Toxicology 218 (2006) 186–196

Induction of lupus-related specific autoantibodies by non-specificinflammation caused by an intraperitoneal injection

of n-hexadecane in BALB/c mice

Yoshiki Kurodaa, Nobutaka Onoc, Jun Akaogia, Dina C. Nacionalesa, YoshiokiYamasakia, Tolga T. Barkera, Westley H. Reevesa,b, Minoru Satoha,b,∗a Division of Rheumatology and Clinical Immunology, Department of Medicine, ARB-R2-156, 1600 SW Archer Road,

Box 100221, Gainesville, FL 32610-0221, USAb Department of Pathology, Immunology, and Laboratory Medicine, University of Florida, Gainesville, FL 32610-0221, USA

c First Department of Pathology, Fukushima Medical College, Fukushima 960-1247, Japan

Received 23 August 2005; received in revised form 13 October 2005; accepted 16 October 2005Available online 23 November 2005

Abstract

A single intraperitoneal (i.p.) injection of pristane, incomplete Freund’s adjuvant (IFA), or the adjuvant oil squalene, but not highmolecular weight medicinal mineral oils, induces lupus-related autoantibodies to nRNP/Sm and -Su in non-autoimmune strains ofmice. This ability appears to be associated with the low molecular weight and adjuvanticity of hydrocarbon.

n-Hexadecane (C16H34), which is present in petroleum, has adjuvant activity and induces arthritis in rodents like other lupus-i st, jetf n oils antibodiesi logya xamined3 s within2 plasmica n-specifici eviouslyu ns andi©

K

f

dro-ntal25ted

0

nducing oils. In addition to dietary exposure ton-hexadecane in mineral oils, exposure also occurs via inhalation of oil miuel, or diesel exhaust or by absorption through the skin. Sincen-hexadecane is a low molecular weight adjuvant hydrocarboimilar to other lupus-inducing hydrocarbons, the present study examined whether it can also induce lupus-related auton mice. Female BALB/cJ mice received a single i.p. injection of 0.5 ml ofn-hexadecane, pristane, or saline (control). Pathond serology (immunoglobulin levels, autoantibodies by immunofluorescence, immunoprecipitation, and ELISA) were emonths later. Unexpectedly, alln-hexadecane-treated mice, but none in the other groups, developed inflammatory ascite.5 months.n-Hexadecane induced hypergammaglobulinemia (IgG1, IgG2a), antinuclear (titer > 1:160, 67%) and -cytontibodies (58%) and autoantibodies to nRNP/Sm (25%), Su (33%), ssDNA (83%), and chromatin (100%). Therefore, no

nflammation caused byn-hexadecane resulted in the production of a limited set of specific autoantibodies. These prnrecognized immunological effects ofn-hexadecane may have implications in monitoring human exposure to hydrocarbo

n the pathogenesis of autoimmune diseases.2005 Elsevier Ireland Ltd. All rights reserved.

eywords: Hexadecane; Pristane; Hydrocarbon; Autoantibodies; Lupus

∗ Corresponding author. Tel.: +1 352 846 0695;ax: +1 352 846 1858.

E-mail address: satohm@ufl.edu (M. Satoh).

1. Introduction

Humans are exposed daily to various types of hycarbons via food, medicine, or other environmefactors. Dietary exposure to white mineral oil (>carbons; C25, heavy mineral oil) is well documen

300-483X/$ – see front matter © 2005 Elsevier Ireland Ltd. All rights reserved.doi:10.1016/j.tox.2005.10.011

Y. Kuroda et al. / Toxicology 218 (2006) 186–196 187

(Heimbach et al., 2002) and is believed to be responsiblefor lipogranulomas (follicular lipidosis) seen in the liver,spleen, or lymph nodes of healthy individuals in west-ern countries (Cruickshank, 1984; Dincsoy et al., 1982).Low molecular weight hydrocarbons (C15–C25, lightmineral oil) generally have stronger adjuvant activityand are more inflammatory than high molecular weighthydrocarbons (Whitehouse et al., 1974). Incomplete Fre-und’s adjuvant (IFA), which contains the light mineraloil (C15–C25) Bayol F, and the C19 isoprenoid alkanepristane (2,6,10,14-tetramethylpentadecane) have adju-vant activity and induce chronic arthritis in mice andrats (Vingsbo et al., 1996; Wooley et al., 1989). In addi-tion, a single intraperitoneal (i.p.) injection of pristaneinduces lupus-like autoimmune syndrome in virtuallyany non-autoimmune strains of mice (Satoh et al., 1995;Satoh and Reeves, 1994). We have recently reported thatIFA and the adjuvant oil squalene can also induce lupus-related autoantibodies to nRNP/Sm and Su in BALB/cmice, suggesting that the ability of hydrocarbons toinduce these autoantibodies is associated with their adju-vanticity (Satoh et al., 2003a).

n-Hexadecane (C16H34) is an adjuvant (Whitehouseet al., 1974) that induces chronic arthritis in rodents(Kleinau et al., 1995), similar to the effects of pristaneor IFA. Exposure of humans ton-hexadecane occursthrough ingestion, inhalation, or dermal absorption ofmineral hydrocarbons in food products, tobacco smoke,gasoline, diesel smoke, jet fuel, and other sources (Mannet al., 1977; Riviere et al., 1999). We report here thatn-h andp es al Sm,S on.T ectso esist ar-b avei trialh

2

2

romJ d ina d-i thso jec-t aCt were

left untreated. Since these two groups showed very similarresults in autoantibody analysis, they were analyzed as a sin-gle untreated/PBS group. The mice were euthanized 3 monthsafter injection and the blood, spleen, liver, kidneys, and lungswere harvested and processed. Peritoneal cells were harvestedby lavaging with Dulbecco’s modified Eagle’s media (DMEM)containing 10% fetal calf serum and 1 U/ml heparin. This pro-tocol was approved by the IACUC of University of Florida andthe Malcom Randall VA Medical Center, Gainesville, FL.

2.2. Cell surface staining by flow cytometry

Cell surface phenotypes were characterized by flow cytom-etry as described, using the following mAbs conjugatedto fluorescein-isothianate (FITC), phycoerythrin (PE), orallophycocyanin (APC): CD3 (17A2), CD4 (GK1.5), CD8(53–6.7), CD11b (M1/70), Gr-1 (RB6-8C5), and CD19(1D3) (BD Bioscience, San Jose, CA). Species, isotype, andconjugation-matched irrelevant mAbs were used as controls.Data were collected with a FACScaliber flow cytometer andCellQuest software (Becton Dickinson) (Satoh et al., 2003a)and were analyzed using FlowJo 4.2 software program (TreeStar Inc., Ashland, OR).

2.3. Histology

Tissues were fixed with 4% paraformaldehyde and 4�msections were stained with hematoxylin–eosin (H&E) asdescribed (Satoh et al., 1996a).

2.4. Antinuclear and anti-cytoplasmic antibodies byimmunofluorescence

s ineter-use

eusingula-

erathsken

ously

ere2ed.

exadecane induces chronic peritoneal inflammationolyclonal hypergammaglobulinemia, but also induc

imited set of lupus-related autoantibodies to nRNP/u, and chromatin within 3 months of an i.p. injectihese previously unrecognized immunological efff n-hexadecane are consistent with our hypoth

hat many low molecular weight adjuvant hydrocons can induce autoimmunity. The data may h

mplications in monitoring human exposure to indusydrocarbons.

. Materials and methods

.1. Treatment of mice

Four-week-old female BALB/cJ mice were purchased fackson Laboratory (Bar Harbor, ME), and were housevirus free animal facility at the Malcom Randall VA Me

cal Center (Gainesville, FL) in barrier cages. At 3 monf age, 16–45 mice per group received a single i.p. in

ion (0.5 ml) of pristane orn-hexadecane (both from Sigmhemical Co., St. Louis, MO) (Satoh and Reeves, 1994). Con-

rol mice received phosphate buffered saline (PBS) or

Levels of antinuclear and anti-cytoplasmic antibodiesera from BALB/cJ mice 3 months after treatment were dmined by indirect immunofluorescence using L929 mofibroblast cells as described (Satoh et al., 1996a). Sera werscreened at a 1:40 dilution and the titers were estimatedImage Titer (RhiGene Inc., De Plaines, IL), a titration emtion system (Satoh et al., 2003b).

2.5. Anti-ssDNA, anti-chromatin, anti-nRNP/Sm, andanti-Su antibody ELISA

Levels of anti-ssDNA and -chromatin antibodies in s(1:250 dilution) from BALB/cJ mice were tested 3 monafter treatment by ELISA, using calf thymus ssDNA or chicchromatin as the antigen (Satoh et al., 2000). Anti-nRNP/Smand -Su antigen-capture ELISAs were performed as previdescribed (Satoh et al., 2000).

2.6. Radiolabeling and immunoprecipitation

Autoantibodies to cellular proteins in murine sera wanalyzed by immunoprecipitation of35S-radiolabeled K56cell extract using 5�l murine serum per sample as describ

188 Y. Kuroda et al. / Toxicology 218 (2006) 186–196

Autoantibody specificities were verified by SDS-PAGE usingreference sera (Satoh and Reeves, 1994).

2.7. Immunoglobulin ELISA

Serum immunoglobulin levels were measured 3 monthsafter treatment by sandwich ELISA as described (Satoh et al.,2000). Levels of IgG1, IgG2a, IgG2b, IgG3, IgA, and IgMwere determined using 1:200,000 and/or 1:500,000 sera.

2.8. Statistical analysis

Frequencies of autoantibodies were compared by Fisher’sexact test. Levels of immunoglobulins and autoantibodies werecompared by the Mann–Whitney test.

3. Results

3.1. Pathological examination

Pristane-induced plasmacytomas in BALB/cAn andother susceptible strains of mice are usually recog-nized by the development of ascites after 6 months orlater of treatment (Potter and Wax, 1983). The develop-ment of ascites within 6 months of an i.p. injection ofpristane or other hydrocarbon oils is extremely rare inBALB/cJ mice. Unexpectedly, alln-hexadecane-treatedmice developed massive ascites within 2–2.5 months, butnone of the pristane-treated nor the untreated mice devel-oped ascites. Whether this is due to early plasmacytomadevelopment, another malignancy, or inflammation was

e of

ednealre

hilspsy,ragm

and on the surface of the enlarged liver and spleen(Fig. 1A), as in pristane-treated mice (Shaheen etal., 1999). Although pristane treatment significantlyincreased spleen weight (p < 0.01 versus untreated,Mann–Whitney) consistent with previous studies (Satohet al., 2000), the spleen ofn-hexadecane-treated miceweighed twice as much as that of pristane-treated mice(p < 0.01, by Mann–Whitney). Similarly, the liversof n-hexadecane-treated mice were much larger thanthose of pristane-treated mice (p < 0.01 versus untreatedor pristane, by Mann–Whitney), consistent with astrong inflammatory reaction caused byn-hexadecane(Table 1). Mediastinal lymphadenopathy was apparentin most of the pristane orn-hexadecane-treated mice(p < 0.01 versus untreated, pristane). The peritoneumwas thickened with inflammatory tissue and the adhe-sion of the peritoneum to internal organs. The presenceof solid tumors was not apparent macroscopically.

Upon histological examination, neutrophils andmacrophages were abundant in the inflammatory peri-toneal tissue inn-hexadecane-treated mice, however,no signs of malignancy were seen. Granuloma forma-tion and diffuse inflammatory tissue were seen in theperitoneum, diaphragm, liver, and spleen (Fig. 1B–E).Therefore, we concluded that ascites was not dueto the development of malignancy, but resulted fromthe chronic peritoneal inflammation induced byn-hexadecane. A small amount of inflammatory asciteswas often seen in pristane-treated mice upon autopsy,but this was considerably less than what was seen in

ng or

3

G1

examined.n-Hexadecane treated mice had an averag

10.0 ml of ascites (Table 1) and weighed∼10 g morethan other groups (Table 1) (p < 0.01 versus untreator pristane-treated group). The number of peritocells increased∼10-fold over controls. These cells wepredominantly CD11b (+) Gr1 (+) mature neutrop(∼60%) and macrophages (not shown). On autowhite granulomatous tissue was seen on the diaph

Table 1Body and organ weights and number of peritoneal cells

Untreated (n = 12)

Body weight (g) 29.0± 4.3Liver (g) 1.47± 0.16Kidneys (g) 0.41± 0.02Spleen (g) 0.10± 0.02Mediastinal lymph node (g) 0Ascites (ml) 0Peritoneal cells (×106) 13.7± 13.3

* p < 0.01 vs. untreated, by Mann–Whitney.** p < 0.05 vs. untreated, by Mann–Whitney.

n-hexadecane-treated mice, suggesting that a strodifferent type of inflammatory reaction is caused byn-hexadecane.

3.2. Serum immunoglobulin levels

Serum immunoglobulin levels were measuredmonths after treatment by ELISA (Fig. 2). Likepristane, n-hexadecane treatment increased Ig

Pristane (n = 12) n-Hexadecane (n = 12)

27.0± 3.1 36.5± 6.2*

1.68± 0.19** 2.10± 0.41*

0.38± 0.06 0.46± 0.060.14± 0.05** 0.28± 0.09*

0.04± 0.05* 0.09± 0.08*

0 10.0± 6.2*

35.6± 25.5* 121.9± 77.0*

Y. Kuroda et al. / Toxicology 218 (2006) 186–196 189

Fig. 1. (A) Peritoneal cavity of a BALB/c mouse 3 months aftern-hexadecane treatment. White granulomatous lesions around the liver and spleenand adherent to the diaphragm are seen. (B) The liver and inflammatory tissue from ann-hexadecane-treated mouse. H&E staining of the liver froma mouse treated withn-hexadecane 3 months earlier. A granuloma and inflammatory tissue are diffusely attached to the liver. Original magnification40×. (C) and (D) Oil granuloma attached to the liver. H&E staining of the oil granuloma attached to the liver from ann-hexadecane-treated mouse.Inflammatory cells are infiltrating into liver tissue. Original magnification: (C) 100× and (D) 200×. (E) Fibrous thickening of the liver capsule.H&E staining, original magnification 200×.

(p < 0.0001 versus untreated, Mann–Whitney test),IgG2a (p < 0.0001), IgG2b (p < 0.0001), IgA (p < 0.003),and IgM (p < 0.0001) with predominant effects on IgG1.This was in contrast to the predominant increase in IgG2a

by pristane treatment (p < 0.0001 versus untreated group,by Mann–Whitney).

Reflecting the predominant increase of IgG2a in thepristane-treated group, a surrogate marker of Th1/Th2

190 Y. Kuroda et al. / Toxicology 218 (2006) 186–196

Fig. 2. Serum immunoglobulins levels. Total serum IgG1, IgG2a,IgG2b, IgG3, IgA, and IgM levels were measured by ELISA 3months after treatment. Mean values plus standard deviation for eachgroup (untreated, pristane,n = 20, n-hexadecane,n = 12) is shown.(a) p<0.0001; (b)p = 0.003; (c)p<0.02, pristane vs. untreated orn-hexadecane vs. untreated by Mann–Whitney test. No statistical sig-nificance was observed between the pristane vs. then-hexadecanegroup for any of the isotypes or subclasses.

cytokine balance in vivo, IgG2a (IFN-� dependent)/IgG1(IL-4 dependent) ratios (Satoh et al., 2003a,b), weredramatically increased in the pristane-treated group butnot in the hexadecane-treated group (average± S.D.,hexadecane 1.87± 1.55, pristane 4.00± 2.77, untreated1.77± 1.52, hexadecane versus pristanep < 0.05; hex-adecane versus untreated not significant; pristane versusuntreatedp < 0.01, by Mann–Whitney).

3.3. Antinuclear and anti-cytoplasmic antibodies byimmunofluorescence

Many sera from mice treated withn-hexadecane hadantinuclear and/or anti-cytoplasmic antibodies similarto pristane-treated mice.Fig. 3 shows representativenuclear (A and B) and cytoplasmic (C and D) staining

Fig. 3. Antinuclear and anti-cytoplasmic antibodies by immunofluorescetreatment, showing nuclear (A and B) and cytoplasmic (C and D) stainiSerum dilution 1:40, original magnification 400×.

nce. L929 cells stained with serum from mice 3 months aftern-hexadecaneng. Negative staining by a serum from an untreated mouse is shown in (E).

Y. Kuroda et al. / Toxicology 218 (2006) 186–196 191

by sera fromn-hexadecane-treated mice and negativestaining (E) by a serum from an untreated mouse.

The titers of antinuclear antibodies in sera fromBALB/cJ mice 3 months after pristane orn-hexadecanetreatment or age-matched untreated mice were exam-ined by indirect immunofluorescence with a titrationemulation system. Sixty-seven percent (8/12) ofn-hexadecane-treated mice had antinuclear antibodies attiters≥ 1:160 versus 3% (1/34) in the untreated group(p < 0.0001, by Fisher’s exact test). Anti-cytoplasmicantibodies at titers≥ 1:160 were found in 58% (7/12)of n-hexadecane-treated mice versus 7% (2/34) in theuntreated group (p = 0.0004, Fisher’s exact test).

3.4. Anti-ssDNA antibody and anti-chromatinantibodies

Levels of anti-ssDNA antibodies in sera (1:250 dilu-tion) from BALB/cJ mice were tested by ELISA 3months after treatment. The levels of anti-ssDNA anti-bodies inn-hexadecane treated mice were higher than inuntreated mice (p < 0.01, by Mann–Whitney) or pristane-treated mice (p < 0.01) (Fig. 4A). Similarly, levels ofanti-chromatin antibodies in sera fromn-hexadecane-treated mice were significantly higher than untreated orpristane-treated mice (Fig. 4B). These results suggestthat n-hexadecane induces anti-ssDNA and -chromatinantibodies even more efficiently than pristane. Data forpristane-treated mice and untreated mice were previ-ously published (Kuroda et al., 2004), but are includedh

3a

inedb el -t diesa -t heses nti-b tane-t 6,7 no-p -i d)a iesw bleta d bya( lso

Fig. 4. IgG anti-ssDNA and anti-chromatin antibodies by ELISA.Levels of IgG anti-ssDNA and -chromatin antibodies in sera (1:250dilution) from BALB/cJ mice 3 months after treatment (untreated, pris-tane,n-hexadecane) were examined by ELISA. (A) IgG anti-ssDNAantibodies. Mice treated withn-hexadecane had significantly higherlevels than the pristane or untreated group (p<0.01, by Mann–Whitneytest). (B) IgG anti-chromatin antibodies. Then-hexadecane-treatedgroup had significantly higher levels than the pristane or untreatedgroup. *p < 0.01, **p < 0.05, U: vs. untreated group, P: vs. pristane-treated group, median in each group is indicated by a horizontal line.

produced antibodies to nuclear factor (NF) 45/90/110(Fig. 5B, arrow heads), which have been previouslyreported in pristane-treated mice on C57BL/6 or B10background (Satoh et al., 1999, 2000).

While anti-nRNP/Sm antibodies inn-hexadecane-treated mouse sera immunoprecipitated nRNP/Sm veryweakly compared with those by pristane-treated mice(Fig. 5A), the intensity of Su bands immunoprecipitatedby sera from both groups was comparable (Fig. 5Aand B). Therefore, the levels of anti-nRNP/Sm and-Su antibodies were examined by ELISA (Fig. 6A andB). Sera from mice treated with pristane 3 monthsearlier contain high levels of anti-nRNP/Sm and -Su

ere for comparison.

.5. n-Hexadecane treatment inducesnti-nRNP/Sm and -Su antibodies in BALB/c mice

Specific autoantibodies in sera were determy immunoprecipitation of 35S-methionine/cystein

abeled K562 cell extract (Fig. 5A and B). Pristanereated mice produced anti-nRNP/Sm and -Su antibos described (Satoh and Reeves, 1994). n-Hexadecane

reated mice also produced autoantibodies with tpecificities, though the levels of anti-nRNP/Sm aodies appeared to be much lower than those in pris

reated mice (compareFig. 5A lanes 1, 2, and 5 versus, and 9). The frequency of autoantibodies by immurecipitation is summarized inTable 2. n-Hexadecane

nduced anti-nRNP/Sm (25%,p < 0.05 versus untreatend -Su (33%,p < 0.01 versus untreated) antibodithin 3 months. The characteristic 100 kDa dound 200 kDa proteins commonly immunoprecipitatenti-Su sera (Satoh et al., 1994) are indicated inFig. 5Barrows). Two of the12n-hexadecane-treated mice a

192 Y. Kuroda et al. / Toxicology 218 (2006) 186–196

Table 2Frequency of autoantibodies by immunoprecipitation 3 months after treatment

Treatment n Anti-nRNP/Sm (%) Anti-Su (%) Anti-nRNP/Sm or anti-Su (%) Anti-NF45/90/110 (%)

n-Hexadecane 12 25a 33b 42c 17Pristane 20 50d 35c 65d 5Untreated 28 0 0 0 0

a p < 0.05 vs. untreated group by Fisher’s exact test.b p < 0.01 vs. untreated group by Fisher’s exact test.c p < 0.005 vs. untreated group by Fisher’s exact test.d p < 0.0001 vs. untreated group by Fisher’s exact test.

Fig. 5. Immunoprecipitation using sera 3 months after treatment.35S-labelemice 3 months after pristane (panel A, lanes 1–5),n-hexadecane (panel A18–21). (A) 12.5% SDS-PAGE. In addition to pristane,n-hexadecane induceindicated by arrow heads) and -Su (100 kDa Su protein, arrow head Suand 6, 7, and 9 are weakly positive. Samples 2, 5, 7, 9, 12, and 15 aproteins immunoprecipitated by BALB/c mice 3 months aftern-hexadecanedoublet and 200 kDa proteins immunoprecipitated by anti-Su sera areproduced antibodies to NF45/90/110 (arrow heads, NF45, NF90, and Npositive. Sample 2 does not immunoprecipitate the upper band of the 1to those in lanes14–17, 10–13, and 6–9, of panel A, respectively.

d K562 cell extract was immunoprecipitated with sera from BALB/cJ, lanes 6–17, panel B, lanes 1–12), or PBS (untreated, panel A, lanesd anti-nRNP/Sm (components of UsnRNPs A, B′/B, C, D, E/F, and G are

) antibodies. Samples in lanes 1, 2, 5, are strongly positive for anti-nRNP/Smre anti-Su positive. (B) Eight percent SDS-PAGE. Large molecular weighttreatment were fractionated on 8% SDS-PAGE. Characteristic 100 kDa

indicated (arrows, Su 100 and 200 k). Two of 12n-hexadecane-treated miceF110 from the bottom, lanes 5 and 10). Samples 2, 7, 10, and 12 are anti-Su00 kDa doublet. Samples in lanes 1–4, 5–8, and 9–12 in panel B correspond

Y. Kuroda et al. / Toxicology 218 (2006) 186–196 193

Fig. 6. Levels of anti-nRNP/Sm and -Su antibodies by ELISA. Sera from mice without treatment or treated with pristane orn-hexadecane 3 monthsearlier were tested by ELISA (1:250 dilution). (A) Anti-nRNP/Sm ELISA. (*)n-Hexadecane vs. untreated,p < 0.01, (**) pristane vs. untreated,p < 0.01, pristane vs.n-hexadecane,p < 0.05, by Mann–Whitney test. (B) Anti-Su ELISA. Pristane vs. untreated,p < 0.05,n-hexadecane vs. untreated,p < 0.05, pristane vs.n-hexadecane, not significant. (C) Relationship between anti-nRNP/Sm vs. -Su antibodies. The levels of anti-nRNP/Sm vs.anti-Su in individual mice were compared in pristane-treated (left) vs.n-hexadecane (right) -treated mice.

antibodies (Fig. 6A and B). The levels of anti-Suin n-hexadecane-treated mice were comparable tomost of pristane-induced anti-Su. However, levels ofanti-nRNP/Sm in sera fromn-hexadecane-treated micewere∼10–100-fold lower than those in pristane-treated

mice (p < 0.05, by Mann–Whitney), consistent with theweak signal by immunoprecipitation (Fig. 6A). Therelationship between the levels of anti-nRNP/Sm versusanti-Su in individual mice was compared in pristaneor n-hexadecane-treated mice (Fig. 6C). There were

194 Y. Kuroda et al. / Toxicology 218 (2006) 186–196

mice that produced high levels of both specificities.However, other mice produced anti-nRNP/Sm withoutanti-Su (dots onY-axis) and others produced anti-Subut not anti-nRNP/Sm (dots onX-axis) in both pristaneand n-hexadecane-treated groups, consistent with thedifferent mechanisms of regulation in anti-nRNP/Smversus anti-Su antibodies.

4. Discussion

Low molecular weight hydrocarbons such as IFA(C15–C25) and pristane (C19) have adjuvant effects andinduce chronic arthritis in mice and rats (Whitehouse etal., 1974). We have recently reported that i.p. injectionof IFA and squalene (MF59 adjuvant, 2,6,10,15,19,23-hexamethyl-2,6,10,14,18,22-tetracosahexaene), but nothigh molecular weight (C25–C40) medicinal mineraloils, induces lupus-related autoantibodies to nRNP/Smand Su in BALB/c mice (Satoh et al., 2003a) simi-lar to the effects of pristane (Satoh et al., 1995; Satohand Reeves, 1994). Although the amount of pristanein IFA was ∼9–67-fold more than that in medici-nal oils, it was only∼0.17% of the total amount ofhydrocarbons in IFA (Kuroda et al., 2004). There-fore, we suspected that the reason why IFA inducesanti-nRNP/Sm and -Su is because other hydrocarbonsfound in the IFA can also induce lupus autoantibod-ies. In the present study, we confirmed that at leastone additional adjuvant hydrocarbon induces autoim-munity, namely the C16 hydrocarbonn-hexadecane

,

in-deis

tes-,

d,edper-rys;

n-ntainanes

(Singh et al., 2003) have received increased attentionin recent years (Harris et al., 2002; Peden-Adam et al.,2001; Ullrich and Lyons, 2000). Jet fuels can be inhaledor absorbed through skin and cause local as well as sys-temic toxicity (Allen et al., 2001; Ramos et al., 2002;Singh et al., 2003; Ullrich, 1999). n-Hexadecane wasidentified as one of the toxic components in JP-8 jetfuel (Singh et al., 2003). However, other than its adju-vant effects (Crowle and Hu, 1966) and arthritogenic-ity (Whitehouse et al., 1974), little is known about theimmunological effects ofn-hexadecane.

In the present study, non-specific inflammationcaused byn-hexadecane induced autoantibodies to alimited set of cellular proteins and protein–nucleic acidcomplexes such as snRNPs (nRNP and Sm), Su, andchromatin. It should be noted that the cell extract thatcontains virtually all cellular proteins radiolabeled with35S-methionine is immunoprecipitated by mouse serumin immunoprecipitation assay. Sera from pristane orhexadecane-treated mice immunoprecipitated only a fewselected proteins or protein–nucleic acids complexes(snRNPs, Su, NF45/90/110) out of thousands of pro-teins as shown inFig. 5. Furthermore, other autoan-tibodies that are readily detectable by this techniqueand are found frequently in human lupus, such as anti-Ro and anti-La, are absent in hexadecane-treated micewhile anti-NF45/90/110 that is very rare (<0.1%) inhuman lupus (Satoh et al., 1999) is found in 17% ofthe mice in this group. All these observations are consis-tent with the interpretation thatn-hexadecane induces

ownucedtaneadev-velsoup.

ibod-ns,

,

c-leriti-car-hory

ctson-an

oan-

(Whitehouse et al., 1974), which was found∼4–40-foldmore in IFA than medicinal mineral oils (Kuroda et al.2004).

Like pristane,n-hexadecane from petroleum and meral oils (Kuroda et al., 2004) contaminates soil anmarine water (Coates et al., 1997). Dietary exposurto mineral hydrocarbons containingn-hexadecanealso quite common (Heimbach et al., 2002). Whentaken orally,n-hexadecane is absorbed from the intine and is found in the circulation (Mann et al.1977; McCarthy, 1964). Non-dietary exposure ton-hexadecane includes mineral oil mist (Svendsen anHilt, 1997), gasoline, diesel exhaust (Chan et al.1981), jet fuel (Singh et al., 2003), and tobacco smok(Van Duuren and Goldschmidt, 1976). When absorbethrough skin, it causes local inflammation and hykeratosis (Rossmiller and Hoekstra, 1966). Inhaled ointradermally injectedn-hexadecane preferentially stain lymph nodes for a prolonged period (Chan et al., 1981Kleinau et al., 1995). The immunotoxicity and potetial health hazards of jet fuels such as JP-8 that con-hexadecane and other low molecular weight alk

a limited set of autoantibodies selected via unknmechanisms. The specificities of autoantibodies indbyn-hexadecane was similar to those induced by prisin general. However,n-hexadecane-treated mice hhigher levels of anti-chromatin antibodies but lower lels of anti-snRNPs while producing comparable leof anti-Su antibodies versus the pristane-treated grThese data suggest that different types of autoanties are regulated differently by various hydrocarboconsistent with our previous studies (Kuroda et al.2004).

It is highly likely that biological changes or fators from peritoneal inflammation play a critical roin the autoimmune process in this model. One ccal step is probably the adjuvant effects of hydrobon oils involving toll-like receptors (TLRs), whicinduce proinflammatory cytokines and costimulatmolecules (Kaisho and Akira, 2002). This may beinduced either by the hydrocarbon itself or by produor changes resulting from hydrocarbon-induced nspecific inflammation. The next step must involveantigen-specific process since the specificity of aut

Y. Kuroda et al. / Toxicology 218 (2006) 186–196 195

tibodies produced in these mice is highly limited tosnRNPs (nRNP and Sm) and Su, despite exposure tovarious self-proteins from dead cells at the site of inflam-mation. The mechanisms of antigen selection may beantigen-specific post-translational modifications as sug-gested in the murine model of induction of autoantibod-ies specific for fibrillarin by mercuric chloride (Pollard etal., 2000). Certain ribonucleoproteins complex may behighly immunogenic because the nucleic acid compo-nent of ribonucleoprotein complex stimulates TLRs ofantigen-presenting cells while allowing processing andpresentation of the protein components of the ribonu-cleoproteins (Leadbetter et al., 2002). It should alsobe noted that the specificity of the autoantibodies pro-duced in these mice is very similar to those in humansand mice with spontaneous lupus (Satoh et al., 1994,1996b). Determining the mechanisms of autoantibodyspecificity in hydrocarbon-induced autoimmunity mod-els may provide insight into the clinical associations ofautoantibodies.

One limitation of the mouse model is simulating thelifetime chronic human hydrocarbon exposure by theshort-term exposure in mice that live only for 2 years.The route of administration is another important factorthat needs to be considered. i.p. injection was used in thepresent study to compare the immunological effects ofn-hexadecane with those of pristane and other hydro-carbons in our previous studies (Satoh et al., 2003a;Satoh and Reeves, 1994), although it was not a physi-ological route when considering human exposure. Orala ll asi n ofi ies.D therl oseo ro-c e isq t thei mt thes manyh ateda he

hen on-a arer iref no-l nsi oni-t

Acknowledgements

This study was supported by NIH grants R21-AR050661, R01-AR44731 and AI44074. We would liketo thank Ms. Minna Honkanen-Scott for technical assis-tance and Ms. Gina D. Tonogbanua for editorial assis-tance. This material is the result of work supported withresources and the use of facilities at the Malcom RandallVA Medical Center, Gainesville, FL.

References

Allen, D.G., Riviere, J.E., Monteiro-Riviere, N.A., 2001. Analysis ofinterleukin-8 release from normal human epidermal keratinocytesexposed to aliphatic hydrocarbons: delivery of hydrocarbons tocell cultures via complexation with alpha-cyclodextrin. Toxicol. InVitro 15, 663–669.

Chan, T.L., Lee, P.S., Hering, W.E., 1981. Deposition and clearance ofinhaled diesel exhaust particles in the respiratory tract of Fischerrats. J. Appl. Toxicol. 1, 77–82.

Coates, J.D., Woodward, J., Allen, J., Philp, P., Lovley, D.R., 1997.Anaerobic degradation of polycyclic aromatic hydrocarbons andalkanes in petroleum-contaminated marine harbor sediments.Appl. Environ. Microbiol. 63, 3589–3593.

Crowle, A.J., Hu, C.C., 1966. A comparison ofn-hexadecane and min-eral oil emulsions in induction of hypersensitivity in mice. Proc.Soc. Exp. Biol. Med. 123, 94–97.

Cruickshank, B., 1984. Follicular (mineral oil) lipidosis. I. Epidemi-ologic studies of involvement of the spleen. Human Pathol. 15,724–730.

Dincsoy, H.P., Weesner, R.E., MacGee, J., 1982. Lipogranulomas innon-fatty human livers: a mineral oil induced environmental dis-ease. Am. J. Pathol. 78, 35–41.

002.hich

hen,min-tates.

tors.

is in-

N.J.,une

. 78,

.C.,IgGl-like

n in

sat-

, J.,n of

dministration simulating dietary exposure as wenhalation and skin exposure simulating the conditiondustrial workers will be considered in future studetermining the physiologically relevant dose is ano

imitation since little data is available on route and df human exposure ton-hexadecane and other hydarbons. Although the 0.5 ml i.p. dose given to micuite large, the present study clearly showed tha

mmunological effects ofn-hexadecane are distinct frohose of medicinal mineral oils when given to mice iname manner. It also supports our hypothesis thatydrocarbons have the ability to induce lupus-relutoantibodies like pristane (Kuroda et al., 2004; Satot al., 2003a).

This is the first report showing the ability of t-hexadecane to induce lupus-like autoimmunity in nutoimmune strains of mice. Whether our findingselevant to humann-hexadecane exposure will requurther studies. The previously unrecognized immuogical effects ofn-hexadecane may have implication the pathogenesis of autoimmune diseases and moring human exposure to hydrocarbons.

Harris, D.T., Sakiestewa, D., Titone, D., Young, R.S., Witten, M., 2JP-8 jet fuel exposure results in immediate immunotoxicity, wis cumulative over time. Toxicol. Ind. Health 18, 77–83.

Heimbach, J.T., Bodor, A.R., Douglass, J.S., Barraj, L.M., CoS.C., Biles, R.W., Faust, H.R., 2002. Dietary exposures toeral hydrocarbons from food-use applications in the United SFood Chem. Toxicol. 40, 555–571.

Kaisho, T., Akira, S., 2002. Toll-like receptors as adjuvant recepBiochim. Biophys. Acta 1589, 1–13.

Kleinau, S., Dencker, L., Klareskog, L., 1995. Oil-induced arthritDA rats: tissue distribution of arthritogenic14C-labelled hexadecane. Int. J. Immunopharmacol. 17, 393–401.

Kuroda, Y., Akaogi, J., Nacionales, D.C., Wasdo, S.C., Szabo,Reeves, W.H., Satoh, M., 2004. Distinctive patterns of autoimmresponse induced by different types of mineral oil. Toxicol. Sci222–228.

Leadbetter, E.A., Rifkin, I.R., Hohlbaum, A.M., Beaudette, BShlomchik, M.J., Marshak-Rothstein, A., 2002. Chromatin-complexes activate B cells by dual engagement of IgM and tolreceptors. Nature 416, 603–607.

Mann, M.D., Pirie, D.J., Wolfsdorf, J., 1977. Kerosene absorptioprimates. J. Pediatr. 91, 495–498.

McCarthy, R.D., 1964. Mammalian metabolism of straight-chainurated hydrocarbons. Biochim. Biophys. Acta 84, 74–79.

Peden-Adam, M.M., Eudaly, J., Eudaly, E., Dudley, A., ZeiglerLee, A., Robbs, J., Gilkeson, G., Keil, D.E., 2001. Evaluatioimmunotoxicity induced by single or concurrent exposure toN,N-

196 Y. Kuroda et al. / Toxicology 218 (2006) 186–196

diethyl-m-toluamide (DEET), pyridostigmine bromide (PYR), andJP-8 jet fuel. Toxicol. Ind. Health 17, 192–209.

Pollard, K.M., Pearson, D.L., Bluthner, M., Tan, E.M., 2000. Prote-olytic cleavage of a self-antigen following xenobiotic-induced celldeath produces a fragment with novel immunogenic properties. J.Immunol. 165, 2263–2270.

Potter, M., Wax, J.S., 1983. Peritoneal plasmacytomagenesis in mice:comparison of different pristane dose regimens. J. Natl. CancerInst. 71, 391–395.

Ramos, G., Nghiem, D.X., Walterscheid, J.P., Ullrich, S.E., 2002. Der-mal application of jet fuel suppresses secondary immune reactions.Toxicol. Appl. Pharmacol. 180, 136–144.

Riviere, J.E., Brooks, J.D., Monteiro-Riviere, N.A., Budsaba, K.,Smith, C.E., 1999. Dermal absorption and distribution of topicallydosed jet fuels jet-A, JP-8, and JP-8(100). Toxicol. Appl. Pharma-col. 160, 60–75.

Rossmiller, J.D., Hoekstra, W.G., 1966. Hexadecane-induced hyperk-eratinization of guinea pig skin. J. Invest. Dermatol. 47, 39–43.

Satoh, M., Hamilton, K.J., Ajmani, A.K., Dong, X., Wang, J., Kanwar,Y., Reeves, W.H., 1996a. Induction of anti-ribosomal P antibodiesand immune complex glomerulonephritis in SJL mice by pristane.J. Immunol. 157, 3200–3206.

Satoh, M., Kumar, A., Kanwar, Y.S., Reeves, W.H., 1995. Antinuclearantibody production and immune complex glomerulonephritis inBALB/c mice treated with pristine. Proc. Natl. Acad. Sci. U.S.A.92, 10934–10938.

Satoh, M., Kuroda, Y., Yoshida, H., Behney, K.M., Mizutani, A.,Akaogi, J., Nacionales, D.C., Lorenson, T.D., Rosenbauer, R.J.,Reeves, W.H., 2003a. Induction of lupus-related autoantibodies byadjuvants. J. Autoimmun. 21, 1–9.

Satoh, M., Langdon, J.J., Chou, C.H., McCauliffe, D.P., Treadwell,E.L., Ogasawara, T., Hirakata, M., Suwa, A., Cohen, P.L., Eisen-berg, R.A., Reeves, W.H., 1994. Characterization of the Su antigen,a macromolecular complex of 100/102 and 200 kDa proteins rec-ognized by autoantibodies in systemic rheumatic diseases. Clin.

nka,uneto U197,

J.,ves,pro-

duce lupus-related anti-RNA helicase A autoantibodies, but areresistant to pristane-induced lupus. Int. Immunol. 15, 1117–1124.

Satoh, M., Reeves, W.H., 1994. Induction of lupus-associated autoan-tibodies in BALB/c mice by intraperitoneal injection of pristane.J. Exp. Med. 180, 2341–2346.

Satoh, M., Shaheen, V.M., Kao, P.N., Okano, T., Shaw, M., Yoshida,H., Richards, H.B., Reeves, W.H., 1999. Autoantibodies define afamily of proteins with conserved double- stranded RNA-bindingdomains as well as DNA binding activity. J. Biol. Chem. 274,34598–34604.

Satoh, M., Weintraub, J.P., Yoshida, H., Shaheen, V.M., Richards, H.B.,Shaw, M., Reeves, W.H., 2000. Fas and Fas ligand mutations inhibitautoantibody production in pristane-induced lupus. J. Immunol.165, 1036–1043.

Shaheen, V.M., Satoh, M., Richards, H.B., Yoshida, H., Shaw,M., Jennette, J.C., Reeves, W.H., 1999. Immunopathogenesis ofenvironmentally-induced lupus in mice. Environ. Health Perspect.107 (Suppl. 5), 723–727.

Singh, S., Zhao, K., Singh, J., 2003. In vivo percutaneous absorption,skin barrier perturbation, and irritation from JP-8 jet fuel compo-nents. Drug Chem. Toxicol. 26, 135–146.

Svendsen, K., Hilt, B., 1997. Exposure to mineral oil mist and respira-tory symptoms in marine engineers. Am. J. Ind. Med. 32, 84–89.

Ullrich, S.E., 1999. Dermal application of JP-8 jet fuel induces immunesuppression. Toxicol. Sci. 52, 61–67.

Ullrich, S.E., Lyons, H.J., 2000. Mechanisms involved in the immuno-toxicity induced by dermal application of JP-8 jet fuel. Toxicol.Sci. 58, 290–298.

Van Duuren, B.L., Goldschmidt, B.M., 1976. Cocarcinogenic andtumor-promoting agents in tobacco carcinogenesis. J. Natl. CancerInst. 56, 1237–1242.

Vingsbo, C., Sahlstrand, P., Brun, J.G., Jonsson, R., Saxne, T., Holm-dahl, R., 1996. Pristane-induced arthritis in rats. A new modelfor rheumatoid arthritis with a chronic disease course influ-enced by both major histocompatibility compledx and non-major

75–

Fre-icity

989.tureshritis

Immunol. Immunopathol. 73, 132–141.Satoh, M., Langdon, J.J., Hamilton, K.J., Richards, H.B., Pa

D., Eisenberg, R.A., Reeves, W.H., 1996b. Distinctive immresponse patterns of human and murine autoimmune serasmall nuclear ribonucleoprotein C protein. J. Clin. Invest.2619–2626.

Satoh, M., Mizutani, A., Behney, K.M., Kuroda, Y., Akaogi,Yoshida, H., Nacionales, D.C., Hirakata, M., Ono, N., ReeW.H., 2003b. X-linked immunodeficient mice spontaneously

histocompatibility complex genes. Am. J. Pathol. 149, 161683.

Whitehouse, M.W., Orr, K.J., Beck, F.W.J., Pearson, C.M., 1974.und’s adjuvants: relationship of arthritogenicity and adjuvantin rats to vehicle composition. Immunology 27, 311–330.

Wooley, P.H., Seibold, J.R., Whalen, J.D., Chapdelaine, J.M., 1Pristane-induced arthritis. The immunologic and genetic feaof an experimental murine model of autoimmune disease. ArtRheum. 32, 1022–1030.