Occupational and Environmental Medicine 1996;53:387-393

Wool and grain dusts stimulate TNF secretion byalveolar macrophages in vitro

D M Brown, K Donaldson

AbstractObjective-The aim of the study was toinvestigate the ability of two organicdusts, wool and grain, and their solubleleachates to stimulate secretion oftumournecrosis factor (TNF) by rat alveolarmacrophages with special reference to therole oflipopolysaccharide (LPS).Methods-Rat alveolar macrophageswere isolated by bronchoalveolar lavage(BAL) and treated in vitro with wholedust, dust leachates, and a standard LPSpreparation. TNF production was mea-sured in supernatants with the L929 cellline bioassay .Results-Both wool and grain dust sam-ples were capable of stimulating TNFrelease from rat alveolar macrophages ina dose-dependent manner. The standardLPS preparation caused a dose-depen-dent secretion of TNF. Leachates pre-pared from the dusts contained LPS andalso caused TNF release but leachableLPS could not account for the TNFrelease and it was clear that non-LPSleachable activity was present in the graindust and that wool dust particles them-selves were capable of causing release ofTNF. The role of LPS in wool dustleachates was further investigated bytreating peritoneal macrophages fromtwo strains of mice, LPS responders(C3H) and LPS non-responders(C3H/HEJ), with LPS. The non-respon-der mouse macrophages produced verylow concentrations of TNF in response tothe wool dust leachates compared withthe responders.Conclusions-LPS and other unidentifiedleachable substances present on the sur-face of grain dust, and to a lesser extenton wool dust, are a trigger for TNFrelease by lung macrophages. Wool dustparticles themselves stimulate TNF. TNFrelease from macrophages could con-tribute to enhancement of inflammatoryresponses and symptoms of bronchitisand breathlessness in workers exposed toorganic dusts such as wool and grain.

(Occup Environ Med 1996;53:387-393)

Keywords: organic dust; macrophage; tumour necrosisfactor; endotoxin; bronchitis

Tumour necrosis factor (TNF) is a proinflam-matory cytokine produced by macrophages ormonocytic phagocytes in response to various

stimuli.'-3 In the non-stimulated state, alveolarmacrophages do not secrete TNF. However,triggers of macrophage activation includingthe phagocytosis of organic and inorganicdusts, phorbol esters and bacterial lipopolysac-charide (endotoxin) (LPS) stimulate secretionofTNF both in vivo and in vitro.45 The effectsofTNF on other cells are wide ranging, and aswell as causing differentiation of myeloid celllines and growth of B-lymphocytes, TNF alsostimulates mitogenesis of lung fibroblasts.' 2 6An important proinflammatory action ofTNFand other cytokines is the activation of celladhesion molecules, such as the leukocyteintegrins (CD 1 1 a b c/CD 18), and intercellularadhesion molecule-1 (ICAM-1).78 Endothelialcell adhesion molecules are also activated byTNF.910 The increased adhesivity of theendothelial cell adhesion molecules enhancesneutrophil adherence, an essential first step inthe migration of leucocytes from the micro-vasculature into the tissues to sites of inflam-mation.

Organic dusts include a range of airbornematerial comprising plant components, pol-lens, and spores. Dust which is deposited inthe alveolar region of the lung may cause acti-vation of complement, which is present in thealveolar lining fluid, with generation of leuco-cyte chemotaxins."1 These events have previ-ously been shown with extracts of organicdusts, endotoxin, and activation of serum. 12 13

In the lung after deposition of organic dust, asequence of events may be envisaged wherealveolar macrophages become activated,secrete inflammatory mediators, and act in asynergistic manner with activated complementfragments to enhance inflammation.

In a previous study, a relation was shownbetween the airborne mass concentration ofwool dust and symptoms of airway irritation inexposed workers in wool textile mills in thenorth of England.'4 1' Dusts collected from theair of wool mills were shown to be contami-nated with bacterial endotoxin and to be ableto produce an acute inflammatory responsewhen instilled into the lungs of rats'6 but werenot notably toxic to macrophages and epithe-lial cells.'7 Endotoxin has been implicated insymptoms of byssinosis in workers in a woolcarpet factory in Turkey.'8 The wool dust col-lected from the air of English wool mills wasalso capable of up regulating homotypic adhe-sion of alveolar macrophages'9 by activationof LFA-1/ICAM-1 ligand-receptor binding.Wool dust also caused granulomatous lesionsafter instillation into the lungs of rats.20 In thisstudy, we have examined the ability of

Institute ofOccupationalMedicine, CityHospital, GreenbankDrive, Edinburgh,ScotlandD M BrownK DonaldsonNapier University,Department ofBiological Sciences,10 Colinton Road,Edinburgh, ScotlandD M BrownK DonaldsonCorrespondence to:Dr D M Brown, NapierUniversity, Department ofBiological Sciences,10 Colinton Road,Edinburgh EH10 5DT, UK.Accepted 16 January 1996

387

on 17 May 2018 by guest. P

rotected by copyright.http://oem

.bmj.com

/O

ccup Environ M

ed: first published as 10.1136/oem.53.6.387 on 1 June 1996. D

ownloaded from

Brown, Donaldson

inspirable wool dusts, sieved grain dust, andleachates from the dusts to produce TNFsecretion by rat alveolar macrophages in vitro.

Materials and methodsANIMALSMale SPF HAN rats (Charles River, Margate,UK) were used throughout.

COLLECTION OF WOOL DUSTSDusts were collected from two wool processmills in the north of England designated S(start) and M (middle) which representedopening-blending and combing processesrespectively. A series of six Institute ofOccupational Medicine static inspirable dustsamplers,2' were placed at each site in the dusti-est zones. Samplers were operated for a fullwork shift and an inspirable sample of dust wascollected on Gelman GLA filters witha 5 jin pore size (Gelman Hawksley,Northampton). Dust was removed from filterswith a soft brush. Dust from each mill waspooled into a tube, weighed, and mechanicallyrotated for 24 hours to ensure mixing and thesesamples were stored at - 20'C until required.The wool dust used for the study wasinspirable-that is, dust which could pass intothe upper airways. This fraction of dust alsoincluded a large proportion of respirable dustwhich could penetrate to the terminal bronchi-oles and proximal lung parenchyma.

COLLECTION OF GRAIN DUSTSamples of dust were collected from the ledgesof a barn which stored wheat and barley. Thedust was sieved through a 200 pm mesh, fol-lowed by a 45,um mesh, by shaking mechani-cally for 30 minutes. The dust so obtained fromthis process and used in all subsequent assayswas less than 45 pm diameter and included asubstantial proportion of respirable dust.

PREPARATION OF WOOL AND GRAIN DUSTS ANDPREPARATION OF LEACHATESSeparate samples of leachate (materials leachedfrom organic dust into aqueous solution) wereprepared by mixing wool dust and grain dust inphosphate buffered saline (PBS) at a concen-tration of 5 mg/ml at room temperature for 24hours. Solutions were then centrifuged at 3000rpm for 15 minutes to remove large fragmentsand then finally filtered through 0-22 pm filtersto sterilise the leachates. Half of each samplewas then depleted of endotoxin by passingdown Polymyxin B columns (Flow, HighWycombe, Bucks) according to the manufac-turer's instructions.

MEASUREMENTS OF THE ENDOTOXIN CONTENTOF WOOL AND GRAIN DUST LEACHATESThe "Coatest" (ICN, Flow, High Wycombe,Bucks) Limulus amoebocyte lysis (LAL) basedspectrophotometric assay was used to assess theendotoxin content of the wool and grainleachates.

BRONCHOALVEOLAR LAVAGE

Animals were killed with a single intraperi-

toneal injection ofNembutal, the thoracic cavitywas opened and the lungs were cannulated andremoved. The lungs were then sequentiallyravaged with four 8 ml aliquots of saline at370C and pooled into a single tube. Cells werecentrifuged at 1000 rpm in a refrigerated benchcentrifuge for 10 minutes, and resuspended inF-I0 medium (Gibco, Paisley) containing 0 2%bovine serum albumin (BSA) (Fraction VSigma, Poole, Dorset) and kept on ice. Totalcell numbers were evaluated and cytocentrifugesmears, which were stained with May-Grunwald Giemsa, were prepared. One hun-dred cells per cytospin were evaluated to obtaina differential cell count. Cells were adjusted to aconcentration of 1 x 106 cells/ml in mediumcontaining the dusts at various concentrationsand 1 ml was pipetted into wells of a 24 wellplate (Greiner Labortechnik, Cam, Dursley).After incubation for one hour to allow adher-ence the dusts were added, suspended in thesame medium. Macrophages were incubatedwith dust overnight (16 hours) at 370C in 5%CO, after which the medium was removed andcentrifuged at 3000 rpm for 15 minutes toremove dust particles. The supernatants werestored at - 70°C until assay for TNF.

PREPARATION OF SUPERNATANTS FROM C3H ANDC3H/HEJ MOUSE PERITONEAL MACROPHAGESA separate group of dust leachates was pre-pared for use in the TNF assay in F-10 medium(Gibco, Paisley) containing 0-2% BSA, (FlowLabs, High Wycombe). This medium con-tained negligible amounts of LPS, as defined bythe manufacturer's assay.

Macrophages from C3H/HEJ mice are unre-sponsive to the effect of LPS,"1 and the relatedC3H mice respond normally. Peritoneal exu-date cells were treated with dust leachates toprepare supernatants, which were then assessedfor their TNF content. This allowed theinvolvement of endotoxin present in theleachates to be estimated.

Mice were treated with a single intraperi-toneal injection of 0.5 ml of 1% thioglycollatebroth. Three days later, these mice were killedwith ether, and the peritoneal cavity wasravaged with 3 x 2 ml volumes of PBS con-taining heparin at a final concentration of 100U/ml. Lavage fluid was centrifuged at 1000rpm for five minutes, resuspended in F-10+ 0-2% BSA and cells were counted. Cells wereadjusted to 1 x 106 cells/ml and 1 ml added toeach well in 24 well plates which were incu-bated at 37°C for one hour. The medium wasthen removed, adherent cells were washedtwice with PBS, and the dust leachates (pre-pared in F-10 + 0-2% BSA) added to thewells. Plates were incubated at 37°C overnight,supernatants removed, spun at 3000 rpm for 15minutes, and stored at - 70°C until required.

ENDOTOXIN DOSE RESPONSEEndotoxin from Escherichia coli serotype011:B4 (Sigma, Poole, Dorset) was made upto 1 mg/ml in PBS, and stored at - 70°C.Some experiments comprised treating alveolarmacrophages with LPS at various concentra-tions. Supernatants from leachate treated

388

on 17 May 2018 by guest. P

rotected by copyright.http://oem

.bmj.com

/O

ccup Environ M

ed: first published as 10.1136/oem.53.6.387 on 1 June 1996. D

ownloaded from

Wool and grain dusts stimulate TNF secretion by alveolar macrophages in vitro

macrophages were prepared by treating adher-ent macrophages with 1 ml of leachate andendotoxin depleted leachate.

TNF ASSAYThis assay relies on the fact that cells of themouse fibroblast cell line L929 are specificallysusceptible to the lethal effects of TNF if thecells are pretreated with the protein synthesisinhibitor actinomycin D. After treatment withTNF, the surviving cells are stained and theabsorbence of stain is taken as a quantitativemeasure of survival, from which the amount ofcell death can be calculated. The specificity ofthe assay for TNF in rat macrophage super-natants has been confirmed with specific anti-body against TNF and is calibrated withrecombinant TNF.The L929 cells were removed from continu-

ous culture with 0* 1% trypsin-EDTA (Gibco,Paisley, Scotland), and resuspended in mini-mal essential medium (Gibco, Paisley),containing 10% heat inactivated foetal calfserum, penicillin, and streptomycin (completemedium). Cells were adjusted to 3 x 1 05cells/ml in complete medium, and 100 ,ul wasadded to each well in 96 well plates (GreinerLabortechnik, Cam, Dursley), which wereincubated at 370C overnight in a humidifiedatmosphere of 5% CO2.On day 2 the medium in each well was

replaced with minimal essential medium con-taining 5% fetal calf serum, 1 yg/ml actino-mycin D (Sigma, Poole, Dorset), withoutantibiotics. The top row of each plate receivedan additional 50 ul of medium containingactinomycin D at a concentration of 2 jg/ml.The bottom row of each plate served as thecontrol wells (six wells) and the remaining setof six wells were used as blanks to set the spec-trophotometer.The top row of each plate received a final

50 p1 of test supernatant, each supernatantbeing measured in triplicate. A multichannelpipette was used to double dilute 100 yl fromthe top row of wells down the plate, endingbefore the bottom control wells. The final 100yul of medium was discarded. One set of wellscontained a TNF standard (National Institutefor Biological Standards and Control, PottersBar, Hertfordshire) at a concentration of 100U/ml. Plates were then incubated for a further24 hours.On day 3, the medium and supernatant in

each well was removed and replaced with 501 of 1% crystal violet stain in 20% methanolfor two minutes. Each plate was washed withtap water and allowed to dry, after which eachwell received 50 p1 of 20% acetic acid to makesoluble the stained cells. Any bubbles wereremoved and plates were read at 540 nm withan automatic plate reader interfaced with apersonal computer.

TNF CALCULATIONSThe mean optical density (OD) was calculatedfor each sample dilution as a percentage of theuntreated control. The sample dilutions wereconverted to log, values and regressions per-formed with the sample OD% v log, dilution.

The log, dilution value that gave an OD of50% of the control cells was calculated withthe equation:

OD% = intercept + slope x log, dilutionlog, dilution (50%) = (50-A)/B

The number of units of TNF was finallycalculated from the TNF standard, which wasincluded in each experiment. As in the samplewells, the OD, which was 50% of the OD ofthe control wells, was obtained and relatedback to the known concentration of TNF inthe standard.

STATISTICAL ANALYSISData were examined with analysis of variance.

Results(1) TNF PRODUCTION BY RAT ALVEOLARMACROPHAGES TREATED WITH WOOL ANDGRAIN DUSTS IN VITROAlveolar macrophages from untreated ratswere obtained by BAL and treated in vitrowith whole dust, dust leachates, leachatestreated with polymyxin, and a standard endo-toxin preparation. The supernatants so pro-duced were estimated for TNF content withthe L929 assay. The data in fig 1 show TNFproduction by macrophages treated withwhole wool and grain dust samples.

This figure shows that spontaneous produc-tion of TNF in untreated macrophages wasvirtually nil. In the dust treated cells, at a con-centration of 50 ug/ml dust, wool dust S pro-duced about 750 U TNF/ml, dust M 650 UTNF/ml, and grain 550 U TNF/ml. At a con-centration of 100,g/ml the TNF productionwas about 1750 U TNF/ml in all three cases.

Analysis of variance showed that there wasno difference between dusts (F = 0-274;P > 0 05). There was, however, a significanteffect of dust concentration (F = 55 40;P < 0.01), but no significant treatment-doseinteraction (F = 1-080; P > 0-05).

Both wool and grain dusts are thus potenttriggers of TNF production, there was a dosedependent effect of dust and, on an equalmass basis, the dusts did not differ in theirability to stimulate secretion ofTNF.

2500

2000

E 1500

F-z 1000

500

A

= 50 jig/mI M 100 jig/mI

Medium Wool S Wool M Grain

Figure 1 TNF production by rat alveolar macrophagesuntreated or treated in vitro with wool dusts S and M, andgrain dust. The data are the mean (SEM) of the numberof units of TNF contained in supernatants prepared onthree separate occasions.

389

on 17 May 2018 by guest. P

rotected by copyright.http://oem

.bmj.com

/O

ccup Environ M

ed: first published as 10.1136/oem.53.6.387 on 1 June 1996. D

ownloaded from

Brown, Donaldson

Figure 2 TNFproduction by rat alveolarmacrophages treated invitro with wool dustleachate S and M, andgrain dust leachate. Thegraph also shows the effectofendotoxin depletion onTNF production. The dataare the mean (SEM) ofthe number of units ofTNF in supernatantprepared on three separateoccasions.

2500 r

2000 KZ Before polymyxinM After polymyxin

1500 [-E

z 1000 e

500

0 KmMedium Wool S Wool M

The endotoxin content of wool and grain dust leachatesbefore and after depletion with a polymyxin column

Endotoxin (nglml)

Leachate Before column After column

Wools S 21-99 0 14Wool M 19-76 0-16Grain 14-99 0-14LPS 22-45 0-13

(2) THE EFFECT OF DUST LEACHATES ON TNFPRODUCTION BY RAT ALVEOLAR MACROPHAGESThe possible role of endotoxin associated withdust in causing release of TNF by alveolarmacrophages treated with dust was tested bydust leachate with or without depletion ofendotoxin on polymyxin columns.

Figure 2, which summarises these results,shows that untreated macrophages producedno TNF. Wool dust leachates stimulated amodest increase in TNF release, wool Sleachate producing about twice as much(about 300 U TNF/ml) as leachate M. Thegrain leachate, stimulated substantially moreTNF secretion than wool leachates, in therange of 1500 U TNF/ml. Thus, in this assaysystem, soluble material from grain was morepotent than wool in stimulating TNF produc-tion by macrophages, and there was a signifi-cant effect of leachate treatment (F = 81 65;P < 0001).A striking effect of LPS was seen in super-

natants which had been depleted of endotoxinby passing through a polymyxin column(table). Supernatant depleted of endotoxinreleased substantially less TNF from alveolarmacrophages than untreated leachates (fig 2).Analysis of variance showed in these resultsthat there was a significant effect of depletion

2

Lipopolysaccharide (yg/ml)Figure 3 TNF production by rat alveolar macrophagestreated in vitro with various concentrations oflipopolysaccharide. The data are the mean (SEM) of thenumber of units of TNF in supernatant prepared on threeseparate occasions.

of LPS by polymyxin (F = 17-08; P < 0-05).There was no significant interaction betweendifferent leachates and polymyxin (F = 0-62;P > 0 05) indicating that the contribution ofLPS was consistent between the differentleachates.The results suggest that a substantial part of

the TNF production was due to the presenceof bacterial endotoxin.

(3) EFFECT OF PURE ENDOTOXIN ON TNFPRODUCTION BY RAT ALVEOLAR MACROPHAGESProduction ofTNF by macrophages treated invitro with a commercially available endotoxinpreparation was assessed to further confirmthe role of LPS (fig 3).

There was virtually no TNF produced inonly the medium, but this increased to 633 UTNF/ml at 200 ng/ml, the lowest dose used;the response only about doubled when theconcentration increased to 800 ng/ml. There-after there was a very shallow increase withdose (data not shown).

(4) EFFECT OF DUST LEACHATES ANDDEPLETED LEACHATES ON TNF PRODUCTION BYMACROPHAGES FROM C3H/HEJ MICEMice of the strain C3H/HEJ do not respond toendotoxin, C3H mice respond normally toendotoxin. Peritoneal macrophages from boththese strains were treated with dust leachatesand endotoxin depleted leachates to comparethe role of endotoxin in TNF secretion.

= Beforepolymyxin

M Afterpolymyxin

-T_

Medium Wool S Wool M Medium Wool S Wool M

Figure 4 TNF production by C3H (A normal endotoxin responders) and C3HIHEJ7 (B endotoxin non-responders)peritoneal macrophages treated with dust leachates and leachates depleted of endotoxin with polymyxin. The data are themean (SEM) of the number of units of TNF in supernatant prepared on three separate occasions.

IGrain

A70

60

50

40

30

20

10

n

B

-EU-z

80

70

60

50

40 H

30 F-

20

10

cz ~

n

390

on 17 May 2018 by guest. P

rotected by copyright.http://oem

.bmj.com

/O

ccup Environ M

ed: first published as 10.1136/oem.53.6.387 on 1 June 1996. D

ownloaded from

Wool and grain dusts stimulate TNF secretion by alveolar macrophages in vitro

Figure 4 summarises these effects andshows that dust leachates can trigger therelease of TNF from normal responder C3Hmice, as is found with rat alveolar macro-

phages, although the overall response was

lower with rat cells. Removal of endotoxincaused a dramatic reduction in the TNF pro-

duced (35 U TNF/ml reduced to 3 U TNF/mlfor leachate S. and 55 U TNF/ml reduced to 4U TNF/ml for leachate M). Grain leachatewas not assessed because of the difficulty inpreparing a sterile supernatant.By contrast, in the non-responder C3H/

HEJ mouse peritoneal macrophages, very

small amounts of TNF were produced, themaximum being 10 U TNF/ml in untreatedleachate M, and about 3 U TNF/ml inuntreated leachate S. Removal of endotoxinreduced these responses still further, to 2 UTNF/ml for both leachates.

After analysis of variance, there was a signif-icant effect of treatment (medium v leachate,F = 15-89; P < 0 05), a significant effect ofpolymyxin (F = 57 70; P < 0O05), and a sig-nificant difference between the two strains ofmice (F = 80-43; P < 0O05).

DiscussionTumour necrosis factor is a member of thecytokine family of proteins which are secretedby macrophages, monocytes, and lympho-cytes.724 Secretion is triggered by stimuli suchas inflammation, infection or injury, and more

specifically by other molecules such as inter-feron-y.25 Microorganisms and viruses can alsocause TNF production.4 One of the mostpotent triggers of TNF and other cytokines isthe bacterial endotoxin (LPS).'627 Endotoxinhas also been shown to be an activator of thecomplement system as well as a potent triggerin the immune system'2 and a stimulator oflow level chemotaxin release frommacrophages. It was notable that in the Snellastudy, high doses of endotoxin, as found here,were associated with toxicity.28 The wool andgrain dusts used in this study have been shownto be heavily contaminated with LPS, whichwas also present in aqueous leachates of thedusts.

Rat alveolar macrophages treated in vitrowith wool and grain dusts caused significantamounts of TNF to be released in a dosedependent manner. Untreated (medium only)macrophages produced almost undetectableamounts of TNF. In keeping with theseresults, leachates of dust produced secretion ofTNF in modest amounts.

Previous work where guinea pigs inhaledcotton dust which was contaminated with bac-terial endotoxin, showed that TNF was pro-duced in the lung only three hours afterexposure.29 That study concluded that endo-toxin was the main stimulus causing TNFsecretion. Similarly, wool and grain dustdepositing in the airways and alveolar regionof the lung may potentially cause TNF releasefrom macrophages and epithelial cells medi-ated mainly by LPS. Tumour necrosis factormay then cause inflammatory cell recruitment

into the lung and may stimulate fibroblast pro-liferation. In a lung chronically exposed to res-pirable dust on a daily basis, this regularinitiation of release of TNF could lead tochronic inflammation with predictable patho-logical sequelae. Although it is difficult toextrapolate from the acute high dose stimula-tion of TNF, shown here in vitro, and thechronic effects in someone inhaling dust, aclear role for TNF in the chronic effects of aninhaled dust has been shown.30 In that studyquartz was shown to cause fibrosis (silicosis) inmice but this could be prevented by an anti-body to TNF. We do not suggest that TNF isthe only mediator that is likely to be involvedin the pulmonary response to inhaled wooldust but the importance ofTNF in promotinginflammation and leading to chronic lung dis-ease30 makes these findings of potential impor-tance.The role of endotoxin in the production of

TNF was further examined with the use ofleachates of dust which had been treated withpolymyxin to remove LPS. The leachates wereused to treat alveolar macrophages in vitro andthe resulting supernatants were tested forTNF content. Grain and wool dust leachateshad about the same amounts of LPS in them,as measured in the Limulus amoebocyte lysateassay. However, in the case of wool dust, theleachates had modest TNF stimulating activitywhich was lowered by polymixin treatmentthat removed the LPS, whereas grain dustleachate had significantly more TNF stimulat-ing activity, which was only halved by removalof LPS in the polymixin column. This couldbe interpreted to mean that, as well as a sub-stantial amount of contaminating LPS,another stimulant ofTNF secretion is releasedfrom grain dust. We have no information onthe identity of this material, which could bederived from the grain or contaminatingmicoorganisms. The fact that the whole wooland grain dusts had similar TNF stimulatingactivities suggests that the wool dust has anadditional associated component that is notleachable but is stimulator to macrophagesecretion of TNF. This could be related to thebiochemistry, or most likely to the fibrousshape, of the grain dust particles and deservesfurther study.With reagent grade purified endotoxin, we

were able to show that rat alveolar macro-phages released substantial quantities of TNFin response to endotoxin. Additional confir-mation of the partial role of endotoxin in TNFproduction by wool dust leachates was shownwith mice which are non-sensitive to LPS.Mice of the C3H/HEJ strain have a geneticalteration which renders them non-responsiveto LPS. The normal counterpart, C3H,responds normally.3' Experiments were carriedout in which dust leachates and leachatestreated with polymyxin to deplete endotoxinwere used to trigger release ofTNF from peri-toneal macrophages from the two strains ofmice. The results clearly showed that the non-responder mouse macrophages produced verylow concentrations ofTNF compared with thenormal responder mouse macrophages in

391

on 17 May 2018 by guest. P

rotected by copyright.http://oem

.bmj.com

/O

ccup Environ M

ed: first published as 10.1136/oem.53.6.387 on 1 June 1996. D

ownloaded from

Brown, Donaldson

response to leachates from wool dust.Removal of endotoxin resulted in even smalleramounts of TNF production by the cells.Unfortunately, due to contamination prob-lems in the preparation of leachate, we wereunable to repeat these studies with grain dust.Tumour necrosis factor plays an important

part in the inflammatory response and affectstarget cells in different ways. One of the maineffects of TNF is that it can enhance neu-trophil recruitment to the site of inflammationby activating adhesion molecules on the cellsurface causing more cells to adhere to theendothelium of capillaries. The chemotacticability of TNF for inflammatory neutrophilshas been shown in vitro32 although a separatestudy showed that TNF could cause inhibitionof neutrophil chemotaxis.33 These effects sug-gest that TNF has two roles in the chemotaxisof cells. Firstly, TNF can recruit neutrophilsto inflammatory sites. Secondly, when cellsarrive at this site, TNF acts in the oppositeway, causing migration to be inhibited, andthe cells to be held at the inflammatory site.The accumulation of neutrophils withincreased phagocytic and bacteriocidal capa-bilities and with greater potential for release ofproteinases, superoxide anion, and other mole-cules is clearly beneficial to efficient hostdefence.34 However, in the lungs of peopleinhaling the organic dusts used here, the sameproinflammatory events can be envisaged tooccur at sites of deposition in the airways andterminal bronchioles. In the long term thesecould lead to tissue damage and stimulation ofmesenchymal cell secretion and proliferation.Symptoms of bronchitis have been

described in workers exposed to wool dust andin grain handlers.'535 In obstructive lung dis-ease, the role of TNF in mediating the inflam-matory response and its consequences hasbeen widely reported.3637 Studies show thatduring exacerbations of chronic bronchitisthere is an increase in the number of cells pos-itive for TNFa in the bronchial mucosa.38 Thedata presented here suggest a possible role forLPS in causing TNF accumulation that couldbe a factor in bronchitis in these workers.

In a different study where an organic parti-cle, yeast cell wall 1-3-,f3D-glucan, wasinfused intravenously into rats, granulomaswere produced in the lung.39 The presence ofchemoattractant protein-1 (MCP-1) derivedfrom monocytes is required for granuloma for-mation,39 and this can be induced by TNF andinterleukin-1 (IL-1). Flory et al39 showed thatinfusion of antibodies against TNF and IL-1considerably reduced the size of the granulo-mas and indicated that these cytokines played aregulatory part in the formation of granulomasproduced by organic particles. The wool andgrain dusts used in the present study were alsocapable of producing granulomas in rat lungsafter intratracheal instillation.20 Although pro-duction of granulomas is very likely to be aresult of the instillation mode of delivery, thisprovides indirect evidence for the role of TNFin the genesis of granulomas.The implication of these findings and the

results of the dust and leachate experiments,

where TNF was produced by rat alveolarmacrophages, indicate that the symptoms ofbronchitis and breathlessness reported byworkers exposed to organic dust may resultfrom similar events in their lungs. Theseresults would indicate that LPS, very likelyoriginating from contaminating bacteria pre-sent on the dust surface, could be the triggerfor release of TNF leading to the enhance-ment and maintenance of a chronic inflamma-tory response in people exposed to respirableorganic dust on a daily occupational basis.

We thank Dr M Topping of the Health and Safety Executivefor providing grain dust samples. This project was funded bythe Health and Safety Executive.

1 Kelley J. Cytokines of the lung. Am Rev Respir Dis 1990;141:765-88.

2 Wakefield PE, James WD, Samaska CP, Meltzer MS.Tumor necrosis factor. J Am Acad Dermatol 1991;24:675-85.

3 Matic M, Simon SR. Tumor necrosis factor release fromlipopolysaccharide-stimulated human monocytes: lipo-polysaccharide tolerance in vitro. Cytokine 1991 ;3:576-83.

4 Bienhoff SE, Allen GK, Berg JN. Release of tumor necrosisfactor-alpha from bovine alveolar macrophages stimu-lated with bovine respiratory viruses and bacterial endo-toxins. Vet Immunol Immunopathol 1991;30:341-57.

5 De Rochemonteix-Galve B, Marchat-Amoruso B, DayerJM, Rylander R. Tumor necrosis factor and interleukin-1activities in free lung cells after single and repeatedinhalation of bacterial endotoxin. Infect Immun 1991;59:3646-50.

6 Kovacs EJ. Fibrogenic cytokines: the role of immune medi-ators in the development of scar tissue. Immunol Today1991;21:17-23.

7 Warren JS. Interleukins and tumor necrosis factor ininflammation. CrtRev Clin Lab Sc 1990;28:37-59.

8 Webb DSA, Mostowski HS, Gerrard TL. Cytokine-induced enhancement of ICAM-1 expression results inincreased vulnerability of tumor cells to monocyte-medi-ated lysis. _l Immunol 199 1; 146:3682-6.

9 Luscinskas FW, Cybulsky MI, Kiely JM, Peckins CS,Davis VM, Gimbrone MA Jr. Cytokine-activated humanendothelial monolayers support enhanced neutrophiltransmigration via a mechanism involving both endothe-lial-leukocyte adhesion molecule-i and intercellularadhesion molecule-1. JfImmunol 199 1;146: 1617-25.

10 Pober JS, Gimbrone MA, Jr, Lapierre LA, Mendrick DL,Fiers W, Rothlein R, Springer TA. Overlapping patternsof activation of human endothelial cells by interleukin-1,tumor necrosis factor, and immune interferon. 7Immunol1986;137:1893-6.

11 Edwards JH, Wagner JC, Seal RM. Pulmonary responsesto particulate materials capable of activating the alternativepathway of complement. Clin Allergy 1976;6:155.

12 Wilson MR, Sekul A, Ory R, Salvaggio JE, Lehrer SB.Activation of the alternative complement pathway byextracts of cotton dust. Clin Allergy 1980;10:303-8.

13 Olenchock SA, Mull JC, Major PC. Extracts of airbornegrain dusts activate alternative and classical complementpathways. Ann Allergy 1980;44:23-8.

14 Love RG, Smith TA, Gurr D, Soutar CA, Scarisbrick DA,Seaton A. Respiratory and allergic symptoms in wool tex-tile workers. BrJ Ind Med 1988;45:727-41.

15 Love RG, Muirhead M, Collins HPR, Soutar CA. Thecharacteristics of respiratory ill health of wool textileworkers. BrJ Ind Med 1991;48:221-8.

16 Donaldson K, Brown GM, Brown DM, Slight J, CullenRT, Love RG, Soutar CA. Inflammation in the lungs ofrats after deposition of dust collected from the air of woolmills: the role of epithelial injury and complement activa-tion. BrJ Ind Med 1990;47:231-8.

17 Brown DM, Donaldson K. Injurious effects of wool andgrain dusts on alveolar epithelial cells and macrophagesin vitro. BrJ'IndMed 1991;48:196-202.

18 Ozesmi Aslan H, Hillerdal G, Rylander R, Ozesmi C, BarisYI. Byssinosis in carpet weavers exposed to wool contam-inated with endotoxin. BrJfIndMed 1987;44:479-83.

19 Brown DM, Dransfield I, Wetherill GZ, Donaldson K.LFA-1 and ICAM-1 in homotypic aggregation of ratalveolar macrophages: organic dust-mediated aggregationby a non-protein kinase C-dependent pathway. Am JRespir CellMol Biol 1993;9:205-12.

20 Brown DM, Donaldson K. Activity of wool mill dust invitro and in vivo; cytotoxicity, cytokine production,lymph node stimulation and histopathology. Ann OccupHyg 1994;38:887-94.

21 Mark D, Vincent JH, Gibson H, Lynch G. A new staticsampler for airborne total dust in workplaces. Am IndHygAssocJ 1985;46:127-33.

22 Rosenstreich DL, Vogel SN, Jacques AR, Wahl LM,Oppenheim JJ. Macrophage sensitivity to endotoxin:genetic control by a single codominant gene. Y Immunol1978;121:1664-70.

392

on 17 May 2018 by guest. P

rotected by copyright.http://oem

.bmj.com

/O

ccup Environ M

ed: first published as 10.1136/oem.53.6.387 on 1 June 1996. D

ownloaded from

Wool and grain dusts stimulate TNF secretion by alveolar macrophages in vitro

23 Donaldson K, Li X-Y, Dogra S, Miller BG, Brown GM.Asbetos-stimulated tumour necrosis factor release fromalveolar macrophages depends on fibre length and opson-isation. JPathol 1992;168:243-8.

24 Larrick JW, Kunkel SL. The role of tumor necrosis factorand interleukin-l in the immunoinflammatory response.Pharmacol Res 1988;5:129-39.

25 Jones A. Tumor necrosis factor. Transfusion Science 1991;12:67-73.

26 Beutler B, Krochin N, Milsark IW, Leudke C, Cerami A.Control of cachectin (tumor necrosis factor) synthesis:mechanisms of endotoxin resistance. Science 1986;232:977-80.

27 Feist W, Ulmer AJ, Musehold J, Brade H, Kusumoto S,Flad HD. Induction of tumor necrosis factor-alpha release by lipopolysaccharide and defined lipo-polysaccharide partial structures. Immunobiology 1989;179:293.

28 Snella MC. Production of a neutrophil chemotactic factorby ebdotoxin-stimulated alveolar macrophages in vitro.BrJ Exp Pathol 1986;67:801-7

29 Ryan LK, Karol MH. Release of tumor necrosis factor inguinea pigs upon acute inhalation of cotton dust. Am J7Respir Cell Mol Biol 1991;5:93-8.

30 Piguet PF, Collart MA, Grau GE, Sappino AP, Vassalli P.Requirement of tumour necrosis factor for the develop-ment of silica-induced pulmonary fibrosis. Nature 1990;344:245-7.

31 Adi S, Pollock AS, Shigenage JK, Moser AH, Feingold KR,Grunfeld C. Role for monokines in the metabolic effects ofendotoxin. Interferon-gamma restores responsiveness ofC3H/HEJ mice in vivo. J Clin Invest 1992;89:1603-9.

32 Figari IS, Palladino MA. Stimulation of neutrophil chemo-

taxis by recombinant tumor necrosis factor alpha andbeta [abstract]. Fed Proc 1987;46:562.

33 Stephens K, Ishizaka A, Basilico L, Larrick J, Raffin T.Human tumor necrosis factor inhibits neutrophil chemo-taxis [abstract]. Am Rev Respir Dis 1987;135(suppl):A338.

34 Nelson S, Bagby GJ, Bainton BG, Wilson LA, ThompsonJJ, Summer WR. Compartmentalization of intraalveolarand systemic lipopolysaccharide-induced tumor necrosisfactor and the pulmonary inflammatory response. JI InfectDis 1989;159:189-94.

35 Corey P, Hutcheon M, Broder I, Mintz S. Grain elevatorworkers show work-related pulmonary function changesand dose-effect relationships with dust exposure. Br IndMed 1982;39:330-7.

36 Distefano A, Maestrelli P, Roggeri A, Turato G, Calabro S,Potena A. Up-regulation of adhesion molecules in thebronchial mucosa of subjects with chronic obstructivebronchitis. Am J Respir Cell Mol Biol 1994;149:803-10.

37 Tosi MF, Stark JM, Smith CW, Hamedani A, GruenertDC, Infeld MD. Induction of ICAM-1 expression onhuman airway epithelial cells by inflammatorycytokines-effects on neutrophil-epithelial cell adhesion.Am J Respir Cell Mol Biol 1992;7:214-21.

38 Saetta M, Distefano A, Maestrelli P, Turato G, RuggieriMP, Roggeri A, et al. Airway eosinophilia in chronicbronchitis during exacerbations. Am Jf Respir Crit CareMed 1994;150:1646-52.

39 Flory CM, Jones ML, Miller BF, Warren JS. Regulatoryroles of tumor necrosis factor-alpha and interleukin-l-beta in monocyte chemoattractant protein-l-mediatedpulmonary granuloma formation in the rat. Am J Pathol1995;146:450-62.

Instructions to authorsThree copies of all submissions should besent to: The Editor, Occupational andEnvironmental Medicine, BMJ PublishingGroup, BMA House, Tavistock Square,London WC1H 9JR, UK. All authorsshould sign the covering letter as evidenceof consent to publication. Papers reportingresults of studies on human subjects must beaccompanied by a statement that thesubjects gave written, informed consentand by evidence of approval from theappropriate ethics committee. These papersshould conform to the principles outlinedin the Declaration of Helsinki (BMJ 1964;ii: 177).

If requested, authors shall produce thedata on which the manuscript is based, forexamination by the Editor.Authors are asked to submit with

their manuscript the names andaddresses of three people who theyconsider would be suitable independentreviewers. They will not necessarily beapproached to review the paper.Papers should include a structured

abstract of not more than 300 words,under headings of Objectives, Methods,Results, and Conclusions. Pleaseinclude up to three keywords or keyterms to assist with indexing.

393

on 17 May 2018 by guest. P

rotected by copyright.http://oem

.bmj.com

/O

ccup Environ M

ed: first published as 10.1136/oem.53.6.387 on 1 June 1996. D

ownloaded from