg, Suadicani et al. 1994 ). Other
parameters such as particles in the dust, the macromolecular dust, and TVOC were
correlated with BRS in this study.
28
Evaporation
Binding to dust
breaking down

Airborne dust sampled from the breathing zone gives a picture of the dust that has an actual
potential of being inhaled at the time of sampling. Floor dust may also contain earth; sand
and other particles dragged in from the outside. Surface dust from horizontal or near
horizontal surfaces has been airborne once. It may consist of large aggregates and is older
than the floor dust. (figure 5). Both floor dust and surface dust have the potential of
becoming airborne again e.g. by mechanical action or by a change in physical factors e.g. the
ventilation rate. As mentioned VOCs, e.g. from cleaning agents may adhere to dust, and
microorganisms and other biological derived constituents can be found in the dust.
The size of particulates range from 0.0005 to 5000 µm (from the size of a molecule to the
size of sand on the beach), but only particles smaller than 10 µm in size remain suspended in
the air for long periods of time. These airborne particulates can themselves inflict health by
penetrating the airways or act as vehicles for the transport of other toxic agents (Hines,
Ghosh et al. 1993). The size of the inhaled dust particle determines the deposition within the
different parts of the airways. Larger particles, ranging from 30 to 60 µm in aerodynamic
diameter are filtered out by the nasal vibrissae (Hines, Ghosh et al. 1993). Thoracic dust is
smaller than 25 µm and may reach the upper airways. Respirable dust may reach the alveoli
and is smaller than 10 µm. Bacteria, smaller fungal spores, and aerosols generated by
talking, coughing and sneezing make up the group of truly respirable particles, ranging 1-5
µm. Particles smaller than 1 µm may also be expelled from the airways by exhalation
(Seltzer 1995). These small particles, however, can also be deposited in the alveolar wall by
diffusion (Owen and Ensor 1992). The ultra fine particles (< 0.1 µm) are emitted into the
atmosphere by combustion prosesses including burning of coal and oil, mealting of metals
and diesel exhaust. A human lung deposition model predicts that 0.02 µm particles have
50% deposition efficiency in the alveolar region (Amdur, 1996). For particles greater than
0.5 µm in diameter, sedimentation and impaction are the primary deposition mechanisms.
For particles less than 0.5 µm, Brownian diffusion becomes a dominant mechanism (Utell,
Samet, 1996).
Respirable dust is found only in small concentrations in the indoor environment. Kildesø et
al. (1998) found over a period of 29 weeks a mean total dust concentration on 63.5 µg/m 3 in
a room of an administration building, 267 µg/m 3 in a room in a kindergarten and 174 µg/m 3
29

in a class-room (measured in the daytimes). The respirable fraction was a bit lower 54.5, 133
and 102 µg/m 3 respectively (Kildesø, Tornvig et al. 1998). These values are well below the
Danish exposure limit of organic dust, 3 mg/m 3 during an 8-hour working day, defined by
the Danish Labour Protection Agency (Grænseværdier for stoffer og materialer (in Danish)
1996). The sampling of respirable dust with a median diameter of 4.25 µm is either by
person borne filter cassettes or stationary samplers. (European Committee for
Standardisation 1993). Long sampling periods or high volume samplers are needed to collect
enough airborne dust to perform the chemical or biological analyses of the dust. Hence,
many technical problems are connected to sampling of airborne dust in the indoor
environment. Systems using sedimented dust as a proximeasure for airborne dust have not
received much attention. Sampling of sedimented dust is easier and faster than sampling of
airborne dusts, but the particle distribution may not be the same in the two kinds of samples.
Furthermore, models for the whirling of sedimented dust (see figure 5) need to be developed.
Models and methods for interpretation of sedimented dust parameters in relation to the actual
inhaled doses of respirable dust are lacking. This Ph.D.-study was initiated to investigate
whether sampling and analysis of surface dust can be used to obtain relevant information on
the indoor environment (see section 1.). If the results on surface dust parameters correlate
with symptoms related to indoor environments, a basis is laid for a screening tool that can be
used in hazard evaluation of dust from the indoor environment.
The composition of dust, including both biologic and non-biologic components, varies
according to the environment. Even within a building the composition can vary. It is
therefore difficult to point out one or just a few chemical factors or components of microbial
origin that may cause or add to the multitude of symptoms related to the indoor environment.
The cause of BRS is believed to be multifactorial in origin (Lahtinen, Huuhtanen et al. 1998;
Hodgson 1991), and thus caused by a combination of factors (i.e. microbial components,
proteins from pets and house dust mites, VOCs, dust and chemical constituents from
building materials). Inflammation is common to many of the symptoms and illnesses related
to bad indoor climate. Inflammation could therefore be considered an overall parameter that
integrates the effect of the total exposure load from an indoor environment, affecting the
immunesystem (Nielsen, Alarie et al. 1995).
30

2.2.4. Measurements of inflammatory reactions of dust
An inflammatory response e.g. in the nose of persons exposed to different xenobiotics may
be indicated by markers in nasal lavage or by swelling of the nasal mucosa measured by
rhinometry or acoustic rhinometry. Douwes et al. (1997) measured different inflammatory
markers (e.g. myeloperoxidase, eosinophil cationic protein, interleukin-8, and nitric oxide) in
compost workers (Douwes, Dubbeld et al. 1997). General endpoints in nasal lavage are
inflammatory cell influx, eiosanoid mediators, neuropeptide release, nasal glandular
products, and vascular products released into the nasal airway. This is a result of increased
vascular permeability, inflammatory cytokines (Interleukin-6, IL-8 and Tumor necrosis
factor-α), and other products from cells as mast cells, neutrophils, and eosinophils (Peden
1996). Inflammatory markers such as IL-8, soluble TNF receptor (sTNF-R75) and albumin
are also suggested to be associated with repeated airborne endotoxin exposure in cotton
workers (Keman, Jetten et al. 1998). Acoustic rhinometry is used to demonstrate a
significant swelling of the mucosa after inhalation of swine dust (Larsson, Palmberg et al.
1997), but also as an indicator of exposure to indoor air pollutants in buildings with low
ventilation (Wålinder, Norback et al. 1997a; Wålinder, Norbäck et al. 1997b). In some
studies total nasal resistance measured by spirometry and peak flow measurements is used to
investigate nasal and bronchial reactivity (Godnic-Cvar, Plavec et al. 1999). These tests on
nasal inflammation could be combined with provocation tests or analysis of lachrymal fluid.
Chamber studies by Mølhave et al (1995) were set out to clarify the effect of dust from the
indoor environment on experimental subjects. Measurements of nasal volume, peak flow,
eye effects (redness, lachrymal fluid, epithelial cells), and skin moisture were combined and
used to study the exposure effects (Mølhave, Kjærgaard et al. 1995). Bronchoalveolar lavage
(BAL) was obtained in some studies to determine different cell populations such as
macrophages, mast cells, eosinophils, neutrophils, lymphocytes, and epithelial cells and
different cellular products in the BAL fluid (Djukanovic, Roche et al. 1990). Investigators
with formal medical training can only carry out this type of investigation. Furthermore, the
test is unpleasant for the participant and therefore not well suited as a routine tool.
Investigations on nasal lavage require a large group of participants to compensate for
biological variation and confounding causes of inflammation. Thus, there is a need for
standardised in vitro methods to screen the inflammatory potential of dust.
31

2.3. In vitro models and cytokines
In vitro models using cells and cell lines are widely used as a substitute for in vivo models,
and where possible in vivo models are replaced by in vitro. In vivo models have the
advantage of including interactions of the total organism, whereas the in vitro models are
attractive because of their simplicity e.g. easy cultivation of cells and interpretation of
results, they are time saving, cheap, and easy to perform.
Table 2. Examples of different cell types used in the in vitro methods in immunotoxicology and their
production of different proinflammatory cytokines.
Cell types
Human epithelial cells
Examples Cytokines measured Reference
Lung epithelial cell line A549 IL-8, IL-6 1), 2), 3), 4) and 5)
Bronchial epithelial cell line BET-1A IL-8
6), 7), 8) and 9)
HS-24 IL-8
BEAS-2B IL-6, IL-8, TNFα
Primary human bronchial
epithelial cell
IL-6, IL-8, TNF-α, GM-CSF 10), 11) and 12)
Epithelial cell lines from colon T80 IL-8
13)
Caco-2 IL-8
HT-29 IL-8
SW620 IL-8
Epithelial cell lines from the A498 IL-6, IL-8
14) and 15)
urinary system
J82 IL-1α, IL-6, IL-8,
Primary tubular epithelial cells
Human monocytes/macrophages
IL-8 16)
Langerhans cells IL-1, IL-8 17)
Lymphoma cell lines U937 IL-1β, IL-6, IL-8, TNF 18), 19), 20), 21), 22), 23), and 24)
THP-1 IL-1, IL-6, IL-8, TNF
HL-60 IL-6
Primary mononuclear cells IL-1α and β, IL-6, IL-8, IL-
10, TNF-α
25) and 26)
Primary alveolar macrophages
Murin monocytes/macrophages
IL-1, IL-6, IL-8, TNF 18), 20), 27) and 28)
Lymphoma cell lines RAW264.7 IL-1, IL-6, TNF 29)
Primary peritoneal macrophages IL-1, IL-6 29)
1) Paper I, 2) (Wang 1997), 3) (Standiford, Kunkel et al. 1990), 4) (Bianchi, Fantuzzi et al. 1993), 5) (Crestani,
Cornillet et al. 1994), 6) (Nakamura, Yoshimura et al. 1992), 7) (Nakamura, Yoshimura et al. 1991), 8)
(Veronesi, Oortgiesen et al. 1999), 9) (Devlin, McKinnon et al. 1994), 10) (Cromwell, Hamid et al. 1992), 11)
(Khair, Devalia et al. 1994), 12) (Carter, Ghio et al. 1997), 13) (Schürer-Maly, Eckmann et al. 1994), 14)
(Hedges, Svensson et al. 1992), 15) (Agace, Hedges et al. 1993), 16) (Gerritsma, Heimstra et al. 1996), 17)
(Loré, Sönnerborg et al. 1998), 18) (Miller, Nagao et al. 1996), 19) (Kuschner, D'Alessandro et al. 1998), 20)
(Deaton, McKellar et al. 1994), 21) (Vowels, Yang et al. 1995), 22) (Taimi, Defacque et al. 1993), 23)
(Friedland, Shattock et al. 1993a), 24) (Zhang, Doerfler et al. 1993), 25) (Johnston, Papi et al. 1997), 26) (de
Waal Malefyt, Abrams et al. 1991), 27) (Wang 1997), 28) (Vanhee, Gosset et al. 1995), 29) (Adachi, Okazaki
et al. 1994).
32

Many different cell types are used in the in vitro models to investigate the production and
secretion of cytokines (Table 2). According to the investigated topic, cells from the target
organ are usually the preferred cells. The in vitro models with cultivated primary cells from
organs should include information about susceptibility and influence from previous exposure
to xenobiotics. In contrast, secondary cells (cell lines) all originate from the same stem cell
and have the potential to respond equally to the same stimuli.
2.3.1. Cytokines
Cytokines are multifunctional mediator molecules of low molecular mass (

mice a peak level of TNF was seen after one hour, with a rapid decline to undetectable levels
after 8 hours, indicating that TNF is under strict regulation (Strieter and Kunkel 1994). This
decline has been attributed to binding of TNF to high affinity receptors, systemic release of
TNF binding proteins and renal clearance (Tracey and Cerami 1993). In LPS stimulation of
human whole blood in vitro, two waves of cytokine gene activations were observed. The first
wave included IL-6, TNF (peak between 2-4 hours), and IL-8 (reaching a plateau after 6-12
hours). In the second wave IL-8 continued to rice until the end of the 24-hour study
(DeForge and Remick 1991).
In the selection of a marker of the inflammatory potential of the test substances the marker
should be a proinflammatory cytokine. Hence, the proinflammatory cytokines IL-1, IL-6, IL-
8 and TNF-α was considered. Larsson et al. have shown that the secretion of IL-8 from the
lung epithelial cell line was larger than the secretion of IL-6 in the background values and
after exposure to different bacteria strains (Larsson, Larsson et al. 1999). This was also
found for exposure with house dust (Saraf, Larsson et al. 1999). The IL-8 is a central
cytokine in the inflammatory process and is secreted in large quantities from the chosen cell
lines. TNF-α and IL-1 is not secreted from the lung epithel cell line (Spriggs, Imamura et al.,
1988; Wang 1997).
2.3.2. Interleukin 8
IL-8 belongs to a group of cytokines called chemokines with four conserved cysteines. Two
subgroups of chemokines are recognised, the CXC chemokines, with the first two cysteines
separated by one amino acid, and the CC chemokines with adjacent cysteines (figure 6)
(Baggiolini, Dewald et al. 1994; Wuyts, Proost et al. 1998). CXC chemokines attract and
activate mainly neutrophils, whereas the CC chemokines attract and activate monocytes,
lymphocytes, basophils, eosinophils, natural killer cells and dendritic cells (Wuyts, Proost et
al. 1998). IL-8 belongs to the group of CXC chemokines and is produced as a 99 amino acid
precursor with a signal sequence of 22 amino acids, which is cleaved to yield the 77-residue
mature protein. The protein can be further truncated to yield different analogues (77-, 72-,
71-, 70-, 69-amino acid forms). The two major analogues found are the 77- and the 72-amino
acid forms (Baggiolini, Dewald et al. 1994; Wuyts, Proost et al. 1998). The gene for IL-8 is
located on chromosome 4 together with many of the other CXC chemokines and consists of
34

four introns and three exons (Baggiolini, Dewald et al. 1994; Matsushima, Baldwin et al.
1992). The CXC chemokines exhibit 20% to 50% homology in the amino acid sequence
(Strieter and Kunkel 1994). The three-dimensional structure of IL-8 is a dimer of two
identical subunits stabilised by six hydrogen bonds and by other side chain interactions. The
monomer, however, is probably the active form of the protein (Wuyts, Proost et al. 1998).
IL-8 is stable at extremes of pH (pH 2.0 and 9.0), under oxidising and reducing conditions, at
heating to 100 °C and freezing, at treatment with detergents and organic solvents, and to
plasma peptidases (Baggiolini, Dewald et al. 1994; Wuyts, Proost et al. 1998). This stability
shows that the production of IL-8 at sites of inflammation prolong the biological activity of
the protein. Expression of IL-8 is seen in many different cells such as alveolar macrophages,
monocytes, neutrophils, endothelial cells, epithelial cells, and fibroblasts (Strieter and
Kunkel 1994), but also cells as smooth muscle cells, astrocytes, hepatocytes, synovial cells,
amnion cells etc. (Wuyts, Proost et al. 1998).
Figure 6. Amino acid sequence of the CXC and the CC chemokines. The number of amino acids in the mature
protein is shown in parentheses. CXC chemokines attracts and activate mainly neutrophils whereas the CC
chemokines attracts and activate monocytes, lymphocytes, basophils, eosinophils, natural killer cells and
dendritic cells. The structures of IL-8, a CXC chemokine, and MCP-1, a CC chemokine are shown. From
Baggioloni et al (1994).
35

2.3.3. Airway inflammation
The epithelial cells of the lung is not only a physical barrier against inhaled xenobiotics, they
also contribute to the inflammatory reaction by production of proinflammatory cytokines
(cytokines produced in the initiation of an inflammatory reaction), inflammatory
eicosanoids, and specific cell adhesion molecules (Devalia and Davies 1993). Macrophages
have a central place in the inflammatory reaction, e.g. in the production of cytokines,
reactive oxygen metabolites, enzymes, and phagocytosis of foreign agents (Nielsen, Alarie et
al. 1995). The epithelial cells and the macrophages constitute the first line of defence against
xenobiotics in the lung. When an inflammatory agent, is inhaled a cascade of
proinflammatory cytokines are secreted from e.g. the macrophages and epithelial cells of the
lungs (figure 7) (Nielsen, Alarie et al. 1995; Wang 1997). Macrophages have the capacity to
produce a panel of cytokines including IL-1, IL-6, IL-8, IL-10, IL-12, TNF-α, IFN-α, IFN-γ,
monocyte chemotactic protein (MCP)-1 and 3, colony stimulating factors (CSF) and TGF-β.
The de novo synthesis of the cytokine mRNA and secretion of the mature proteins in
macrophages are very fast (Cavaillon 1994). As shown in table 2 the epithelial cells are
capable the production of cytokines including: IL-1, IL-6, IL-8, GM-CSF, G-CSF, and TNF-
α (Cromwell, Hamid et al. 1992; Mullol, Xaubet et al. 1995; Devalia and Davies 1993).
The IL-8 bioassay is a way of measuring the proinflammatory potential of different
compounds. Miller et al. (1996) found that IL-8 was secreted from the monocytic cell line
U937 and from human alveolar macrophages stimulated with Staphylococcal toxin,
enterotoxin A or LPS. The secreted IL-8 possessed neutrophil chemotactic activity (Miller,
Nagao et al. 1996). This was also found as a response to hyperoxia (Deaton, McKellar et al.
1994). IL-8 seems to be a indicator of inflammation whereas IL-1 and TNF also have many
other functions causing e.g. fever, anorexia, hypotension, shock, and increased expression of
IL-6. There are now developed bioassays for assessing the potency of organic dusts in which
the secretion of IL-8 from lung epithelial cells is measured after in vitro stimulation (Paper
I), (Palmberg, Larsson et al. 1998).
36

Allergens
Viruses
Air pollution Microorganisms
IL-6
Tissue macrophage
TNF, IL-1
Figure 7. The airway epithelial cells and the alveolar macrophages in the initiation process of inflammation.
IL-8 is produced after exposure of pollutants in the inhaled air. A chemotaxic gradient of IL-8 attracts
neutrophils from the blood stream to the site of exposure. Also the tissue macrophages are stimulated and a
phagocytosis of foreign bodies and xenobiotics commences. A cascade of secreted cytokines is initiated and
attraction of other cells from the immunesystem begins. Modified from Devalia and Davis (1993).
37
IL-8
TNF, IL-1
IL-6 and IL-8
Chemotaxi
Neutrophil
Alveolar epithelium
Venule
Initiation of the inflammatory cascade

3. Materials and methods
3.1. The school investigation
The Copenhagen school investigation was performed as an epidemiologic cross sectional
study including employees and students from the 8 th grade and up (>13 years) from 75
schools in Copenhagen. They were given a questionnaire regarding individual perception of
physical conditions (temperature, air quality, space etc.), symptoms of the eyes, nose, throat,
skin, lungs, and general symptoms (headache, fatigue, difficulties in concentration etc.).
Also questions regarding hay fever, asthma and other symptoms diagnosed by a doctor were
included. Psychosocial questions such as working load, degree of influence on the work and
victimisation were also included. No measurements were included at this stage. Meyer et al.
from the Clinic of Occupational and Environmental Medicine in Copenhagen (Meyer,
Nielsen et al. 1996; Meyer 2000) performed these investigations. 7884 questionnaires were
returned giving a response rate of 66%. A BRS index for each school was calculated using
the mean of 8 symptoms: eye irritation, nose irritation, congested nose, irritation of the
throat, blushing of the skin, unusual tiredness, difficulty in concentrate and headache.
Investigations of the 75 schools were combined with information provided by each of the
schools and from a technical inspection of the schools. These informations explain how the
schools were operated, kept up and organised. Furthermore, a register of schools with water
damages was also included in the investigation of the schools. In this phase of the
investigation no measurements of physical parameters, or airborne chemical and
microbiological parameters were included. Questionnaires and scheme of the buildings are
included in Meyer et al. (1996).
From the above BRS index 20 schools were selected: 10 schools with the lowest prevalence
of BRS and 10 schools with the highest prevalence of BRS. In each of the 20 schools two
class rooms with high and two with low prevalence of BRS was chosen for further
examination. These rooms were examined for appearance of the room, number of persons
occupying the room, heating and ventilation system, cleaning status, visible damp stains,
visible mould growth etc. Furthermore, a thorough examination of the area under the roof
(regarding ventilation, smell, damp, mould etc.) and the ventilation room (regarding the
38

operation of the ventilation system, appearance of the filter, smell, damp, mould etc.) was
performed. Measurements of temperature, humidity, CO2 and CO were performed.
From the chosen class-rooms plus the staff room of the 20 schools, dust samples from the
floor, the horizontal surfaces and the ventilation shafts were collected. The status (“good” or
“bad”) of the schools was blinded for the investigator during sampling and analysis of dust
samples to avoid any prejudgement of the hazard of the dust. The selection of 20 schools
from the originally 75 for the sampling of dust limits the strength of the study, but this was a
necessity to make the sampling protocol feasible for one person with limited assistance from
laboratory technicians.
A cross sectional study describes the exposure and the effect at the same time. Hence, it
gives a cross section through the study base visualising merely the immediate effect of
exposure. No longitudinal effects were studied in this design. In the epidemiologic study no
control group is included, because the main purpose is to compare schools with low versus
high prevalence of BRS. However, some investigators would use the low prevalence
schools as the control group.
3.2. Sampling and handling of dust
Dust from the 20 schools was collected from the floor, the horizontal surfaces (shelves,
window frames, tables, boards etc.), and ventilation shafts. Dust samples were taken from
three to four classrooms and one staff room on each school. Dust samples from ventilation
shafts were only taken, if shafts were present, and if it was possible to open the shafts. The
dust samples were named regarding school number, type of sample, and room. For instance
4Gc, meaning school number 4, a floor sample (G), and room number 3 (c) sampled on that
school.
3.2.1. Floor dust
Floor dust was sampled with the HVS-3 sampler (ASTM Designation: D 5438-94, build by
AMI and SBI) connected to a normal “Miele” vacuum cleaner (figure 8) (ASTM 1994).
Dust is sucked into a cyclone and gravity makes it fall into a 250 ml jar made of glass. The
39

cyclone collects particles > 5 µm mean aerodynamic diameter. Sampling was done between
all the furniture, none of the furniture was moved (figure 9). After sampling the cyclone was
dissembled and cleaned with 70% ethanol. After returning to AMI the dust samples were
weighed and stored in the refrigerator until further processing the next day (Appendix 1).
Figure 9. The floor was vacuumed without moving any of the furnitures.
Figure 8. The HVS-3 connected to a normal “Miele” vacuum
cleaner. A bicycle computer measures the vacuumed distance.
40

3.2.2. Surface dust
Collection of surface dust was done by a portable vacuum cleaner with a special nozzle
(figure 10). The nozzle contains a filter in a filter box which retains 100% of the particles >
10 µm, 95 % of the particles between 1 – 10 µm, 81% of the particles between 0.5 – 1.0 µm,
and 74 % of the particles between 0.3 – 0.5 µm (ALK Abelló, Hørsholm, Denmark)
(Johansen, Heinig et al. 1990). The cleaning standard of the room was noted and collection
was made on all horizontal- and near horizontal surfaces. A new clean nozzle with a new
filter was used for sampling in the next room. On return to AMI each filter box with the dust
sample was weighed and stored in the refrigerator until further processing the next day. The
Nozzle was cleaned in lukewarm soap water and 70% ethanol (Appendix 1).
Figure 10. Collection of surface dust from horizontal or near horizontal surfaces with a portable vacuum
cleaner with a special nozzle and a filter inside the nozzle.
3.2.3. Ventilation shafts
The same portable vacuum cleaner and nozzle as described above was used in sampling
from ventilation shafts. The appearance of the shaft was noted together with the appearance
of the dust in the shaft (figure 11). Dust was collected on the floor of the shaft and on the
sides until about 0.5 m inside the shaft. On return to AMI each filter box with the dust
41

sample was weighed and stored in the refrigerator until further processing the next day
(Appendix 1).
3.2.4. Sieving and fractionation of the dust
42
Figure 11. Collection of dust from exhaust ducts
with a portable vacuum cleaner with a special
nozzle and a filter inside the nozzle. Some exhaust
ducts was in the sealing others in the wall.
The collected dust samples were at first torn in a knife mill (Type A10, Funkenstört, KB
5/10, Janke & Kunkel GmbH & Co.KG) for one minute, and sieved on a sieving machine
(Retsch, Type VE 1000, F. Kurt Retsch GmbH & Co.KG, Haan, Germany) for 15 minutes
at amplitude of 1.5 mm. The samples were sieved on a 300 µm filter (figure 12). The dust
sample less than 300 µm was brushed into a container, weighed and stored at –20 °C. The
fraction > 300 µm was stored at –20 °C. The knife mill and the 300 µm filter were cleaned
between sievings. Storage and sieving was modified from ASTM D 5438-94 (ASTM 1994).

Knife mill
Figure 12. Sieving of dust samples
Each dust sample < 300 µm was divided into six aliquots. Two samples of 50 mg dust for
the in vitro methods. One sample of 5 mg for endotoxin analysis. One sample of 100 mg for
viable counts of microorganisms. One sample of 500 mg for mite, dog and cat allergen
determination (by the ALK Abelló, Hørsholm, Denmark). One for determination of organic
fraction (figure 13). The rest was stored at -20°C. (Samples for viable counts of
microorganisms were divided about half a year after the others, and samples for
determination of organic fraction were divided about a year and a half after the others).
When dividing the dust samples, the fibre fractions (hair, fibres from clothes etc.) were
weighed and the percentage of the total sample was calculated. This percentage was used in
the division of each dust sample, and dust was randomly picked out from the originally
sample. Each dust fraction was stored in 10 ml centrifuge tubes at -20°C.
The 50 mg samples for the in vitro methods were sterilised by γ-irradiation at a dose of 35
kGy (Risø, Denmark).
43
Sieving machine

10 mg for
organic content
Figure 13. The dust samples were divided in six different samples. The rest was stored at -20°C.
3.3. Test compounds
Pure lipopolysaccharides from different Gram negative bacteria, glucans from yeast or a
Gram negative bacterium, pure chemicals with known allergenic effect, and different
surfactants and cytokines were used to characterise the in vitro methods.
3.3.1. Lipopolysaccharides (LPS):
LPS from Escherichia coli O55:B5 (Sigma, L 2637), used as model compound in the
developmental phase of the bioassays and as positive control. LPS preparations from
Klebsiella pneumoniae (Sigma, L 1770), Pseudomonas aeruginosa (Sigma L 8643),
Salmonella enteritidis (Sigma L 2012) and detoxified LPS from E. coli (Sigma L 9023)
were used as test compounds.
3.3.2. Glucans
A549
2 x 50 mg for
bioassays
Irradiation
Dust sample < 300 um
5 mg for endotoxin
mesurements
Pyrogene free
water
Glucan from bakers yeast, Saccharomyces cerevisiae (Sigma G 5011), Zymosan, cell wall
preparation also from bakers yeast (Sigma Z 4250) and Curdlan, β-[1-3]-D-glucan from the
Gram negative bacteria Alcaligenes faecalis (Sigma C 7821). The glucans are in general
insoluble in water. However a suspension of glucan in growth media was prepared.
44
100 mg for
microbiology
Tween 80
Tween 80
glycerol
Monocyte LAL Total counts Viable counts
500 mg for allergen
mite, dog and cat

3.3.3. Pure compounds
Nickel sulphate NiSO4• 6H2O, min. 99% (Merk, Art. 6727), Methyl metacrylate (MMA)
CH3-C (=CH2)-COOCH3 (Heraesus Kulzer via Dansk Hollandsk Ædelmetal A/S,
Denmark), and formaldehyde HCOH 37 - 38% (Merk).
3.3.4. Cytokines
Human recombinant Tumor Necrosis Factor-α (TNF-α) (Genzyme via Bie and Berntsen A-
S, Denmark, GENTNF-H) was used as positive control in the A549-bioassay. Granulocyt
macrophage colony stimulating factor (GM-CSF) (Genzyme via Bie and Berntsen A-S,
Denmark, GENRH-CSF-C) was tested as a possible positive control in the THP-1 and the
U937 bioassay.
3.3.5. Surfactants
Dodecylbenzene sulfonic acid (SDBS) (>99% pure, C18H30O3S, Acros, CAS no. 27176-
87-0), a representative of the linear alkylbenzene sulfonates (LAS-compounds) was
neutralised with NaOH (Merck 1.06498). Sodium dodecyl sulphate (SDS) (>99% pure,
C12H25NaO4S, Fluka Biochemika, EEC no. 2057881) a representative of the alkyl
sulphates. Coconut oil fatty acid sodium salt was obtained from the corresponding fatty acid
mixture (Domoclean, Horsens, Denmark) by neutralisation with NaOH. The sodium salt of
coconut oil fatty acids (primarily with 12 and 14 carbon atoms) was selected as the
representative of soaps. SDBS, SDS and the coconut soap are anionic surfactants. Genapol
X-80, 10% solution (Calbiochem, CAS no. 9002-92-0), has a lipophilic dodecyl alcohol
(nC12) group connected to a chain of eight ethoxylate (EO) units.
3.3.6. Dust
Dust samples were suspended in cell culture media and solicited in a water bath (Branson
2200 E3, 47 kHz ± 6%) for three times one minute just before addition to the cells.
45

Day 1:
Friyday
Day 2:
Monday
Day 3:
Tuesday
Day 4:
Wensday
Day 5:
Thursday
Figure 14. The A549 bioassay.
3.4. A549 bioassay
Setting up
6
1*10 cells/T75 Culture flask
Changing the media
in the culture flask
Trypsination
Seeding in Wells
5
1*10 cells/well
Addition of Agents
Start of the experiment
Cytokin ELISA
on the culture media
The human lung epithelial cell line A549 (ATCC no. CCL-185) was grown in Ham's F12
supplemented with 100 IU/ml penicillin and 100 µg/ml streptomycin, 2 mM L-Glutamine
and 10% heat inactivated Foetal Bovine Serum (FBS) (all reagents from GIBCO BRL, Life
Technologies, Denmark). The cells were grown at 36.5 °C and 5% CO2 in tissue culture
flasks (GIBCO BRL Life Technologies, Denmark). After treatment of the monolayer with
trypsin (GIBCO BRL Life Technologies, Denmark) 1*10 5 cells were transferred to each
well in 24 well multi-dishes (GIBCO BRL, Life Technologies, Denmark) and incubated for
48 hours at 36.5 °C and 5% CO2, with a change of medium after 24 hours. The media was
removed and supplemented with one ml of the test sample suspended in fresh media. Six
concentrations of each test sample in triplets were tested at the same time as two positive
control samples (100 µg LPS (E. coli)/ml and 10 ng TNF/ml) (See Appendix 2a). The IL-8
46
24 hours
48 hours
24 hours
Harvest and freeze
of the culture media

secretion was measured in the media after 24 hours incubation by ELISA technique. The
method is also described in Paper I, Paper II and Paper III (figure 14). Before the media was
harvested the morphology of the cells were viewed in a microscope, to see if the test
compound or the dust sample had any visual cytologic effect on the cells (rounded, detached
or lysed cells).
3.5. Monocyte assays (U937 and THP-1 bioassay)
The human premonocytic cell line U937 (ATCC no. CRL-1593.2) was grown in RPMI
1640 with 25 mM Hepes and glutamax-I supplemented with 100 IU/ml penicillin and 100
µg/ml streptomycin, and 10% heat inactivated Foetal Bovine Serum (FBS) (all reagents
from GIBCO BRL, Life Technologies, Denmark). The cells were grown at 36.5 °C and 5%
CO2 in tissue culture flasks (GIBCO BRL Life Technologies, Denmark). The cells were
centrifuged at 131g, 5 °C for 10 minutes and suspended to 6*10 5 cells/ml and incubated at
36.5 °C and 5% CO2. The next day the cell suspension was diluted to 8*10 5 cells/ml and
500 µl (4*10 5 cells) were transferred to each well in 24 well multi-dish (GIBCO BRL, Life
Technologies, Denmark). A total of one ml of the test sample suspended in fresh media was
added. Six concentrations of each test sample in triplets were tested at the same time as two
positive control samples (50 ng LPS (E. coli)/ml 5 ng TNF-α/ml). See appendix 2a. The IL-
8 secretion to the media was measured after 24-hours incubation by ELISA techniques. Se
also (Hansen, Nexø et al. 1997) (figure 15). The cells were viewed in a microscope after the
24 hours incubation time.
The monocytic cell line THP-1 (ATCC no. TIB-202) was grown in RPMI 1640 with 25 mM
Hepes and glutamax-I supplemented with 100 IU/ml penicillin and 100 µg/m streptomycin,
10% heat inactivated Foetal Bovine Serum (FBS) (all reagents from GIBCO BRL, Life
Technologies, Denmark), 50 µM 2-mercaptoethanol (Merck 15433), and 1 mM sodium
pyrovate (Sigma, SIGS-8636). The cells were grown at 36.5 °C and 5% CO2 in tissue
culture flasks (GIBCO BRL Life Technologies, Denmark). The cells were centrifuged at
131g, 5 °C for 10 minutes and suspended to 8*10 5 cells/ml and incubated at 36.5 °C and 5%
CO2. The next day the cell suspension was diluted to 8*10 5 cells/ml and 500 µl (4*10 5 cells)
were transferred to each well in 24 well multi-dish (GIBCO BRL, Life Technologies,
Denmark). A total of one ml of the test sample suspended in fresh media was added
47

Stimulation and harvest of the cell culture media were as described for the U937 cell line
(figure 15) (See appendix 2a).
Unless stated specifically, the cells of the A549 and the monocytic assays appeared normal
in the microscope at the end of the experiment.
Figure 15. The monocytic bioassays
3.6. ELISA
Day 1:
Friyday
Day 2:
Monday
Day 3:
Tuesday
Day 4:
Wensday
Setting up
6
1.5*10 cells/T75 Culture flask
Changing the media
by spinning
24 hours
Seeding in Wells
5
4*10 cells/well
Addition of Agents
Start of the experiment
24 hours
Harvest and freeze
of the culture media
Cytokin ELISA
on the culture media
Measurements of the secreted IL-8 were performed by standardised ELISA kits (Genzyme
and R&D systems via Bie & Berntsen, Denmark). See appendix 2b. The repeatability and
the limit of detection (LOD) of the ELISA are discussed later in section 4.4.2. A test of IL-8
secretion after LPS stimulation of the A549 cells in both assays, showed no difference
between the measured concentrations of IL-8 in the two assays.
48

3.7. Microbiological parameters
For the measurement of endotoxin content, 5 mg dust sample was suspended in 5 ml
pyrogen free water (BioWhittaker, Biotech, Denmark). The suspension was rotated for 2
hours at room temperature in a blood turner, and centrifuged at 1000 g for 10 min. The
supernatant was transferred to three 1.8 ml cryo tubes (Life Technologies, Denmark) and
stored at –80 °C. The concentration of endotoxin in the dust samples was measured by the
standardised Limulus Amebocyte Lysate (LAL) assay (Kinetic-QCL kit, BioWhittaker,
Biotech, Denmark). See appendix 2b. A standard curve was obtained from E. coli O55:B5
reference endotoxin and used to measure concentrations in the form of endotoxin units (EU)
(1 ng endotoxin = 12.5-15.5 EU, due to different lots). Correlation coefficients for the
standard curve was ≥ 0.98, and CV for the samples analysed was ≤ 20%. Blank samples
may not be over the value of the lowest standard. The Limit of detection (LOD) of the
method was 0.01 EU/ml = approx. 0.5 EU/mg dust (depending on the weight of the dust
sample).
For preparation of dust for total counts and cultivation of microorganisms, approx. 100 mg
of dust was suspended in 8.0 ml pyrogen free water containing 0,05% Tween 80 (Merck,
Struers, Denmark). Two times 1.5 ml dust suspension was transferred to two cryo-tubes,
containing 0.75 ml 75% glycerol (Bie & Berntsen, Denmark), for viable counts. The tubes
were stored at –80 °C until assaying. Bacteria (incl. actinomycetes) were cultivated on
nutrient agar plus actidione (200 mg/l) (Merck, Struers, Denmark), actinomycetes were
cultivated on 10% nutrient agar plus actidione (200 mg/l) and fungi were cultivated on
dichloran glycerol (DG 18) agarbase (Oxoid, Struers, Denmark) with chloramfenicol (100
mg/l) (Oxoid, Struers, Denmark). The plates were incubated at 25 °C for 7 days. See
appendix 2c. The LOD of the method was 120 cfu/100 mg dust (depending on the weight of
the dust sample).
The microbiological section at AMI performed the microbiological analysis (LAL assay and
the cultivations), using the standardised routine methods described above. For further details
se Appendix 2 and the references Douwes and Versloot et al, 1995; Würtz and Kildesø et al,
1999.
49

3.8. Allergens
The content of allergens from the mites Dermatophagonides peteronyssinus (Der p), D.
farinae (Der f) and D. microceras (Der m), cats Felis domesticus (Fel d) and Dogs Canis
familiaris (Can f) in the samples from floor dust and from exhaust ducts were measured by
ALK Abelló, Hørsholm, Denmark with ELISA techniques.
3.9. Organic content of the dust samples
The content of organic matter was determined by incineration. Whatman® CF/C glasfiber
filters of 25 mm (Frisenette Aps, Denmark) were weighed and placed on 40 mm dishes of
glass. About 10 mg of dust was placed on the filter. Filter plus dust were weighed and
acclimatised for 24 hours in the weighing room. The next day the filter plus dust were
weighed again and placed in a preheated muffle furnace (Carbolite GM3) at 480°C for four
hours. When the oven was cooled down the filters were placed in the weighing room for
acclimatisation. The next day the filters were weighed and the content of organic fraction
calculated.
3.10. Statistical models and methods
From the A549 bioassay and the monocyte bioassays the inflammatory potential of the test
compound (endotoxin, glucans, pure chemical samples, dust), termed the potency factor
(PF), was expressed as the slope of the initial linear part of the dose response curve, i.e. the
released IL-8 versus the concentration of dust. To reduce day-to-day variation in the
bioassay the PF was standardised against the value of a positive control, relative to the mean
of all the positive controls in the study (Paper I):
PF = α/(⎺X/T)
Where α is initial linear slope of the dose response curve,⎺X is the mean of three control
samples measured on the same day, and T is the target value being the mean of multiple
control measurements on different days. These values were calculated in Excel (Windows
95).
50

The PF reflects the potency of the test substance to stimulate the cells to an IL-8 response
(PF>0). No response above the background value equals no potency of the test substance
e.g. the slope of the dose response curve is zero (PF=0). A test substance with an inhibitory
effect is not likely to be seen in this experimental design as the background values are close
to zero and values below zero will not be significantly different from the background. Thus
negative PF will not appear. If inhibitory effects of a substance are sought, an inhibition of a
positive signal obtained by a known test substance could be performed. This was not
investigated in this study. The PF was calculated in ng IL-8/µg test substance for better
comparison between the PF of the different cell lines and between PF of different test
substances including dust.
Correlation test of data not following a normal distribution was performed by the non-
parametric Hotelling-Pabst rank correlation test, with α = 0.05%. Test for difference in rank
sums was tested by the Mann-Whitney U test.
3.11. Quality control and method evaluation
3.11.1. Bioassays
The analytical performance of the method was evaluated using the method evaluation
design (Hansen, Olsen et al. 1991) based on linear regression analysis of the relationship
between measured and known concentrations of endotoxin in a series of test samples. A
standard curve with LPS from E. coli O55:B5 was prepared in triplicate (A549 bioassay in
the concentrations 0, 25, 50, 100, 150, and 200 µg/ml, U937 and THP-1 bioassay in the
concentrations of 0, 10, 25, 50, 100 and 250 ng/ml). The standard curve equation was
obtained by linear regression analysis of all the 18 values. At the same time two series of
test samples, also with LPS from E. coli O55:B5, were analysed (A549 bioassay in the con-
centrations of 0, 30, 60, 90, 120, 150, and 180 µg/ml, U937 and THP-1 bioassay in the
concentrations of 0, 15, 30, 70, 120, and 200 ng/ml). The test samples were tested randomly
in triplicate. The concentration of IL-8 in the supernatant of all standard and test samples
was measured by an IL-8 ELISA. The mean values of the triplicates of each test sample
were read on the standard curve to convert the measured concentration of IL-8 to a
corresponding concentration of LPS (figure 16). Linear regression analysis of the measured
versus spiked concentration of LPS in the test samples provides an estimate of the
51

uncertainty (δ) of the method. The LOD was estimated as 3δ (Long and Winefordner 1983),
in the AMIQAS programme. Furthermore the two series of test samples permitted a
statistical test for linearity of the method (Hansen, Olsen et al. 1991).
IL-8
Measured IL-8 secretion
from the test sample
Figure 16. Schematic presentation of how the mean values of the triplicates of each test sample were read on
the standard curve to convert the corresponding concentration of IL-8 to a measured concentration of LPS. The
measured versus the known concentration of added LPS were then compared in the method evaluation.
Two positive control samples (LPS from E. coli and TNF-α) from all the experiments
performed were included in a control chart to assure that the method stayed in control. In
short, a control chart consists of a central line (target value) and upper and lower control
limits. Control results plotted in the chart should fall between the control limits. Results
exceeding the control limit may be taken as a “probable” analytical problem and it is
assumed that the method is “out of control”. The IL-8 secretion from the two controls (LPS
and TNF) is keyed in the AMIQAS programme and an X-R control chart is plotted. The X-
R control chart is described in the international standard ISO 8258 (Hjort, Nielsen et al.
1992).
3.11.2. Microbiological parameters
Corresponding LPS concentration
Microbiological parameters were measured after standard procedures and by commercial
kits.
52
St. Curve
LPS

4. Development of in vitro methods
In the present study the human lung epithelial cell line A549 was selected for the in vitro
assay because it has retained many of the original characteristics of a lung type II cell
(Lieber, Smith et al. 1976). Studies have shown that the A549 lung epithelial cell line has the
capacity of IL-8 secretion after stimulation with dust (Allermann and Poulsen 2000;
Palmberg, Larsson et al. 1998; Saraf, Larsson et al. 1999).
Mononuclear cells (MNC) freshly isolated from blood were previously shown to be much
more sensitive to LPS than the A549 cell line, but a large individual variation in the IL-8
secretion from the MNC was also found (Hansen, Nexø et al. 1997). A human macrophage
cell line may be the preferred cell type, but at that time no such cell line existed for
commercial use. The mouse macrophage cell line, RAW264.7, is used by other researchers
(Hirvonen, Nevalainen et al. 1997a; Ruotsalainen, Hirvonen et al. 1998; Hirvonen,
Nevalainen et al. 1997b). However, a human cell line for comparison with human data was
preferred. The next choice was to use a human monocyte cell line. Two cell lines were
chosen: The U937 and the THP-1 monocytic cell lines. Results from method evaluation and
test of pure agents for all three cell lines are shown below.
In the development of the in vitro bioassays for evaluation of dust from the indoor
environment, single components, generally found in the dust, were used to evaluate the
usability of the assays. The test substances chosen were: LPS from four different Gram
negative bacteria, three glucans from different yeast and a Gram negative bacteria, three
known contact allergens, and four irritants. Endotoxins (LPS) are known from the indoor
environment, but especially in the cotton industry clear correlations between endotoxin
exposure and pulmonary health symptoms are found (Castellan, Olenchock et al. 1987;
Rylander 1987). The effect of endotoxins on the lungs is verified in human exposure
experiments where inhaled endotoxin is shown to e.g. reduce the forced vital capacity in one
second (FEV1), raise the airway resistance and induce both local and systemic inflammatory
responses (Michel 1997). Correlations between the concentration of endotoxin in dust from
the indoor environment and BRS are also shown (Gyntelberg, Suadicani et al. 1994; Teeuw,
Vandenbroucke-Grauls et al. 1994). Earlier experiments have also shown that LPS induces
53

IL-8 secretion from lung epithelial cells (Paper I and II), and macrophages in vitro
(Palmberg, Larsson et al. 1998). Glucans are also known pollutants in the indoor
environment (Fogelmark, Goto et al. 1991; Rylander, Williams et al. 1993). The importance
of glucans for development of airway problems is not as well documented as the effect of
endotoxin (Rylander 1998). However, correlation between the content of glucans in the
indoor air and some BRS is found (Rylander, Persson et al. 1992; Rylander 1997b).
Macrophages have specific receptors for glucan on their surface (Goto, Yuasa et al. 1994),
also leukocytes may be able to bind glucans (Rylander, Williams et al. 1993 ). The allergens
tested are nickel sulphate, formaldehyde and methyl metacrylate (MMA). Nickel sulphate is
a strong contact allergen found in e.g. materials of stainless steel and materials with alloy of
nickel such as paper clips, keys, buttons, buckles etc. In a study of hospital cleaners, the
nickel content of the cleaning water was found to increase during the cleaning process
(Clemmensen, Menne et al. 1981). This could be explained by nickel being found in dust
from the indoor environment (Lemus, Abdelghani et al. 1996). Formaldehyde is also a
relevant pollutant of the indoor air, as it can evaporate from building materials, furnitures
and office machines (WHO 1989b). Formaldehyde is a known strong allergen also used as
preservative in cosmetics and consumer products (WHO 1989b). MMA is by Van der Walle
(1982) defined as a moderate sensitiser (Van der Walle, Klecak et al. 1982), however, only a
small number of persons working with MMA showed a positive patch test towards MMA
(Mürer 1996). MMA is not relevant in the indoor environment but health problems are
known from e.g. the dental industry (Kanerva, Estlander et al. 1993). In this study MMA
was used as a model of a test substance irrelevant for the indoor environment, but may
posses contact allergenic properties. Surfactants are known irritants found in e.g. dust from
the indoor environment (Clausen, Wilkins et al. 1997). They are suspected to alter the tear
film in the eyes (Franck 1991) and correlations between surfactants from carpet cleaning and
BRS are found (Robinson, Tauxe et al. 1983). As surfactants are found in dust from the
indoor climate the question is whether surfactants do induce cytokine release from cells in
vitro or just bring on cytotoxic effects on the cell membrane.
The test substances were tested at least twice over time to give an impression of the
robustness of the methods.
54

Log TNF control (ng IL-8/ml)
100
10
1
1 10
Log LPS control (ng IL-8/ml)
Figure 17. The LPS control values plotted against the TNF control values from the A549 bioassay. Each point
is the mean of a triplet and the mean of all the values are shown together with the standard deviation. The LPS
control resulted in a mean value of 3.8 ± 2.1 ng IL-8/ml and a relative mean square error RMSE = 0.55. The
TNF control resulted in a mean of 12.4 ± 3.2 ng IL-8/ml and a relative mean square error RMSE = 0.26.
Log mean control value (ng IL-8/ml) -
background
100
10
1
0.1
LPS mean
TNF mean
0 10 20 30 40 50 60
Figure 18. Shown are the mean of the controls within an experimental day of the A549 bioassay with the
background subtracted. Each value represents the mean of 5 to 14 control values, tested in triplets in the
experimental series analysed on that day. The experiments were performed from week 38 (1997) until week 18
(1998). The standard variation of experiments performed on the same days range from 0.22 to 1.26 for the LPS
control and 0.99 to 3.83 for the TNF control.
55
week

4.1. Results from the A549 bioassay
4.1.1. Growth experiment
The A549 cell line had a generation time of 30.7 hours (1.3 days), this means that one cell
turns to 1.8 cells per day (Appendix 2b).
4.1.2. Quality control
The method evaluation was performed as described in section 3.11.1.
Method evaluation of the A549 bioassay with LPS from E. coli O55:B5 showed a linear dose
response curve up to 180 µg LPS per ml, with a slope not significantly different from 1,
indicating that the method is unbiased, when correcting against the positive control TNF. The
LOD defined as 3δ was approx. 17 µg LPS per ml (Hansen, Nexø et al. 1997). Control chart of
the LPS controls and the TNF controls indicate that the potency of the LPS standard is
depended on the individual batches, whereas the TNF control produced more homogenous
results (figure 17 and 18).
Small variations in the fitness of the cells and number of cells in each well, on the day of the
experiments (the day-to-day variation of the assay), are reflected in the background values
with a mean of 0.4 ± 0.34 ng IL-8/ml, and in the control values (figure 17). The standard
variation of experiments range from 0.02 in one experimental day to 0.2 in an other, for the
background values, 0.22 to 1.26 respectively for the LPS control and 0.99 to 3.83
respectively for the TNF control (figure 18). The relative mean square error (RMSE =
standard deviation/mean) of the background values was 0.85, for the LPS control it was 0.55
and 0.26 for the TNF control (figure 17) indicating that the variation are larger for the LPS
control than the TNF control.
4.1.3. Endotoxins
Four different endotoxins were tested and each had a characteristic dose response curve.
Figure 19 shows typical dose response curves of A549 epithelial cells stimulated with LPS
56

from K. pneumoniae, P. aeruginosa, E. coli, S. enteritidis and detoxified E. coli. The K.
pneumoniae curve increased rapidly up to 50 µg LPS/ml, at which the highest IL-8
concentration was measured, then the curve slowly decreased and levelled off. The P.
aeruginosa curve increased until stimulation with 100 µg LPS/ml where it reached a plateau.
The E. coli curve increased linearly over the whole stimulation range expressing highest IL-8
secretion at 200 µg LPS/ml. The S. enteritidis curve showed a small linear increase up to 200
µg LPS/ml. The maximal IL-8 secretion was found to be 4.1, 7.2, 4.1 and 1.5 ng IL-8/ml for
LPS from K. pneumoniae, P. aeruginosa, E. coli and S. enteritidis respectively (figure 19).
The detoxified endotoxin from E. coli did not result in stimulation of IL-8 secretion above
the background value. In earlier experiments the same shape of the curves with K.
pneumoniae, P. aeruginosa, and E. coli was seen (Hansen, Nexø et al. 1997).
IL-8 (ng/ml)
8
7
6
5
4
3
2
1
0
K. pneumoniae
P. aeruginosa
E. coli
S. enteritides
E. coli detox
0 50 100
LPS (ug/ml)
150 200
Figure 19. Dose response curves from the lung epithelial cell line A549 stimulated with LPS from four
different Gram negative bacteria and one detoxified LPS from E. coli. Each point is the mean of a triplet shown
together with the standard deviation.
57

The fall in IL-8 secretion when stimulating the lung epithelial cells with K. pneumoniae
could not be explained by cell death, as all the cells appeared normal in the microscope. In
table 3 the potency of LPS from the different bacteria is shown. LPS from K. pneumoniae
seems to have a greater potency than LPS from the other bacteria tested on the A549 cell
line. Next comes LPS from P. aeruginosa, E. coli and S. enteritidis.
Table 3. The potency factor PF of LPS from different Gram negative bacteria used to stimulate the A549 lung
epithelial cell line. The repeated measurements were performed over a period of 2 and a half year. Three
different batches of LPS from K. pneumoniae, P. aeruginosa, and E. coli and two different batches of LPS
from S. enteritidis were used.
Compound IL-8 induction
ng IL-8/µg
LPS
(K. pneumoniae)
LPS
(P. aeruginosa)
LPS
(E. coli)
LPS
(E. coli detoxified)
LPS
(S. enteritidis)
n.d. Not determined.
4.1.4. Glucans
0.12
0.38
0.38
0.12
0.16
0.04
0.09
0.16
0.03
0.06
0.04
0.05
0.06
0.08
0.03
0.02
58
LPS corrected
IL-8 induction
ng IL-8/µg
0.44
0.23
0.25
0.19
0.15
0.08
0.06
0.10
0.07
0.07
0.09
0.07
0.07
0.06
0.05
0.02
TNF corrected
IL-8 induction
ng IL-8/µg
n.d.
n.d.
n.d.
0.19
0.15
n.d.
n.d.
n.d.
0.06
0.04
0.03
0.03
0.04
0.03
0.04
0.02
0.001 0.001 0.001
0.02
0.005
0.04
0.002
0.04
0.003
Stimulation of the A549 cells with glucans is shown in figure 20. The only glucan
stimulating the IL-8 secretion from the A549 epithelial cells above the background was
glucan from the Gram negative bacteria Alcaligenes faecalis (Curdlan). The maximal IL-8
secretion from stimulation with Curdlan was 11.9 ng IL-8/ml. In table 4 the potency of
glucans is shown.

IL-8 (ng/ml)
14
12
10
8
6
4
2
0
Zymosan
Glucan
Curdlan
0 50 100 150 200 250 300 350 400 450 500
Glucan (ug/ml)
Figure 20. Dose response curves from the lung epithelial cell line A549 stimulated with glucans. Each point is
the mean of a triplet shown together with the standard deviation.
Table 4. The potency factor PF of glucans used to stimulate the A549 lung epithelial cell line.
Compound IL-8 induction
ng IL-8/µg
Curdlan 0.02
0.004
0.01
59
LPS corrected
IL-8 induction
ng IL-8/µg
0.01
0.02
0.01
TNF corrected
IL-8 induction
ng IL-8/µg
n.d.
0.01
0.01
Glucan 0 0 n.d.
Zymosan 0 0 n.d.
n.d. Not determined.
4.1.5. Chemical compounds
Three contact allergens (nickel sulphate, methyl metacrylate (MMA), and formaldehyde)
were tested on the A549 cells in the doses of 0-0.1% for nickel sulphate and 0-0.1% for
MMA. Formaldehyde stimulation of the A549 cells was done at very small doses (0-10.9
µg/ml (0.001%)) due to cytotoxic effects of the compound. The IL-8 secretion after
stimulation with nickel sulphate generally increased over the whole concentration range with

a maximal secretion of 0.74 ng IL-8/ml. The morphology of the cells was gradually changed
from normal flat and rounded to more edgy appearance as the dose increased (figure 21).
Stimulation with MMA resulted in IL-8 secretion not significantly above the background
(not shown). The cells appeared normal at the doses of 0 to 283 µg/ml (0.3 %) MMA, but
many dead cells was observed at the 9430 µg/ml (1%) dose. With formaldehyde the maximal
IL-8 secretion was 0.62 ng IL-8/ml. Stimulation in the dose range of 0-2.18 µg/ml (0-
0.0002%) the cells looked normal in the light microscope. Many cells in the wells stimulated
with doses of 10.9 µg/ml (0.001%) many of the cells were dead or dying. At doses of 32.7
µg/ml (0.003%) formaldehyde were dead, and also in the wells of. The Potency of these pure
chemicals is shown in Table 5.
A B
Figure 21. A) The normal appearance of the A549 lung epithelial cell line magnified 400 times. The arrow
pointing at cell division in the phase of cytokinesis where the two new cells are separating. B) The A549 cell
after 24 hours incubation with 0.1 % nickel sulphate magnified 400 times. The arrows are pointing at
pseudopodia remaining as the cells are rounding up.
Table 5. The potency factor of selected chemical compounds tested in the A549 bioassay.
Compound IL-8 induction
ng IL-8/µg
Nickel sulphate 0.0005
0.0008
Methyl metacrylate <
<
<
Formaldehyde 0.08
0.17
0.12
0.28
< The maximal IL-8 secretion is below the background
60
LPS corrected
IL-8 induction
ng IL-8/µg
0.0012
0.0011
<
TNF corrected
IL-8 induction
ng IL-8/µg
0.0002
0.0011
<
<
0.11
0.14

4.1.6. Surfactants
The three anionic surfactants, applied in the dose range of 0-0.01%, showed similar response
curves. The maximal IL-8 secretion was 2.02 (SDS), 1.1 (coconut oil), and 0.36 (SDBS) ng
of IL-8/ml (figure 22). The dose response curve (0-0.01%) of the non-ionic surfactant
(Genapol X-80) reached a peak at 10 µg/ml (0-0.01%) surfactant with an IL-8 secretion of
1.03 ng/ml. Only the non-ionic surfactant was tested twice and this showed that the
surfactant peaked at 50 µg/ml (0.005%) (not shown). All the cells appeared normal at the
blank and at the lowest concen-tration of surfactant, after the 24 hours of incubation. Some
of the cells stimulated with the anionic surfactants were dead at the concentration of 100
µg/ml (0.01%), and at the 1000 µg/ml (0.1%) all the cells were dead. The cells stimulated
with the non-ionic surfactant were normal at the concentration of 10 µg/ml (0.001%) and 50
µg/ml (0.005%) surfactant, but all cells were dead at concentrations above 100 µg/ml
(0.01%). Table 6 summarises the potency of the surfactants, with the Genapol X-80 being
most potent in the A549 assay, but with SDS resulting in the highest IL-8 secretion.
IL-8 (ng/ml)
2,5
2
1,5
1
0,5
0
SDS
LAS (C12)
Coconut soap
Nonionic
0 0,002 0,004 0,006
Surfactant (%)
0,008 0,01 0,012
Figure 22. Dose response curves from the lung epithelial cell line A549 stimulated with four different
surfactants. SDS is an alkyl sulphate, SDBS is a LAS compound. The coconut soap is a natural anionic
surfactants, whereas Genapol X-80 is a nonionic surfactant. Each point is the mean of a triplet shown together
with the standard deviation. The curves of SDS, SDBS and coconut soap are going towards zero for
concentrations above 0,01%.
61

Table 6. The potency factor (PF) of selected surfactants tested in the A549 bioassay.
Compound IL-8 induction
ng IL-8/µg
62
LPS corrected
IL-8 induction
ng IL-8/µg
SDS 0.018 0.031 0.018
SDBS 0.002 0.004 0.002
TNF corrected
IL-8 induction
ng IL-8/µg
Coconut oil 0.009 0.009 0.008
Genapol X-80 0.077 a
0.017 b
0.093 a
0.076
missing
a
0.013 b
a
Peak at 10 µg/ml (0.001%)
b
Peak at 50 µg/ml (0.005%)
4.2. Results from the U937 bioassay
4.2.1. Growth experiment
The U937 cell line had a generation time of 24.6 hours (1.03 days), this means that one cell
turns to 2 cells per day (Appendix 2b).
4.2.2. Quality control
Method evaluation with LPS (E. coli O55:B5) revealed dose response linearity up to 200
ng/ml with a slope not significantly different from 1 when correcting data with the LPS- or
the TNF control (p=0.27 and p=0.16 respectively). However, data were not normally
distributed hence the results of the method evaluation should only be considered indicative,
as a normal distribution is a condition for making the linear regression analysis. The
background values was 0.195 ± 0.15 ng IL-8/ml. The IL-8 secretion after stimulation of the
U937 cells with the two controls is shown in figure 23. The TNF control seemed more stable
than the LPS control so the TNF control was used for further data analysis. LOD was 47 ng
LPS/ml when correcting against the TNF control. A larger variation was observed with the
Granulocyte macrophage colony-stimulating factor (GM-CSF) than with TNF, and the GM-
CSF was not used further as a control.

Log TNF control (ng IL-8/ml)
10
1
0.1 1
Log LPS control (ng IL-8/ml)
10
Figure 23. The IL-8 secretion of the LPS control is plotted against the IL-8 secretion of the TNF control from
the U937 bioassay. Each point is the mean of a triplet and the mean of all the values are shown together with
the standard deviation. The LPS control of all the experiments performed including the ones shown above,
resulted in a mean of 1.6 ± 1.5 ng IL-8/ml and a RMSE = 0.94. The TNF control resulted in a mean of 2.1 ±
0.74 ng IL-8/ml and a RMSE = 0.35.
IL-8 (ng/ml)
1.8
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
0 50 100 150 200 250
LPS (ng/ml)
Figure 24: Example of dose response curves from the monocytic cell line U936 stimulated with LPS from four
different Gram negative bacteria. Each point is the mean of a triplet shown with the standard deviation.
63
E. coli
S. enteritidis
K. pneumoniae
P. aeruginosa

4.2.3. Endotoxins
In figure 24 dose response curves from the U937 cell line stimulated with LPS from E. coli,
S. enteritidis, K. pneumoniae and P. aeruginosa are shown. The curves of LPS from E. coli
and S. enteritidis increased up to stimulation with 100 ng LPS then the curve decreased or
levelled of. The maximal IL-8 secretion was 1.38 and 1.57 ng IL-8/ml, respectively.
Stimulation with LPS from K. pneumoniae and P. aeruginosa resulted in increased IL-8
secretion over the whole concentration range. The maximal IL-8 secretion was 1.0 and 0.27
ng IL-8/ml, respectively. Table 7 shows the potency of LPS from the different Gram
negative bacteria, ranking the endotoxins as follows: S. enteritidis > E. coli > K. pneumoniae
> P. aeruginosa.
Table 7. The potency of LPS from different Gram negative bacteria used to stimulate the U937 monocytic cell
line.
Compound IL-8 induction
ng IL-8/ug
LPS
(K. pneumoniae)
LPS
(P. aeruginosa)
LPS
(E. coli)
LPS
(S. enteritidis)
n.d. Not determined.
4.2.4. Glucans
3.43
10.7
1.86
7.61
12.6
15.5
41.5
30.8
64
LPS corrected
IL-8 induction
ng IL-8/ug
3.90
5.35
1.92
3.46
12.5
7.32
79.0
20.1
TNF corrected
IL-8 induction
ng IL-8/ug
n.d.
n.d.
n.d.
n.d.
n.d.
Stimulation of the U937 cell line with different glucans only produced a slightly higher IL-8
secretion than the background, with a maximal secretion below 0.3 ng IL-8/ml (not shown).
The cells all appeared normal after 24 hours of incubation with different glucans.
4.2.5. Chemical compounds
The U937 monocyte cell line was stimulated with three contact allergens (nickel sulphate,
MMA, and formaldehyde). The dose range for nickel sulphate and MMA was 0-0.1%, and 0-
0.01% for formaldehyde. The dose response curves for nickel sulphate gave a bell shape
appearance with a maximal IL-8 concentration of 1.90 and 3.21 ng IL-8/ml at 500 µg/ml
n.d.
n.d.
n.d.

(0.05%) nickel sulphate (not shown). The morphology of the cells after 24 hours incubation
with nickel sulphate changed gradually as the concentration increased. In the range of 50-
100 µg/ml (0.005-0.01%) about ½ of the cells were dead, and at higher concentration the
morphology of the cells changed to more oblong appearance, and about ¾ of the cells were
dead. Stimulating the U937 cells with MMA gave an increased IL-8 secretion over the whole
stimulation range with a maximal IL-8 secretion of 0.44 and 0.2 ng IL-8/ml at 943 µg/ml
(0.1%) MMA (not shown). The appearance of the cells was normal after 24 hours of
incubation. Stimulation of the U937 cells with formaldehyde (0-0.01%) did not result in an
IL-8 secretion above the background, and cytotoxic effect was observed after stimulation
with 500 µg/ml (0.0005%) formaldehyde. Table 8 summarises the potencies of these
different chemicals on the U937 cell line, with nickel sulphate being the most potent of the
three test substances.
Table 8. The potency factor of the chemical compounds used to stimulate the U937 monocytic cell line.
Compound IL-8 induction
ng IL-8/µg
65
LPS corrected
IL-8 induction
ng IL-8/µg
0.0086
0.0044
0.0003
Nickel sulphate 0.0064
0.0037
n.d.
Methyl metacrylate 0.0003
n.d.
0.00004
0.00004
n.d.
Formaldehyde
n.d. Not determined.
< < n.d.
< The maximal IL-8 secretion is below the background.
4.2.6. Surfactants
TNF corrected
IL-8 induction
ng IL-8/µg
n.d.
None of the different surfactants (SDS, SDBS, coconut oil and Genapol X-80) induced an
IL-8 secretion above the background, when applied to the U937 monocytic cells in the range
of 0-0,1%.
4.3. Results from the THP-1 bioassay
4.3.1. Growth experiment
The THP-1 cell line had a generation time of 40.5 hours (1.7 days), this means that one cell
turns to 1.6 cells per day (Appendix 2b).

4.3.2. Quality control
Method evaluation with LPS from E. coli O55:B5 revealed a slope significantly different
from 1 (p=0.01). The curve was linear (p=0.98), but the data were not normally distributed,
when data were corrected against the TNF control. However, when correcting the data
against the LPS control the slope was not significantly different from 1 (p=0.74), the curve
was linear (p=0.54), and normal distributed. The LOD of the THP-1 bioassay was 2.1 ng
LPS/ml, when correcting against the LPS control. The IL-8 secretion after stimulating the
THP-1 cells with the two controls are shown in figure 25. Both controls seemed stable, but
because of the small IL-8 secretion resulting from stimulation with TNF and the improved
method evaluation after correcting against the LPS control, the LPS control was used for
further data analysis. Stimulation with Granulocyte macrophage colony stimulating factor
(GM-CSF) as a control, gave a smaller response than TNF and was not used further. The
background values for the THP-1 bioassay was 0.04 ± 0.02 ng IL-8/ml, however below the
LOD of the ELISA kit (Se section 4.4.2.).
Log TNF control (ng IL-8/ml)
1
0.1
0.01
0.1 1 10
Log LPS control (ng IL-8/ml)
Figure 25. The IL-8 secretion of the LPS control is plotted against the IL-8 secretion of the TNF control from
the THP-1 bioassay. Each point is the mean of a triplet and the mean of all the values are shown together with
the standard deviation. The LPS control resulted in a mean of 1.3 ± 0.5 ng IL-8/ml and a RMSE = 0.38. The
TNF control resulted in a mean of 0.14 ± 0.04 ng IL-8/ml and a RMSE = 0.29.
66

4.3.3. Endotoxins
In figure 26 dose response curves from the THP-1 cell line stimulated with LPS from E. coli,
S. enteritidis, K. pneumoniae and P. aeruginosa are shown. All the curves increased over the
whole concentration range with LPS from S. enteritidis being the most potent, then E. coli
and K. pneumoniae and P. aeruginosa giving only a small increase in the IL-8 secretion. The
maximal IL-8 secretion after stimulation with LPS from S. enteritidis was 5.6 ng IL-8/ml,
and for LPS from E. coli it was 2.9 ng IL-8/ml. The maximal IL-8 secretion after stimulation
with LPS from K. pneumoniae and P. aeruginosa was 0.3 and 0.2 ng IL-8/ml, respectively.
Table 9 shows the potency of LPS from the different Gram negative bacteria. The rank S.
enteritidis > E. coli > K. pneumoniae > P. aeruginosa was observed.
IL-8 (ng/ml)
6
5
4
3
2
1
0
E. coli
S. enteritidis
K. pneumoniae
P. aeruginosa
0 50 100 150 200 250
LPS (ng/ml)
Figure 26: Example of dose response curves from the monocytic cell line THP-1 stimulated with LPS from
four different Gram negative bacteria. Each point is the mean of a triplet shown with the standard deviation.
67

Table 9. The potency of LPS from different Gram negative bacteria used to stimulate the THP-1 monocytic
cell line.
Compound IL-8 induction
ng IL-8/ug
LPS
(K. pneumoniae)
LPS
(P. aeruginosa)
LPS
(E. coli)
LPS
(S. enteritidis)
n.d. Not determined.
4.3.4. Glucans
1.65
1.80
0.87
0.91
12.4
11.1
13.1
20.3
68
LPS corrected
IL-8 induction
ng IL-8/ug
1.83
2.14
1.14
1.22
20.1
18.9
16.8
24.6
TNF corrected
IL-8 induction
ng IL-8/ug
n.d.
1.50
n.d.
0.96
n.d.
Stimulation of the THP-1 monocytic cell line with different glucans gave a linear increase of
the IL-8 secretion, as the concentration of glucan increased (not shown). However, large
doses were required (µg) to stimulate the cells, and low potencies in the range of 0.58x10 -6
ng IL-8/µg to 1.6x10 -4 ng IL-8/µg were observed. The glucans ranked Zymosan > Curdlan >
Glucan.
4.3.5. Chemical compounds
The THP-1 monocyte cell line was stimulated with three contact allergens (nickel sulphate,
MMA and formaldehyde). The cells were stimulated in the range of 0-0.1% for nickel
sulphate and MMA, and 0-0.01% for formaldehyde. The two dose response curves for nickel
sulphate gave a bell shape appearance with a maximal IL-8 concentration of 29.9 and 23.7
ng IL-8/ml at 500 µg nickel sulphate /ml (0.05%) (not shown). The morphology of the cells
after 24 hours incubation with nickel sulphate changed gradually as the concentration
increased. The cells stimulated with 0-100 µg/ml (0-0.01%) appeared normal. In the wells
with 500 µg/ml (0.051%) the morphology of the cells changed to more oblong appearance.
At the higher concentrations of nickel sulphate all the cells were dead. Stimulating the THP-
1 cells with MMA did not result in IL-8 secretion above the background (Not shown). The
appearance of the cells was normal after 24 hours of incubation with MMA. Stimulation of
the THP-1 cells with formaldehyde in the dose of 0-0.01% resulted in a bell shaped dose
response curve with a maximal IL-8 secretion of 0.37 and 0.41 ng IL-8/ml at 500 µg
formaldehyde /ml (0.0005%). The cells appeared normal at stimulation with up to 500-1000
8.15
n.d.
16.3

µg formaldehyde/ml (0.0005-0.001%), after 24 hours of incubation. Table 10 summarises
the potencies of these different chemicals on the THP-1 cell line, with nickel sulphate being
the most potent of the three chemicals.
Table 10. The potency of different pure test substances used to stimulate the THP-1 monocytic cell line.
Compound IL-8 induction
ng IL-8/ug
Nickel sulphate 0.049
0.062
Methyl metacrylate 0
0
Formaldehyde 0.001
0.001
4.3.6. Surfactants
69
LPS corrected
IL-8 induction
ng IL-8/ug
0.069
0.060
0
0
0.0012
0.001
TNF corrected
IL-8 induction
ng IL-8/ug
0.084
0.050
0
0
0.0014
0.001
Stimulation of the THP-1 monocytic cells with SDS, SDBS, Genapol X-80 and coconut oil
in the range of 0-0.005%, all resulted in an IL-8 secretion around the background for the IL-
8 ELISA. The cells all appeared normal at the concentrations below 10 µg/ml (0.001%),
whereas many cells were dead after 24 hours of incubation at higher concentrations.
4.4. Discussion of development of in vitro methods
4.4.1. Growth experiment
The growth experiments were performed with cells in log phase of the growth curve (figure
27). The A549 cell line had a doubling time of 30.7 hours; the THP-1 cell line had a
doubling time of 40.5 hours and the U937 cell line a doubling time of 24.6. From a simple
experimental viewpoint the fastest growing cells are most attractive since the acquired
number of cells for an experiment is easier accessible. In the literature the U937 monocytic
cell line is often denoted premonocytic and the THP-1 cell line as a more mature monocyt,
than the existing monocytic cell lines (Verhaegen, Verschueren et al. 1989).

Figure 27: Growth curve of exponentially growing cells
4.4.2. Quality control
Cell number
Lag phase Log phase Plateau
Two positive controls were tested in each experiment. The LPS control was chosen because
it is a biological component of organic dust and IL-8 secretion after stimulation with LPS
has been shown earlier in the A549 cells (Paper I) (Palmberg, Larsson et al. 1998).
Stimulation of cells with endotoxin is believed to be mediated trough the endotoxin receptor
(CD14) via the cross-linking of LPS to the membrane bound receptor mCD14, mediated by
LPS binding protein (LBP) (Tobias, Soldau et al. 1993) (figure 28). Epithelial and
endothelial cells do not express the membrane bound mCD14 but a soluble form sCD14 has
been shown to participate in the activation of these cells (Ulevitch and Tobias 1995; Ulmer
1997; Rietschel, Brade et al. 1996). Even though epithelial cells seem to lack specific receptors
for endotoxin (Pugin, Schürer-Maly et al. 1993), endotoxin efficiently stimulates IL-8 secretion
from A549 lung epithelium cells (Paper I and II) (Palmberg, Larsson et al. 1998). To my
knowledge the monocytic cell lines U937 and THP-1 are CD14 negative. However, transfection
or activation with e.g. vitamin D3 can induce CD14 expression (Fan, Stelter et al. 1999;
Poussin, Foti et al. 1998; Tobias, Soldau et al. 1993; Kitchens, Wang et al. 1998). As shown in
this study LPS are capable of inducing IL-8 secretion from the two monocytic cell lines,
perhaps through the phagocyte receptor CD11/CD18. This receptor belongs to the integrin
family and has been reported to be a transmembrane receptor for endotoxin (Ingalls and
70
Time

Golenbock 1995; Morrison, Lei et al. 1993). Studies of these receptors, however, indicate that
the CD14 receptor induces a stronger response than the CD11/CD18 receptor (Ingalls and
Golenbock 1995).
Figure 28: Activation of epithelial cells and monocytic cells through the LPS receptor CD14. LPS binding
protein (LBP) cross-links to LPS and this complex binds to the membrane bound CD14 (mCD14) on
monocytes or the LPS-LBP complex delivers LPS to soluble CD14 (sCD14) through which an activation of
epithelial and endothelial cells is initiated. Modified from Ulevich and Tobias (1995).
TNF-α is an early response cytokine often initiating the cascade of produced cytokines in the
process of inflammation. Secretion of IL-8 from cells in vitro after TNF-α stimulation is
observed in several studies (Standiford, Kunkel et al. 1990; Kunkel, Standiford et al. 1991;
Strieter, Chensue et al. 1990; Nakamura, Yoshimura et al. 1991). TNF stimulates the cells
through transmembrane receptors (TNF-R) containing both extra cellular and intracellular
regions (Carpenter, Evans et al. 1995). Two distinct but structurally homologous types of
receptors are identified, and distinguished on the basis of their molecular mass, TNF-R1 (60
kDa) and TNF-R2 (80 kDa) (Zhang and Tracey 1998). Both receptors are present on
virtually all cell types except for red blood cells, but TNF-R1 is most ubiquitous and TNF-
R2 is more abundant on endothelial cells and cells of the hemapoietic lineage (Zhang and
Tracey 1998). All three cell lines used in this study are therefore expected to express the
TNF-R. The role of the TNF-R2 has been speculated to capture TNF on the cell surface and
pass it to TNF-R1, which mediate the specific functions of TNF (Zhang and Tracey 1998).
Soluble forms of the TNF-R is also found, derived from both TNF-R1 and TNF-R2 after a
71

proteolytic cleavage, and may serve to modulate TNF bioactivity (Carpenter, Evans et al.
1995).
GM-CSF was tested as an alternative to TNF-α as positive control in the monocytic cell
lines. GM-CSF promotes differentiation of monocytes to large macrophage like cells,
increases the metabolism and promotes their function as antigen presenting cells. GM-CSF
also induces the production of various cytokines including IL-8 (Quesniaux and Jones 1998).
Because of larger variation in the IL-8 secretion, than after TNF stimulation, this cytokine
was not used as control.
The day-to-day variation of the assays is reflected in the variation of the background values
and in the values of the controls (figure 17). The within-day standard deviations are lower
than the standard deviations of all the control values (figure 18), indicating that the inter
assay variation are larger than the intra assay variation. The RMSE value of LPS controls
was larger than the RMSE of TNF controls. Indicating that the TNF control samples yield a
more stable assay.
In the calculation of the PF the slope of the initial linear part of the dose response curves was
used. In this way the PF value reflects the induction of IL-8 secretion at the lowest
concentrations of the test compounds, and in case of a bell shaped dose response curve the
PF value is not influenced by cytotoxic effects at the higher concentrations of test
compound. The initial linear part of the dose response curves was judged subjectively. An
alternative might be to include a test for linearity (Hansen, Olsen et al. 1991). This however
could reject some of the values. In a linear regression model the data from figure 30A
(Section 5.2.1.) revealed linearity of the surface dust sample until 1 mg/ml (p=0.01), the
floor dust sample and dust from the exhaust ducts until 1 mg/ml (p

It may be argued that results in bioassays are frequently lognormal distributed. It may
therefore be more correct to plot the logarithm of the IL-8 secretion in the dose response
curves. However, each point in the dose response curves is the mean of three values, which
is insufficient for a statistical test of normal distribution. In a lognormal distribution the
standard deviation increases with increasing response (i.e. the same standard deviation is
obtained after transformation). This does not seem to be the case for this data material, see
for example figure 19. Hence, the data do not justify the use of log transformation in
plotting the dose response curves, and estimation of PF from the dose response curves
without log transformation is easier. In other studies stimulation of the A549 lung epithelial
cells and alveolar macrophages with bacteria, endotoxin and dust, no log transformation was
used in the presentation of the cytokine response of different doses of the test compound
(Larsson, Larsson et al. 1999b; Palmberg, Larsson et al. 1998; Saraf, Larsson et al. 1999).
The IL-8 secretion of the controls (LPS and TNF) was keyed in the AMIQAS PC
programme and X-R control charts were plotted (figure 29). The LPS control in the A549
bioassay resulted in a mean of 3.8 ± 2.1 ng IL-8/ml (at 100 µg LPS/ml), in the U937 bioassay
the mean was 1.6 ± 1.5 ng IL-8/ml (at 50 ng LPS/ml) and in the THP-1 bioassay the mean was
1.3 ± 0.5 ng IL-8/ml ml (at 50 ng LPS/ml) (figure 17, 23 and 25). A large variation was seen in
the A549 bioassay and the U937 bioassay for the LPS control but the control seemed stable in
the THP-1 bioassay. The TNF control in the A549 bioassay resulted in the mean of 12.4 ± 3.2
ng IL-8/ml ml (at 10 ng TNF/ml), in the U937 bioassay the mean was 2.1 ± 0.7 ng IL-8/ml (at 5
ng TNF/ml) and in the THP-1 bioassay the mean was 0.14 ± 0.04 ng IL-8/ml (at 5 ng TNF/ml)
(figure 17, 23 and 25). The TNF control seemed stable for the A549 and the U937 bioassay. In
addition to control of the stability of the methods the positive controls are also used for
correction of day-to-day variation of the assay. Especially table 3 illustrates the reduced
variation of the PF when correcting against the positive controls. In the A549 and the U937
bioassay the TNF control was used for correction of the PF, whereas in the THP-1 bioassay
the LPS control was used. Generally the monocytic cell lines were more sensitive to the LPS
and the TNF controls than the epithelial cell line.
73

Average(ng/mL)
Range(ng/mL)
27.00
24.00
21.00
18.00
15.00
12.00
9.00
6.00
3.00
14.00
12.00
10.00
8.00
6.00
4.00
2.00
Set & Date
Project: TNF control A549
X-R Chart Sample: Date: 17-12-99
0
0 5 10 15 20 25 No.
0
980416-1
980416-2
980416-3
980416-4
980416-5
980416-6
980423-7
980423-8
980423-9
980423-10
980423-11
980423-12
980423-13
980423-14
980423-15
Figure 29: Example of X-R control charts for some of the TNF controls from the A549 bioassay. In the Xchart
the mean of the control results in the subgroup are shown. The reference value is the mean of the control
material. The upper and lower control limits (UCL and LCL, respectively) are calculated by the AMIQAS
programme, and are defined as ±3δ/√n, where δ is the standard deviation and n is the number of control results.
In the R-chart the difference between the largest and the smallest control result (i.e. the range) in the subgroup
are shown. The centre line and control limits for the R-charts are calculated from the values of δ and n. The
control limits are 3δ limits, indicating the limits of extreme deviations from the centre line. Filled circles
indicate results out of control. In this case the experiment was performed again, and the original data
discharged.
Method evaluations were performed for all three bioassays, revealing that the methods were
in statistical control especially after correction against one of the positive controls. The IL-8
secretion after stimulation with the positive controls reflected the sensitivity of the cell line.
In the A549 and the U937 bioassay the TNF control seem more stable, properly because of
batch variations in the LPS. However, in the THP-1 bioassay the controls were equally
stable, but the TNF control gave only a small IL-8 secretion, and method evaluation after
correlation with the LPS control gave a better result. The LOD of the three bioassays were
17 µg LPS/ml (A549), 47 ng LPS/ml (U937) and 2.1 ng LPS/ml (THP-1). As described later
74
980423-16
980423-17
980423-18
980423-19
980423-20
980430-21
980430-22
980430-23
980430-24
980430-25
980430-26
980430-27
980506-28
980605-29
991217-30
XUCL: 22.148
XCL: 12.425
XLCL: 2.702
RUCL: 11.946
RCL: 3.656

the concentration of endotoxin found in dust from the schools is below 200 EU/mg ≈ 12.9 ng
endotoxin/mg dust (at 15.5 EU/ng endotoxin). This concentration is below the LOD of the
A549 bioassay and the U937 bioassay. The lower concentrations of endotoxin in dust will
not be measured in the THP-1 bioassay either. The three bioassays therefore are too crude
for detection of this biological component in the dust samples.
Experiments on the importance of adding serum to the media in the A549 bioassay were
performed by Palmberg et al. 1998, who found that stimulation with LPS without serum in
the media resulted in a small increase in the IL-8 secretion compared to LPS stimulation
with serum in the media. Palmberg and colleagues suggested that the serum in combination
with LPS could have an inhibitory effect on the IL-8 secretion (Palmberg et al. 1998).
However, the observed difference was within the uncertainty of the method used in the
present study. Serum was included in the media during growth of the cells and during the
stimulation experiments in the present study, to maintain the same growth conditions of the
cells. The content of LPS in serum was not measured.
During the use of two different ELISAs from Genzyme and R&D systems the Limit of
detection (LOD) was found according to the equation (described by IUPAC, calculated on
20 blank samples or samples with a low concentration of the test substance) (Poulsen, Holst
et al, 1993):
LOD =⎺X + 3δ (without correction for blank)
LOD = 3δ (with correction for blank)
Using the lowest value from the standard curve the LODs were as follows: 28,9 pg IL-8/ml
(Genzyme, using 15.9 pg IL-8/ml), 65,8 pg IL-8/ml (R&D systems, using 31,25 pg IL-8/ml).
The 20 OD values ware read on a standard curve chosen by random to convert the OD to pg
IL-8/ml. The ELISA kit from Genzyme was used in the A549 bioassay and half of the U937
bioassay the kit from Genzyme was used in the THP-1 bioassay. The background values of
the A549 and the U937 bioassay were therefore within the LOD and both assays. The
background of the THP-1 bioassay, however, was lower than the LOD of the ELISA kit.
The LOD under reproducibility conditions of the ELISA kits was not estimated as variables
as equipment, technician and laboratory was the same throughout the study. The LOD of the
ELISA was estimated under repeatability conditions.
75

4.4.3. Pure test substances
For validation of the three bioassays a panel of test substances were used. The test panel
included substances reacting through presumed receptors, known contact allergens and
substances with known or suspected irritative effect. Test of pure endotoxins in the epithelial
cell line and the monocytic cell lines resulted in different ranking of the endotoxins. The
A549 cell line ranks the different LPS: K. pneumoniae > P. aeruginosa > E. coli > S.
enteritidis (table 3 and Paper II). The monocytic cell lines, U937 and THP-1 however,
ranked S. enteritidis > E. coli > K. pneumoniae > P. aeruginosa (Table 7 and 9). This may
indicate different mechanisms of endotoxin stimulation of epithelial cells compared with
monocytic cells. Comparing the PF of the different LPS in the three bioassays, the U937
bioassay was about 150 times more sensitive to the different LPS than the A549 bioassay,
and about 2 fold more sensitive than the THP-1 bioassay. The THP-1 bioassay was about 80
fold more sensitive to the different LPS than the A549 bioassay. In the monocytic assays,
however, four times more cells were seeded in the wells prior to stimulation. The difference
in results between the assays cannot directly be explained by expression of mCD14, as none
of the cell lines seem to express this receptor. However differences in stimulation via sCD14
or the CD11/CD18 receptor could be a possible explanation. Another possibility is a
different pathway in unspecific stimulation of the cells. It may also be speculated that the
different potency of the different endotoxins, with respect to the induction of IL-8 secretion
from the same cell line, reflect differences in the binding affinity of endotoxin to other
unknown receptors than the ones mentioned above. These differences in the general potency
of LPS on the cell lines could likewise be caused by different activation systems (unspecific
and specific). For instance, differences in IL-6 and IL-8 secretion between one Gram
negative (E. coli) and four Gram positive (Bacillus subtilis, Staphylococcus hominis,
Staphylococcus and Micrococcus luteus) bacteria tested on the A549 cell line were found,
with E. coli as the most potent, then S. hominis, B. subtilis, S. lentus and M. luteus (Larsson,
Larsson et al. 1999a). Analysis of extracts of the bacteria gave almost the same rank, but
now the S. hominis was more potent than E. coli. Spores of the Gram-positive bacteria
Streptomyces sp. also stimulated the mouse macrophage cell line RAW264.7 to production
of TNF-α and IL-6 (Hirvonen, Nevalainen et al. 1997a).
Differences in the LPS response of the A549 bioassay as seen in figure 1, Paper I and figure
2, Paper II could be ascribed to batch to batch variations of the LPS and differences in the fit
76

of the cells of the day of the experiment. These differences are smoothening out when the PF
of the dose response curves are corrected with the positive controls according to table 2
Paper II.
Exposure of the A549 cells with pure glucans from different sources resulted in an IL-8
response only from Curdlan (a β 1-3 D-glucan from Gram negative bacteria Alcaligenes
faecalis). However, stimulation of the monocytic cell lines (U937 and THP-1) with the three
different glucans resulted in a very small, if any IL-8 secretion from the cells. Palmberg et al
(1998) also found only a small IL-8 secretion after stimulation of the A549 lung epithelial
cell line with a β 1-3 D-glucan from an unknown source (Palmberg, Larsson et al. 1998). As
mentioned earlier the β 1-3 D-glucan can stimulate macrophages through specific receptors
(Goto, Yuasa et al. 1994). Adachi et al found IL-1, IL-6 and TNF production after
stimulation of the mouse macrophage cell line RAW264.7 with grifolan, a purified β 1-3 D-
glucan (Adachi, Okazaki et al. 1994). Also whole viable spores of fungi are shown to induce
cytokine secretion from macrophages in vitro. Stimulation of the mouse macrophage cell
line RAW264.7 with spores from several strains of stachybotrys induced increased reactive
oxygen species production and secretion of TNF-α and IL-6 (Ruotsalainen, Hirvonen et al.
1998). Kauffman et al. (2000) showed that proteases in extracts of different fungi stimulate
epithelial cells to production of IL-6 and IL-8. The authors’ conclude that proteases may
stimulate the cells through a protease-activated receptor type 2-driven mechanism
(Kauffman, Tomee et al. 2000). The reasons for the lack of IL-8 response in the monocytic
cell lines, despite the responsiveness of macrophage cell lines remain obscure. Glucans are
insoluble in water, which could add to the lack of stimulation observed.
After stimulation of the three cell types with three pure test compounds (nickel sulphate,
MMA and formaldehyde) clear toxic effects were seen. Nickel sulphate resulted in the
strongest IL-8 secretion in all three cell types in the tested dose range. However, the potency
of formaldehyde ranged highest in the A549 bioassay. In the U937 and in the THP-1
bioassay the nickel sulphate shows the highest potency. The potencies of these compounds
were generally lower than the potencies of LPS in the same cell type. No clear correlation
between the allergenic potential and the inflammatory potential of these three substances
could be found. Nickel sulphate and formaldehyde gave a clear response at non-toxic
concentrations, perhaps reacting through specific receptors. They are both strong contact
77

allergens but no consistency in potency was found, when tested in the different cell lines.
The cytotoxic effect of formaldehyde is well documented and maybe some of the cell lines
(e.g. the U937) are more sensitive to cytotoxic substances than others. The reaction to MMA
was low in all three cell lines. Only few persons exposed to MMA through their work
showed a positive patch test, indicating that this moderate sensitiser perhaps only should be
classified as an irritant (Mürer 1996). Surfactants used to stimulate the three cell lines
resulted in clear dose response curves for all four surfactants in the A549 bioassay but no
clear response of any of the surfactants in the monocytic bioassays. Different mechanisms in
the stimulation pathway between the different cell lines could be an explanation, or it may be
that the concentrations in the monocytic bioassays were to high, inducing a toxic response
rather than an IL-8 response.
The concentrations of test substances in the A549 bioassay are high, compared to exposures
of humans in the working environment, because of the low sensitivity of the cell line. The
concentrations used in the monocytic assays are more biological relevant.
A screening method for evaluation of dust from the indoor environment must be stable, in
control, fast and easy to perform. All three in vitro models developed and tested with
different pure test substances are shown by the method evaluation to meet these criteria. The
bioassays react with IL-8 secretion towards biological active compounds maybe through
specific receptors, and dust contains many different biological active components. This
indicates that these methods may be used for hazard evaluation of dust. In the development
of the bioassays for evaluation of dust, experiments on preparation of the dust were
performed. A preliminary study suggested that the activity of dust in the A549 bioassay was
related to the particulate fraction of the dust (Paper I). An explanation could be that the
activation of the cells is dependent on the presentation of the biological active compounds on
the dust particles, or that not all of the compounds are soluble in the same solution. The most
relevant approach was to investigate the total dust sample, both the particles and the soluble
part of the sample. Palmberg et al. (1998) used dust from a swine confinement building
suspended in cell culture media to stimulate IL-8 secretion from the A549 cell line, human
bronchial epithelial cells and alveolar macrophages for 24 hours (Palmberg, Larsson et al.
1998). In this study no growth of microorganisms was reported. In a study of the IL-8 and
IL-6 secretion from monocytes, dust from the indoor and outdoor environment was extracted
78

in water. The extract was freeze-dried, and resolublised in cell culture media. The study
indicated that the cytokine secretion was associated to the soluble extract of outdoor particles
smaller than 10 µm (Monn and Becker 1999). In a cytotoxic assay performed by Roepstorff
et al. (1997) aqueous and alkaline extracts of different dust samples were used (Roepstorff
and Sigsgaard 1997).
To avoid growth of microorganisms in the cell culture media, which could activate the cells,
different methods of sterilisation of the dust, were tested. Sterile filtered extracts were
already shown to give a lower potency than the total dust sample (Paper I). Formalin
treatment of the total dust sample did not always kill the microorganisms (unpublished
results). Sterilisation of the dry dust by γ radiation at 35 kGy was found to give the best
results and was used on all the dust samples in the present study. For sterilisation of medical
equipment doses of 20 – 40 kGy are used, and for radiation of food doses of 5 – 10 kGy are
used (Henriksen, 1995). Treatment of dust with γ radiation will to some extent lead to the
formation of free radicals. These will, however, quite fast react with other compounds and
become stable again (Pers. Comm. Gunnar Damgaard, Nat. Inst. Occup. Health., Denmark).
Therefore to my knowledge the treatment of dust with γ radiation have only small impact on
the composition of the dust other than killing all living organisms. The dust sample
suspended in cell culture media was sonicated for three times one minute for detachment of
particles from each other.
In the monocytic assays 500 µl old growth media, out of a total of 1 ml, are included in the
experiments to avoid spinning of the cells. This will transfer any pre secreted IL-8 into the
experiment. The concentration will, however, be the same in all the wells in an experiment.
Earlier experiments has shown that spinning of the cells in it self stimulate an higher IL-8
secretion than the amount secreted in the chosen way (Unpublished data).
79

5. Dust samples from schools tested in the in vitro methods
The lung epithelial bioassay with the cell line A549 was developed and evaluated first. All
the collected dust samples were tested using this method in concentrations 0, 0.1, 0.5, 1, 3,
and 5 mg dust/ml (see below). Initially only some of the dust samples were tested on the
U937 and the THP-1 monocytic cell lines, to see which responded best. The concentrations
of dust used in the monocytic assays were 0, 5, 10, 50, 150, and 250 µg dust/ml. The results
are shown below. To reduce interassay variation the corrected PF are used in the calculations
of the correlations described below. For the A549 bioassay and the U937 bioassay correction
with the TNF-control was used, as this control produced more homogenous results in these
assays. For the THP-1 bioassay correction with the LPS control was used, because the TNF
control gave a very low IL-8 secretion in this assay. A Hotelling-Pabst test was used for
correlations at α = 0.05, when nothing else is stated.
5.1. Results from dust collection
A total of 158 dust samples were collected from 20 schools (96 floor dust samples, 96
surface dust samples pooled to 21 samples, 41 dust samples from exhaust ducts). One floor
dust sample was too small for analysis and was excluded from the study. This gives a total of
157 dust samples. The median and range for the amount of collected dust samples are shown
in table 11 together with the number of samples and the mean values.
Table 11: Median and range values for the amount of collected dust samples.
Median and range (g) n Mean (g)
Floor samples 3.2 (0.2-120) 96 6.3
Floor samples after sieving 2.0 (0.1-97.2) 96 4.6
Surface samples 0.20 (0.06-1.01) 96 0.25
Pooled surface samples 0.88 (0.48-2.34) 21 1.14
Pooled surface samples after sieving 0.49 (0.2-1.93) 21 0.69
Exhaust duct samples 3.13 (0.04-75.47) 46 11.6
Exhaust duct samples after sieving 3.21 (0.17-58.20) 41 8.3
Dust samples collected from floor range from 0.2 to 120 g, with the highest values from
floors covered with carpets. The amount of surface dust from horizontal or near horizontal
surfaces ranged from 0.06 to 1.01 g. Because of this small amount, most of the samples were
80

pooled to gain one sample for each school. The same was done for some of the exhaust dust
samples, ranging from 0.04 to 75.47 g. Exhaust dust samples taken from the naturally
ventilated schools had generally a higher weight than the samples from mechanically
ventilated exhaust ducts. This could be due to a larger fraction of sand and plaster from the
duct, giving weight to the sample.
5.2. Results from dust tested in the in vitro assays
5.2.1. Dust tested in the A549 bioassay
Dust from horizontal or near horizontal surfaces tested in the A549 bioassay showed typical
cytotoxic dose response curves with increase of the IL-8 secretion until stimulation with 0.5
mg or 1 mg of dust, and then a decline towards zero with stimulation at higher dust
concentrations (figure 30A and 30B). The maximal IL-8 secretion ranged from 2.8 ng to
19.0 ng IL-8 with background values of 0.11 ng to 0.8 ng IL-8. In the microscope some of
the cells in the wells stimulated with 0.5 to 1 mg of dust were normal and some were
rounded up. At stimulation with 3 to 5 mg of dust the cells could not be seen because of dust
lying on top of the cells.
Stimulation in the A549 bioassay with dust from the floors also showed cytotoxic dose
response curves with a maximum IL-8 secretion at stimulation with 0.5 mg or 1 mg dust
(figure 30A and 30B). The maximal IL-8 secretion ranged from 0.97 ng to 5.9 ng IL-8 with a
background value between 0.11 and 1.8 ng IL-8. The cells in the wells stimulated at the
concentrations 0 to 1 mg of dust appeared normal in the microscope after 24 hours of
incubation. Only in the sample 4Gc a change of shape of the cells was noted in the wells
stimulated with 0.5 mg of dust. In the wells stimulated with 1 mg of dust many of the cells
were rounded up and detached from the bottom of the well. In the samples of 4Ga, 4Gc,
4Gd, 9Gd and 17Ga the cells were rounded up in the wells stimulated with 3 and 5 mg of
dust. Stimulation with the samples of 9Gc, 10Ga and, 18Gc resulted in lysis and clearly dead
cells in the wells stimulated with 5 mg of dust.
81

IL-8 (ng/ml)
IL-8 (ng/ml)
9
8
7
6
5
4
3
2
1
0
9
8
7
6
5
4
3
2
1
0
0 0.5 1 1.5 2 2.5
Dust (mg/ml)
3 3.5 4 4.5 5
0 0.5 1 1.5 2 2.5
Dust (mg/ml)
3 3.5 4 4.5 5
Figure 30. Example of the A549 lung epithelial cells stimulated with dust from surfaces, the floor and the
exhaust ducts. The mean of a triplet is shown together with the standard deviation. A: School 6 (PFcorr TNF =
0.012; 0.006 and 0.009 ng IL-8/µg dust respectively), and B: school 17 (PFcorr TNF = 0.009; 0.002 and 0.0008
IL-8/µg dust respectively).
82
Surface
Floor
Exhaust duct
Surfaces
Floor
Exhaust duct
A
B

Dust from exhaust ducts was also used to stimulate the A549 lung epithelial cells to
secretion of IL-8. In some samples the cytotoxic dose response course as mentioned above
was seen. In others a linear increase over the whole dose spectra was observed (figure 30A
and 30B). The maximal IL-8 secretion found, range from 0.34 ng to 13.9 ng IL-8, with
background values between 0.16 ng and 0.37 ng IL-8. Most of the cells in the wells
stimulated with 1 mg of dust appeared normal, but many could not be seen because of the
dust. The cells in the wells stimulated with the two largest concentrations (3 and 5 mg of
dust) could mostly not be seen but appeared normal when visible. The cells in the wells
stimulated with control samples (LPS and TNF) appeared normal in all the above
experiments. The inflammatory potential of the different dust samples to elicit an IL-8
secretion from the A549 lung epithelial cells was calculated as the potency factor and
summarised in table 12 as the median and the range.
Table 12. The potency factor (PF) of the different dust samples tested in the A549 bioassay. The median is
shown together with the range. Because of large variations in the LPS control (figure 17 and 18) corrected
values are only shown for the TNF control.
Dust samples from IL-8 induction
ng IL-8/µg
Surface 0.0083
(0.0008 - 0.0385)
Floor 0.0024
(0.0001 - 0.0083)
Exhaust ducts 0.0086
(0.0002 - 0.0285)
83
TNF corrected
IL-8 induction
ng IL-8/µg
0.0110
(0.0011 - 0.0286)
0.0028
(0.0001 - 0.0118)
0.0074
(0.0002 – 0.0305)
A significant positive correlation between the PF of the floor samples and the PF from the
surface dust samples (rs=0.43) was found. No correlation was found between the PF of the
floor samples and the PF of the samples from exhaust ducts, or between the PF from the
surface dust samples and the PF of the samples from exhaust ducts.
Grouping the samples regarding their origin from a “good” or a “bad” school and using the
Mann-Whitney test revealed highly significant (p

5.2.2. Dust tested in the U937 bioassay
To evaluate the two monocytic bioassays against each other only a selected part of the dust
samples were tested in the assays at first. These samples originated from the same school and
included a surface dust sample (1Oa), a floor dust sample (1Ga) and a dust sample from an
exhaust duct (1Ke). Surface dust samples from three schools with a low potency in the A549
bioassay (2Oa, 3Oa, 16Oa) and three surface dust samples with a high potency (6Oa, 13Oa,
14Oa) were also tested.
In the U937 bioassay the dose response curves showed an increase over the whole dose
spectra, with the potency of dust from the exhaust ducts being most potent (Table 13). The
maximal IL-8 was found to be 0.78 ng IL-8 for the surface dust sample, 0.65 ng IL-8 for the
floor dust sample and 0.8 ng IL-8 for the dust sample from the exhaust duct. The surface
dust samples from three schools with low and three schools with high potency in the A549
bioassay gave dose response curves increasing over the dose range in the U937 bioassay.
The maximal IL-8 secretion was 0.36, 0.57 and 0.41 ng IL-8, respectively for the low
samples, and 0.6, 0.67 and 0.64 ng IL-8, respectively for the high potency samples. The
background value for all the dust samples tested in the U937 bioassay range between 0.14
and 0.41 ng IL-8/ml. The potency of the dust samples tested in the U937 bioassay is shown
in table 13.
Table 13. The potency factor of selected dust samples used to stimulate the U937 monocytic cell line.
Dust sample IL-8 induction
ng IL-8/µg
1Oa a
1Ga b
1Ke c
2Oa a
3Oa a
16Oa a
6Oa a
13Oa a
14Oa a
a
Surface dust samples
b
Floor dust sample
c
Dust from exhaust ducts
0.0024
0.0019
0.0041
0.0060
0.0012
0.0006
0.0010
0.0015
0.0013
84
LPS corrected
IL-8 induction
ng IL-8/µg
0.0020
0.0009
0.0026
0.0011
0.0014
0.0004
0.0011
0.0014
0.0009
TNF corrected
IL-8 induction
ng IL-8/µg
0.0023
0.0015
0.0042
0.0008
0.0014
0.0006
0.0011
0.0014
0.0012

5.2.3. Dust tested in the THP-1 bioassay
The THP-1 cell line was, at first, tested with three different dust samples from the same
school: A surface dust sample (1Oa), a floor dust sample (1Ga) and a dust sample from an
exhaust duct (1Ke). The dose response curves showed an increase over the whole dose
spectra, with the surface dust sample being most potent (Table 14). The maximal IL-8
secretion was 13.7 ng IL-8/ml for 1Oa, 4.64 ng IL-8/ml for 1Ga, and 2.0 ng IL-8/ml for 1Ke.
Table 14. The potency factor of selected dust samples used to stimulate the THP-1 monocytic cell line.
Dust sample IL-8 induction
ng IL-8/µg
1Oa a
2Oa a
3Oa a
4Oa a
5Oa a
6Oa a
7Oa a
8Oa a
9Oa a
10Oa a
11O a
12Oa a
13Oa a
14Oa a
15Oa1 a
15Oa2 a
16Oa a
17Oa a
18Oa a
19Oa a
20Oa a
1Ga b
1Ke c
a
Surface dust samples
b
Floor dust sample
c
Dust from exhaust ducts
0.0567
0.0082
0.0592
0.0558
0.0363
0.0773
0.0345
0.0443
0.1592
0.1432
missing
0.0510
0.0852
0.0771
0.0524
missing
0.0246
0.0641
0.0428
0.0655
0.0776
0.0180
0.0125
85
LPS corrected
IL-8 induction
ng IL-8/µg
0.0820
0.0109
0.0615
0.0843
0.0641
0.0917
0.0611
0.0522
0.2003
0.1441
missing
0.0553
0.0832
0.0637
0.0683
missing
0.0228
0.0857
0.0586
0.0790
0.1096
0.0223
0.0150
Because the THP-1 cell line showed a larger IL-8 response than the U937 cell line when
stimulated with dust, the THP-1 cell line was selected for test of all the surface dust samples.
Within the time frame of this study it was not possible to analyse the samples from floor and
exhaust ducts in the monocytic assays.
When stimulating the THP-1 monocytic cell line with surface dust samples from the schools
the IL-8 concentration in the media increased as the concentration of dust increased (figure

31). The maximal IL-8 secretion ranged between 2.17 and 23.1 ng IL-8/ml. The potencies of
the dust samples tested in the THP-1 cell line are shown in table 14. The background value
for all the dust samples tested in the THP-1 bioassay ranged from below the detection limit
to 0.08 ng IL-8/ml.
IL-8 (ng/ml)
25
20
15
10
5
0
16Oa
17Oa
18Oa
19Oa
20Oa
0 50 100 150 200 250
Dust (ug/ml)
Figure 31. Example of dose response curves from the THP-1 monocytic cell line stimulated with pooled
surface dust from five different schools. The mean is shown together with the standard deviation.
Grouping the samples regarding their origin from a “good” or a “bad” school and using the
Mann-Whitney test revealed no significant difference in the PF from a “good” or a “bad”
school in the surface dust samples.
A test for positive correlation between the PF of the surface dust samples from the THP-1
and the A549 bioassay did not indicate any significant correlation (n=19) (figure 32). No
correlation was found between the U937 bioassay and the A549 bioassay or between the two
monocytic assays (not shown).
86

PF corr TNF of dust samples tested in the A549 bioassay
0,03
0,03
0,02
0,02
0,01
0,01
0,00
0 0,02 0,04 0,06 0,08 0,1 0,12 0,14 0,16 0,18
PF corr LPS of dust samples tested in the THP-1 bioassay
Figure 32: Correlation between PF of surface dust samples tested in the A549 bioassay and in the THP-1
bioassay
5.3. Discussion of test of dust samples in the in vitro methods
To meet the purpose of this study a large difference between the samples from school to
school was desired. The chosen 20 schools lie with 10 schools in each end of a symptom
scale of all the schools in Copenhagen, and may therefore meet this demand. Whether these
schools are representative for all the schools in Denmark are not important for this study, but
with the large difference between e.g. old and new buildings and types of ventilation this
properly reflects the different types of schools in Denmark.
Looking at the PF of surface dust samples in the A549 bioassay (Table 12) and the THP-1
bioassay (Table 14), the THP-1 bioassay is 10-20 times more sensitive than the A549
bioassay. Mononuclear cells isolated from blood were previously found to be up to 300
times more sensitive towards LPS than the A549 bioassay, when looking at the IL-8
secretion (Hansen, Nexø et al. 1997).
87

A problem with the sampling of floor dust with the HVS-3 was that the sampler only retains
particles above 5µm. The effectiveness of dust collection by the HVS-3 was determined in
laboratory tests (ASTM 1994). Smaller particles are on the other hand present in the surface
dust samples and the samples from the exhaust ducts, which retained particles as small as
0.3µm. However, despite the difference in particle size distribution a significant correlation
on a school level between the PF of floor dust versus the PF of surface dust was observed in
the A549 bioassay, indicating that the same biological active components were present in
these different types of dust. The PF of floor dust was in mean found to be about four times
lower than the PF of surface dust (Table 12), which could be explained by the higher amount
of inert sand, found in the floor dust samples. Sand consists mainly of silica particles in
different forms and most minerals are build upon a polymerised framework of (SiO4) 4-
tetrahedral units (Heaney, Banfiels, 1993). How much this fraction contributes to the PF was
not studied. However, increased IL-8 secretion and increased IL-8 mRNA transcription after
stimulation of human pulmonary epithelial cells with quartz and titanium oxide indicates that
particles in it self has an stimulatory effect (Schins, McAlinden et al, 2000).
Regarding the dust samples origin from a “good” or a “bad” school a correlation was found
for floor dust but not for surface dust or dust from exhaust ducts. The small number of
surface dust samples, and the large difference in the samples from the exhaust ducts may
explain this lack of correlation. PF of dust from the exhaust ducts did not correlate with the
PF of floor dust and surface dust, which may be explained by the different types of
ventilation. Samples from natural ventilation contained visibly more sand and plaster than
samples from mechanical ventilation ducts. As explained above the sand and plaster may
give weight to the sample but not add to the inflammatory potential. Silica in the form of
quartz is an accessory in e.g. concrete, industrial sand and gypsum (Heaney, Banfiels, 1993).
Nineteen samples from surfaces were tested in the THP-1 bioassay and no correlation was
found with the A549 bioassay. This lack of correlation may be due to the small sample
number of the THP-1 bioassay. Another explanation may be that the cell lines are stimulated
through different receptors or pathways and possibly measure different effects of the dust. In
section 4 the PF of the different endotoxins tested in the two bioassays received different
ranks (table 3 and 9), which is also seen for the chemical test substances (table 5 and 10).
This indicates that the two cell types use different stimulatory pathways.
88

The course of the dose response curves (figure 30) was often an increase of the IL-8
secretion until a maximum about 0.5 or 1 mg/ml and then the curve falls to values some
above the background. This could be explained by a cytotoxic effect on the cells at the
higher concentrations or an interference of the secreted IL-8 with dust. The interference of
dust particles with the measured IL-8 concentration was studied by adding in duplicate
increasing amounts of dust to samples with a known concentration of IL-8 (Appendix 3).
Significant interference was only observed at the highest dust concentration. At a
concentration of 1 mg/ml of dust, a 38% reduction in the measured IL-8 concentration was
observed. The reduction increased to 80-90% at concentrations of 3 to 5 mg/ml of dust. This
problem was attempted solved by adding salt to the wells immediately before harvesting the
media or changing the pH of the media, but with no luck (Appendix 3). The interference may
be due to either binding of IL-8 to the dust particles, breaking down of the IL-8 protein, or to
quenching of the ELISA. Kauffman et al. 2000 found that the proteases from four different
fungi used to stimulate the A549 epithelial cell line to cytokine secretion, also degraded the
produced IL-6 and IL-8 protein (Kauffman, Tomee et al, 2000). Since the potency factor
(PF) was calculated from the initial linear part of the dose response curve, this interference
had limited impact on the calculation of the PF, at the most a small underestimation of the
PF. Hence, no further attempts were made to reduce the interference. The observation of
cells rounded up in the wells with the highest concentration of dust could be caused by a
toxic effect induced on the cells or by breakdown of the cells adhesion proteins by proteases.
Kauffman et al. 2000 found that both shrinking of the cells or desquamation of the cell layer
were depended upon proteases in an extract from different fungi (Kauffman, Tomee et al,
2000).
The concentrations used in the A549 bioassay are high compared to concentrations found in
the indoor air. Kildesø et al. (1998) found respirable dust concentrations of 54.4 µg/m3 in an
administration building, 133 µg/m3 in a kindergarten and 102 µg/m3 in a school (Kildesø,
Tornvig et al, 1998). The exposure during an eight-hour working day gives 156.7 µg in the
administration building, 383.0 µg in the kindergarten and 293.8 µg in the school (Calculated
on the basis of 12 inhalations/min., 0.5 l/inhalation, assuming that all the inhaled dust
reaches the lungs). These values are within the range of doses used in the monocytic assays.
In other studies with the A549 cell line and a human alveolar macrophages exposure for 24
89

hours of 1 to 100 µg/ml of swine dust, LPS, grain dust, and glucans was used (Palmberg,
Larsson et al. 1998). 5 to 200 µg of house dust, LPS or peptidoglycan-polysaccharide
suspension was used to stimulate the A549 cell line (Saraf, Larsson et al. 1999). These
concentrations are in the lower range of the doses used in the present study.
A comparison of the obtained PF from the A549 bioassay with PF obtained from floor dust
samples from waste handling plants (paper sorting plants and bottle sorting plants) showed
that the PF from the school floor dust samples are at the same level as the PF from the waste
handling plants. In contrast, the PF from the surface dust samples were in some cases much
higher than the PF from floors of waste handling plants (Allermann and Poulsen 2000). The
severe health problems related to the handling of waste are often related to the exposure to
organic dust (Sigsgaard, Abell et al. 1994; Poulsen, Breum et al. 1995b). Some of the most
common symptom of the BRS, i.e. irritation of the mucus membranes and the upper
respiratory tract, headache, and lethargy (Redlich, Sparer et al. 1997), are also part of the
multitude of symptoms observed among workers handling organic household waste. House
dust has also been shown to be more potent in electing secretion of IL-6 and IL-8 from the
A549 cell line than dust from swine confinement buildings (Saraf, Larsson et al. 1999). Dust
from swine confinements has been shown to induce acute inflammatory reaction with
increased IL-6 concentration in nasal lavage and peripheral blood (Larsson, Larsson et al.
1999b; Wang, Malmberg et al. 1996). Also increased IL-8 concentration in nasal lavage and
BAL, increased number of neutrophils in nasal lavage and increased number of
macrophages, lymphocytes and eosinophils in BAL could be observed (Larsson, Palmberg et
al. 1997). Hence, the comparison of the potency of dust from the indoor environment in a
non industrial work environment, as the schools, with the potency of dust from other work
places indicates that house dust may be highly potent with respect to eliciting inflammatory
reactions of e.g. the mucus membranes or the lungs.
The correlation between the PF of airborne dust and the PF of sedimented dust could not be
made in this study, as no airborne samples was collected. Recalling the high potential of
surface dust versus dust from floors the same picture may be expected for airborne dust,
giving a higher PF than surface dust. The airborne dust is assumed to contain smaller particle
size and only a few inert particles, than floor and surface dust, and therefore more fine
particles per weight unit. Earlier studies of dust from an incineration plant showed PF from
90

dust from the air being about two times higher than dust from the floor (Allermann and
Poulsen 2000). The PF of surface dust and PF of dust from the floors correlated (section
5.2.1.). A correlation between PF of airborne dust sample and PF of surface dust would
therefore also be expected.
5.4. Evaluation of the A549 bioassay
In the A549 bioassay the PF of the floor dust samples from the “bad” schools were found to
be significantly higher than the PF of the “good” schools. This is also verified in Paper IV
and in the thesis of Harald Meyer (Meyer 2000).
In the evaluation of the A549 bioassay a cut-off value was defined as a PF value separating
the samples from a “bad” school and the samples from a “good” school. An estimate of the
best cut-off values (CO) could be made subjectively by looking at the number of samples
above the cut-off and originating from a “bad” school. Likewise for the samples below the
cut-off and originating from “good” schools. Two cut-off values were estimated (CO1 and
CO2) from the criteria:
1. The likelihood of a measurement from a “good” school was in the interval [0;CO1]
should be high.
2. The likelihood of a measurement from a “bad” school was in the interval [CO2;∞]
should be high.
Therefore samples where PFCO2 are defined as samples having a high potency
(e.g. being a “bad” school). For PF in between (CO1

CO1 78% of the schools were “good” and 22% were “bad”, likewise, of the schools in the
area above the CO2, 9% were “good” schools and 91% were “bad” schools. About 42% of
the samples from “good” and “bad” schools were between CO1 and CO2.
0 47,5% “Good”
2.0 4.5 9% “Good”
52,5% “Bad”
CO1 CO2
Low PF High PF
Figure 33: Cut-off values CO1 and CO2 indicated on the PF scale. PF < CO1 are defined as samples with a
low PF and PF > CO2 are defined as samples with a high PF.
In the evaluation of dust from the indoor environment the A549 bioassay could find 43% of
the samples originating from a “bad” school (PF > 4.5 ng IL-8/mg dust). As factors causing
the BRS are believed to be multifactorial in origin and also physical and psychological
factors are inflicting on the experience of BRS, the method refinding of 43% of the “bad”
schools is not poor. Factors not necessarily reflected in the dust such as VOCs are also
shown to cause BRS (Ohman and Eberly 1998; Hodgson 1995). Differences and correlation
of epidemiological data to data on the PF will be discussed in section 7.1.
As mentioned in section 3.1., ten “good” and ten “bad” schools were selected in each end of
a BRS index scale, from the original 75 schools. Schools with a BRS index ≤ 11 were
defined as “good” and schools with a BRS index ≥ 19.6 defined as “bad”. In between lies 55
schools that were not analysed in this study. Figure 34 gives a picture of the BRS index of
the schools versus the PF of the floor samples from the schools tested in the A549 bioassay.
A trend is seen that the “bad” schools have a higher PF than the “good” schools, which was
also verified by the statistical analysis.
92
91% “Bad”
PF
(ng IL-8/mg)

PF (ng IL-8/mg dust)
12
10
8
6
4
2
0
0 5 10 15 20 25 30 35
"good" schools BRS index
"bad" schools
Figure 34: BRS index of the original 75 schools versus the PF of the floor dust samples from the 20 schools
tested in the A549 bioassay. Cut-off values separating the “good” schools from the “bad” were defined as ten
schools in each end of the scale (BRS index ≤ 11 were defined as “good” and schools with an BRS index ≥
19.6 were defined as “bad”). The cut-off values of the floor dust samples from these ten “good” and ten “bad”
schools were sat to CO1 < 2 ng IL-8/mg dust and CO2 > 4.5 ng IL-8/mg dust, with PF < CO1 defined as low
and PF > CO2 defined as high.
93
PF>4.5
PF

6. Correlation of results from in vitro methods with parameters in the dust
For the A549 bioassay and the U937 bioassay correction with the TNF-control was used. For
the THP-1 bioassay correction with the LPS control was used. A Hotelling-Pabst test was
used at α = 0.05, when nothing else is stated.
As expected from earlier experiments on mononuclear cells the monocytic assays were more
sensitive to dust than the epithelial cell line (Hansen, Nexø et al. 1997) (see section 5.2.).
However, the epithelial cell line produced stable results why most effort was placed in the
work with this cell line. This is why not all the dust samples were tested on the monocytic
models (see section 5.2.).
6.1. Correlation to content of organic fraction
The mean organic content of the surface dust samples was 52.3% with a range of 20.1 –
76.8%. For the floor dust samples the mean content of organic dust in the sample was 35.5%
(7.9 – 73,1%) and 33.7% (2.3 – 81.5%) for the dust samples from exhaust ducts. The
mechanical exhaust ducts had a mean content of organic dust of 56.2% (18.3 – 81.5%) and
the natural ventilated ducts had 17.1% (2.3 – 64.7%).
6.1.1. A549 bioassay
Using the Hotelling Pabst rang correlation test, a significant positive correlation was found
between the TNF corrected PF of all the dust samples in the A549 bioassay and the content
of organic dust, with a Spearman rank correlation coefficient rs = 0.66 (figure 35). The
surface dust coefficient was rs = 0.90, for the floor dust rs = 0.53, and for the dust samples
from exhaust ducts rs = 0.74. When correlating the PF calculated from the total dust sample
versus PF calculated from the fraction of organic dust, a significant positive correlation was
found with rs = 0.59.
A Mann-Whitney test showed a significant difference (p

PF corr TNF (ng/mg dust)
35
30
25
20
15
10
5
0
0 10 20 30 40 50 60 70 80 90
Organic fraction (%)
Figure 35: The PF corrected with TNF of dust samples tested in the A549 bioassay versus the organic fraction
in the sample.
PF corr TNF (ng/mg dust)
4.5
4
3.5
3
2.5
2
1.5
1
0.5
0
Figure 36: The PF corrected with TNF dust samples tested in the U937 bioassay versus the organic fraction in
the sample.
95
Surface
Floor
Exhaust duct
0 10 20 30 40 50 60 70 80 90
organic fraction (%)

6.1.2. U937 bioassay
Using the Hotelling-Pabst test to correlate the TNF corrected PF with the content of organic
dust for the samples tested in the U937 bioassay no correlation was found (n=9), however a
trend towards a correlation was observed (figure 36).
6.1.3. THP-1 bioassay
No correlation was found between the LPS corrected PF of the THP-1 bioassay and the
content of organic matter in the dust samples (n=19), however, a trend was observed (figure
37).
PF corr LPS (ng/mg dust)
200
180
160
140
120
100
80
60
40
20
0
0 10 20 30 40
Organic fraction (%)
50 60 70 80
Figure 37. The PF corrected with LPS of surface dust samples tested in the THP-1 bioassay versus the organic
fraction in the sample.
6.2. Correlation to microbiological parameters
Data from the microbiological analysis of the dust samples are shown in Appendix 4a and
4b.
96

6.2.1. A549 bioassay
The PF from the A549 bioassay and the concentration of endotoxin in the sample gave a
significant positive correlation for the surface dust samples (rs = 0.31) and for the floor dust
samples (rs = 0.07) (figure 38). However, correcting the PF with the content of organic
matter in the dust samples, a significant positive correlation for the floor dust samples and
the samples from exhaust ducts versus the PF emerged, with rs = 0.10 and 0.14 respectively.
PF (ng IL-8/mg dust)
35
30
25
20
15
10
5
0
0 50 100 150 200
Endotoxin (EU/mg)
Figure 38. Correlation between the PF of the different dust samples tested in the A549 bioassay and the
concentration of endotoxin in the sample.
In table 15 an overview of the content of viable microorganisms in the dust samples and the
Spearman rank correlation coefficients are shown. Mann Whitney tests revealed a significant
difference (p

was found for moulds in the floor dust and the PF of organic fraction. A statistically positive
correlation was also found between all tested microorganisms and the PF of organic fraction
from exhaust ducts.
Table 15. Viable counts (cfu/mg dust) are shown as median and range. The Spearman rank correlation
coefficients (rs) are given where significant positive correlation is found between viable counts and PF in the
Hotelling-Pabst rank correlation test.
Microorganism Surface Floor Exhaust ducts
Actinomycetes 8.109 (0.080-356.8)
No correlation a)
No correlation b)
Total bacteria 1185 (334.8-4790)
No correlation a)
No correlation b)
Mould 69.15 (16.19-410.8)
No correlation a)
No correlation b)
98
28.05 (0.599-2402)
No correlation a)
No correlation b)
2072 (10.10-21840)
rs=0.14 a)
No correlation b)
31.29 (1.078-400.0)
No correlation a)
rs=0.01 b)
0.600 (0.598-20.54)
rs=0.14 a)
rs=0.10 b)
64.74 (0.598-33348)
rs=0.01 a)
rs=0.18 b)
a) PF of total dust corrected with the TNF control versus microorganisms in the dust.
b) PF of organic fraction corrected with the TNF control versus microorganisms in the dust.
6.2.2. U937 bioassay
11.88 (0.599-1310)
rs=0.46 a)
rs=0.06 b)
Correlation to microorganisms was not made because of the small number of samples tested
(n=9).
6.2.3. THP-1 bioassay
No correlation was found between the PF of the THP-1 bioassay and the content of
endotoxin in the samples (n=19), or between the PF and the content of microorganisms
(n=18). However, X-Y plots of PF versus endotoxin or versus the content of microorganisms
indicated a trend towards a correlation (not shown).
6.3. Correlation to content of allergens
Measurements of allergens in the dust were made by ALK Abelló and are shown in
Appendix 4c.

6.3.1. A549 bioassay
The concentration of allergens in the dust and the correlation with the PF of the dust from
the A549 bioassay is shown in table 16. The results for the three mite allergens were pooled
to “total mite” because of the small amount of the single mite allergens. A correlation was
found between the mite, cat and dog allergen in the floor dust samples and the PF, and also
between cat and dog allergens in dust from the exhaust ducts and the PF. When correcting
the PF with the content of organic matter in the dust samples, a correlation was only found
between the dog allergens and the PF of the organic fraction in dust from exhaust ducts.
Table 16. Total mite, cat, and dog allergen concentration (ng/g dust) showed as median and range. If
significant positive correlation was found from the rank correlation test between allergen concentration and the
PF the Spearman rank correlation coefficient (rs) is given.
Allergen Floor Exhaust ducts
Total mite 0 (0-123)
rs=0.07 a)
No correlation b)
Cat 78 (3-2625)
rs=0.25 a)
No correlation b)
Dog 212 (10-2340)
rs=0.39 a)
No correlation b)
99
0 (0-122)
No correlation a)
No correlation b)
131 (0-2102)
rs=0.84 a)
No correlation b)
77.5 (0-2813)
rs=0.94 a)
rs=0.23 b)
a) PF of total dust corrected with the TNF control versus allergens in the dust.
b) PF of organic fraction corrected with the TNF control versus allergens in the dust.
6.3.2. U937 bioassay
Correlation to allergens was not performed because of the small number of samples tested.
6.3.3. THP-1 bioassay
As the content of allergens was only measured in floor dust and in dust from exhaust ducts
and only the surface dust samples were tested in the THP-1 bioassay this correlation could
not be made.
6.4. Discussion of correlation of in vitro results to parameters in the dust
A statistical significant correlation of the Hotelling Pabst test indicates if the two variables
follows each other, meaning if one increases the other does to. The spearman rank

correlation coefficient, indicating how closely the variables of the graph follow each other,
therefore could be very low even though a significant correlation was found.
As mentioned in section 2.2.3. dust contains a complex mixture of organic and inorganic
materials, with the organic part characterised as unspecific samples of both living and dead
organic material of vegetable, animal and microbial origin (Chan-Yeung, Clark et al. 1994).
One of the hypotheses of this study is that the potency of the dust can be explained by the
organic fraction of the dust, and maybe some parameters (chemical or microbiological) can
be characterised as the responsible agents of the observed inflammatory potencies.
The organic fraction of the dust sample correlated with the PF of the dust, indicating that the
inflammatory agent was to be found in the organic fraction of the dust. Dust from
mechanical ventilation systems contained a higher content of organic dust than dust from the
natural ventilated systems. This was also reflected in the A549 bioassay, where significant
higher PF were found from the dust samples from mechanical ventilation ducts compared
with the samples from natural ventilation ducts. Also schools defined as “bad” had
significantly more rooms with mechanical ventilation (Meyer 2000).
Previous studies have indicated that endotoxin or microorganisms in house dust may
contribute to SBS or aggravation of allergic asthma (Teeuw, Vandenbroucke-Grauls et al.
1994). Gyntelberg et al. (1994) found endotoxin concentrations, measured by the LAL assay,
of 24.8 EU/g in floor dust from 12 town halls (Gyntelberg, Suadicani et al. 1994), which is
3000 times less than found in the floor dust of this study (mean 76.4 EU/mg). An
explanation could be that children drag more contaminants into the rooms than adults.
However, Kildesø et al. (1998) measured the endotoxin concentration in an administration
building, a kindergarten and a school and found 3.3 EU/mg, 1.65 EU/mg and 2.84 EU/mg
dust respectively (Kildesø, Tornvig et al. 1998). These levels are approx. 30 times lower than
the levels found in the present study. In the waste handling industry, endotoxin
concentrations are generally low in measured air samples (0.1-30 ng/m 3 corresponding to
about 1.25-375 EU/m 3 ), but comparable to the levels found in the present study (Poulsen,
Breum et al. 1995b). In waste samples as compost and biowaste the concentration of
endotoxin was 4,000-190,000 ng/g (corresponding to about 50,000-2,375,000 EU/g)
(Nielsen 1998), being much higher than found in the indoor climate. The endotoxin
100

concentration in the different dust samples (surface, floor and ventilation duct) was not
significantly different in “good” and “bad” schools (Meyer 2000).
In this study a significant, but weak positive correlation was observed for the PF of the
surface dust samples and endotoxin (Paper III). An even stronger correlation emerged when
correction was made for the content of organic matter of the surface dust samples and the
dust samples from the ventilation shafts. Hence, the results may indicate, that endotoxin
contributes to some extent to the inflammatory potential of dust from schools. As mentioned
in section 4.4.2. the concentrations of endotoxin found in the dust was below the detection
limit of practically all three assays. Correlations between endotoxin and potency of school
dust may therefore indicate that endotoxin is a proximeasure for other components as the
presence of Gram-negative bacteria. Microbial growth in it self could lead to production and
secretion of xenobiotics leading to the observed symptoms. Meaning that endotoxin could be
a marker for e.g. several stimulatory agents. Alternatively, it could be that the potency of
endotoxin in house dust is enhanced by co-presence of other biological active components.
We have in Paper II used LPS purified from four different Gram negative bacteria strains to
stimulated the A549 cells, and found that the potency of the endotoxins were different in the
A549 bioassay. This was also shown for the monocytic assays (see 4.2.3. and 4.3.3.). Taking
the above facts in consideration the content of endotoxin should not be considered of major
importance in governing the inflammatory potency of dust from schools.
It has been questioned to what extent the bioavailability of endotoxin measured by the LAL
assay, correlates with bioavailability in the lung. When comparing different cell wall
preparations from Enterobacter agglomerans with respect to the ability to induce acute
reductions in forced expiratory volume in one second (FEV1) in human inhalation
experiments, the LAL assay underestimated the endotoxin content up to three-fold
(Rylander, Bake et al. 1989). This suggests that the LAL assay only detects one third of the
biologically active endotoxin in the lungs, and the rest remains inside dust particles or on
bacterial cell fragments, but is still able to induce endotoxic effects (Rylander 1997a).
Hence, the LAL assay may not be the best method for measurement of endotoxic activity,
but it is one of the most used methods for quantitative endotoxin measurements today. In a
study by Saraf et al. (1999) the content of 3-hydroxylated fatty acids (3-OH) in house dust, a
marker for LPS, was measured by gas chromatography-mass spectrometry (GC-MS) or GC-
101

tandem MS (GC-MS-MS) and endotoxin activity of the same samples was measured by the
LAL assay. The correlation coefficient between the LAL assay and the GM-MS methods
was 0.6 for (3-OH) of 10, 12 or 14 carbon chain lengths, for 16 and 18 carbon chain lengths
the coefficient was poorer (Saraf, Park et al. 1999). This study shows that other methods for
endotoxin measurements are available. In another study dust samples, from a villa and an
apartment, induced secretion of IL-6 and IL-8 from the A549 cell line. The cytokine
secretion did not correlate with the measured amount of 3-OH and the results of the LAL
assay (Saraf, Larsson et al. 1999). Potencies of floor dust samples from waste handling
facilities also did not show any correlation to endotoxin (Allermann and Poulsen 2000). The
endotoxin concentrations found in dust from these waste-handling facilities (1.5-753 EU/mg
floor dust) are in the same range as found in the dust samples from the schools.
Bacteria and fungi growing saprotrophically on damp surfaces in buildings are found in the
air of the indoor climate (Flannigan 1992), and in the dust. Both content of bacteria and
fungi have been shown to correlate to BRS (Gyntelberg, Suadicani et al. 1994; Koskinen,
Husman et al. 1994; Norbäck, Edling et al. 1994). Also in the in vitro studies are bacteria,
fungi and components derived from these microorganisms used to induce inflammatory
response. Larsson et al. (1999) used different bacteria strains (one Gram negative and four
Gram positive) and extracts from these bacteria to induce IL-6 and IL-8 secretion from A549
epithelial cells and primary human alveolar macrophages (Larsson, Larsson et al. 1999a).
Spores of Actinomycetes sp. and Streptomyces sp. isolated from mouldy houses have been
shown to induce nitric oxide, TNF-α and IL-6 secretion from RAW364.7 mouse
macrophages (Hirvonen, Ruotsalainen et al. 1997; Hirvonen, Nevalainen et al. 1997a). Fungi
associated with ODTS have also been shown to induce the production of superoxide anion,
leukotriene B4, IL-1 and complement component C5a from cultures of rat and guinea pig
alveolar macrophages (Sorenson, Shahan et al. 1994).
In this study the amount of viable fungi, bacteria and actinomycetes was measured from the
dust samples without differentiation into species. The number of bacteria in the dust samples
was generally higher than the number of actinomycetes and fungi. Flannigan and Miller
(1994) also noted that the number of bacteria in indoor air is frequently considerably higher
than fungi (Flannigan and Miller 1994). The highest number of bacteria (incl. actinomycetes)
was found in the floor dust and the highest number of fungi was found in the surface dust
102

(Table 15). This was also reported in Paper III. The concentration of microorganisms,
however, was expressed in mg of dust, and dust from the floors contains more sand and dirt,
giving weight to the sample, than in surface dust samples. This could explain the higher
concentration of fungi in surface dust. In a study of 12 town halls floor dust contained in
median 0.015 cfu microfungi/mg dust (0.005-0.13 cfu/mg) and 0.0074 cfu bacteria/mg dust
(0.0022-0.009 cfu/mg) (Gyntelberg, Suadicani et al. 1994). These concentrations are very
low compared to the concentrations found in this study. Kildesø et al. (1998) found
comparable levels of fungi in floor dust from an administration building, a kindergarten and
a school, the amount of bacteria found was 114 cfu/mg, 315 cfu/mg and 70 cfu/mg
respectively (Kildesø, Tornvig et al. 1998). This is 6.5-30 times less than found in the floor
dust of the present study. In other working environments the concentration of bacteria and
fungi are much higher. People occupied in agriculture, waste collection and other similar
occupations are exposed to organic dust containing bacteria and fungi in the range of 10-10 9
cfu/mg (Dutkiewicz 1997). Settled dust from animal sheds contained 10 3 -10 7 cfu/mg of
mesophilic bacteria, which was 10 2 -10 4 fold higher than found in dust from schools and daycare
centres (Andersson, Weiss et al. 1999). Airborne concentrations of mesophilic bacteria
were 10 5 -10 9 cfu/m 3 in animal sheds and 10 2 -10 3 cfu/m 3 in schools and day-care centres
(Andersson, Weiss et al. 1999). Hyvärinen et al. (1993) found 98 cfu/mg dust in buildings
with mould problems and 72 cfu/mg dust in reference buildings, in the fall. In the winter the
concentrations were slightly higher (Hyvärinen, Reponen et al. 1993), which is comparable
to the concentrations found in the present study. Nevalainen (1990) has suggested that the
actinomycetes may be the real indoor air problem, as she found actinomycetes in 70% of the
air samples from problem rooms but only in less than 10% of the samples from control
rooms (Nevalainen, Kotimaa et al. 1990). However, the concentration of actinomycetes
found was low (below 240 cfu/m 3 ) and no difference between actinomycetes in “good” or
“bad” schools was observed. In the present study the concentration of actinomycetes found
in the dust was highest for the floor dust (28.05 cfu/mg dust ranging 0.6-2402 cfu/mg), and
could indicate that these soil bacteria was dragged in from the outside. In the thesis by
Harald Meyer the concentration of bacteria (median value) in the floor dust was statistically
higher (p=0.009) in samples from the “good” schools (Meyer 2000). The mean score of the
extension of moist and mould in the 20 schools was found to be higher for individuals with
BRS than those without.
103

No correlation was observed between the viable counts of microorganisms and the PF from
the surface dust samples (Paper III). The strongest correlation was observed with moulds
from dust samples from the exhaust ducts, but after correction to the organic fraction in the
dust the correlation coefficient dropped. Possibly the correlation reflects the large difference
in the organic fraction observed between dust from natural exhaust ducts and dust from
mechanical exhaust ducts. Consequently, the concentration of viable microorganisms does
not seem to contribute significantly to the inflammatory potential of dust in schools.
Important indoor allergen sources are mites, cats, dogs, cockroaches, moulds, and rat urine
(Munir 1995). In this study allergens from mites (Der p, Der f, and Der m), cat (Fel d) and
dog (Can f) in dust samples from the floor and from exhaust ducts were measured. Factors as
climate, season, dampness, humidity, ventilation, carpets and upholsteries are all influencing
the concentration of allergens found in buildings (Munir 1995). In a Swedish investigation
mite allergen (Der p and Der f) was detected in all the investigated classrooms. The highest
concentration was found in dust from the tables, which was significantly higher than the
concentration in dust samples from the floors and chair surfaces (Einarsson, Munir et al.
1995). It was suggested that cat and dog owners bring cat and dog allergens to public areas
e.g. schools and day-care centres, and that textiles and upholstery function as reservoirs for
these allergens (Berge, Munir et al. 1998; Munir, Einarsson et al. 1995). In the present study
the concentration of mite allergens was generally less than 123 ng/g dust, which is
considered to be non to low contamination (0-2000 ng/g) (Data sheet from ALK Abelló).
The concentration of cat allergens was generally considered to be none to low (0-1000 ng/g);
only seven samples of the 82 samples were given the contamination degree of middle (1001-
8000 ng/g) (Data sheet from ALK Abelló). Dog allergens were also generally considered to
be none to low (0-2000 ng/g), only three samples contained concentrations above 2000 ng/g
(Data sheet from ALK Abelló). Concentrations of allergen above 2000 ng allergen/g dust is
according to WHO considered to contain a risk for development of an IgE mediated
response. In allergic patients a concentration of 10000 ng of allergen/g dust is considered a
risk for development of acute asthma (Data sheet from ALK Abelló). In this study only a
few samples contained allergen concentrations above 2000 ng/g dust, and the risk for
development of an IgE-mediated asthma due to exposure to these allergens from going to
school must be considered to be low. It is interesting that the content of allergens in dust
samples correlated with the results of the A549 bioassay, measuring the inflammatory
104

potential of the dust. Airway inflammation is one of the major components of asthma, and
the eosinophils are considered as proinflammatory cells with capacity to be recruited to the
lungs in response of IgE-depended signals (Djukanovic, Roche et al. 1990). Hoover and
Platts-Mills (1995) argue that exposure of the lungs to allergens may not only trigger acute
asthma attacks, but also produce inflammation, late reactions, and persistent bronchial hyperresponsiveness
(Hoover and Platts-Mills 1995). The production of IL-8 from lung epithelial
cells therefore may be a part of the inflammatory process seen in e.g. asthma, as asthma is an
inflammatory disease characterised by epithelial damage, mediator release, and cellular
infiltration (Hoover and Platts-Mills 1995). Tomee et al. 1998 found that extracts of mite
(Lepidoglyphus destructor and Dermatophagoides pteronyssinus) and pollen allergens
(Phleum pratensis and Betula verrucosa) caused a dose-dependent IL-8 secretion from A549
cells, reaching significant difference at 8 µg/ml (mite and grass pollen) or at 40 µg/ml (birch
pollen). At concentrations above 40 µg mite allergen/ml cell detachment was seen (Tomee,
Van Weissenbruch et al., 1998). Extracts of the mite allergen Der p1 from
Dermatophagoides pteronyssinus was also found to promote activation of the transcriptional
factor NF-kBα by interference of its cytosolic inhibitor IkBα, leeding to transcription of m-
RNA for GM-CSF and RANTES (Stacey, Sun et al., 1997). These experiments however,
used concentrations well above the concentrations found in the dust samples.
Correcting the PF with the organic fraction did generally not result in better correlations with
the microbiological data or with allergens, indicating that none of the tested parameters are
quantitatively dominating factors in the dust. The content of these single parameters in the
dust themselves contributes very little to the potency of a total dust sample. Hence, when
performing the A549 bioassay a correction of data against the content of organic matter
seems not necessary, which will simplify the use of the method in future research and
screening.
As mentioned in section 2.1. many different factors may cause the BRS. The question is,
whether these multifactorial causes are reflected in the PF from dust? Endotoxin correlates
weakly with the PF. The content of microorganisms in the dust revealed generally very low
correlation coefficients with the PF. The same was true for the allergens in the floor dust.
The obtained potencies of the dust may therefore have multifactorial causes with none of the
measured single parameters being of major importance.
105

7. Correlation of results from the in vitro methods with epidemiological
data
In summary, the results of the thesis of Harald Meyer are mentioned below (Meyer 2000).
The response rate of the questionnaire sent to employees and pupils (8’Th grade and up) of
75 schools in Copenhagen was 66%. Technical investigation was performed and dust
samples were collected from ten schools with the lowest mean prevalence of BRS (“good”
schools) and ten schools with the highest (“bad” schools).
The prevalence of BRS in the questionnaire survey was fatigue 36%, headache 25% and lack
of concentration 22%. The prevalence of symptoms from the mucous membrane was about
12%. Higher prevalences were found in women, smokers, in participants with asthma or hay
fever and in participants with problems in their psychosocial work environment. These test
variables were all associated with BRS together with the building parameter “flat roof”. The
employees experienced more symptoms of the mucous membrane than the pupils did, but the
pupils reported more often CNS-symptoms than the employees did.
The 10 “bad” schools had significantly higher mean room temperature, less room volume per
person and a higher proportion of rooms with mechanical ventilation. Strong associations
between BRS and mechanical ventilation were found. The prevalence of headache was 38%
and nasal irritation 9% in “mouldy rooms” versus 27% and 9% in “non mouldy rooms”. In
models of multiple logistic regression the dominating exposure variables were mechanical
ventilation, but also the concentration of dog allergens in floor dust, and the PF was highly
significantly associated with BRS. No association was found between building factors and
asthma, and only week associations were observed between building factors and hay fever.
7.1. Correlation of results to parameters of the technical investigation and selected
epidemiological data
The correlations below are based on the PF of the total dust sample tested in the A549
bioassay.
106

A statistical analysis (by Harald Meyer) on the PF against different parameters from the
technical investigation of the school, using a Mann-Whitney test or multiple linear
regression, gave a significant correlation (p

No difference was found for symptoms such as asthma, dry throat, itching of the eyes,
running eyes, positive index of at least two out of five symptoms of the mucous membrane
and the skin, headache and difficulty in concentration for “bad” schools above and below
CO2. In the same test headache and itching eyes on both “good” and “bad” schools were
significant higher above the CO2 than below. A positive index of at least two out of five
symptoms of the mucous membrane and the skin were found in 84% of all schools above the
CO2 but this was not statistically significant (p=0.1).
7.2. Discussion of correlations to parameters of the technical investigation and selected
epidemiological data
The chosen design of the school study was a cross sectional study. In contrast to a cohort
study the cross sectional study could be performed within a limited time span, and include a
large study population. Information of exposure and symptoms would be available. The main
disadvantage of the chosen design is the difficulties of knowing if the exposure variables and
the variables of the outcome are included in a exposure-effect association. The knowledge of
symptoms before exposure are also lacking in this study design. 20 schools of the original 75
schools were chosen, which was within the economic limit of the project. To include a large
variation between symptoms of “good” and “bad” schools 10 schools in each end of a
symptom scale were chosen.
A positive correlation between the roof construction and the PF indicates that buildings with
a flat roof have some components in the dust that is not present or present in smaller
concentrations in the dust from buildings with a pitched roof. These components may be
influenced by more cases of water damage in the buildings with a flat roof, with subsequent
decomposition of building materials, growth of microorganisms and liberation of
components from microorganisms or VOCs. Meyer 2000 also found an association between
the roof construction and BRS, and between the year of construction of the building and
BRS (Meyer 2000; Meyer, Nielsen et al. 1996). The volume per person was negatively
correlated with the PF and to symptoms of the CNS and the BRS index (Meyer, Nielsen et
al. 1996; Meyer 2000). This indicates a larger accumulation of potentially harmful
compounds tracked in or emitted by the users in the smaller rooms.
108

Significant correlations were found between the PF and symptoms of the mucous membrane
and the skin. The inflammatory potential of the lung epithelial cells therefore may be an
indicator of inflammatory processes in other tissues than the lungs, e.g. the mucosa of the
nose and eyes, and the skin. In the school study, 12% of the participants experienced
symptoms of the mucous membrane (Meyer, Nielsen et al. 1996; Meyer 2000). Menzies et al
(1996) reported increased symptoms of the mucosa with higher total VOCs, nitrogen dioxide
and higher total contaminant load of the office building (Menzies, Tamblyn et al. 1996).
They also reported eye symptoms with higher concentrations of dust and nitrogen dioxide. In
the Danish “Town Hall Study” a significant association between the prevalence of particles
in the dust (determined by microscopy evaluation of the dust) and symptoms of the mucous
membranes was found (Gyntelberg, Suadicani et al. 1994). I the same study symptoms of the
mucous membranes were also correlated to the concentration of Gram-negative bacteria in
the dust. In conclusion contaminents of the indoor air are included in or adhered to the dust
of the room and may therefore add to the observed symptoms of e.g. the mucous membrane
via exposure to dust.
Headache is a symptom of the CNS, but an inflammatory reaction may induce some types of
headache where the inflammatory markers IL-1 and TNF-α have been shown to participate
(Martelletti, Stirparo et al. 1999). IL-8 is a proinflammatory cytokine as IL-1 and TNF and
could therefore also be involved in development of headache. This may explain the observed
correlation between headache and PF of floor dust. In the school study there were higher
report rate of headache in rooms with mould extension ≥ 0.25 m 2 (Meyer 2000). Gyntelberg
et al. (1994) found that the prevalence of headache and general malaise/dizziness was
significantly correlated to macromolecular dust (Gyntelberg, Suadicani et al. 1994).
Lack of concentration did not show any association with the PF, and may be caused by
exposures, which are not likely to be reflected in dust, e.g. a high level of CO2, high
temperature, high humidity and high noise level in the room. In the school study many of the
participants were often annoyed by stuffy air, narrow space and unpleasant smell (Meyer,
Nielsen et al. 1996; Meyer 2000). These factors could influence the symptoms of headache
and lack of concentration, as Meyer (2000) also found 36% with the feeling of abnormal
tiredness, 25% reported headache and 22% with difficulty in concentration. In the school
study no difference in temperature was found between naturally and mechanically ventilated
109

ooms, but the naturally ventilated rooms had a significantly higher level of CO2 (Meyer
2000). The naturally ventilated rooms also had a smaller volume per person than the
mechanically ventilated rooms. Only one of the naturally ventilated rooms and 1 /3 of the
mechanically ventilated rooms had levels of CO2 below the recommended maximal value of
1000 ppm. In many rooms CO2 levels above 2000 ppm were found, which is the limit value
set by the labour inspection agency in Denmark, for improvement of the air change in the
room (Meyer 2000). Rooms with natural ventilation also had a higher level of background
noise maybe because of their central location in the city, where the noise from traffic is high.
This indicating that, the observed CNS symptoms may be caused by exposures that are not
reflected in the dust, i.e. CO2 and noice. Stratifying the data on “good” and “bad” schools,
the “bad” schools had significantly higher levels of mean temperature and air change and
significantly lower volume per person (Meyer 2000), which further supports the hypothesis
that physical parameters are important for the sensation of the indoor climate.
Atopics, persons with allergic disorders or persons with bronchial hyperresponsivness react
to pollutants at lower levels or more vigorous than normal subjects, and an association
between some of chemical and physical parameters could therefore be expected for these
individuals. However, no correlation was found between asthma and PF. Meyer 2000 found
no correlation between asthma and room measurements on temperature, CO2 concentration,
volume per person, humidity, noise level and air change. For hay fever a week association
was found to the parameters CO2 concentration and volume per person (Meyer 2000).
Asthma was reported by 11%, hay fever by 23% and 20% reported to have or have had
“child eczema”, giving a study base of about 800-1600 persons (Meyer, Nielsen et al. 1996;
Meyer 2000), which maby is to small to show such associations.
No difference in the tested symptoms from the “bad” schools was found above and below
the CO2. This may indicate that the symptoms found at the “bad” schools are the same
whether the PF is above or below the CO2. Alternativly, resticting data analysis to the “bad”
schools only may result in to few observations to demonstrate significant defferences. Using
a cut off value of 3.8 ng IL-8/mg (Meyer 2000), includes more “good” schools and should
therefore not give a larger difference in the symptoms above and below the chosen cut off
value. Symptoms in buildings below the CO2 could be a result of exposures or factors not
110

included in the measurements of the dust e.g. air humidity, light, noice or psycosocial
problems, but factors influencing the individual perception of the indoor air.
All the above correlations between symptoms and other parameters are made on a group of
adults and older children (>13 years). The response of smaller children could therefore not
be estimated from this study. The first years of life seems to be an especially vulnerable
period, as an association between early exposure to outdoors and indoor allergens and
development of sensitisation and development of allergic diseases were found (Høst 1997).
An increase in the prevalence of atopic diseases in children and young adults are seen (Høst
1997). Ulrik et al. 1998 found, in a six-year follow-up study of Danish school children (age
7-17 years), an increased prevalence of asthma from 5.4% to 15%. A significant increase in
positive house dust mite skin-prick-test was also found (Ulrik and Backer et al. 1998). These
results imply that the future generation of school children may be more sensitive to
xenobiotics in their surroundings compared to adults and school children to day.
111

8. Conclusion
In this Ph.D.-study three in vitro methods for evaluation of dust from the indoor environment
were developed using the cell lines A549 (lung epithelial cell line), U937 and THP-1
(monocytic cell lines). All three in vitro models developed are shown by the method
evaluation to be stable, in control, fast and easy to perform. Based on the hypothesis that
monocytes are more sensitive than the epithelial cells, the question was whether a monocytic
assay could replace the A549 bioassay, as a more sensitive screening tool. The U937 and the
THP-1 cell lines are clearly more sensitive than the A549 cell line when comparing the PF
obtained on pure chemical components and dust samples. However, the data material on the
monocytic cell lines in the present study is not large enough to make powerful correlations to
individual parameters in the dust or to the health outcomes in the questionnaires. Hence, at
present time no firm conclusion can be made on the monocytic assays compared with the
A549 bioassay.
The potencies of the dust samples (PF) tested in the three bioassays were tested for
correlation to content of viable microorganisms, endotoxin, and content of organic fraction
and allergens in the dust. Also correlation tests to selected epidemiological data from Harald
Meyer (AMK) were made. Statistical significant positive correlation was found between the
TNF corrected PF of all the dust samples in the A549 bioassay and the content of organic
dust. This was also found for the surface dust (rs = 0.90), for the floor dust (rs = 0.53), and
for the dust samples from exhaust ducts (rs = 0.74), indicating that the organic fraction of the
dust samples are important for the IL-8 secretion. However, no correlation was found
between the PF of the two monocytic bioassays and the content of organic dust, perhaps
because of the small number of samples tested, but a tendency was seen.
The PF of the A549 bioassay was significantly correlated to endotoxin in the surface dust
and the content of allergens in the floor dust and in dust from exhaust ducts. However, the
concentrations of endotoxin in the dust samples were below the detection limit of the
bioassays. Hence endotoxin in itself does not add much to the obtained stimulation of the
cells and one may speculate that endotoxin acts as a proximeasure of some other unknown
biological active components in the dust samples. The content of viable bacteria in the floor
112

dust and the PF of floor dust were found to correlate. This was also the case for all tested
microorganisms versus the PF of dust from exhaust ducts. A positive correlation was also
found between all tested microorganisms and the PF of organic fraction from exhaust ducts.
However, these correlations were weak, indicating that microorganisms do not contribute
significantly to the inflammatory potential of dust in schools.
A significant difference between the content of the different types of microorganisms in the
three types of dust samples (the surface dust, floor dust and dust from the exhaust ducts) was
found, with the content of actinomycetes and total bacteria being highest in the floor dust
and mould in the surface dust. The actinomycetes are soil bacteria and could have been
dragged in from the outside, which may explain the higher content in floor dust than the
other types of dust. Spores of fungi are designed to spread by air and therefore may be
airborne for longer time than e.g. bacteria. This could explain the higher content of mould in
the surface dust than the floor dust. This could, on the other hand, not explain the low
content in the exhaust dust samples.
The PF from floor dust and surface dust tested in the A549 bioassay correlated, indicating
that the same biological active compounds are reflected in both samples, and an exchange of
particles takes place between the floor dust and the surface dust.
PF from “good” schools was significantly lower than the PF of the “bad” schools. Statistical
positive associations to the PF of floor dust were found to some of the individual parameters:
itching of the eyes, positive index of at least two out of five symptoms of the mucous
membrane and the skin, and headache. These are all parameters associated with the BRS,
indicating that the potency of dust measured by IL-8 secretion from the A549 bioassay may
be used to predict a “bad” indoor climate on the basis of the dust from the building.
In schools with high BRS index (“bad” schools) a significantly higher mean room
temperature, less room volume per person and a higher proportion of rooms with mechanical
ventilation was found (Meyer 2000). The PF of the exhaust duct samples from rooms with
mechanical ventilation were significantly higher than samples from rooms with natural
ventilation. Buildings with flat roof had a significant higher PF than buildings with a pitched
roof, and less volume per person was associated with a higher PF. The picture of a building
113

with a “bad” indoor climate could therefore be a building with mechanical ventilation, a flat
roof, many persons per m 2 , and high mean temperature; all factors associated with a high PF
of the dust and a high BRS index.
In an evaluation of the A549 bioassay two cut-off values of 2.0 (low potency) and 4.5 ng IL-
8/ mg (high potency) were chosen. Differences above and below the calculated cut-off
values of 4.5 ng IL-8/mg floor dust (CO2) were tested. A significant difference in
distribution of “good” and “bad” schools above and below the CO2 was found with the
“good” schools below and the “bad” schools above the CO2. A smaller volume per person
was found for “bad” schools above the CO2 than below the CO2, which was also found for
all the schools (“good” and “bad”) above the CO2. Rooms with no recirculation were
associated with PFs above the CO2. The area covered by mould was found to be larger
below the CO2 than above, indicating that the aerea covered by visible growth of mould
does not have to correlate with observed biological effect, as othere factors as high water
content could inflict on the liberation of spores and mycotoxins (Wilkins, Larsen et al.,
1998).
Symptoms of headache and itching eyes in all the schools were found to be higher above the
CO2 than below, but no difference was found for positive index of at least two out of five
symptoms of the mucous membrane and the skin. These three symptom scores however,
correlated with PF of the floor dust. The CO2 value, therefore, is not sensitive in predicting
prevalence of the individual BRS.
The A549 bioassay seems to be able to differentiate between the “good” and the “bad”
schools, and therefore this bioassay may be a useful screening tool in the evaluation of
indoor air problems. The THP-1 bioassay needs to be evaluated further, before final
conclusion can be drawn.
114

9. References
(1996). Grænseværdier for stoffer og materialer. At-anvisning Nr. 3.1.0.2. Report no. 3.1.0.2. Andersen, J.
(eds). Arbejdstilsynet, Copenhagen. (Report in Danish)
Adachi, Y., Okazaki, M., Ohno, N., and Yadomae, T. (1994). Enhancement of cytokine production by
macrophages stimulated with (1→3)-beta-D-glucan, grifolan (GRN), isolated from Grifola frondosa. Biol
Pharm Bull. 17 (12), 1554-1560.
Agace, W., Hedges, S., Andersson, U., Andersson, J. et al. (1993). Selective Cytokine Production by Epithelial
Cells following Exposure to Escherichia coli. Infect Immun. 61 (2), 602-609.
Allermann, L. and Poulsen, O. M. (2000). Inflammatory potential of dust samples from waste handling
facilities measured as IL-8 secretion from lung epithelial cells in vitro. Ann Occup Hyg, 44, 259-269.
Amdur M. (1996). Animal toxicology. (cap.5), In: Particles in our air. Concentration and health effects. 85-
121. Harward University Press. ISBN: 0-674-24077-4.
Andersen, I. and Lundquist, G. R. (1966). Indendørsklima i skoler. SBI-Rapport 57. Teknisk Forlag,
København. (Report in Danish)
Andersson, A. M., Weiss, N., Rainey, F., and Salkinoja-Salonen, M. S. (1999). Dust-borne bacteria in animal
sheds, schools and children's day care centres. J Appl Microbiol. 86 (4), 622-634.
ASTM. (1994). Standard Practice for Collection of Floor Dust for Chemical Analysis. Designation: D 5438 -
94. in: Annual Book of ASTM Standards. 570-576.
Bachmann, M. O. and Myers, J. E. (1995). Influences on sick building syndrome symptoms in three buildings.
Soc Sci Med. 40 (2), 245-251.
Baggiolini, M., Dewald, B., and Moser, B. (1994). Interleukin-8 and Related Chemotatic Cytokines - CXC and
CC chemokines. in: Advances in Immunology, Vol 55. Academic Press, Inc., 97-179.
Becher, R., Hongslo, J. K., Jantunen, M J., and Dybing, E. (1996). Environmental chemicals relevant for
respiratory hypersensitivity: the indoor environment. Toxicology Letters. 86 155-162.
Berge, M., Munir, A. K., and Dreborg, S. (1998). Concentrations of cat (Fel d1), dog (Can f1) and mite (Der f1
and Der p1) allergens in the clothing and school environment of Swedish schoolchildren with and without
pets at home. Pediatr Allergy Immunol. 9 (1), 25-30.
Bianchi, M., Fantuzzi, G., Bertini, R., Perin, L., Salmona, M., and Ghezzi, P. (1993). The Pneumotoxicant
Paraquat Induces IL-8 mRNA in Human Mononuclear Cells and Pulmonary Epithelial Cells. Cytokine. 5
(5), 525-530.
Bischoff, E. (1989). Sources of pollution of indoor air by mite-allergen-containing house dust. Environment
International. 15 181-192.
Brock, T. D., Madigan, M. T., Martinko, J. M, and Parker, J. (1994a). The cell Wall of bacteria: Structure and
function. in: Biology of Microorganisms. Prentice-Hall, Inc., Englewood Cliffs, New Jersey. 58-66.
ISBN:0-13-042169-3.
Brock, T. D., Madigan, M. T., Martinko, J. M, and Parker, J. (1994b). Filamentous Actinomycetes. in: Biology
of Microorganisms. Prentice-Hall Inc., Englewood Cliffs, New Jersey. 811-814. ISBN:0-13-042169-3.
Brock, T. D., Madigan, M. T., Martinko, J. M, and Parker, J. (1994c). Fungi. in: Biology of microorganisms.
Prentice-Hall, Inc., Englewood Cliffs, New Jersey. 846-850. ISBN:0-13-042169-3.
115

Brooks, S. M. (1994). Host susceptibility to indoor air pollution. J Allergy Clin Immunol. 94 (2 Pt 2), 344-351.
Carpenter, A., Evans, T. J., Buurman, W. A., Bemelmans, M. H., Moyes, D., and Cohen, J. (1995). Differences
in the shedding of soluble TNF receptors between endotoxin-sensitive and endotoxin-resistant mice in
response to lipopolysaccharide or live bacterial challenge. J Immunol. 155 (4), 2005-2012.
Carter, J. D., Ghio, A. J., Samet, J. M., and Devlin, R. B. (1997). Cytokine Production by Human Airway
Epithelial Cells after Exposure to an Air Pollution Particle Is Metal-Dependent. Toxicol Appl Pharmacol.
146 180-188.
Castellan,R.M., Olenchock,S.A., Kinsley,K.B. and Hankinson,J.L. (1987). Inhaled endotoxin and decreased
spirometric values. N Engl J Med 317, 605-610.
Cavaillon, J. M. (1994). Cytokines and macrophages. Biomed Pharmacother. 48 (10), 445-453.
Chan-Yeung, M., Clark, C. S., Donham, K. J., do Pico, G. A. et al. (1994). Organic Dust. Exposure, Effects,
and Prevention. Rylander, R. and Jacobs, R. R. (eds). Lewis Publishers CRC Press, Inc., 3-283. ISBN:0-
87371-966-X.
Clausen, P. A., Wilkins, C. K., Wolkoff, P., and Jeppesen, P. (1997). Fedtsyresalte (sæbe) i gulvstøv fra
kontorbygninger. NAM '97. 45 , 129-130. Denmark. (Conference Proceeding in Danish).
Clemmensen, O. J., Menne, T., Kaaber, K., and Solgaard, P. (1981). Exposure of nickel and the relevance of
nickel sensitivity among hospital cleaners. Contact Dermatitis. 7 (1), 14-18.
Crestani, B., Cornillet, P., Dehoux, M., Rolland, C. et al. (1994). Alveolar Type II Epithelial Cells Produce
Interleukin-6 In Vitro and In Vivo. Regulation by Alveolar Macrophage Secretory Products. J Clin Invest.
94 731-740.
Cromwell, O., Hamid, Q., Corrigan, C. J., Barkans, J. et al. (1992). Expression and generation of interleukin-8,
IL-6 and granulocyte- macrophage colony stimulating factor by bronchial epithelial cells and enhancement
by IL-1beta and tumor necrosis factor-alfa. Immunology. 77 330-337.
de Waal Malefyt, R., Abrams, J., Bennett, B., Figdor, C. G., and de Vries, J. E. (1991). Interleukin 10(IL-10)
inhibits cytokine synthesis by human monocytes: an autoregulatory role of IL-10 produced by monocytes. J
Exp Med. 174 (5), 1209-1220.
Deaton, P. R., McKellar, C. T., Culbreth, R., Veal, C. F., and Cooper, J. A., Jr. (1994). Hyperoxia stimulates
interleukin-8 release from alveolar macrophages and U937 cells: attenuation by dexamethasone. Am J
Physiol. 267 (2 Pt 1), L187-92.
DeForge, L. E. and Remick, D. G. (1991). Kinetics of TNF, IL-6, and IL-8 gene expression in LPS- stimulated
human whole blood. Biochemical and Biophysical Research Communications. 174 (1), 18-24.
Devalia, J. L., Bayram, H., Rusznak, C., Calderon, M., Sapsford, R. J., Abdelaziz, M. A., Wang, J., and Davies,
R. J. (1997). Mechanisms of pollution-induced airway disease: in vitro studies in the upper and lower
airways. Allergy, 52 45-51.
Devalia, J. L. and Davies, R. J. (1993). Airway epithelial cells and mediators of inflammation. Resp Med. 87
405-408.
Devlin, R. B., McKinnon, K. P., Noah, T., Becker, S., and Koren, H. S. (1994). Ozone-induced release of
cytokines and fibronectin by alveolar macrophages and airway epithelial cells. Am J Physiol. 266 (6 Pt 1),
L612-L619.
Dinarello, C. A. (1992). The Biology af Interleukin-1. in: Interleukines: Molecular Biology and Immunology.
Kishimoto, T. (eds). Chem Immunol, Basel, Karger. 1-32.
116

Djukanovic, R., Roche, W. R., Wilson, J. W., Beasley, C. R., Twentyman, O. P., Howarth, R. H., and Holgate,
S. T. (1990). Mucosal inflammation in asthma. Am Rev Respir Dis. 142 (2), 434-457.
Douwes, J., Dubbeld, H., van Zwieten, L., CA, Doekes, G. et al. (1997). Work Related Acute and (Sub-
)Chronic Airways Inflammation Assessed by Nasal Lavage in Compost Workers. Ann Agric Environ Med.
4 149-151.
Douwes, J., Versloot, P., Hollander, A., Heederik, D. and Doekes, G. (1995). Influence of Various Dust
Sampling and Extraction Methods on the Measurement of Airborne Endotoxin. Applied and Environmental
Microbiology, 61, 1763-1769.
Dutkiewicz, J. (1997). Bacteria and Fungi in Organic Dust as Potential Health Hazard. Ann Agric Environ Med.
4 11-16.
Dutkiewicz, J., Skorska, C., Sitkowska, J., Ochalska, B., and Kaczmarski, F. (1988). Properties of the
Endotoxins Produced by Various Gram-Negative Bacteria Present in Occupational Environments. Cotton
Dust. 187-189. New Orleans. (Conference proceeding)
Einarsson, R., Munir, A. K., and Dreborg, S. K. (1995). Allergens in school dust: II. Major mite (Der p I, Der f
I) allergens in dust from Swedish schools. J Allergy Clin Immunol. 95 (5 Pt 1), 1049-1053.
European Committee for Standardization. (1993). Workplace atmospheres - Size fraction definitions for
measurement of airborne particles. European Standard. Report no. EN 481. CEN, Brussels.
Fan, X., Stelter, F., Menzel, R., Jack, R., Spreitzer, I., Hartung, T., and Schutt, C. (1999). Structures in Bacillus
subtilis are recognized by CD14 in a lipopolysaccharide binding protein-dependent reaction. Infect Immun.
67 (6), 2964-2968.
Flannigan, B. (1992). Indoor microbiological pollutants - Sources, species, characterisation and evaluation. in:
Chemical, microbiological, health and comfort aspects of indoor air quality - State of the art in SBS.
Knöpel, H. and Wolkoff, P. (eds). Kluwer Academic Publishers, Dordrecht. 73-98. ISBN:0-7923-1703-3.
Flannigan, B. and Miller, D. (1994). Health Implications of fungi in indoor environments - an overview. in:
Health implications of fungi in indoor environments. Samson, R. A., Flannigan, B., Flannigan, M. E.,
Verhoeff, A. P. et al. (eds). Elsevier Science B.V., Amsterdam, The Netherlands. 3-28. ISBN:0-444-81997-
5.
Fogelmark, B., Goto, H., Itoh, W., Kojima, T. et al. (1991). First Glucan Lung Toxicity Workshop. Report no.
4. Rylander, R. and Goto, H. (eds). Comittee on Organic Dusts, ICOH, Sweden. 1-20.
Franck, C. (1991). Fatty layer of the precorneal film in the 'office eye syndrome'. Acta Ophthalmol. 69 (6), 737-
743.
Friedland, J. S., Shattock, R. J., Johnson, J. D., Remick, D. G., Holliman, R. E., and Griffin, G. E. (1993).
Differential cytokine gene expression and secretion after phagocytosis by a human monocytic cell line of
Toxoplasma gondii compared with Mycobacterium tuberculosis. Clin Exp Immunol. 91 282-286.
Friedland, J. S., Shattock, R. J., Remick, D. G., Griffin, G. E., and Shattock, R. (1993). Mycobacterial 65-kD
heat shock protein induces realease of proinflammatory cytokines from human monocytic cells
Mycobacterial 65-kD heat shock protein induces release of proinflammatory cytokines from human
monocytic cells. Clin Exp Immunol. 91 (1), 58-62.
Gerritsma, J. S. J., Heimstra, P. S., Gerritsen, A. F., Prodjosudjadi, W. et al. (1996). Regulation and production
of IL-8 by human proximal tubular epithelial cells in vitro. Clin Exp Immunol. 103 289-294.
Godnic-Cvar, J., Plavec, D., Somogyi-Zalud, E., and Tudoric, N. (1999). Non-specific nasal and bronchial
reactivity are not correlated in non- asthmatic subjects occupationally exposed to irritants and in healthy
subjects. Am J Ind Med. 35 (4), 426-431.
117

Goto, H., Yuasa, K., and Rylander, R. (1994). (1-3)-b-D-glucan in indoor air, its measurement and in vitro
activity. Am J Ind Med. 25 81-83.
Gyntelberg, F., Suadicani, P., Nielsen, J. W., Skov, P. et al. (1994). Dust and the sick building syndrome.
Indoor Air. 4 223-238.
Hansen, L. A., Nexø, B. A., Borg, L., and Poulsen, O. M. (1997). Inflammatory potential of airbornr
microorganisms. Phase1: Development of a measurement method (UK title). Luftbårne mikroorganismers
evne til at udløse betændelse. Fase 1: Udvikling af en målemetode. (DK title). Arbejdsmiljøfondet,
Copenhagen. 1-36 ISBN: 87-7359-838-0 (Report in Danish, with English summary).
Hansen, Å. M., Olsen, I. L. B., Holst, E., and Poulsen, O. M. (1991). Validation of a High-Performance Liquid
Cromatography/Fluroscence Detection Method For the Simultaneous Quantification of Fifteen Polycyclic
Aromatic Hydrocarbons. Ann Occup Hyg. 35 (6), 603-611.
Heaney,P.J. & Banfiels,J.A. (1993) Structure and chemistry of silica, metal oxides and phosphates. Health
effects of mineral dusts (ed. by G. D. Guthrie & B. T. Mossman), Mineralogical society of America. pp.
185-206. ISBN: 0-939950-33-2
Hedge, A., Sterling, T. D., Sterling, E. M., Collett, C. W. et al. (1989). Indoor air quality and health in two
office buildings with different ventilation systems. Environment International. 15 115-128.
Hedges, S., Svensson, M., and Svanborg, C. (1992). Interleukin-6 Response of Epithelial Cell Lines to
Bacterial Stimulation In Vitro. Infect Immun. 60 (4), 1295-1301.
Henriksen,T. (1995) Stråling og helse, 2. edn, internetudgave, Fysisk institut, Universitetet i Oslo. pp. 1-126.
ISBN: 82-992073-2-0 (Book in Norwegian).
Hines, A. L., Ghosh, T. K., Loyalka, S. K., and Warder, R. C. Jr. (1993). Respirable particulates. in: Indoor Air.
Quality and Control. PTR Prentice Hall, Englewood Cliffs, New Jersey 07632. 116-142. ISBN:0-13-
463977-4.
Hirvonen, M.R., Nevalainen, A., Makkonen, N., Mönkkönen, J., and Savolainen, K. (1997a). Streptomyces
spores from moldy houses induce nitric oxide, TNF-alfa and IL-6 secretion from RAW264.7 macrophage
cell line without causing subsequent cell death. Environmental Toxicology and Pharmacology. 3 57-63.
Hirvonen, M. R., Nevalainen, A., Makkonen, N., Monkkonen, J., and Savolainen, K. (1997b). Induced
production of nitric oxide, tumor necrosis factor, and interleukin-6 in RAW 264.7 macrophages by
streptomycetes from indoor air of moldy houses. Arch Environ Health. 52 (6), 426-432.
Hirvonen, M. R., Ruotsalainen, M., Savolainen, K., and Nevalainen, A. (1997). Effect of viability of
actinomycete spores on their ability to stimulate production of nitric oxide and reactive oxygen species in
RAW264.7 macrophages. Toxicology. 124 (2), 105-114.
Hjort, E. L., Nielsen, H. B., and Thyregod, P. (1992). DSIF 66 ISO/DIS 8258, Shehart kontrolkort. in: SPC
Statistisk processtyring. Kontrolkort og proceskapabilitet. Dansk Standard, Holstebro. 113-164. ISBN:87-
7310-143-5. (Book in Danish)
Hodgson, M. (1995). The Sick-building syndrome. in: Effects of the Indoor Environment on Health. Seltzer, J.
M. (eds). Hanley & Belfus, Inc., Philadelphia. 167-174.
Hodgson, M. J. (1991). Indoor air quality. Allergy Proc. 12 (6), 371-383.
Hoover, G. E. and Platts-Mills, T. A. E. (1995). What the Pulmonologist Needs to Know About Allergy.
Clinics in Chest Medicine. 16 (4), 603-620.
Hyvärinen, A., Reponen, T., Husman, K., Ruuskanen, J., and Nevalainen, A. (1993). Characterizing mold
problem buildings - Concentrations and Flora of viable fungi. Indoor Air. 3 337-343.
118

Høst, A. (1997). Development of atopy in childhood. Allergy, 52, 695-697.
Ingalls, R. R. and Golenbock, D. T. (1995). CD11c/CD18, A Transmembrane Signaling Receptor for
Lipopolysaccharide. J Exp Med. 181 1473-1479.
Jaakkola, J. J. K., Heinonen, O. P., and Seppänen, O. (1989). Sick building syndrome, sensation of dryness and
termal comfort in relation to room temperature in an office building: Need for intividual control of
temperature. Environment International. 15 163-168.
Jacobs, R. R. (1997). Analysis of Endotoxins. Int J Occup Environ Health. 3 (1), S42-S48.
Johansen, N., Heinig, J. H., and Mosbech, H. (1990). A New Equipment for Standardized Dust Sampling.
XVII . Nordic Congress of Allergology, Aarhus, Denmark. (Conference proceeding)
Johnston, S. L., Papi, A., Monick, M. M., and Hunninghake, G. W. (1997). Rhinoviruses induce interleukin-8
mRNA and protein production in human monocytes. Journal of Infectious Diseases. 175 (2), 323-329.
Kanerva, L., Estlander, T., Jolanki, R., and Tarvainen, K. (1993). Occupational allergic contact dermatitis
caused by exposure to acrylates during work with dental prostheses. Contact Dermatitis. 28 268-275.
Kauffman, H. F., Tomee, J. F. C., Van de Riet, M. A., Timmerman, A. J. B. and Borger, P. (2000). Proteasedependent
activation of epithelial cells by fungal allergens leads to morphologic changes and cytokine
production. J. Allergy. Clin. Immunol. 105 1185-1193.
Keman, S., Jetten, M., Douwes, J., and Borm, P. J. A. (1998). Longitudinal changes in inflammatory markers in
nasal lavage of cotton workers. Int Arch Occup Environ Health. 71 131-137.
Khair, O. A., Devalia, J. L., Abdelaziz, M. M., Sapsford, R. J., Tarraf, H., and Davies, R. J. (1994). Effect of
Haemophilus influenzae on the synthesis of IL-6, IL-8, TNF-alfa and expression of ICAM-1 in cultured
human bronchial epithelial cells. Eur Respir J. 7 2109-2116.
Kildesø, J., Tornvig, L., Skov, P., and Schneider, T. (1998). An Intervention Study of the Effect of Improved
Cleaning Methods on the Concentration and Composition of Dust. Indoor Air. 8 12-22.
Kitchens, R. L., Wang, P., and Munford, R. S. (15-11-1998). Bacterial lipopolysaccharide can enter monocytes
via two CD14-dependent pathways. J Immunol. 161 (10), 5534-5545.
Klein, J. (1991). Lymphokiens and their receptors. in: Immunology. Blackvell Scientific Publications, Inc.,
Cambridge, Massachusetts. 227-245. ISBN:0-86542-151-X.
Koren, H. S., Graham, D. S., and Devlin, R. B. (1992). Exposure of humans to a volatile organic mixture. III. -
Inflammatory response. Arch Environ Health. 47 (1), 39-44.
Koskinen, O., Husman, T., Hyvärinen, A., Reponen, T. et al. (1994). Respiratory symptoms and infections in
mouldy dwellings and day-care centres. in: Health implications of fungi in indoor environments. Samson,
R. A., Flannigan, B., Flannigan, M. E., Verhoeff, A. P. et al. (eds). Elsevier Science B.V., Amsterdam, The
Netherlands. 211-214. ISBN:0-444-81997-5.
Kunkel, S. L., Standiford, T. J., Kasahara, K., and Strieter, R. M. (1991). Interleukin-8 (IL-8): The Major
Neutrophil Chemotactic Factor in the Lung. Exp Lung Res. 17 17-23.
Kuschner, W. G., D'Alessandro, A., Hambleton, J., and Blanc, P. D. (1998). Tumor necrosis factor-alpha and
interleukin-8 release from U937 human mononuclear cells exposed to zinc oxide in vitro. Mechanistic
implications for metal fume fever. J Occup Environ Med. 40 (5), 454-459.
Lacey, J. (1997). Actinomycetes in Compost. Ann Agric Environ Med. 4 113-121.
119

Lahtinen, M., Huuhtanen, P, and Reijula, K. (1998). Sick building syndrome and psychosocial factors - a
literature review. Indoor Air. Suppl. 4 71-80.
Larsson, B-M., Larsson, K., Malmberg, P., and Palmberg, L. (1999a). Gram positive bacteria induce IL-6 and
IL-8 production in human alveolar macrophages and epithelial cells. Inflammation. 23 (3), 217-230.
Larsson, B-M., Larsson, K., Malmberg, P., Martensson, L., and Palmberg, L. (1999b). Airway responses in
naive subjects to exposure in poultry houses: comparison between cage rearing system and alternative
rearing system for laying hens. Am J Ind Med. 35 (2), 142-149.
Larsson, B-M., Palmberg, L., Malmberg, P. O., and Larsson, K. (1997). Effect of exposure to swine dust on
levels of IL-8 in airway lavage fluid. Thorax. 52 (7), 638-642.
Lemus, R., Abdelghani, A. A., Akers, T. G., and Horner, W. E. (1996). Health risks from exposure to metals in
household dusts. Rev Environ Health. 11 (4), 179-189.
Levin, J. (1985). The History of the Development of the Limulus Amebocyte Lysate Test. Prog Clin Biol Res.
189 3-28.
Li, C-S., HSU, C-W., and Tai, M-L. (1997). Indoor Pollution and Sick Building Syndrome Symptomes among
Workers in Day-Care Centers. Arch Environ Health. 52 (3), 200-207.
Lieber, M., Smith, B., Szakal, A., Nelson-Rees, W., and Todaro, G. (15-1-1976). A continuous tumor-cell line
from a human lung carcinoma with properties of type II alveolar epithelial cells. Int J Cancer. 17 (1), 62-
70.
Long, G. L. and Winefordner, J. D. (1983). Limit of detection. A closer look at the IUPAC definition. Analyt
Chem. 55 (7), 712A-724A.
Loré, K., Sönnerborg, A., Spetz, A-L., Andersson, U., and Andersson, J. (1998). Immunocytochemical
detection of cytokines and chemokines in Langerhans cells and in vitro derived dendritic cells. J Immunol
Methods. 214 97-111.
Magnus, O. (1555). Historia De Gentibvs Septentrionalibvs. De Viottis, J. M. (eds). Rome. (Reprinted by
Rosenkilde and Bagger International Book sellers and Publidhers, Copenhagen 1972)
Martelletti, P., Stirparo, G., and Giacovazzo, M. (1999). Proinflammatory cytokines in cervicogenic headache.
Funct Neurol. 14 (3), 159-162.
Matsushima, K., Baldwin, E. T., and Mukaida, N. (1992). Interleukin-8 and MCAF: Novel Leukocyte
Recruitment and Activating Cytokines. in: Interleukines. Molecular Biology and Immunolgy. Kishimoto, T.
(eds). Chem Immunol, Basel, Karger. 236-265.
Menzies, D., Tamblyn, R. M., Nunes, F., Hanley, J., and Tamblyn, R. T. (1996). Exposure to Varying Levels of
Contaminants and Symptoms among Workers in Two Office Buildings. Am J Public Health. 86 (11), 1629-
1633.
Meyer, H. W. (2000). Skoleundersøglesen i København. Et indeklimastudie. Arbejds- og Miljømedicinsk
Klinik. (Ph.D.-Thesis in Danish)
Meyer, H. W., Nielsen, P., Hansen, M. Ø., Nielsen, P. A. et al. (1996). Skoleundersøgelsen i København I.
(Report in Danish)
Michel, O. (1997). Human Challenge Studies with Endotoxins. Int J Occup Environ Health. 3 (1), S18-S25.
Miller, E. J., Nagao, S., Carr, F. K., Noble, J. M., and Cohen, A. B. (1996). Interleukin-8 (IL-8) is a major
neutrophil chemotaxin from human alveolar macrophages stimulated with staphylococcal enterotoxin A
(SEA). Inflamm Res. 45 (8), 386-392.
120

Monn, C. and Becker, S. (15-3-1999). Cytotoxicity and induction of proinflammatory cytokines from human
monocytes exposed to fine (PM2.5) and coarse particles (PM10-2.5) in outdoor and indoor air. Toxicol Appl
Pharmacol. 155 (3), 245-252.
Morrison, D. C., Lei, M-G., Kirikae, T., and Chen, T-Y. (1993). Endotoxin Receptors on Mammalian Cells.
Immunobiology. 187 212-226.
Mullol, J., Xaubet, A., Gaya, A., Roca-Ferrer, J., Lopez, E., Fernandez, J. C., Fernandez, M. D., and Picado, C.
(1995). Cytokine gene expression and release from epithelial cells. A comparison study between healthy
nasal mucosa and nasal polyps. Clinical and Experimental Allergy. 25 607-615.
Munir, A. K. (1995). Environmental factors influencing the levels of indoor allergens. Pediatr. Allergy
Immunol. 6 (Suppl 7), 13-21.
Munir, A. and Björkstén, B. (1992). Indoor pollution and allergic sensitization. in: Chemical, microbiological,
health and comfort aspects of indoor air quality - State of the art in SBS. Knöpel, H. and Wolkoff, P. (eds).
Kluwer Academic Publishers, Dordrecht. 181-199. ISBN:0-7923-1703-3.
Munir, A. K., Einarsson, R., and Dreborg, S. K. (1995). Mite (Der p I, Der f I), cat (Fel d I) and dog (Can f I)
allergens in dust from Swedish day-care centres. Clin Exp Allergy. 25 (2), 119-126.
Mürer, A. J. L. (1996). Hudproblemer blandt tandteknikere. Arbejdsmiljøinstituttet, København. (Ph.D.-Thesis
in Danish)
Mølhave, L., Kjærgaard, S. K., Atterman, J., and Pedersen, O. F. (1995). Husstøv og indeklima - Et
klimakammerforsøg med menneskers reaktioner på luftbårent husstøv. AMF-Projekt 1990-71. Institut for
Miljø- og Arbejdsmedicin, Aarhus Universitet, Aarhus. (Report in Danish)
Nakamura, H., Yoshimura, K., Jaffe, H. A., and Crystal, R. G. (1991). Interleukin-8 Gene Expression in
Human Bronchial Epithelial Cells. The Journal of Biological Chemistry. 266 (29), 19611-19617.
Nakamura, H., Yoshimura, K., McElvaney, N. G., and Crystal, R. G. (1992). Neutrophil Elastase in
Respiratory Epithelial Lining Fluid of Individuals with Cystic Fibrosis Induces Interleukin-8 Gene
Expression in a Human Bronchial Epithelial Cell Line. J Clin Invest. 98 1478-1484.
Nevalainen, A. and Jantunen, M J. (1988). The indoor air quality of Finnish homes with mold problems.
Planing, physics and climate technology for helathier buildings. Healthy Buildings 88'. 2 , 319-323.
Stockholm, Sweden. (Conference Proceeding)
Nevalainen, A., Kotimaa, M., Pasanen, A-L., Pellikka, M. et al. (1990). Mesophilic actinomycetes - The real
Indoor air problem? The fifth international conference on indoor air quality and climate. Indoor Air. 90 ,
203-206. Toronto, Canada. (Conference Proceeding)
Nielsen, B. H. (1998). Mikroorganismer i bioaffald. Arbejdsmiljømæssige aspekter ved håndtering af "grønt"
husholdningsaffald og kompost. Sikkerhed og sundhed ved affald og genanvendelse. Report no. 16.
Arbejdsmiljøinstituttet, København. (Ph.D-Thesis in Danish)
Nielsen, G. D., Alarie, Y., Poulsen, O. M., and Nexø, B. A. (1995). Possible mechanisms for the respiratory
tract effects of noncarcinogenic indoor-climate pollutants and bases for their risk assessment. Scand J work
Environ Health. 21 165-178.
Norbäck, D., Bjornsson, E., Janson, C., Widstrom, J., and Boman, G. (1995). Asthmatic symptoms and volatile
organic compounds, formaldehyde, and carbon dioxide in dwellings. Occup Environ Med. 52 (6), 388-395.
Norbäck, D., Edling, C., and Wieslander, G. (1994). Asthma symptoms and the sick building syndrome - The
significance of microorganisms in the indoor environment. in: Health implications of fungi in indoor
environments. Samson, R. A., Flannigan, B., Flannigan, M. E., Verhoeff, A. P. et al. (eds). Elsevier Science
B.V., Amsterdam, The Netherlands. 229-239. ISBN:0-444-81997-5.
121

Norbäck, D., Torgen, M., and Edling, C. (1990). Volatile organic compounds, respirable dust, and personal
factors related to prevalence and incidence of sick building syndrome in primary schools. Br J Ind Med. 47
(11), 733-741.
Norbäck, D. and Torgén, M. (1989). A longitudinal study relating carpeting with sick building syndrome.
Environment International. 15 129-135.
Nowotny, A. (1990). Section I: Structure and function studies. Introduction. in: Cellular and molecular aspects
of endotoxin reactions. Nowotny, A., Spitzer, J. J., and Ziegler, E. J. (eds). Elsevier Science Publishers
B.V., 3-14.
Ohman, P. A. and Eberly, L. E. (1998). Relating sick building symptoms to environmental conditions and
worker charateristics. Indoor Air. 8 172-179.
Ott, W. R and Roberts, J. W. (1998). Everyday Exposure to Toxic Pollutants. Environmental regulations have
improved the quality of outdoor air. But problems that persists indoors have received too little attention.
Scientific American. Feburary 86-91.
Owen, M. K. and Ensor, D. S. (1992). Airborne particle sizes and sources found in indoor air. Atmospheric
Environment. 26A (12), 2149-2162.
Palmberg, L., Larsson, B-M., Malmberg, P., and Larsson, K. (1998). Induction of IL-8 production i human
alveolar macrophages and human bronchial epithelial cells in vitro by swine dust. Thorax. 53 (4), 260-264.
Pearson, F. C., Weary, M. E., Bohon, J., and Dabbah, R. (1982). Relative Potency of "Environmental"
Endotoxin as Measured by the Limulus Amebocyte Lysate Test and the USP Rabbit Pyrogen Test. Progress
in Clinical and Biological Research. 93 65-77.
Peden, D. B. (1996). The Use of Nasal Lavage for Objective Measurement of Irritant-Induced Nasal
Inflammation. Regul Toxicol Pharmacol. 24 S76-S78.
Poulsen, O. M., Breum, N. O., Ebbehøj, N., Hansen, M. et al. (1995a). Collection of domestic waste. Review of
occupational health problems and their possible causes. Science of the Total Environment. 170 1-19.
Poulsen, O. M., Breum, N. O., Ebbehøj, N., Hansen, M. et al. (1995b). Sorting and recycling of domestic
waste. Review of occupational helath problems and their possible causes. Science of the Total Environment.
168 33-56.
Poulsen,O.M., Holst,E., & Christensen,J.M. (1993) Metodeevaluering og kvalitetsstyring af kemiske
måleprocesser, Arbejdstilsynet, Arbejdsmiljøinstituttet, København. pp. 1-135.. ISBN: 87-7534-440-8
(Book in Danish)
Poussin, C., Foti, M., Carpentier, J. L., and Pugin, J. (7-8-1998). CD14-dependent endotoxin internalization via
a macropinocytic pathway. J Biol Chem. 273 (32), 20285-20291.
Pugin, J., Schürer-Maly, C-C., Leturcq, D., Moriarty, A. et al. (1993). Lipopolysaccharide activation of human
endothelial and epithelial cells is mediated by lipopolysaccharide-binding protein and soluble CD14. Proc
Natl Acad Sci USA. 90 2744-2748.
Quesniaux, V. F. J. and Jones, T. C. (1998). Granulocyte-macrophage colony stimulating factor. in: The
cytokine handbook. Thomson, A. (eds). Academic Press, San Diego. 637-669. ISBN:0-12-689662-3.
Rautiala, S., Reponen, T., Hyrvärinen, A., Nevalainen, A. et al. (1996). Exposure to airborne microbes during
the repair of moldy buildings. American Industrial Hygiene Association Journal. 57 279-284.
Redlich, C. A., Sparer, J., and Cullen, M. R. (1997). Sick-building syndrome. Lancet. 349 1013-1016.
Rietschel, E. T. and Brade, H. (1992). Bacterial Endotoxins. Scientific American. August 26-33.
122

Rietschel, E. T., Brade, H., Holst, O., Brade, L. et al. (1996). Bacterial Endotoxin: Chemical Constitution,
Biological Recognition, Host Response, and Immunological Detoxification. in: Pathology of Septic Shock.
Rietschel, E. T. and Wagner, H. (eds). CTMI, 40-81. ISBN:3-540-61026-X.
Rietschel, E. T., Kirikae, T., Schade, F. U, Ulmer, A. J. et al. (1993). The Chemical Structure of Bacterial
Endotoxin in Relation to Bioactivity. Immunobiology. 187 169-190.
Robinson, P. A., Tauxe, R. V., Winkler, W. G., and Levy, M. E. (1983). Respiratory illness in conference
participants following exposure to rug shampoo. Infect Control. 4 (3), 158-160.
Roepstorff, V. and Sigsgaard, T. (1997). Cytotoxic Effect of Organic Dust Extract from Different Working
Environments: An In Vitro Assay. Ann Agric Environ Med. 4 195-201.
Ruotsalainen, M., Hirvonen, M.-R., Hyvärinen, A., Meklin, T. et al. (1998). Cytotoxicity, production of
reactive oxygen species and cytokines induced by different strains of Stachybotrys sp. from moldy
buildings in RAW264.7 macrophages. Environmental Toxicology and Pharmacology. 6 193-199.
Rylander, R. (1987). The Role of Endotoxin for Reactions After Exposure to Cotton Dust. Am J Ind Med. 12
687-697.
Rylander, R. (1997a). Evaluation of the Risks of Endotoxin Exposure. Int J Occup Environ Health. 3 (1), S32-
S36.
Rylander, R. (1997b). Airborne (1-3)-beta-D-Glucan and Airway Disease in a Day-Care Center before and after
Renovation. Arch Environ Health. 52 (4), 281-285.
Rylander, R. (1998). Microbial cell wall constituents in indoor air and their relation to disease. Indoor Air.
Suppl. 4 59-65.
Rylander, R., Bake, B., Fischer, J. J., and Helander, I. M. (1989). Pulmonary Function and Symptoms after
Inhalation of Endotoxin. Am Rev Respir Dis. 140 981-986.
Rylander, R., Persson, K., Goto, H., Yuasa, K., and Tanaka, S. (1992). Airborne Beta-1,3-Glucan May Be
Related to Symptoms in Sick Buildings. Indoor Environ. 1 263-267.
Rylander, R., Williams, D. L., McWilliams, A., Yuasa, K. et al. (1993). Second Glucan Inhalation Toxicity
Workshop. Report no. 1. Rylander, R. and Peterson, Y. (eds). Committee on Organic Dusts, ICOH, 1-53.
Salvaggio, J. E. (1994). Psychological aspects of "environmental illness", "multiple chemical sensitivity", and
building-related illness. J Allergy Clin Immunol. 94 (2), 366-370.
Saraf, A., Larsson, L., Larsson, B-M., Larsson, K., and Palmberg, L. (1999). House dust induces IL-6 and IL-8
response in A549 epithelial cells. Int J Indoor Air Qual Clim. (in Press)
Saraf, A., Park, J-H., Milton, D K., and Larsson, L. (1999). Use of quadrupole GC-MS and ion trap GC-MS-
MS for determining 3-hydroxy fatty acids in settled house dust: relation to endotoxin activity. J Environ
Monit. 1 163-168.
Schata, M., Jorde, W., Elixmann, J. H., and Linskens, H. F. (1989). Allergies to molds caused by fungal spores
in air conditioning equipment. Environment International. 15 177-179.
Schins, R. P., McAlinden, A., MacNee, W., Jimenez, L. A., Ross, J. A., Guy, K., Faux, S. P., and Donaldson,
K. (2000). Persistent depletion of I kappa B alpha and interleukin-8 expression in human pulmonary
epithelial cells exposed to quartz particles. Toxicology and Applied Pharmacology, 167, 107-117.
Schürer-Maly, C-C., Eckmann, L., Kagnoff, M. F., Falco, M. T., and Maly, F.-E. (1994). Colonic epithelial cell
lines as a source of interleukin-8: stimulation by inflammatory cytokines and bacterial lipopolysaccharide.
Immunology. 81 85-91.
123

Seltzer, J. M. (1995). Biological contaminants. in: Effects of the Indoor Environment on Health. Seltzer, J. M.
(eds). Hanley & Belfus, Inc., 1-26.
Sigsgaard, T., Abell, A., Jensen, L. D., and Nielsen, B. H. (1994). Work-Related Symptoms and Lung Function
Measurements in Paper Mill Workers Exposed to Recycled Water. Occup Hyg. 1 177-189.
Skov, P. (1992). The Sick Building Syndrome. in: Sources of Indoor air Contaminants: Characterizing
Emissions and Health Impacts. Tucker, W. G., Leaderer, B. P., Molhave, L., and Cain, W. S. (eds). The
New York Academy of Sciences, New York. 17-20.
Skov, P., Valbjørn, O., and Pedersen, B. V. (1989a). Influence of personal characteristics, job-related factors
and psychosocial factors on the sick building syndrome. Danish Indoor Climate Study Group. Scand J work
Environ Health. 15 (4), 286-295.
Skov, P., Valbjørn, O., Gyntelberg, F., and DISG. (1989b). Rådhusundersøgelsen. Indeklima i kontorer.
Arbejdsmiljøfondet, København. (Report in Danish)
Smith, J. E., Anderson, J. G., Lewis, C. W., and Murad, Y. M. (1992). Cytotoxic fungal spores in the indoor
atmosphere of the damp domestic environment. FEMS Microbiol Lett. 79 (1-3), 337-343.
Sorenson, W. G., Shahan, T. A., and Lewis, D. M. (1994). Activation of alveolar macrophages by conidia of
common fungi associated with organic dust toxic syndrome. in: Health implications of fungi in indoor
environments. Samson, R. A., Flannigan, B., Flannigan, M. E., Verhoeff, A. P. et al. (eds). Elsevier Science
B.V., Amsterdam, The Netherlands. 325-358. ISBN:0-444-81997-5.
Spriggs D.R., Imamura, K., Rodriguez, C., Sariban, E., and Kufe, D.W. (1988). Tumor necrosis factor
expression in human epithelial tumor cell lines. J Clin Invest. 81 455-460.
Stacey M.A., Sun, G., Vassalli, G., Marini, M., Bellini, A., and Mattoli, S. (1997). The allergen Der p1 induces
NF-kappaB activation through interference with IkappaB alpha function in asthmatic bronchial epithelial
cells. Biochem Biophys Res Commun. 236 (2), 522-526.
Standiford, T. J., Kunkel, S. L., Basha, M. A., Chensue, S. W., Lynch III, J. P., Toews, G. B., Westwick, J., and
Strieter, R. M. (1990). Interleukin-8 Gene Expression by a Pulmonary Epithelial Cell Line. A Model for
Cytokine Networks in the Lung. J Clin Invest. 86 1945-1953.
Strieter, R. M., Chensue, S. W., Basha, M. A., Standiford, T. J. et al. (1990). Human Alveolar Macrophage
Gene Expression of Interleukin-8 by Tumor Necrosis Factor-alfa, Lipopolysaccharide, and Interleukin-
1beta. Am J Respir Cell Mol Biol. 2 321-326.
Strieter, R. M. and Kunkel, S. L. (1994). Acute Lung Injury: The Role of Cytokines in the Eliction of
Neutrophils. J Invest Med. 42 (4), 640-651.
Taimi, M., Defacque, H., Commes, T., Favero, J. et al. (1993). Effect of retinoic acid and vitamin D on the
expression of interleukin-1beta, tumour necrosis factor-alfa and interleukin-6 in the human monocytic cell
line U937. Immunology. 79 229-235.
Teeuw, K. B., Vandenbroucke-Grauls, C. M. J. E., and Verhoef, J. (1994). Airborne Gram-negative Bacteria
end Endotoxin in Sick Building Syndrome. A Study in Dutch Governmental Office Buildings. Arch Intern
Med. 154 2339-2345.
Tomee J.F., van Weissenbruch, R., de Monchy, J.G., and Kauffman, H.F. (1998). Interactions between inhalant
allergen extracts and airway epithelial cells: effect on cytokine production and cell detachment. J Allergy
Clin Immunol. 102 (1), 75-85.
Thriene, B., Sobottka, A., Willer, H., and Weidhase, J. (1996). Man-made mineral fibre boards in buildings -
health risks caused by quality deficiencies. Toxicology Letters. 88 (1-3), 299-303.
124

Thörn, Å., Lewné, M., and Belin, L. (1996). Allergic alveolitis in a school environment. Scand J work Environ
Health. 22 (4), 311-314.
Tobias, P. S., Soldau, K., Kline, L., Lee, D.-J., Kato, K., Martin, T. P., and Ulevitch, R. J. (1993). Cross-
Linking of Lipopolysaccharide (LPS) to CD14 on THP-1 Cells Mediated by LPS-Binding Protein. J
Immunol. 150,7 3011-3021.
Tracey, K. J. and Cerami, A. (1993). Tumor necrosis factor, other cytokines and disease. Annu Rev Cell Biol. 9
317-343.
Ulevitch, R. J. and Tobias, P. S. (1995). Receptor-dependent mechanisms of cell stimulation by bacterial
endotoxin. Ann Rev Immunol. 13 437-457.
Ulmer, A. J. (1997). Biochemistry and Cell Biology of Endotoxins. Int J Occup Environ Health. 3 (1), S8-S17.
Ulrik, C. S., Backer, V., Hesse, B., and Dirksen, A. (1998). Risikofaktorer for udvikling af astma hos børn og
unge. Resultater fra en longitudinel populationsundersøgelse. Ugeskr Læger, 160, 2243-2248. Article in
Danish with English summery.
Utell M., Samet, J. (1996). Airborne particles and respiratory disease: Clinical and pathogenic considerations.
(cap. 8), In: Particles in our air. Concentration and health effects. 169-188. Harvard University Press.
ISBN: 0-674-24077-4.
Van der Walle, H. B., Klecak, G., Geleick, H., and Bensink, T. (1982). Sensitizing potential of 14 mono (meth)
acrylates in the guinea pig. Contact Dermatitis. 8 223-235.
Vanhee, D., Gosset, P., Boitelle, A., Wallaert, B., and Tonnel, A. B. (1995). Cytokines and cytokine network in
silicosis and coal workers' pneumoconiosis. Eur Respir J. 8 (5), 834-842.
Verhaegen, S., Verschueren, H., Dekegel, D., and de Baetselier, P. (1989). The in vitro invasive behavior of
human myelomonocytic cell lines is modulated by HuIFN-gamma and pertussis toxin. J Biol Response
Mod. 8 (1), 97-107.
Veronesi, B., Oortgiesen, M., Carter, J. D., and Devlin, R. B. (1999). Particulate matter initiates inflammatory
cytokine release by activation of capsaicin and acid receptors in a human bronchial epithelial cell line.
Toxicol Appl Pharmacol. 154 106-115.
Vowels, B. R., Yang, S., and Leyden, J. J. (1995). Induction of proinflammatory cytokines by a soluble factor
of Propionibacterium acnes: implications for chronic inflammatory acne. Infect Immun. 63 (8), 3158-3165.
Waegemaekers, M., Van Wageningen, N., Brunekreef, B., and Boleij, J. S. (1989). Respiratory symptoms in
damp homes. A pilot study. Allergy. 44 (3), 192-198.
Wang, Z., Malmberg, P., Larsson, B. M., Larsson, K., Larsson, L., and Saraf, A. (1996). Exposure to bacteria
in swine-house dust and acute inflammatory reactions in humans. Am J Respir Crit Care Med. 154 (5),
1261-1266.
Wang, Z. (1997). Acute Cytokine Responses to Inhaled Swine Confinement Building Dust. Arbete Och Hälsa.
23 1-60.
Weary, M. E., Donohue, G., Pearson, F. C., and Story, K. (1980). Relative Potencies of Four Reference
Endotoxin Standards as Measured By the Limulus Amoebocyte Lysate and USP Rabbit Pyrogen Test. Appl
Environ Microbiol. 40 (6), 1148-1151.
Whicher, J. T. and Evans, S. W. (1990). Cytokines in Disease. Clinical Chemistry. 36 (7), 1269-1281.
WHO. (1983). Indoor air pollutants: Exposure and health effects. EURO Reports and Studies. Report no. 78.
WHO, Denmark.
125

WHO. (1989a). Indoor air quality: organic pollutants. EURO reports and studies. Report no. 111. WHO,
Denmark.
WHO. (1989b). Environmental Heath Criteria 89: formaldehyde. WHO, Geneva.
Wilkins C.K., Larsen, S.T., Hammer, M., Poulsen, O.M., Wolkoff, P., and Nielsen, G.D. (1998). Respiratory
effects in mice exposed to airborne emissions from Stachybotrys chartarum and implications for risk
assessment. Pharmacology & Toxicology. 83 112-119.
Wolkoff, P. and Kjærgaard, S. K. (1992). Indeklimasyndromet i kontormiljøet. State-of-the-art oversigt.
Bygge- og Boligstyrelsen, Copenhagen. 1-43. (Report in Danish).
Wuyts, A., Proost, P., and Van Damme, J. (1998). Interleukin-8 and other CXC chemokines. in: The Cytokine
Handbook. Thomson, A. (eds). Academic Press, San Diego. 271-311. ISBN:0-12-689662-3.
Würtz, H., Kildesø, J., Meyer, H. W., and Nielsen, J. B. (1999). A pilot study on airborne microorganisms in
danish classrooms. pp. 316-320. Indoor Air 99', Edinburg, Skotland. (Conference proceeding).
Wålinder, R., Norback, D., Wieslander, G., Smedje, G., and Erwall, C. (1997a). Nasal congestion in relation to
low air exchange rate in schools. Evaluation by acoustic rhinometry. Acta Otolaryngol. 117 (5), 724-727.
Wålinder, R., Norbäck, D., Wieslander, G., Smedje, G., and Erwall, C. (1997b). Nasal Mucosal Swelling in
Relation to Low Air Exchange Rate in Schools. Indoor Air. 7 198-205.
Zhang, M. and Tracey, K. J. (1998). Tumor Necrosis Factor. in: The cytokine handbook. Thomson, A. (eds).
Academic Press, San Diego. 517-548. ISBN:0-12-689662-3.
Zhang, Y., Doerfler, M., Lee, T. C., Guillemin, B., and Rom, W. N. (1993). Mechanisms of Stimulation of
Interleukin-1beta and Tumor Necrosis Factor-alfa by Mycobacterium tuberculosis Components. J Clin
Invest. 91 2076-2083.
126

10. Appendix for protocols and raw data (in Danish)
127

11. Papers
128