free radical activity of synthetic vitreous fibers: iron chelation inhibits hydroxyl radical...

This article was downloaded by: [University of Delaware]On: 04 October 2014, At: 16:02Publisher: Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH,UK

Journal of Toxicology andEnvironmental Health, Part A:Current IssuesPublication details, including instructions for authorsand subscription information:http://www.tandfonline.com/loi/uteh20

FREE RADICAL ACTIVITY OFSYNTHETIC VITREOUS FIBERS:IRON CHELATION INHIBITSHYDROXYL RADICAL GENERATIONBY REFRACTORY CERAMIC FIBERDavid M. Brown Carolyn Fisher Ken DonaldsonPublished online: 30 Nov 2010.

To cite this article: David M. Brown Carolyn Fisher Ken Donaldson (1998) FREERADICAL ACTIVITY OF SYNTHETIC VITREOUS FIBERS: IRON CHELATION INHIBITSHYDROXYL RADICAL GENERATION BY REFRACTORY CERAMIC FIBER, Journal ofToxicology and Environmental Health, Part A: Current Issues, 53:7, 545-561, DOI:10.1080/009841098159132

To link to this article: http://dx.doi.org/10.1080/009841098159132

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all theinformation (the “Content”) contained in the publications on our platform.However, Taylor & Francis, our agents, and our licensors make norepresentations or warranties whatsoever as to the accuracy, completeness, orsuitability for any purpose of the Content. Any opinions and views expressedin this publication are the opinions and views of the authors, and are not theviews of or endorsed by Taylor & Francis. The accuracy of the Content shouldnot be relied upon and should be independently verified with primary sourcesof information. Taylor and Francis shall not be liable for any losses, actions,claims, proceedings, demands, costs, expenses, damages, and other liabilities

http://www.tandfonline.com/loi/uteh20

http://www.tandfonline.com/action/showCitFormats?doi=10.1080/009841098159132

http://dx.doi.org/10.1080/009841098159132

whatsoever or howsoever caused arising directly or indirectly in connectionwith, in relation to or arising out of the use of the Content.

This article may be used for research, teaching, and private study purposes.Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expresslyforbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

Dow

nloa

ded

by [

Uni

vers

ity o

f D

elaw

are]

at 1

6:02

04

Oct

ober

201

4

http://www.tandfonline.com/page/terms-and-conditions

http://www.tandfonline.com/page/terms-and-conditions

FREE RADICAL ACTIVITY OF SYNTHETIC VITREOUS FIBERS: IRON CHELATION INHIBITS HYDROXYL RADICAL GENERATION BY REFRACTORY CERAMIC FIBER

David M. Brown, Carolyn Fisher, Ken Donaldson

Biomedicine Research Group, Department of Biological Sciences, Napier University, Edinburgh, Scotland

Synthetic vitreous fibers are in widespread use but the parameters that dictate their car-cinogenicity are still a matter of scientific debate. The free radical activities of a panelcomprising an asbestos sample and five different respirable synthetic vitreous fiber sam-ples were determined, to address the hypothesis that carcinogenic fibers have greater freeradical activity than noncarcinogenic fibers. On the basis of recent inhalation studies, thesix samples were divided into three carcinogenic fibers—amphibole asbestos, silicon car-bide, and refractory ceramic fiber 1 (designated with the abbreviation RCF 1)—and threenoncarcinogenic fibers—man-made vitreous fiber 10 (a glass fiber sample designated withthe abbreviation MMVF 10), Code 100/475 glass fiber, and RCF4. All experiments werecarried out with equal fiber numbers. Of the two assays of free radical activity used, theplasmid assay of DNA scission showed only amosite asbestos to have free radical activity,while the salicylate assay of hydroxyl activity showed that both amosite asbestos andRCF1 release hydroxyl radicals; silicon carbide fibers had no free radical activity in eitherof the assays. None of the noncarcinogenic fibers demonstrated free radical activity ineither of the assays. The differences in the two assays in demonstrating free radical activitywith RCF1 may be due to increased release of Fe from RCF1 under the more acid condi-tions of the salicylate assay, which was confirmed by the fact that soluble iron causedhydroxylation of salicylate. Presence of an iron chelator inhibited the ability of the RCF1fibers to cause hydroxylation of salicylate, demonstrating that RCF1 generates hydroxylradical by Fenton chemical reaction in the same way as amphibole asbestos.

Exposure to asbestos carries with it an increased risk of pleuralmesothelioma and bronchial carcinoma (Donaldson et al., 1993;Mossman, 1994). Studies have emphasized the importance of the physi-cal properties of the fibers including fiber length (Davis et al., 1996;Donaldson et al., 1989; Hesterberg & Barrett, 1990) and solubility(Morgan et al., 1977). Other fibers such refractory ceramic fibers (RCFs)and silicon carbide have been shown to be carcinogenic in animal studies(Mast et al., 1995a, 1995b; Davis et al., 1996). However, some studiesthat have examined the lung fiber burden in relation to carcinogenicityindicated that carcinogenic and noncarcinogenic fibers cannot be clearly

545

Journal of Toxicology and Environmental Health, Part A, 53:545–561, 1998Copyright © 1998 Taylor & Francis

0098-4108/98 $12.00 + .00

Received 26 May 1997; sent for revision 30 June 1997; accepted 19 August 1997.This research was funded by the Health and Safety Executive. We also acknowledge an equip-

ment grant from the British Occupational Health Research Foundation. K. Donaldson holds theBritish Lung Foundation/British Gas Transco Fellowship in Air Pollution and Respiratory Health.

Address correspondence to Professor Ken Donaldson, Biomedicine Research Group, Departmentof Biological Sciences, Napier University, Edinburgh EH10 5DT, Scotland. E-mail: [email protected]

Dow

nloa

ded

by [

Uni

vers

ity o

f D

elaw

are]

at 1

6:02

04

Oct

ober

201

4

delineated on the basis of dose, dimension, or durability (Hesterberg et al.,1993; Davis et al., 1996). In these studies a substantial proportion of longglass fibers persisted in the lungs of exposed rats but no pathologicalchanges occurred. This has prompted the suggestion that a further fiber fac-tor, surface reactivity, could play a role in the ability of fibers to cause lungdisease (Donaldson, 1995). The case of erionite further argues for a role ofthe fiber surface. This fiber was exceptionally active in causing mesothe-lioma in rats, yet was no more durable nor longer than amphibole asbestos,which caused far fewer mesotheliomas (Wagner et al., 1985).

Cell and molecular studies suggested that the free radical activity ofthe fiber surface could be a factor in carcinogenicity. For example, amajor factor in the cellular effects of asbestos is induction of active oxy-gen species derived from the fiber surface and mediated in part throughthe redox cycling of iron (Fubini, 1996; Ghio et al., 1994; Mossman &Marsh, 1989; Kamp et al., 1992). Iron is important in catalyzing a varietyof reactions, in particular the formation of hydroxyl radicals (Fubini et al.,1995). Iron, through the formation of active oxygen species from the sur-face of fibers, may therefore be important in damaging biomolecules.Gilmour et al. (1995) demonstrated that asbestos possesses strong hydroxylradical activity that is mediated by iron and found the release of iron froma variety of fibers.

The fibers used in the present study were designated as carcinogenicor noncarcinogenic based upon recent in vivo carcinogenicity studies(Hesterberg et al., 1993; Glass et al., 1995; Davis et al., 1996). The car-cinogenic fibers were long-fiber amosite, silicon carbide, and refractoryceramic fiber 1 (RCF1); noncarcinogenic fibers were refractory ceramicfiber 4 (RCF4), special-purpose glass fiber Code 100/475, and man-madevitreous fiber 10 (MMVF10). It is important to point out that althoughRCF4 is classified here as nontumorigenic based on the results of theinhalation studies carried out by RCC (Mast et al., 1995a, 1995b), theresults do not show that the RCF4 composition is noncarcinogenic. TheRCF4 is substantially shorter than RCF1 and the other fibers used in thestudy, despite the fact that it is derived from it, so its lack of carcinogenic-ity could be due to its relative shortness. RCF4 did cause fibrosis, soalthough it is noncarcinogenic it is pathogenic, causing fibrosis in thehighest exposure group. The assays used here involve interactionsbetween the fibers and various molecular detecting systems (not cells)that could not be fiber length sensitive, so the results here are relevant tothe RCF4 composition. The purpose of the study was to determine anycorrelation between known carcinogenic fibers and their ability to gener-ate free radicals at their surfaces.

MATERIALS AND METHODS

The intrinsic hydroxyl radical activity of each fiber type was assessedby plasmid DNA scission and by high-performance liquid chromatogra-

546 D. M. BROWN ET AL.

Dow

nloa

ded

by [

Uni

vers

ity o

f D

elaw

are]

at 1

6:02

04

Oct

ober

201

4

phy (HPLC) using a hydroxyl radical trap, salicylate. Fibers depositing inthe lung become coated with lung lining material, which may modify thefiber surface reactivity and hence the fiber’s oxidant-generating ability(Brown & Donaldson, 1998). We therefore used rat lung lining fluid tocoat the fibers to determine whether the oxidant-generating ability couldbe modulated. The role of iron in mediating hydroxyl radical productionwas assessed by the use of the chelator desferrioxamine-B (DSF-B), andthe hydroxyl radical scavenger mannitol was utilized in some assays.

Fibers and ReagentsThe fibers used in this study were:

1. Long-fiber amosite asbestos (Davis et al., 1986).2. Silicon carbide fiber (Advanced Composite Materials Corporation).3. Code 100/475 glass fibers, originally made by Johns Manville.4. The remainder of the fiber samples were obtained from the TIMA fiber

repository.

All of the fibers in the repository are synthetic vitreous fibers. However,the refractory ceramic fibers are designated RCF and the glass fibers asMMVF; these prefixes are followed by a suffix to indicate the particularsample. In the present study the following synthetic vitreous fiber samplesfrom the TIMA repository were used: MMVF10, a sample of 901 glass:RCF 1, a kaolin-based refractory ceramic fiber; and RCF4, a heated orafter-service sample of RCF1.

f X174 RF closed circular, supercoiled plasmid DNA was obtainedfrom Gibco Europe (Paisley, UK) and was used in all the plasmid assays.The hydroxyl radical scavenger D-mannitol, iron chelator DSF-B, salicylicacid, and 2,3-dihydroxybenzoic acid (2,3-DHBA) were purchased fromSigma (Poole, UK). Iron sulfate, iron chloride, sodium acetate, andtrisodium citrate were purchased from BDH Chemicals, Glasgow.

Length distributions of the different fiber preparations used here areshown in Table 1.

The dose used in experiments to assess the free radical activity of thefibers was adjusted to equal 8.24 × 107 fibers/ml. This dose was used inour previous studies with cells and is a nontoxic dose of fibers to cells inculture, as we have shown in previous studies using dose responses(Brown & Donaldson, 1998; Hill et al., 1996).

HPLC ReagentsThe citrate acetate buffer for HPLC, pH 3.85–3.9, was filtered through

a 0.45-µm filter before use. Salicylic acid was made up to a stock 1 Msolution in methanol and diluted with buffer to give a final concentrationof 25 mM. A 1 M stock solution of 2,3-DHBA in methanol was also pre-pared, and standards ranging from 125 M to 6.25 M were prepared inHPLC buffer without salicylic acid.

FREE RADICAL ACTIVITY OF FIBERS 547

Dow

nloa

ded

by [

Uni

vers

ity o

f D

elaw

are]

at 1

6:02

04

Oct

ober

201

4

Preparation of Surfactant and Pretreatment of FibersFemale Wistar rats approximately 4 mo old weighing approximately

350 g were used throughout. Rats were killed with a single ip injection ofpentobarbitone; the lungs were cannulated and removed and lavagedwith 4 × 8 ml of sterile saline. The lavageate was pooled into a singletube. Tubes were spun at 1200 rpm for 5 min at 4°C and the supernatantwas decanted. The lavageate was transferred and spun at 23,000 rpm for45 min. At the end of the centrifugation period, the supernatant fromeach tube was discarded and the remaining pellet resuspended in 10 mlsterile saline. This surfactant-enriched fraction, which was termed 1×concentrated, was stored at –20°C until required in accordance with themethod of Baughman et al. (1987). All fibers were suspended at a con-centration of 2 mg/ml in sterile saline or 1× concentrated rat surfactant.The suspensions were sonicated briefly to disperse the fibers and incu-bated in a rotary shaker for 1 h at 37°C. After the incubation period, thefibers were washed twice with sterile saline.

Plasmid Assayf X174RF plasmid DNA (Life Technologies, Paisley, UK) was added to

ultrapure sterile distilled water at a concentration of 240 ng/20 µl. Fiberswere suspended in the DNA solution at 924,900 fibers/20 µl. Test conditionswere fibers (coated and uncoated) in DNA/water only, or fibers (uncoated) inDNA/water plus mannitol (4 mM). Each treatment was incubated at 37°C for8 h. Four microliters of tracking dye was then added to each sample and theDNA plasmid was separated by electrophoresis for 16 h at 20 V on 0.8%agarose gel. After staining in ethidium bromide, a photograph of the gelunder ultraviolet (UV) light was taken and the bands indicating damage tothe plasmid were quantified by densitometry. The results were expressed asthe percentage of the treatments to the untreated plasmid.

Hydroxyl Radical Generation AssayFor the hydroxyl radical generation assay, fibers were resuspended in

salicylic acid buffer to give a final concentration of 8.24 × 107 fibers/ml.Fiber treatments were as follows: salicylic acid buffer only (coated anduncoated fibers); or buffer with mannitol (uncoated fibers, 10 mM); orbuffer with desferal (DSF-B, concentrations ranging from 0.01 to 0.1 mM,


TABLE 1. Size distribution of the fibers used in the study

>10 µm >20 µm

LFA 64.75 35.25SiC 60.86 27.6RCF1 77.36 45.27RCF4 59.35 17.96MMVF10 85.24 67.17Code 100/475 50.00 19.32

Dow

nloa

ded

by [

Uni

vers

ity o

f D

elaw

are]

at 1

6:02

04

Oct

ober

201

4

uncoated fibers). Additionally, a separate set of buffers were set up thatdid not contain fibers; these were as follows: buffer only; Fe2+ (sulfate); orFe3+ (chloride) (iron at concentrations 25–100 µM). All samples wereincubated at 37°C for 20 h. After incubation, tubes were spun at 3000rpm to remove the fibers, and the supernatant was collected and filteredthrough a 0.22-µm filter. Samples were stored at –20°C until required.

The HPLC (LKB) was set up and carried out according to the methodof Coudray et al. (1995). The analytical column was 150 × 4.6 mmpacked with octadecyl silane (Spherisorb ODS2, C18) with an averageparticle size of 3 µm (Alltech). The guard column contained 10 µmSpheri-10 PR18 (30 × 4.6 mm ID), also from Alltech. The flow rate wasset at 1 ml/min, and 100 µl of sample was loaded onto the column ineach case. Analysis of the products were controlled by computer(Shimadzu model C-RSA Chromatopac integrator) using a detection sys-tem with a 275-nm fil ter. Standard concentrations of 2,3-DHBA(described earlier) were loaded onto the column and the area under thepeak produced after an elution time of 6.9 min was recorded. A standardcurve was then constructed from these data.

Measurement of Fe3+ Release From FibersTo compare the amount of Fe3+ released from RCF1 fibers in the plas-

mid assay and HPLC assay conditions, the Fe3+ chelator (DSF-B) was used.Fibers were suspended in water or salicylic acid/citrate acetate buffer at8.24 × 107 fibers/ml and incubated in a rotary shaker for 8 h and 20 h. Thefibers were then removed from solution by centrifugation at 4000 rpm for10 min and the supernatant was collected. The concentration of Fe3+ wasdetermined by adding 250 µl of 2 mM DSF-B to 250 µl of sample. Theamount of iron present in the samples was determined by comparing with astandard curve of FeCl3 (Fe

3+) with DSF-B and read at 430 nm. The resultswere expressed as nanomoles Fe3+ per 8.24 × 107 fibers per 250 µl.

StatisticsAll data were analyzed by one-way analysis of variance with Tukey’s

multiple comparison test to detect differences between treatments. Aprobability of <.05 was taken to indicate significance. Data were eitherused in their “raw” form or normalized, as indicated on the figures.

RESULTS

Fiber-Mediated Free Radical Damage to Plasmid DNA

All fibers displayed some free radical activity as assessed by their abil-ity to decrease the percentage of supercoiled plasmid DNA (Figure 1), butthis was only significant in the case of long-fiber amosite asbestos, whichcaused 55% depletion of supercoiled DNA. The remaining fibers had freeradical activity that ranged from 5 to 20%, but these values were not sig-nificantly different from control.


Dow

nloa

ded

by [

Uni

vers

ity o

f D

elaw

are]

at 1

6:02

04

Oct

ober

201

4

Effect of Hydroxyl Radical Scavenger and Surfactant Coating on Fiber-Mediated Damage to Plasmid DNAThe hydroxyl radical scavenger mannitol was used to confirm the

role of hydroxyl radical in causing damage to plasmid DNA. In addition,the effect of coating the fibers with rat lung surfactant was examined.This series of experiments was carried out using only the long-fiberamosite asbestos sample, because this was the only fiber type that pro-duced a significant depletion of supercoiled DNA. Long-fiber amositeasbestos alone caused a significant 35% depletion of DNA (Figure 2).However, when mannitol was included in the assay, DNA depletion wasnot significantly different from control. Similarly, when rat surfactant-coated fibers were included in the plasmid assay, supercoiled DNAdepletion was limited to a level that was not significantly different fromcontrol (Figure 2).

Hydroxylation of Salicylic Acid by FibersThe intrinsic hydroxyl radical production by fibers was additionally

assessed by measuring the amount of 2,3-DHBA formation after incu-bation with salicylic acid using HPLC. Data in Figure 3 show that the


FIGURE 1. Depletion of supercoiled plasmid DNA by 9.249 × 105 fibers of each fiber type com-pared to untreated DNA (DNA). Data are mean + SEM of three separate experiments. LFA, long-fiberamosite; SiC, silicon carbide; RCF, refractory ceramic fibers, MMVF10, man-made vitreous fiber 10;C100/475, Code 100/475 special purpose glass microfiber. Asterisk indicates significantly differentfrom DNA control (p < .05).

Dow

nloa

ded

by [

Uni

vers

ity o

f D

elaw

are]

at 1

6:02

04

Oct

ober

201

4

product was only detected when long-fiber amosite (LFA) and RCF1 weretested, but not with any of the other fibers. Mannitol, desferal, and surfac-tant were included in experimental series to confirm the involvement ofthe surface of the fiber in hydroxyl radical production. Figures 4 and 5show the inhibitory effect of surfactant-coating and mannitol on 2,3-DHBA production by the two free radically active fibers. A desferal doseresponse is shown in Figure 6 with the DHBA production levels normal-ized to 100% for the two fibers; the absolute values for DHBA were [dataas mean (SEM) 2,3-DHBA production pmol/100 µl from 3 separate exper-iments] LFA 0.45 (0.05) and RCF1 1.0 (0.05). In all treatments and forboth fiber types, there was significantly less 2,3-DHBA with mannitol,surfactant coating, and desferal treatment.

Contradictory Result on Free Radical Activity of RCF1 in the Plasmid and the Salicylic Acid AssaysRCF1 was shown to possess free radical activity in the salicylic acid

assay but no activity in the plasmid assay. Data suggest that this differ-ence was due to the difference in release of iron in the two assays. The


FIGURE 2. Depletion of supercoiled plasmid DNA 8.24 × 107 fibers of native long-fiber amosite,long-fiber amosite in the presence of mannitol (LFA/mannitol), and long-fiber amosite coated withsurfactant (LFA/surf). DNA, untreated. Data are mean + SEM of three separate experiments. Asteriskindicates significant difference from control (p < .05) only for LFA. See legend to Figure 1 for expla-nation of abbreviations.

Dow

nloa

ded

by [

Uni

vers

ity o

f D

elaw

are]

at 1

6:02

04

Oct

ober

201

4

release of Fe3+ from RCF1 in the plasmid and salicylic acid assay condi-tions was compared. The data showing release of Fe3+ are illustrated inTable 1. Over a period of 20 h, more Fe3+ was released from the fibers inthe salicylic acid conditions compared with plasmid assay conditions. Atthe time point used to treat fibers for HPLC analysis an average of 52.52nmol Fe3+ was released in the plasmid assay conditions, compared with76.84 nmol Fe3+ released in the salicylic acid condition, an approximate50% increase.

In order to confirm the role of iron in hydroxyl radical production,Fe2+ and Fe3+ ions were added into the salicylic acid buffer and theamount of 2,3-DHBA measured. These data are shown in Figure 7. Bothkinds of iron caused hydroxylation of the substrate, but almost twice theamount of product was produced with Fe2+ ions compared to Fe3+ ions,and this effect was dose related.

DISCUSSION

The focus of this study was the synthetic vitreous fibers, which are inwidespread use and are manufactured in a variety of ways to producefibers with varying properties. In this study three types of synthetic vitre-ous fiber were compared: glass fibers (Code 100/475 and MMVF10),refractory ceramic fibers (RCF1 and RCF4), and silicon carbide. These


FIGURE 3. Hydroxyl radical activity, measured as hydroxylated salicylic acid derivative, 2,3-DHBA,following incubation of 8.24 × 107 fibers with salicylic acid. Data are mean + SEM of three separateexperiments. See legend to Figure 1 for explanation of abbreviations.

Dow

nloa

ded

by [

Uni

vers

ity o

f D

elaw

are]

at 1

6:02

04

Oct

ober

201

4

were compared with amosite asbestos for ability to generate free radicalsin solution to determine whether this was a predictor of carcinogenicity.

The ability of asbestos fibers to cause lung diseases such as fibrosis orcancer has been considered to be related to length of fibers (Davis et al.,1996; Donaldson et al., 1993) and biopersistence (Davis & Donaldson,1993; Donaldson, 1995). However, in recent inhalation studies noncar-cinogenic fibers were found to accumulate in the lung to similar extentsas carcinogenic fibers and were of comparable length distribution andbiopersistence in vivo; examples of noncarcinogenic fibers that showedthis effect were Code 100/475 glass fiber (Davis et al., 1996) andMMVF10 (Hesterberg et al., 1993). Thus long fibers can persist in thelung and no apparent pathological alterations develop, and this suggeststhat another factor associated with carcinogenic fibers is important inleading to disease.

This putative additional factor could be the ability of fibers to gener-ate free radicals at their surface, hydroxyl radical in particular. Argumentsin favor of this as a factor in toxicity are the following: Asbestos causes


FIGURE 4. Production of 2,3-DHBA by 8.24 × 107 fibers of native LFA or RCF1 and by fibers coatedwith lung lining fluid. Data are mean and SEM of four different experiments normalized to the controlvalue. See legend to Figure 1 for explanation of abbreviations.

Dow

nloa

ded

by [

Uni

vers

ity o

f D

elaw

are]

at 1

6:02

04

Oct

ober

201

4

hydroxylated adducts of DNA and DNA breakage in vitro (Leanderson etal., 1988; Gilmour et al., 1995); asbestos causes DNA breakage in cellsin vitro (Libbus et al., 1989), an event that is mediated by hydroxyl radi-cal in vitro (Gilmour et al., 1995); Fe chelators and antioxidants caninhibit production of cytokines stimulated by asbestos in vitro (Simeonova& Luster, 1995); and antioxidants can inhibit the inflammation caused byshort-term in vivo exposure to asbestos (Mossman et al., 1990).

The production of free radicals by pathogenic particles has beenattributed to iron at the particle surface and hydroxyl radical, generatedby Fenton chemical reactions (Lund & Aust, 1991; Kennedy et al., 1989;Ghio et al., 1994; Gilmour et al., 1995; Maples & Johnson, 1992). Usingan assay that measures hydroxyl radical activity as ability of fibers to scis-sion-deplete supercoiled plasmid DNA, long-fiber amosite was the onlyfiber that displayed hydroxyl radical activity, as previously reported(Gilmour et al., 1995). In the salicylic acid assay, however, both long-fiber amosite and RCF1 showed hydroxyl radical activity. The ability ofRCF1 to generate hydroxyl radical but its failure to damage DNA could


FIGURE 5. Production of 2,3-DHBA by 8.24 × 107 fibers of LFA and RCF1 fibers in the presence andthe absence of the hydroxyl radical scavenger mannitol. Data are mean and SEM of 4 different exper-iments normalized to the control value. See legend to Figure 1 for explanation of abbreviations.

Dow

nloa

ded

by [

Uni

vers

ity o

f D

elaw

are]

at 1

6:02

04

Oct

ober

201

4

be explained by differences in the release of Fe3+ from the RCF1 surfacein the salicylic acid medium of the hydroxyl assay, since:

1. Under the conditions of the salicylic acid assay there was morerelease of Fe3+ from RCF1 compared to the plasmid assay.

2. Fe3+ and Fe2+ were shown to hydroxylate salicylic acid assay, with Fe2+

having the greatest activity.

Gilmour et al. (1995) demonstrated that, for a range of synthetic vitre-ous fibers, the simple release of large amounts of soluble Fe3+ is not itselfa reliable indicator of the ability of any fiber sample to generate hydroxylradical in the plasmid assay or to be carcinogenic; an additional factorassociated with the surface of the fiber is necessary for free radical gener-ation. However, as shown here the salicylate assay does respond to Fe2+

and Fe3+. Despite the fact that 100-fold higher concentrations of Fe werenecessary to produce 2,3-DHBA than were found after incubation offibers in buffer, the concentration of Fe close to the fiber, before it diffusesaway into solution, could easily reach the levels necessary to induce


FIGURE 6. Effect of desferal on production of 2,3-DHBA by 8.24 × 107 fibers of LFA and RCF1 fibersincubated with salicylic acid. Data are mean from three separate experiments converted to the per-centage of the control value. SEM of original data was less than 15% of the mean.

Dow

nloa

ded

by [

Uni

vers

ity o

f D

elaw

are]

at 1

6:02

04

Oct

ober

201

4

hydroxylation of salicylate in the normal assay. It is noteable that the con-ditions of the salicylate assay closely mimic the acidity of the macro-phage phagolysosome (Nyberg et al., 1992), so phagolysosomal RCF1fibers may be a major source of oxidative stress on the macrophage.

Confirmation of the role of the fiber surface was demonstrated by thefact that lung surfactant inhibited the generation of hydroxyl radical, andthis is in keeping with our previous results on the inhibitory effects of sur-factant (Brown & Donaldson, 1998). In vitro assays should utilize lung


TABLE 2. Fe(III) released (nmol) from 8.24 × 107 RCF1 fibersafter treatment in conditions that mimic the plasmid or thesalicylate assay

Assay conditionsTime ________________________(h) Plasmid Salicylate

8 33.59 (1.4) 47.83 (3.00)20 52.52 (1.4) 76.84 (5.69)

FIGURE 7. Effect of various concentrations of Fe2+ and Fe3+ on formation of 2,3-dihydroxybenzoicacid from salicylic acid.

Dow

nloa

ded

by [

Uni

vers

ity o

f D

elaw

are]

at 1

6:02

04

Oct

ober

201

4

lining fluid because of the likely modifying role that this has on biologicalactivity of fibers and other particles.

These results are important for development of screening assays for pre-dicting the carcinogenicity of fibers, as well as for understanding the mech-anism of fiber toxicity. Despite the fact that asbestos has the ability to stim-ulate cells (Simeonova & Luster, 1995) and cause inflammation (Mossmanet al., 1990) via free radical mechanisms, this does not appear to be gener-ally true for other carcinogenic fiber types. Silicon carbide has proven to beone of the most carcinogenic fibers to be investigated in experimentalpathology studies (Davis et al., 1996), as well as being as cytotoxic andcytostatic as asbestos (Vaughan et al., 1991). The absence of free radicalactivity of this fiber in the two assays used here suggests either that free rad-icals are not involved in silicon carbide carcinogenicity or that the condi-tions of the assays are not sensitive to detect free radical generation.

From the point of view of screening assays, the salicylic acid assay didshow that long-fiber amosite and RCF1, two of the three carcinogenicfibers, had the ability to generate hydroxyl radical activity that could con-tribute to their carcinogenicity. RCFs have been reported to stimulate ratalveolar macrophages to release tumor necrosis factor (TNF) and eico-sanoids in amounts approximately equivalent to those stimulated by cro-cidolite asbestos at the same mass dose (Leikauf et al., 1995). Our datasuggest that RCF1 may exert its toxicity via its ability to generate Fentonchemistry-derived hydroxyl radicals, in much the same way as amositeasbestos and crocidolite asbestos (Gilmour et al., 1995; Lund & Aust,1992). This is supported by the fact that RCF4 (the heat-treated derivativeof RCF1) exerted no free radical activity and was not carcinogenic ininhalation studies while RCF1 was carcinogenic (Mast et al., 1995a,1995b), and RCF4 was the only completely negative fiber in a large-scaleperitoneal injection mesothelioma study by the author and co-workers,where RCF1 was highly carcinogenic (Jones et al., 1997). It is not possi-ble to rule out the composition of RCF4 as being carcinogenic in theinhalation studies because RCF4 was much shorter and thicker than RCF1following its heat treatment and this could have contributed to its lack ofcarcinogenicity. Our data do, however, tend to support the importance offiber composition, since RCF4 contrasted with RCF1, in producing nohydroxyl radical in vitro. Brown et al. (1992) reported the formation ofmullite on the surface of heated “after-service” RCF fibers, with cristo-balite formation in the core. The formation of the mullite surface layer onheating may be responsible for the difference between RCF1 and RCF4 inability to generate free radicals. The level of Fe released from RCF1 wassurprising, since the original TIMA Repository data sheets describe theproportion of iron (as oxide) in RCF1 as 0.97%. A companion paper tothe present one describes the release of iron from all of the fibers used inthe present study, and RCF1 consistently releases more Fe than any of theother fibers, including LFA (data not shown).


Dow

nloa

ded

by [

Uni

vers

ity o

f D

elaw

are]

at 1

6:02

04

Oct

ober

201

4

In this study the fibers were equalized to the same fiber number.However, to gain the complete information on the specific activity of thefiber surface, it would be necessary to calculate the surface area andequalize to surface area. The necessary data to carry out such calcula-tions are not available, but should be the aim of future studies. In terms ofrisk assessment, however, fiber exposure is assessed as fibers in air andsurface area is not taken into account, so comparison on the basis ofequal fiber number is valid.

Maples and Johnson (1992) compared hydroxyl radical activity in thesalicylate assay and ability to cause mesothelioma after instillation, for apanel of eight fibers. There was a correlation between hydroxyl radical-generating ability and mesothelioma after instillation into the pleuralspace for the naturally occurring fibers, but two glass fibers caused highamounts of hydroxylation of salicylate but negligible mesotheliomas. Twoglass fibers, Code 100/475 and MMVF10, consistently failed to generatehydroxyl radicals in the present study. Data suggest that this may be relatedto the acid pH (3.9) that was used, which closely mimicked the pH of themacrophage phagolysosome, compared to the neutral conditions used byMaples and Johnson (1992). In fact, most fibers are phagocytosed veryrapidly after depositing in the lung, although long fibers may be incom-pletely phagocytosed (Donaldson et al., 1993), and the sections of fiberwithin phagolysosomes are thereafter exposed to the acid phagolysoso-mal milieu. Studies aimed at dissecting the effects on fibers of residencein the lung should therefore consider the highly acid conditions per-taining in the phagolysosome. The results obtained suggest that acid con-ditions would be more likely to release iron but, as pointed out in ourprevious study, the fiber surface can also be involved in catalytic-typeeffects in hydroxyl radical generation (Gilmour et al., 1995). If acid con-ditions modulate this factor—for example, acid conditions favor catalytictype effects in the naturally occurring fibers but inhibit them in the glassfibers—then this could explain differences between the in vitro and the invivo findings. Certainly, the acid microenvironment exerts a differentchemical effect on the surface of different fibers, since acid conditions,similar to those found in the phagosome, favor the persistence of sometypes of glass fiber (Morgan, 1994) while encouraging the dissolution ofothers (Bauer et al., 1994).

This study shows that, for a small panel of three noncarcinogenic andthree carcinogenic fibers, the ability to generate hydroxyl radical in solu-tion is not 100% reliable as a discriminator. However, the salicylic acidassay of hydroxyl radical generation was positive for two of the three car-cinogenic fibers (amosite asbestos and RCF1) and negative for three of thethree noncarcinogenic fibers; silicon carbide was the carcinogenic fiberthat was not active in generating hydroxyl radicals. Hydroxyl radical gen-eration from both amosite and RCF1 involved iron, as shown by the abil-ity to be blocked by iron chelator. Based on these studies, the detection


Dow

nloa

ded

by [

Uni

vers

ity o

f D

elaw

are]

at 1

6:02

04

Oct

ober

201

4

of surface hydroxyl radical activity in a test fiber of unknown toxicityprovides support for the contention that the fiber is carcinogenic; how-ever, failure to detect free radical activity cannot be taken as evidencefor noncarcinogenicity, because of the negativity of silicon carbide inthis assay. A greater range of fibers should be tested in this system toconfirm or refute the usefulness of this approach. This assay, in combina-tion with other mechanism-based endpoints associated with long-termeffects, could provide a battery of tests that would be more reliable thanthe use of any single test in predicting carcinogenicity of fibers (MRC/IEH, 1995).

REFERENCESBauer, J. F., Law, B. D., and Hesterberg, T. 1994. Dual pH durability studies of man-made vitreous

fibre. Environ. Health Perspect. 102(suppl. 5):61–66.Baughman, R. P., Mangels, D. J., Strohofer, S., and Corser, B. C. 1987. Enhancement of macrophage

and monocyte cytotoxicity by the surface active material of lung lining fluid. J. Lab. Clin. Med.109:692–697.

Brown, D. M., Robertson, N., and Donaldson K. 1998. Production of superoxide anion by macro-phages exposed to respirable industrial fibres: modifying effect of lung lining fluid. Toxicol. InVitro 12:15–24.

Brown, R. C., Sara, E. A., Hoskins, J. A., Evans, C. E., Young, J., Laskowski, J. J., Acheson, A., Forder,S. D., and Road, A. P. 1992. The effects of heating and devitrification on the structure and bio-logical activity of aluminosilicate refractory ceramic fibres. Ann. Occup. Hyg. 36:115–129.

Coudray, C., Talla, M., Martin, S., Fatome, M., and Favier, A. 1995. High-performance liquid chro-matography-electrochemical determination of salicylate hydroxylation products as an in vivomarker of oxidative stress. Anal. Biochem. 227:101–111.

Davis, J. M. G., and Donaldson, K. 1993. Respirable industrial fibres: Pathology in animal models.Ann. Occup. Hyg. 37:227–236.

Davis, J. M. G., Bolton, R. E., Donaldson, K., Jones, A. D., and Smith, T. 1986. The pathogenicity oflong versus short fibres of amosite asbestos administered to rats by inhalation and intraperi-toneal injection. Br. J. Exp. Pathol. 67:415–430.

Davis, J. M. G., Brown, D. M., Cullen, R. T., Donaldson, K., Jones, A. D., Miller, B. G., McIntosh, C.,and Searl, A. 1996. A comparison of methods of determining and predicting the pathogenicityof mineral fibers. Inhal. Toxicol. 8:747–770.

Donaldson, K. 1995. Mechanisms of DNA, molecular and cellular damage: An overview In Naturaland Man-Made Mineral Fibres: UK Research Priorities, pp. 51–68. Report commissioned by theMRC Committee on Toxic Hazards in the Environment and the Workplace. Leicester: Institute ofEnvironment and Health.

Donaldson, K., Brown, G. M., Brown, D. M., Bolton, R. E., and Davis, J. M. G. 1989. The inflamma-tion generating potential of long and short fibre amosite asbestos samples. Br. J. Ind. Med.46:271–276.

Donaldson, K., Brown, R. C., and Brown, G. M. 1993. Respirable industrial fibres: Mechanisms ofpathogenicity. Thorax 48:390–395.

Fubini, B. 1996. Use of physico-chemical and cell-free assays to evaluate the potential carcinogenic-ity of fibres. In Mechanisms of fibre carcinogenesis, pp. 35–54. IARC Scientific Publications No140. Lyon: IARC.

Fubini, B., Mollo, L., and Giamello, E. 1995. Free radical generation at the solid/liquid interface iniron-containing minerals. Free Radical Res. 23:593–614.

Ghio, A., Kennedy, T. P., Stoneheurner, J. G., Crumbliss, A. L., and Hoidal, J. R. 1994. DNA strandbreaks following in vitro exposure to asbestos increase with surface complexed [Fe(III)]. Arch.Biochem. Biophys. 311:13–18.


Dow

nloa

ded

by [

Uni

vers

ity o

f D

elaw

are]

at 1

6:02

04

Oct

ober

201

4

Gilmour, P., Beswick, P. H., and Donaldson, K. 1995. Detection of surface free radical activity of res-pirable industrial fibres using supercoiled phi X 174 plasmid DNA. Carcinogenesis 16:2973–2979.

Glass, L. R., Brown, R. C., and Hoskins, J. A. 1995. Health effects of refractory ceramic fibres: scien-tific issues and policy considerations. Occup. Environ. Med. 52:433–440.

Hesterberg, T. W., and Barrett, J. C. 1990. Dependence of asbestos and mineral dust induced trans-formation of mammalian cells on fibres dimension. Cancer Res. 44:115–124.

Hesterberg, T. W., Miller, W. C., McConnell, E. E., Chevalier, J., Hadley, J. G., Bernstein, D. M.,Thevanaz, P., and Anderson, R. 1993. Chronic inhalation toxicity of size-separated glass fibersin Fischer-344 rats. Fundam. Appl. Toxicol. 20:464–476.

Hill, I. M., Beswick, P. H., and Donaldson, K. 1996 Enhancement of the macrophage oxidative burstby immunoglobulin coating of respirable fibres: Fibre-specific differences between asbestos andman-made fibres. Exp. Lung Res. 22:133–148.

Jones, A. D., Miller, B. G., Cullen, R. T., Searl, A., Davis, J. M., Buchanan, D., Donaldson, K., Souter,C. A., and Bolton, R. E. 1997. The colt fibre research programme: Aspects of toxicological riskassessment. Ann. Occup. Hyg. 41(Suppl. 1):244–250.

Kamp, D. W., Gracefa, P., Pryor, W. A., and Weitzman, S. A. 1992. The role of free radicals inasbestos-induced diseases. Free Radical Biol. 12:293–315.

Kennedy, T. P., Dodson, R., Rao, N. V., Henry, K. Y., Hopkins, C., Baser, M., Tolley, E., and Hoidal,J. R. 1989. Dusts causing pneumoconiosis generate �OH and produce hemolysis by acting asFenton catalysts. Arch. Biochem. Biophys. 269:359–364.

Leanderson, P., Soderquist, P., Tagesson, C., and Axelson, O. 1988. Formation of 8-hydroxy-deoxyguanosine by asbestos and man made mineral fibres. Am. J. Ind. Med. 45:309–311.

Leikauf, G. D., Fink, S. P., Miller, M. L., Lockey, J. L., and Driscoll K. E. 1995. Refractory ceramicfibres activate alveolar macrophage eicosanoid and cytokine release. J. Appl. Physiol. 78:164–171.

Libbus, B. L., Illenye, S. A., and Craighead, J. E. 1989. Induction of DNA strand breaks in cultured ratembryo cells by crocidolite asbestos as assessed by nick translation. Cancer Res. 49:5713–5718.

Lund, L. G., and Aust, A. E. 1991. Iron-catalysed reactions may be responsible for the biochemicaland biological effects of asbestos. BioFactors 3:83–89.

Maples, K. R., and Johnson, N. F. 1992. Fiber-induced hydroxyl radical formation: Correlation withmesothelioma induction in rats and human. Carcinogenesis 11:2035–2039.

Mast, R. W., McConnell, E. E., Anderson, R., Chavalier, J., Kotin, P., Bernstien, D. M., Thevenaz, P.,Glass, L. R., Miller, W. C., and Hesterberg, T. W. 1995a. Studies on the chronic toxicity (inhala-tion) of refractory ceramic fiber in male Fischer 344 rats. Inhal. Toxicol. 7:425–467.

Mast, R. W., McConnell, E. E., Hesterberg, T. W., Chevalier, J., Kotin, P., Thevenaz, P., Bernstein,D. M., Glass, L. R., Miller, W., and Anderson, R. 1995b. Multiple-dose chronic inhalation toxic-ity study of size-separated kaolin ceramic fiber in male Fischer-344 rats. Inhal. Toxicol. 7:469–502.

Morgan, A. 1994. In vivo evaluation of chemical biopersistence of man-made mineral fibres. Environ.Health Perspect. 102(Suppl. 5):127–132.

Morgan, A., Evans, J. C., and Holmes, A. 1977. Deposition and clearance of inhaled fibrous mineralsin the rat. Studies using radioactive tracer techniques. In Inhaled particles IV, ed. W. H. Walton,pp. 259–272. Oxford: Pergamon Press.

Mossman, B. T. 1994. Carcinogenesis and related cells and tissue responses to asbestos: A review.Ann. Occup. Hyg. 38:617–624.

Mossman, B. T., and Marsh, J. P. 1989. Evidence supporting a role for active oxygen species inasbestos-induced toxicity and lung disease. Environ. Health Perspect. 81:91–94.

Mossman, B. T., Marsh, J. P., Sesko, A., Shatos, M., Doherty, J., Petruska, J., Adler, K. B.,Hemenway, D., Mickey, R., Vacek, P., and Kagan, E. 1990. Inhibition of lung injury, inflamma-tion and interstitial pulmonary fibrosis by polyethylene glyco-conjugated catalase in a rapidinhalation model of asbestosis. Am. Rev. Respir. Dis. 141:1266–1271.

MRC/IEH. 1995. IEH Report on Natural and Man-Made Mineral Fibres: UK Research Priorities.Report R3. Leicester, UK: Institute of Environment and Health.


Dow

nloa

ded

by [

Uni

vers

ity o

f D

elaw

are]

at 1

6:02

04

Oct

ober

201

4

Nyberg, K., Johansson, V., Johansson, A., and Camner, P. 1992. Phagolysosomal pH in alveolarmacrophages. Environ. Health Perspect. 97:149–152.

Simeonova, P. P., and Luster, M. I. 1995. Iron and reactive oxygen species in the asbestos-inducedtumor necrosis factor-alpha response from alveolar macrophages. Am. J. Respir. Cell Mol. Biol.12:676–683.

Vaughan, G. L., Jordan, J., and Karr, S. 1991. The toxicity, in vitro, of silicon carbide whiskers.Environ. Res. 56:56–67.

Wagner, J. C., Skidmore, J. W., Hill, R. J., and Griffiths, D. M. 1985. Erionite exposure and mesothe-lioma in rats. Br. J. Cancer 51:727–730.


Dow

nloa

ded

by [

Uni

vers

ity o

f D

elaw

are]

at 1

6:02

04

Oct

ober

201

4

free radical activity of synthetic vitreous fibers: iron chelation inhibits hydroxyl radical...

Documents