Lupus-Prone Mice as Models to Study Xenobiotic-Induced Accelerationof Systemic AutoimmunityK. Michael Pollard,1 Deborah L. Pearson,1 Per Hultman,3 Bernhard Hildebrandt,1 and Dwight H. Kono2

1Department of Molecular and Experimental Medicine, 2Department of Immunology, The Scripps Research Institute, La Jolla, California USA;3Division of Molecular and Immunological Pathology, Department of Health and Environment, Linkoping University, Linkoping, Sweden

The linkage between xenobiotic exposures and autoimmune diseases remains to be clearly defined.However, recent studies have raised the possibility that both genetic and environmental factors actsynergistically at several stages or checkpoints to influence disease pathogenesis in susceptiblepopulations. These observations predict that individuals susceptible to spontaneous autoimmunityshould be more susceptible following xenobiotic exposure by virtue of the presence of predisposingbackground genes. To test this possibility, mouse strains with differing genetic susceptibility tomurine lupus were examined for acceleration of autoimmune features characteristic of spontaneoussystemic autoimmune disease following exposure to the immunostimulatory metals nickel andmercury. Although NiCI2 exposure did not exacerbate autoimmunity, HgCI2 significantly acceleratedsystemic disease in a strain-dependent manner. Mercury-exposed (NZB x NZW)F1 mice hadaccelerated lymphoid hyperplasia, hypergammaglobulinemia, autoantibodies, and immune complexdeposits. Mercury also exacerbated immunopathologic manifestations in MRL+/+ and MR -Iprmice. However, there was less disease acceleration in lpr mice compared with MRL+/+ mice, likelydue to the fact that environmental factors are less critical for disease induction when there is stronggenetic susceptibility. Non-major histocompatability complex genes also contributed to mercury-exacerbated disease, as the nonautoimmune AKR mice, which are H-2 identical with the MRL,showed less immunopathology than either the MRL//pr or MRL+/+ strains. This studydemonstrates that genetic susceptibility to spontaneous systemic autoimmunity can be apredisposing factor for HgCI2-induced exacerbation of autoimmunity. Such genetic predispositionmay have to be considered when assessing the immunotoxicity of xenobiotics. Additionalcomparative studies using autoimmune-prone and nonautoimmune mice strains with differentgenetic backgrounds will help determine the contribution that xenobiotic exposure makesin rendering sensitive populations susceptible to autoimmune diseases. Key words: animalmodel, autoimmunity, lupus, mercury, nickel, xenobiotic. - Environ Health Perspect 107(suppl 5):729-735 (1999).http.//ehpnetl.niehs.nih.gov/docs/1999/suppl-5/729-735pollard/abstract.html

Analysis of the role of xenobiotics inautoimmunity, particularly in humans, ishampered by identification of potentiallysensitive populations. Epidemiologic studieshave revealed associations between xenobi-otic exposure and certain systemic autoim-mune diseases (1,2) but do not identifysusceptible individuals. Conversely,although it is appreciated that systemicautoimmune diseases are under multigeniccontrol (3), identification of the genetic ele-ments involved and whether they respondto environmental exposures is in its infancy.However, recent studies have raised the pos-sibility that both genetic and environmentalfactors act synergistically at several stages orcheckpoints of disease pathogenesis (3).This exacerbating role for environmentalagents has been suggested by a number ofstudies that have shown that exogenousagents can accelerate the onset of auto-immunity in animal models (4-7). Thesestudies predict that individuals susceptibleto spontaneous autoimmunity should bemore susceptible following xenobiotic expo-sure by virtue of the presence of predispos-ing background genes. Such accelerated

disease would be expected to more closelyresemble the immunologic features of idio-pathic autoimmunity rather than thatinduced by xenobiotic exposure in a non-autoimmune-prone host.

The availability of animal models ofsystemic autoimmune disease provides avaluable resource with which to studythe responses of a population sensitive toxenobiotic-induced acceleration of auto-immunity. Among the best-known animalmodels of human systemic autoimmune dis-ease are those mouse strains that sponta-neously develop features of systemic lupuserythematosus (SLE) (3,8,9). Each of thelupus-prone strains has been derived fromdifferent genetic backgrounds, which resultsin both quantitative and qualitative differ-ences in disease expression. In some casesspecific features of autoimmune disease havebeen ascribed to individual genetic defects inindividual strains (9). Lupus-prone strainsthus provide unique genetic backgrounds forthe study of xenobiotic influences on theconstellation of immunologic features thatcomprise systemic autoimmunity. Thestrains used in this study include (NZB x

NZW)F1 (NZBWF1) (H-2dIz) female mice,which due to the influence of sex hormonesdevelop systemic autoimmunity earlier inlife than their male counterparts (3,8). Alsoexamined were MRL (H-2k) mice, whichconsist of two substrains, MRLIIpr andMRL+/+ (3,8). In the MRLlIpr, autoimmu-nity occurs very early in life and is associatedwith massive lymphoproliferation due toabnormal Fas (CD95) expression. In con-trast, the MRL+/+ substrain, which has nor-mal Fas expression, develops less severesystemic autoimmune disease, with onsetconsiderably later in life than the MRL//pr.In order to compare the influence of non-major hisocompatability complex (MHC)genes on xenobiotic acceleration of autoim-munity, the nonautoimmune strain AKR(H-2k) from which the MRL substrainsderive their H-2 was also tested.

Xenobiotics selected for study were twometals, mercury and nickel, known to pro-duce distinctly different immunologicresponses. Mercury induces autoimmunityin mice (10) and rats (11) and an immune-mediated kidney disease in humans (12).Mercury-induced autoimmunity (HgIA) inrodents is associated with three major patho-logic sequelae-lymphoproliferation, hyper-gammaglobulinemia, and the developmentof autoimmunity-manifested as autoanti-body production and immune complex dis-ease (10,11). Elicitation of these pathologicfeatures is dependent on genetic back-ground, with lymphadenopathy occurring inmost strains; autoimmunity is morerestricted, being controlled in large part bythe MHC (10). HgIA was initially thought

This article is based on a presentation at the Workshopon Linking Environmental Agents and AutoimmuneDiseases held 1-3 September 1998 in ResearchTriangle Park, North Carolina.

Address correspondence to K.M. Pollard, Dept.of Molecular and Experimental Medicine, SBR6,10550 North Torrey Pines Rd., La Jolla, CA 92037.Telephone: (619) 784-9214. Fax: (619) 784-2131.E-mail: mpollard@scripps.eduWe thank R.L. Rubin, The Scripps Research

Institute, La Jolla, CA, for providing chromatin prepa-rations. This publication was made possible byNational Institute of Environmental Health Sciencesgrants ES0751 1, ES08080, and ES08666; SwedishMedical Research Council grant 09453; and DeutscheForschungsgemeinschaft Fellowship Hi 618-1/1. Thisis publication number 12149 MEM from The ScrippsResearch Institute, La Jolla, CA.

Received 15 January 1999; accepted 28 April 1999.

Environmental Health Perspectives * Vol 107, Supplement 5 * October 1999 729

POLLARD ET AL.

to be a CD4+ T-helper (Th)2 responsebecause of increased levels of predominantlyinterleukin (IL)-4, and IgGI autoantibodiesafter HgCl2 exposure (13). However allthe features of systemic autoimmunity(hypergammaglobulinemia, autoantibodies,immune complex deposits) are now knownto be dependent upon the Thl cytokineinterferon (IFN)-y (14).

In contrast, exposure to nickel is associatedwith allergic rather than autoimmune reac-tions. Nickel elicits increases in nitric oxideproduction (15), activation of nuclear factor-KB (16), and delayed-type hypersensitivity(DTH) reactions in humans (17) and rodents(18). Nickel-induced DTH is characterizedby the presence of IFN-y-producing Thl andThO CD4+ T cells (19).

In this study, particular emphasis wasplaced on the potential of these xenobioticsto exacerbate the natural course of systemicautoimmune disease in lupus-prone mice asrevealed by lymphadenopathy, elevations ofserum immunoglobulin, appearance ofautoantibodies, and immune complexdeposits. Although NiCl2 exposure had lit-tle effect on autoimmunity, mercury signifi-cantly accelerated systemic disease in astrain-dependent manner. Thus, mercuryexposure exacerbated almost all features ofautoimmunity in NZBWFI mice. Mercuryalso exacerbated immunopathologic featuresin both MRL substrains, although less effecton the humoral response was observed inmice with the lpr mutation. Non-MHC andnon-lpr genes contributed to the mercury-exacerbated disease in the MRL, as H-2identical nonautoimmune AKR miceshowed less immunopathology. Thesestudies suggest that autoimmune-pronemice are useful models for nonhumanstudies on the effects of xenobiotics onsensitive populations.

Materials and MethodsMiceFemale NZBWFI (H-2d/z), MRL-FaslPr(MRL/Ipr) (H-2k), MRL/Fas+1+ (MRL+/+)(H-2k), and AKR/J (H~2k) mice wereobtained from The Scripps Research InstituteAnimal Colony (La Jolla, CA) and main-tained under specific pathogen-free condi-tions. All experimental procedures usinganimals followed the guidelines set down inthe National Institutes of Health Guide forthe Care and Use ofLaboratory Animals (20).

Treatment ofMiceGroups of up to eight mice (4 weeks of age)were injected subcutaneously twice perweek for 4 weeks with 100 pL phosphate-buffered saline (PBS) containing 40 pgHgCl2, 40 pg NiCl2, or PBS alone. Mice

were bled before the first injection andsacrificed on day 30; serum was stored at-70°C until use. Autopsies were performedas previously described (21). Samples ofkidney and spleen were harvested for analy-sis of immune complex deposition and his-tology (see "Tissue Complex Deposits" and"Light Microscopy").

Detection ofSerum AntibodyAntibodies nuclear antigens (ANA) weredetected using HEp-2 cell slides (BionEnterprises, Park Ridge, IL) as previouslydescribed (22). Prior to assay, sera werediluted 100-fold in PBS containing 0.1%bovine gamma globulin (BGG), 0.5%bovine serum albumin (BSA), 0.001%gelatin, and 0.05% Tween 20. Goat anti-mouse IgG-fluorescein isothiocyanate(FITC) (Caltag Laboratories, South SanFrancisco, CA) diluted 100-fold in PBS con-taining 0.5% BGG, 0.1% BSA, and 0.05%Tween 20 was used as detecting reagent.Antifibrillarin (nucleolar) monoclonal anti-body 72B9 (23) was used as control.Intensity of fluorescence was graded on ascale of 0-4+, with 1 + considered positive.

Antichromatin antibodies were detectedby enzyme-linked immunosorbent assay(ELISA) as previously described (24). Serawere diluted 100-fold prior to assay, and chro-matin-bound antibodies detected with horse-radish peroxidase (HRP)-conjugated goatantimouse IgG (Caltag Laboratories) diluted2,000-fold. Antichromatin monoclonal anti-body ID12 was used as positive control (25).

Serum Immunoglobulin QuantitationSerum IgG, IgG1, and IgG2a levels werequantified by ELISA, essentially as previ-ously described (26). ELISA plates werecoated with 200 pL of 2 pg/mL goat anti-mouse kappa light chain antibody (CaltagLaboratories) in PBS, and incubatedovernight at 4°C. Plates were postcoated for1 hr with 0.1% gelatin in PBS, followed bythree washes with PBS-0.05% Tween 20.Sera were diluted in serum diluent (26), andincubated at room temperature while shak-ing for 2.5 hr, followed by three washes withPBS-0.05% Tween 20. HRP-conjugatedgoat antimouse IgG, IgG1, or IgG2a anti-bodies (Caltag Laboratories) diluted in anti-immunoglobulin diluent (26) were addedbefore incubation with shaking for 90 min.After three washes with PBS-0.05% Tween20 and four washes with PBS, 2,2'-azinobis(3-ethylbenzthiazolinesulfonic acid)(ABTS) substrate solution was added andoptical density read at 405 nm (OD405).Standard curves were generated by serialdilutions of polyclonal mouse referenceserum containing predetermined levels ofimmunoglobulin isotypes and subclasses

(The Binding Site, Birmingham, UK).Serum IgG, IgGI, and IgG2a concen-trations were quantified by converting theaverage OD405 of duplicate wells toimmunoglobulin concentrations by use ofstandard curves. Only serum dilutions thatgave OD405 values falling within the linearportion of the standard curve were used tocalculate immunoglobulin levels.

Tissue Immune Complex DepositsSections 2- to 3-mm thick of the kidney andspleen were snap frozen in isopentane-C02and prepared for direct immunofluorescenceas previously described (27). Cryostat sections4- to 5-pm thick were fixed in ethanol andincubated with serial dilutions of FITC-con-jugated goat antibodies to IgG (,y-chain spe-cific) or C3 (Southern BiotechnologyAssociates, Birmingham, AL). The end-pointtiter of the immune complex deposits wasdefined as the highest dilution of antibody atwhich specific fluorescence could be detectedand was expressed as the reciprocal titer. Ascore of 0 was recorded when no specific fluo-rescence was detected at a dilution of 1/40.The presence of granular deposits in small andmedium-size arteries was also examined. Allhistology slides were examined withoutknowledge of treatment given or other results.

Light MicrspyA 2- to 3-mm thick section of kidney wasimmersed in Histochoice fixative (Amresco,Solon, IL), embedded in paraplast, and cutinto 1- to 2-pm sections that were stainedwith periodic acid Schiff's reagent and withperiodic acid silver methenamine. Slides werethen examined by light microscopy withoutknowledge of treatment given or otherresults. The types of glomerular alterationwere determined and the degree of endocapil-lary cell hyperplasia was scored for each ani-mal as follows: 0 = normal, 0.5 = justdetectable alteration, 1 = slight, 2 = moderate,3 = strong, and 4 = maximal.

Statistical AnalysisGroups were compared by single-factorANOVA, Mann-Whitney Utest, or Fisher'sexact test, as appropriate. Comparisons are ofHgC12-treated mice with PBS- and NiCI2-treated animals unless described otherwise.p < 0.05 was considered significant.

ResultsEffct ofHlgCI2 and NiC12 on FearsofAutoimmu nit in NZBWF1 MiceMercury-exposed NZBWFI mice had ele-vated serum IgG, IgG1, and IgG2a comparedto PBS- and NiCl2-exposed animals (Table 1,Figure 1). Antibody to nuclear antigens(ANA) (Table 1) of a predominantly dense

Environmental Health Perspectives * Vol 107, Supplement 5 * October 1999730

XENOBIOIC-ACCELERATED AUTOIMMUNITY

fine to homogenous nuclear speckling ofinterphase cells and metaphase chromo-somes was found in all NZBWF1 miceexposed to HgC12. All pretreatment sera aswell as PBS- and NiCI2-treated mice showedless frequent ANA responses. Levels ofantichromatin antibodies (Table 1, Figure 1)in HgCl2-exposed NZBWF1 mice weremarkedly elevated above those found ineither NiCl2- or PBS-exposed animals.

Exposure of NZBWF1 mice to mercuryincreased organ wet weight of spleen andcervical lymph nodes but not mesentericlymph nodes when compared to PBS- andNiCl2-treated animals (Table 2). All HgC12-treated mice developed very high titers ofIgG and C3 deposits in glomeruli (Table 2).Granular IgG deposits in the glomeruli werelocalized to the capillary loops in 3/8 mice,restricted to the mesangium in one mouse,

Table 1. Immunoglobulin levels and autoantibodies in NZBWF1 mice following metal exposure.ab

Immunoglobulin levelStrain lgG IgGl lgG2a ANA Antichromatin(sex) No. Treatment (mg/mL) (mg/mL) (mg/mL) (pos/no.) (OD405)NZBWF1 8 Pre-PBS 1.44 ± 1.07 0.20 ± 0.08 0.47 ± 0.22 0/8 0.02 ± 0.01(female) Post-PBS 2.64 ± 0.94 0.30 ± 0.08 1.50 ± 0.50 0/8 0.06 ± 0.05

8 Pre-NiCI2 1.69 ± 0.65 0.28 ± 0.06 0.74 ± 0.34 0/8 0.04 ± 0.02Post-NiCI2 3.17 ± 1.38 0.30 ± 0.06 2.06 ± 0.86 1/8 0.04 ± 0.03

8 Pre-HgCI2 1.56 ± 0.05 0.41 ± 0.11 0.96 ± 0.25 0/8 0.06 ± 0.04Post-HgCI2 10.58 ± 2.351 3.06 ± 1.54+ 8.10 ± 1 .70 8/8 1.27 ± 0.651

Abbreviations: OD4,s, optical density read at 405 nm; PBS, phosphate-buffered saline; pos/no., positive/number of animals. 'p-valuesare from comparison of mercury-treated group with PBS- and nickel-treated groups. +p < 0.0002; tp < 0.0001. bData are expressed asmean ± standard deviation.

IgG E-E-

IgGl EE

ElgG2a -.E

20-18-16-14-12-,10-8-6-4-2-

5-

4-

* 3-

2-

00

t00

a&A'

1-0

U . ~

10 -

8-

6-

4- 40

2-

~a& w

12 -

10 - b8 - Ja

86-

4- 0

0~~~~~~02- B Ni H

PBS Ni Hg

0

Ii

PBS Ni Hg

Figure 1. Hypergammaglobulinemia andantichromatin antibodies in NZBWF1 mice afterexposure to mercury. OD405, optical densityread at 405 nm; PBS, phosphate-bufferedsaline. Female NZBWF, mice were injectedwith HgCI2, NiCI2, or PBS for 4 weeks asdescribed in "Materials and Methods." SerumIgG, IgGl, and IgG2a (left panels) were deter-mined by enzyme-linked immunosorbent assay(ELISA) before (o) and after (c) metal expo-sure. lgG antichromatin antibodies (right panel)were determined by ELISA before (o) and after(.) metal exposure. See Table 1 for statisticalanalysis.

and present along the capillary loops as wellas in the mesangium in the remaining fourmice. Granular deposits of C3 were found inthe mesangium in 6/8 animals, whereas 2/8mice showed C3 deposits along the capillaryloops as well as in the mesangium. Renal andsplenic vessel walls showed granular depositswith high titers of IgG and C3. In contrast,mice treated with NiCl2 or PBS had muchlower titers of IgG in the glomerularmesangium in 2/8 and 3/8 mice, respec-tively, without any deposits along the capil-lary loops. The titers of C3 in the glomeruliwere also significantly lower in these mice.None of the NiCl2- or PBS-treated animalsshowed deposits of IgG or C3 in the renalor splenic vessel walls. Histologically therewas only slight endocapillary cell hyperpla-sia in the glomeruli, which did not differamong the three treatment groups, probablybecause of the young age of mice at sacri-fice. There were no alterations in theglomerular basement membrane among thethree treatment groups.

Effect ofHgC12 and NiCI2 on FeaturesofAutoimmunity in MRIJWpr MiceMetal exposure did not lead to elevations ofimmunoglobulin in MRL/Ipr mice, as allthree treatment groups developed hyper-gammaglobulinemia during the experimen-tal period (Table 3, Figure 2). Metal- andsaline-exposed MRL/Ipr mice showed nosignificant differences in ANA andantichromatin antibody at the end of thetreatment period; each of the groups devel-oped antichromatin antibodies during theexposure period (Table 3, Figure 3).

Mercury-exposed MRL/Ipr mice hadlarger mesenteric lymph nodes than PBS- orNiCI2-treated animals but showed no differ-ence in cervical node or spleen weights(Table 4). Mercury treatment increased theincidence of glomerular, mesangial IgGdeposits in all MRLIIpr mice compared with5/8 mice treated with NiCl2 and only 2/8PBS-treated animals (Table 4). Mercury sig-nificantly increased the titer of mesangial IgGand C3 deposits (Table 4). No vessel walldeposits were seen in any of the animals.There was a mild endocapillary cell hyperpla-sia in the glomeruli, which tended to be morepronounced in the HgCl2-treated group, butthis did not reach statistical significance.

Table 2. Pathologic changes in NZBWF, mice following metal exposure.abcOrgan wet weight (mg) Kidney immunopathology Spleen immunopathology

Strain Cervical Mesenteric Glomerular Glomerular Vessel Vessel Vessel Vessel(sex) No. Treatment Spleen LN LN lgG C3 IgG C3 lgG C3

NZBWF1 8 PBS 87±3 32±2 107±5 25±37 480±171 0 0 0 0(female) 8 NiCI2 89 ± 3 29 ± 3 107 ± 8 20 ± 37 330 ± 151 0 0 0 0

8 HgCI2 129 ±5 70 ± 61 97 ± 6 1760 ± 6620 1440 ± 4530 1600 ± 5930 1160 ± 3390 1280 ± 593 1400 ± 7460

PBS, phosphate-buffered saline. "p-values are from comparison of mercury-treated group with PBS- and nickel-treated groups. #p < 0.001; tp < 0.0001. hData are expressed as mean ± standard deviation.9gG and C3 titers are expressed as reciprocal titers.

Environmental Health Perspectives * Vol 107. Supplement 5 * October 1999

n u _- _n

731

POLLARD ETAL

Table 3. Immunoglobulin levels and autoantibodies in MRL mice following metal exposure.abImmunoglobulin level Anti-

Strain lgG IgGl IgG2a ANA chromatin(sex) No. Treatment (mg/mL) (mg/mL) (mg/mL) (pos/no.) (OD40)MRL/lpr 8 Pre-PBS 2.71 ± 2.10 1.03 ± 0.67 3.45 ± 2.74 0/8 0.02 ± 0.05(female) Post-PBS 9.58 ± 4.44 3.66 ± 1.25 6.06 ± 2.85 4/8 0.35 ± 0.43

8 Pre-NiCI2 2.11+ 0.71 0.82 + 0.45 1.88 + 0.83 1/8 0.01+ 0.01Post-NiCI2 10.41 ± 3.17 3.45 ± 1.02 4.48 ± 1.56 8/8 0.70 ± 0.46

8 Pre-HgCI2 1.30 ± 0.33 0.79 ± 0.16 1.25 ± 0.54 1/8 0.03 ± 0.01Post-HgCI 8.65 ± 3.38 4.07 ± 2.53 4.62 ± 1.46 7/8 0.56 ± 0.43

MRL+/+ 8 Pre-PBS 2.93 ± 4.32 1.17 ± 2.15 1.45 ± 2.47 5/8 0.10 ± 0.15(female) Post-PBS 4.06 ± 2.71 1.45 ± 1.40 2.01 ± 2.04 7/8 0.17 ± 0.16

8 Pre-NiCI2 1.71 ±1.14 0.46 ± 0.27 0.64 ± 0.31 6/8 0.57 ± 1.20Post-NiCI2 3.92 ± 2.43 1.29 ± 0.86 1.81 ± 1.05 7/8 0.48 ± 0.48

8 Pre-HgCI2 1.44 ± 0.69 0.41 ± 0.14 0.72 ± 0.44 4/8 0.38 ± 0.64Post-HgCI2 7.12 ±1.76§ 3.47 ± 0.87+ 3.68 ±1.84 8/8 1.88 ± 1.71*

Abbreviations: OD4%, optical density read at 405 nm; PBS, phosphate-buffered saline; pos/no., positive/number of animals. 'p-values are from comparison of mercury-treated group with PBS- and nickel-treated groups. *p< 0.05; sp< 0.02; +p< 0.005. &Data are expressed as mean + standard deviation.

20-18-16-14-

i 12

lgG -. 10-E 8-

6-4-

2 -

MRL/lpr

a 0

* a 0

* 0

0

MRL+/+

0

Sb 00

4 oh

$ CD4

0

AKRIJ

CO0

0

8-

6 0

4- a

0 o

0

2- 0 0 t

2 S5t (&bo~o12-

10- O'

8-

00~~~

6- ;

2- 8 0- %0n0 0 02 &

PBS Ni Hg

0

0

S

* a

0

0

0- o

0

- 0 0

0 N

PBS Ni Hg

j

0

0

*0

PBS Hg

Figure 2. Hypergammaglobulinemia in MRL and AKR mice after exposure to mercury. PBS, phosphate-buffered saline. Female MR//pr, MRL+/+ and AKR mice were injected withHgCI2, NiCI2, or PBS for 4 weeks as described in "Materials and Methods." Serum IgG, IgGl, and IgG2a were determined by enzyme-linked immunosorbent assay (ELISA) before(o) and after (c) metal exposure. See Tables 3 and 5 for statistical analysis.

F&c't ofHgC12 and NiCi2 onF s

ofA mmunity in MRL+/+ Mice

Mercury-exposed MRL+/+ mice had elevatedserum IgG and IgGI compared to PBS- andNiCl2-exposed animals, whereas IgG2a levelswere elevated compared to NiCl2-treated mice(Table 3, Figure 2). A variety ofANA patternswere found in MRL+/+ mice before and after

the various treatments (data not shown).Mercury treatment of MRL+/+ mice resultedin elevated levels of antichromatin antibodies,which were significantly different from PBS-and NiCI2-treated animals (Table 3, Figure 3).

Mercury-exposed MRL+/+ mice hadlarger cervical nodes than their controlcounterparts, but spleen weights were not

increased and mesenteric nodes were smaller

(Table 4). All MRL+I+ mice treated withHgCl2 showed glomerular and mesangial IgGdeposits, whereas only 5/8 NiGC2-treated miceand 3/8 PBS-treated mice showed suchdeposits (Table 4). Renal vessel wail depositsof IgG were seen in 7/8 and C3 in 2/8 of theHgC12-treated MRL+/+ mice. Only one

Environmental Health Perspectives * Vol 107. Supplement 5 * October 1999

J

lgG2a EE

1

u

732

XENOBIOT1C-ACCELERATED AUTOIMMUNITY

5 -

4 -

3 -

M 2-

0

0 %

4P

PBS Ni Hg

1 -

0

5-MRL+/+

04-

0

.

0

0

0 A

3-

W 2-0

0

0* 1 -

PBS Ni Hg

AKR/J

PBS Hg

Figure 3. Antichromatin antibodies in MRL and AKR mice after metal exposure. 0D34%, optical density read at 405 nm; PBS, phosphate-buffered saline. Female MRL/Ipr, MRL+/+,and AKR mice were injected with HgCI2, NiCI2, or PBS for 4 weeks as described in "Materials and Methods." lgG antichromatin antibodies were determined by enzyme-linkedimmunosorbent assay (ELISA) before (o) and after (c) metal exposure. See Tables 3 and 5 for statistical analysis.

Table 4. Pathologic changes in MRL mice following metal exposure.abOrgan wet weight (mg) Kidney immunopathologyc

Strain Cervical Mesenteric Glomerular Glomerular Vessel Vessel(sex) No. Treatment Spleen LN [N lgG C3 lgG C3

MRL/Ipr 8 PBS 193±177d 63±21 141±29 60±119 295±406 0 0(female) 8 NiCI2 175±66 83±13 178±11 120±135 220±111 0 0

8 HgCI2 180 ± 89 195 ± 114 227 ± 15§ 470 ± 384*.fe 680 + 267* 0 0MRL+/+ 8 PBS 95±3 57±5 196±17 60±111 270±95 0 0(female) 8 NiCI2 97 ± 6 49 ± 1 182 ± 7d 140 ± 134 300 ± 159 0 0

8 HgCI2 104 ± 5 69 ± 31 135 ± 12+ 300 ± 217*,e 490 ± 398 380+ 410 ,*,e 240 ± 476

PBS, phosphate-buffered saline. Data are expressed as mean ± standard deviation. 'bpvalues are from comparison of mercury-treated group with PBS and nickel-treated groups. *p < 0.05; §p< 0.02; fp <0.01; +p< 0.005; tp< 0.0001. cgG and C3 titers are expressed as reciprocal titers. Sdata from 7 animals. 'p-values for increased incidence of respective pathology.

Table 5. Immunoglobulin levels and autoantibodies in AKR mice following metal exposure.abImmunoglobulin level

Strain lgG IgG1 IgG2a ANA Antichromatin(sex) No. Treatment (mg/mL) (mg/mL) (mg/mL) (pos/no.) (OD405)AKR 8 Pre-PBS 2.05 ± 0.83 0.25 ± 0.12 1.52 ± 0.86 1/8 0.02 ± 0.01(female) 7 Post-PBS 2.84 ± 0.47 0.70 ± 0.32 2.95 ± 0.68 1/7 0.01 ± 0.00

Pre-NiCI2 ND ND ND ND NDPost-NiCI2 ND ND ND ND ND

8 Pre-HgCI2 2.22 ± 0.49 0.38 ± 0.16 1.69 ± 0.30 1/8 0.01 ± 0.017 Post-HgCI2 5.21 ±1.99+ 1.19 ± 0.56 5.44 ± 1.67+ 6/7 0.12 ± 0.01f

Abbreviations: ANA, antibodies to nuclear antigens; ND, not done; 0D14%, optical density read at 405 nm; PBS, phosphate-bufferedsaline; pos/no., position/number of animals. p-values are from comparison of mercury-treated group with PBS and nickel-treatedgroups. fp< 0.01; +p< 0.005. &Data are expressed as mean ± standard deviation.

animal had deposits of Ig and C3 in splenicvessel walls. Vessel wall deposits were notfound in either the kidney or spleen of PBS-and NiCl2-treated animals. Compared withPBS-treated animals, there was significantlyincreased focallsegmental hyperplasia of endo-capillary cells in the mercury group (0.590.23 vs 1.25 ± 0.13; p < 0.001).

EflFct ofHgC02 and NiCI2 on FeaturesofAutoimmunity inAKR Mice

AKR mice also developed hypergamma-globulinemia with increases in IgG andIgG2a following HgCl2 exposure (Table 5).Mercury-treated AKR mice developed ANAbut only low levels of antichromatin antibod-ies (Table 5). Compared to PBS controls,mercury-exposed AKR mice had significantincreases of IgG (51 ± 55 vs 274 ± 78; p <

0.0001) and C3 (34 ± 28 vs 120 ± 52; p <

0.005) deposits in the mesangium of the kid-ney glomeruli. There were no deposits in thekidney vessels or spleen ofAKR mice and no

histologic abnormalities by light microscopy.

DiscussionThe exacerbation by mercury of systemicautoimmunity in lupus-prone mice identifiesa potential model for the study of xenobioticeffects on autoimmune-sensitive populations.The autoimmune features and the extent to

which they were accelerated differed betweenthe stains tested, suggesting that genetic back-ground influenced the response to xenobioticexposure. Much of the disease accelerationappeared to be due to non-MHC lupus-predisposing genes, as an MHC-matchednonautoimmune strain had less severe diseasemanifestations. Strain-dependent responses,as well as the importance of non-MHC genes

in disease severity, suggest that xenobioticexposure led to acceleration of idiopathic dis-ease beyond the induction of HgIA.Acceleration of autoimmunity was dependenton the specific immunostimulant; anothermetal that can activate the immune system,

nickel, had little effect.The most pronounced response to

mercury exposure was observed in femaleNZBWF1 mice, considered to be the proto-

typic murine model for human lupus (11).This strain exhibited lymphadenopathy in thespleen and the cervical lymph nodes, whichdrain the mercury injection site. NZBWF1mice also had pronounced hypergamma-globulinemia. Most striking was the markedelevation in antichromatin autoantibodies-an autoantibody response characteristic ofmurine lupus (28) and idiopathic (29) anddrug-related human lupus (30) but found lessfrequently in murine HgIA (31). Extensivetissue deposits of immunoglobulin and com-

plement, particularly in the vessels of the kid-ney, are consistent with the progressive

glomerulonephritis of idiopathic disease inNZWBF1 mice (32). The higher titer ofglomerular deposits in the NZBWF1 com-

pared to the MRL substrains is also consistentwith the idiopathic picture in these strains(32). These studies with mercury, and thosewith other exogenous agents (4,7), suggestthat the NZBWF1 strain may be particularlysuited to studies on xenobiotic-induced exac-

erbation of systemic autoimmunity.

Environmental Health Perspectives * Vol 107, Supplement 5 - October 1999

MRL /Ipr5 -

4 -

3 -W

cnW 2-

1-

0-1 ..

733

POLLARD ET AL.

The MRL/Ipr has been used extensively instudies on autoimmunity, primarily to under-stand the significance of the lpr gene but alsobecause its accelerated disease and early mor-tality facilitate in vivo studies. However, thelong life of the MRL+/+ offers the advantageof examining the effects of long-term xenobi-otic exposures on the expression of autoim-munity. Previous studies have suggested thatautoimmunity in MRL mice is either acceler-ated (33,34) or unaffected (33) by xenobioticexposure. Both MRL substrains were exam-ined in this study to determine if the acceler-ated disease of the MRL/Ipr, and the delayeddisease of the MRL+/+, would be similarlyaffected by exposure to metals. The presenceof the Ipr gene, which is associated with eleva-tions of IgG and ANA as early as 2 months ofage (32,35), appeared to be the predominantinfluence on immunoglobulin and autoanti-body levels, as neither metal led to increasesabove those of the PBS controls. In contrast,mercury-exposed MRL+/+ mice showed evi-dence of both accelerated hypergamma-globulinemia and elevated antichromatinantibodies compared to nickel- or PBS-treatedgroups. Interestingly, both MRL substrainshad increased glomerular deposits of IgGcompared with those in control mice, andMRL+/+ mice showed more severeimmunopathology with IgG and C3 depositsin kidney vessels as well as hyperplasia ofendocapillary cells. The increased glomerulardisease in Ipr mice exposed to HgC12 despitesimilar levels of immuloglobulins and autoan-tibodies as PBS and NiCI2 controls suggeststhat HgCl2 may also promote disease bymechanisms beyond autoantibody produc-tion, possibly related to autoantibody speci-ficity or effector pathways involved inend-organ damage.

The MHC is a major contributor to bothidiopathic lupus (3) and HgIA (31,36). Itcould therefore be argued that exacerbation ofautoimmunity in lupus-prone mice is primar-ily due to an effect on H-2 genes and notacceleration of idiopathic disease per se. Toexamine this possibility, AKR mice, fromwhich the MRL derives its H-2 genes, wereexposed to mercury. Although the AKR pos-sesses the H-2k haploytpe, which confers sus-ceptibility to HgIA (31), the degree ofautoimmunity elicited differed from that ofMRL mice. AKR mice had predominant ele-vations of IgG2a, whereas MRL+/+ miceshowed increased IgGl. HgC12-exposedMRL+/+ mice had more severe disease, withhigher titers of antichromatin antibodies,increased deposits of IgG and C3 in kidneyvessels, and histologic glomerular changes.This comparison between autoimmune-pronestrains and an H-2 identical nonautoimmunestrain demonstrates the influences of bothnon-MHC and non-Ipr genes on mercury-

accelerated systemic autoimmunity in micepredisposed to spontaneous disease.Furthermore, the data demonstrate how anenvironmental agent, mercury, can act syner-gistically on individuals with intermediatesusceptibility (NZBWFI and MRL+/+) toaccelerate disease, whereas there was less effecton MRLIlpr mice with the highest suscepti-bility. This is consistent with the thresholdliability model (9) and the potential impor-tance of environmental factors in influencingdisease induction, particularly in individualswith intermediate susceptibility.

Mercury, in various forms, elicits systemicautoimmunity in a number of nonauto-immune murine strains (10). Susceptibility toHgIA is genetically controlled, as this xenobi-otic is able to elicit differing degrees ofautoimmunity depending upon the geneticbackground of the host (10,27). Theimmunologic features of HgIA, which includelymphadenopathy, hypergammaglobulinemia,autoantibodies, and immune complex disease,resemble features of SLE. Even so, there aresignificant differences between murine modelsof HgIA and idiopathic SLE. HgIA is tran-sient, with disease features resolving with timeonce exposure to mercury ceases (27). In idio-pathic SLE, disease is progressive, often lead-ing to death (8,9). At the height of diseaseHgIA is associated with increased IL-4 andIgG1 autoantibodies (13). In contrast, murinemodels of idiopathic SLE have autoantibodiesof IgG2a and IgG2b isotypes (28) andelevations in IL-1 and IFN-,y (9).

Several mechanisms have been postulatedto account for mercury's autoimmune-promoting effects, including generation ofcryptic T-cell determinants (37) or produc-tion of T cells reacting against self-class II(38). The finding that mercury did not accel-erate humoral responses in the presence of thelpr gene raises the possibility that mercury,like the Ipr defect, may act through inhibitionof apoptosis. This would be consistent withthe polyclonal lymphoid hyperplasia andimmunoglobulin increases found in theMRL+/+ after mercury treatment. Mercurycould possibly prevent apoptosis by interact-ing with the sulfhydryl moiety of cysteine inthe active site of one or more caspases (39)that are involved in the signaling and effectorpathways responsible for induction ofapoptotic cell death.

Comparison between the MRL and AKRstrains revealed the importance of studies withgenetically related nonautoimmune strains indissecting the contribution of xenobiotic expo-sure to acceleration of idiopathic disease.Additional studies contrasting the geneticallyrelated autoimmune-prone BXSB (H-2b) andnonautoimmune C57BL/6 (H-2b) have alsoshown that mercury exposure exacerbates idio-pathic autoimmunity (40). The significance of

such studies cannot be overstated. Humanstudies have shown that xenobiotic-induceddisease, although exhibiting features of theidiopathic syndrome, may lack key pathologicsequlea and have dinical features not associ-ated with idiopathic disease (2,41). Studiesthat seek to link xenobiotics with auto-immunity must therefore attempt todiscriminate between xenobiotic-inducedautoimmunity and xenobiotic-accelerated idio-pathic disease. Comparative studies usinggenetically related autoimmune-prone andnonautoimmune mice might help determinethe contribution that xenobiotic exposuremakes in rendering sensitive populationssusceptible to autoimmunity.

REFERENCES AND NoTEs

1. Rubin RL. Drug-induced lupus. In: Dubois' Lupus Erythematosus,5th ed (Wallace DJ, Hahn BH, eds). Baltimore, MD:Williamsand Wilkins, 1997;871-901.

2. Silver RM. Scleroderma and pseudoscleroderma. In: SystemicSclerosis (Clements PJ, Furst DE, eds). Baltimore, MD:Williamsand Wilkins, 1996;81-98.

3. Kono DH, Theofilopoulos AN. Genetic contributions to SLE.J Autoimmunity 9:437-452 (1996).

4. Hang L, Slack JH, Amundson C, lzui S, Theofilopoulos AN, DixonFJ. Induction of murine autoimmune disease by chronic poly-clonal B cell activation. J Exp Med 157:874-883 (1983).

5. Ansel JC, Mountz J, Steinberg AD, DeFabo E, Green I. Effects ofUV radiation on autoimmune strains of mice: increased mortal-ity and accelerated autoimmunity in BXSB male mice. J InvestDermatol 85:181-186(1985).

6. Lewis RE, Cruse JM, Johnson WW, Mohammad A.Histopathology and cell-mediated immune reactivity inhalothane-associated lymphomagenesis and autoimmunity inBXSB/Mp and MRL/Mp mice. Exp Mol Pathol 36:378-395 (1982).

7. Carpenter DF, Steinberg AD, Schur PH, Talal N. The pathogene-sis of autoimmunity in New Zealand mice. II: Acceleration ofglomerulonephritis by polyinosinic.polycytidylic acid. Lab Invest23:628-634 (1970).

8. Theofilopoulos AN. Murine models of lupus. In: Systemic LupusErythematosus. 2nd ed (Lahita RG, ed). New York:ChurchillLivingstone, 1992:121-194.

9. Theofilopoulos AN, Kono DH. Murine Lupus Models: Gene-specific and genome-wide studies. In: Systemic LupusErythematosus, 3rd ed (Lahita RG, ed). New York:ChurchillLivingstone, 1999;149-181.

10. Pollard KM, Hultman P. Effects of mercury on the immune sys-tem. In: Metal Ions in Biological Systems. Vol 34: Mercury andIts Effects on Environment and Biology (Sigel H, Sigel A, eds).Basel:Marcel Dekker, 1997;421-440.

11. Bigazzi PE. Animal models of autoimmunity: spontaneous andinduced. In: The Autoimmune Diseases (Rose NR, Mackay IR,eds). New York:Academic Press, 1998;211-244.

12. Barr RD. The mercurial nephrotic syndrome. E Afr Med J67:381-386 (1990)

13. Ochel M, Vohr HW, Pfeiffer C, Gleichmann E. [L-4 is requiredfor the IgE and IgGl increase and IgGl autoantibody formationin mice treated with mercuric chloride. J Immunol146:3006-3011 (1991).

14. Kono, DH, Balomenos D, Pearson DL, Park MS, Hildebrandt B,Hultman P, Pollard KM. The prototypic autoimmunity induced bymercury is dependent on IFN-y and not Thl/Th2 imbalance.J Immunol 161:234-240(1998).

15. Tian L, Lawrence DA. Metal-induced modulation of nitric oxideproduction in vitro by murine macrophages: lead, nickel, andcobalt utilize different mechanisms. Toxicol AppI Pharmacol141:540-547 (1996).

16. Goebeler M, Roth J, Brocker E-B, Sorg C, Schultze-Osthoff K.Activation of nuclear factor-icB and gene expression in humanendothelial cells by the common haptens nickel and cobalt.J Immunol 155:2459-2467 (1995).

17. Bemstein Dl. Allergic reactions to workplace allergens. JAMA278:1907-1913 (1997).

18. Van Hoogstraten IM, Boos C, Boden D, Von Blomberg ME,Scleper RJ, Kraal 0. Oral induction of tolerance to nickel sensi-tization in mice. J Invest Dermatol 101:28-31 (1993).

734 Environmental Health Perspectives * Vol 107, Supplement 5 * October 1999

XENOBIOTIC-ACCELERATED AUTOIMMUNITY

19. Kapsenberg ML, Wieerenga EA, Bos JD, Jansen HM. Functionalsubsets of allergen-reactive human CD4+ T cells. ImmunolToday 12:392-395 (1991).

20. National Research Council. Guide for the Care and Use ofLaboratory Animals. Washington, DC:National AcademyPress, 1996.

21. Kono D, Burlingame R, Owens DG, Kuramochi A, Balderas RS,Balomenos, D, Theofilopoulos AN. Lupus susceptibility loci in NewZealand mice. Proc Nati Acad Sci USA 91:10168-10172 (1994).

22. Takeuchi K, Turley SJ, Tan EM, Pollard KM. Analysis of theautoantibody response to fibrillarin in human disease and murinemodels of autoimmunity J. Immunol 154:691-971 (1995).

23. Reimer G, Pollard KM, Penning CA, Ochs RL, Lischwe MA,Busch H, Tan EM. Monoclonal autoantibody from a (NewZealand black x New Zealand white)F1 mouse and some humanscleroderma sera target an Mr 34,000 nucleolar protein of theU3 RNP particle. Arthritis Rheum 30:793-800(1987).

24. Burlingame RW, Rubin RL. Subnucleosome structures as sub-strates in enzyme-linked immunosorbent assays. J ImmunolMeth 143:187-199 (1990).

25. Kotzin BL, Lafferty JA, Portanova JP, Rubin RL, Tan EM.Monoclonal anti-histone autoantibodies derived from murinemodels of lupus. J Immunol 133:2554-2559 (1984).

26. Bell S, Hobbs M, Rubin R. Isotype-restricted hyperimmunity in amurine model of the toxic oil syndrome. J Immunol148:3369-3376 (1992).

27. Hultman P, Turley SJ, Enestrom S, Lindh U, Pollard KM. Murinegenotype influences the specificity, magnitude and persistenceof murine mercury-induced autoimmunity. J Autoimmunity9:139-149 (1996).

28. Burlingame RW, Rubin RL, Balderas RS, Theofiopoulos AN.Genesis and evolution of antichromatin autoantibodies inmurine lupus implicates T-dependent immunization with selfantigen. J Clin Invest 91:1687-1695 (1993).

29. Burlingame RW, Boey ML, Starkebaum G, Rubin RL. The centralrole of chromatin in autoimmune responses to histones and DNAin systemic lupus erythematosus. J Clin Invest 94:184-192 (1994).

30. Rubin RL, Bell SA, Burlingame RW. Autoantibodies associatedwith lupus induced by diverse drugs target a similar epitope inthe (H2A-H2B)-DNA complex. J Clin Invest 90(1):165-173 (1992).

31. Hultman P, Bell W, Enestrtm S, Pollard KM. Murine susceptibil-ity to mercury. 1: Autoantibody profiles and systemic immunedeposits in inbred, congenic, and intra-H-2 recombinant strains.Clin Immunol Immunopathol 65:98-108 (1992).

32. Andrews BS, Eisenberg RA, Theofilopoulos AN, lzui S, WilsonCB, McConahey PJ, Murphy ED, Roths JB, Dixon FJ.Spontaneous murine lupus-like syndromes: clinical andimmunopathological manifestations in several strains. J ExpMed 148:1198-1215(1987).

33. Rodgers KE. Effects of oral administration of malathion on thecourse of disease in MRL-Ipr mice. J Autoimmunity10:367-373 (1997).

34. Kahn MF, Kaphalia BS, Ansari GA. Time-dependent autoim-mune response of dichloroacetyl chloride in female MRI+/+mice. Immunopharmacol Immunotoxicol 19:265-277 (1997).

35. Theofilopoulos AN, Dixon FJ. Experimental murine systemiclupus erythematosus. In: Systemic Lupus Erythematosus (LahitaRG, ed). New York:John Wiley and Sons, 1987;121-202.

36. Hultman P, Bell LJ, Enestrom S, Pollard KM. Murine susceptibil-ity to mercury. 2: Autoantibody profiles and renal immunedeposits in hybrid, backcross, and H-2d congenic mice. ClinImmunol Immunopathol 68:9-20 (1993).

37. Griem P, Wulferink M, Sachs B, Gonzalez JB, Gleichmann E.Allergic and autoimmune reactions to xenobiotics: how do theyarise? Immunol Today 19:133-141 (1998).

38. Druet P, Pelletier L. Th2 and Thl autoreactive anti-class 11 celllines in the rat suppress or induce autoimmunity.J Autoimmunity 9:221-226 (1996).

39. Alnemri ES, Livingston DJ, Nicholson DW, Salvesen G,Thornberry NA, Wong WW, Yuan J. Human ICE/CED-3 proteasenomenclature [Letter]. Cell 87:171 (1996).

40. Pollard KM, Pearson DL, Hultman P, Hildebrandt B, Kono DH.Unpublished data.

41. Fritzler MJ, Rubin RL. Drug-induced lupus. In: Dubois' LupusErythematosus. 4th ed (Wallace DJ, Hahn BH, eds). London:Leaand Febiger, 1993;442-453.

Environmental Health Perspectives . Vol 107, Supplement 5 * October 1999 735