strickland 2015

6
CD4 þ T cells epigenetically modied by oxidative stress cause lupus-like autoimmunity in mice Faith M. Strickland a, * , YePeng Li a , Kent Johnson b , Zhichao Sun c , Bruce C. Richardson a, d a Department of Internal Medicine, Rheumatology Division, The University of Michigan, Ann Arbor, MI 48109, USA b Department of Pathology, The University of Michigan, Ann Arbor, MI 48109, USA c Department of Biostatistics, School of Public Health, The University of Michigan, Ann Arbor, MI 48109, USA d Department of Medicine, Ann Arbor VA Medical Center, USA article info Article history: Received 21 April 2015 Received in revised form 1 June 2015 Accepted 7 June 2015 Available online 9 July 2015 Keywords: Systemic lupus erythematosus (SLE) CD70 CD40L KirL1 abstract Lupus develops when genetically predisposed people encounter environmental agents such as UV light, silica, infections and cigarette smoke that cause oxidative stress, but how oxidative damage modies the immune system to cause lupus ares is unknown. We previously showed that oxidizing agents decreased ERK pathway signaling in human T cells, decreased DNA methyltransferase 1 and caused demethylation and overexpression of genes similar to those from patients with active lupus. The current study tested whether oxidant-treated T cells can induce lupus in mice. We adoptively transferred CD4 þ T cells treated in vitro with oxidants hydrogen peroxide or nitric oxide or the demethylating agent 5-azacytidine into syngeneic mice and studied the development and severity of lupus in the recipients. Disease severity was assessed by measuring anti-dsDNA antibodies, proteinuria, hematuria and by histopathology of kidney tissues. The effect of the oxidants on expression of CD40L, CD70, KirL1 and DNMT1 genes and CD40L protein in the treated CD4 þ T cells was assessed by Q-RT-PCR and ow cytometry. H 2 O 2 and ONOO decreased Dnmt1 expression in CD4 þ T cells and caused the upregulation of genes known to be sup- pressed by DNA methylation in patients with lupus and animal models of SLE. Adoptive transfer of oxidant-treated CD4 þ T cells into syngeneic recipients resulted in the induction of anti-dsDNA antibody and glomerulonephritis. The results show that oxidative stress may contribute to lupus disease by inhibiting ERK pathway signaling in T cells leading to DNA demethylation, upregulation of immune genes and autoreactivity. © 2015 Elsevier Ltd. All rights reserved. 1. Introduction Systemic lupus erythematosus (SLE) is a chronic relapsing autoimmune disease that primarily affects women. SLE develops and ares when genetically predisposed people encounter certain environmental agents. The genes predisposing to lupus are being identied, and epidemiologic evidence indicates that environ- mental agents which cause oxidative stress, such as infections, sun exposure, silica and smoking, are associated with lupus onset and ares [1]. However, how environmentally induced reactive oxygen species interact with the immune system to trigger lupus ares remains unclear. Our group reported that CD4 þ T cells epigenetically altered with DNA methylation inhibitors like 5-azacytidine (5-azaC), procaina- mide (Pca) or hydralazine (Hyd) cause lupus-like autoimmunity in animal models [2], and that similar epigenetically modied T cells are found in lupus patients during disease ares [1]. We traced the cause of the epigenetic defect to PKCd inactivation, which prevents upregulation of DNA methyltransferase 1 (Dnmt1) during mitosis to copy methylation patterns [3]. Others reported that serum pro- teins are nitrated in patients with active lupus, caused by super- oxide (O 2 - ) combining with nitric oxide (NO) to form peroxynitrite (ONOO ), a highly reactive molecule that nitrates proteins and other molecules [4]. Our group subsequently found that PKCd is similarly nitrated in T cells from patients with active lupus, and that nitrated PKCd is unable to transmit signals that upregulate Dnmt1 to copy DNA methylation patterns in dividing T cells, causing demethylation and overexpression of genes normally suppressed by DNA methylation [5]. We extended these observations by * Corresponding author. Division of Rheumatology, BSRB 3009, 109 Zina Pitcher Pl., Ann Arbor, MI 48109-2200, USA. E-mail address: [email protected] (F.M. Strickland). Contents lists available at ScienceDirect Journal of Autoimmunity journal homepage: www.elsevier.com/locate/jautimm http://dx.doi.org/10.1016/j.jaut.2015.06.004 0896-8411/© 2015 Elsevier Ltd. All rights reserved. Journal of Autoimmunity 62 (2015) 75e80

Upload: vlad-andreea-maria

Post on 29-Jan-2016

228 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Strickland 2015

lable at ScienceDirect

Journal of Autoimmunity 62 (2015) 75e80

Contents lists avai

Journal of Autoimmunity

journal homepage: www.elsevier .com/locate/ jaut imm

CD4þ T cells epigenetically modified by oxidative stress causelupus-like autoimmunity in mice

Faith M. Strickland a, *, YePeng Li a, Kent Johnson b, Zhichao Sun c, Bruce C. Richardson a, d

a Department of Internal Medicine, Rheumatology Division, The University of Michigan, Ann Arbor, MI 48109, USAb Department of Pathology, The University of Michigan, Ann Arbor, MI 48109, USAc Department of Biostatistics, School of Public Health, The University of Michigan, Ann Arbor, MI 48109, USAd Department of Medicine, Ann Arbor VA Medical Center, USA

a r t i c l e i n f o

Article history:Received 21 April 2015Received in revised form1 June 2015Accepted 7 June 2015Available online 9 July 2015

Keywords:Systemic lupus erythematosus (SLE)CD70CD40LKirL1

* Corresponding author. Division of Rheumatology,Pl., Ann Arbor, MI 48109-2200, USA.

E-mail address: [email protected] (F.M. Strickla

http://dx.doi.org/10.1016/j.jaut.2015.06.0040896-8411/© 2015 Elsevier Ltd. All rights reserved.

a b s t r a c t

Lupus develops when genetically predisposed people encounter environmental agents such as UV light,silica, infections and cigarette smoke that cause oxidative stress, but how oxidative damage modifies theimmune system to cause lupus flares is unknown. We previously showed that oxidizing agents decreasedERK pathway signaling in human T cells, decreased DNA methyltransferase 1 and caused demethylationand overexpression of genes similar to those from patients with active lupus. The current study testedwhether oxidant-treated T cells can induce lupus in mice. We adoptively transferred CD4þ T cells treatedin vitro with oxidants hydrogen peroxide or nitric oxide or the demethylating agent 5-azacytidine intosyngeneic mice and studied the development and severity of lupus in the recipients. Disease severity wasassessed by measuring anti-dsDNA antibodies, proteinuria, hematuria and by histopathology of kidneytissues. The effect of the oxidants on expression of CD40L, CD70, KirL1 and DNMT1 genes and CD40Lprotein in the treated CD4þ T cells was assessed by Q-RT-PCR and flow cytometry. H2O2 and ONOO�

decreased Dnmt1 expression in CD4þ T cells and caused the upregulation of genes known to be sup-pressed by DNA methylation in patients with lupus and animal models of SLE. Adoptive transfer ofoxidant-treated CD4þ T cells into syngeneic recipients resulted in the induction of anti-dsDNA antibodyand glomerulonephritis. The results show that oxidative stress may contribute to lupus disease byinhibiting ERK pathway signaling in T cells leading to DNA demethylation, upregulation of immune genesand autoreactivity.

© 2015 Elsevier Ltd. All rights reserved.

1. Introduction

Systemic lupus erythematosus (SLE) is a chronic relapsingautoimmune disease that primarily affects women. SLE developsand flares when genetically predisposed people encounter certainenvironmental agents. The genes predisposing to lupus are beingidentified, and epidemiologic evidence indicates that environ-mental agents which cause oxidative stress, such as infections, sunexposure, silica and smoking, are associated with lupus onset andflares [1]. However, how environmentally induced reactive oxygenspecies interact with the immune system to trigger lupus flaresremains unclear.

BSRB 3009, 109 Zina Pitcher

nd).

Our group reported that CD4þ T cells epigenetically altered withDNA methylation inhibitors like 5-azacytidine (5-azaC), procaina-mide (Pca) or hydralazine (Hyd) cause lupus-like autoimmunity inanimal models [2], and that similar epigenetically modified T cellsare found in lupus patients during disease flares [1]. We traced thecause of the epigenetic defect to PKCd inactivation, which preventsupregulation of DNA methyltransferase 1 (Dnmt1) during mitosisto copy methylation patterns [3]. Others reported that serum pro-teins are nitrated in patients with active lupus, caused by super-oxide (O2

- ) combining with nitric oxide (NO) to form peroxynitrite(ONOO�), a highly reactive molecule that nitrates proteins andother molecules [4]. Our group subsequently found that PKCd issimilarly nitrated in T cells from patients with active lupus, and thatnitrated PKCd is unable to transmit signals that upregulate Dnmt1to copy DNA methylation patterns in dividing T cells, causingdemethylation and overexpression of genes normally suppressedby DNA methylation [5]. We extended these observations by

Page 2: Strickland 2015

F.M. Strickland et al. / Journal of Autoimmunity 62 (2015) 75e8076

demonstrating that treating human CD4þ T cells with the oxidizingagents H2O2 or ONOO� inhibits PKCd activation, thereby decreasingERK pathway signaling, decreasing Dnmt1 levels and causingdemethylation and overexpression of CD70 and Kir genes similar toT cells from patients with active lupus [6]. However, whether theepigenetically modified T cells are sufficient to cause lupus-likeautoimmunity was unknown.

We have now tested if female murine CD4þ T cells treated withH2O2 or ONOO� also overexpress methylation sensitive genes, andif the treated cells cause lupus-like autoimmunity in mice. Theresults demonstrate that the oxidized T cells overexpress the X-linked gene CD40L, one copy of which is silenced by DNA methyl-ation in female human and mouse T cells [7,8], and Kir genes,normally expressed by NK cells but silenced by DNAmethylation inhuman and mouse T cells [9]. The results also demonstrate thatadoptive transfer of the treated cells into syngeneic recipientscauses anti-DNA antibodies and an immune complex glomerulo-nephritis, similar to T cells treated 5-azaC, Pca or Hyd. Togetherthese studies support the contention that environmentally-inducedoxidative stress may trigger lupus flares by epigenetically altering Tcells through effects on T cell signaling and Dnmt1 expression.

2. Materials and methods

2.1. Drugs and reagents

5-azaC and H2O2were purchased from SigmaeAldrich (St. Louis,MO), and ONOO� from Calbiochem (San Diego, CA).

2.2. Mice

Female SJL mice were purchased from Jackson Laboratories,housed in filter-protected cages, and provided with standard irra-diated rodent diet 5053 (Lab Diet; PMI Nutrition International) andwater ad libitum. Urinary protein and hemoglobin were measuredusing Chemstrip 7 dipsticks (Roche, Madison, WI). The mice weremaintained in a specific pathogenefree facility by the Unit forLaboratory Animal Medicine at the University of Michigan inaccordance with the National Institutes of Health and the Associ-ation for Assessment and Accreditation of Laboratory Animal Care(AAALAC) International Guidelines. All procedures were approvedby the University of Michigan Institutional Animal Care and UseCommittee.

2.3. T cell isolation, culture, treatment and injection

Splenocytes, thymocytes and lymph node cells were isolatedfrom female SJL mice, pooled and stimulated in vitro with 5 mg/mlconcanavalin A in RPMI 1640 supplemented with 10% fetal calfserum, 2 mM L-glutamine, 10 mM HEPES, 1000 U Penicillin/1 mgstreptomycin (Fisher Scientific, Pittsburgh, PA) for 18e24 h at 37 �C,5% CO2/balanced air incubator. CD4þ T cells were then isolatedusing magnetic beads (Miltenyi, Auburn, CA), and treated with5 mM 5-azaC, 20 mM H2O2 or 20 mM ONOO� for 3 days in 6 wellplates as previously described [2,6,10]. The cells were then washedand 5 � 106 viable cells injected into the tail vein of each female SJLrecipients, beginning when the mice were 12 weeks of age andcontinuing every 2 weeks for a total of 7 injections. Cells analyzedfor gene expression by PCRwere additionally treated with 50 ng/mlphorbol myristate acetate (PMA) and 500 ng/ml ionomycin duringtheir final 6 h of culture. Blood and urine samples were obtainedevery other week. Two weeks after the last injection the mice weresacrificed, CD4þ T cells isolated for further study and kidneysremoved for histologic analysis.

2.4. Flow cytometric analysis

Spleen cells were washed twice in Standard Buffer (PBS con-taining 1% horse serum and 1 mg/ml sodium azide) at 4 �C. Allincubations were performed on ice. Non-specific binding wasblocked by incubating the cells 1 h on ice in Standard Buffer con-taining 10% horse serum. The cells were then stained in the dark for1 h with PE-Cy5-rat anti-mouse CD4 (BD Pharmingen, Fullerton,CA) together with PE-hamster anti-mouse CD154 (CD40L) ormatching IgG controls washed, then fixed in 2% paraformaldehydeand stored in the dark at 4 �C. The cells were analyzed using aFACSCalibur flow cytometer (BD Biosciences, Franklin Lakes, NJ) aspreviously described [8].

2.5. RT-PCR

Total RNA and DNA were simultaneously isolated from bead-purified CD4þ T cells using a Qiagen RNAEasy Mini Kit (Qiagen,Valencia, CA). RNA was quantified using a NanoDrop 1000 spec-trophotometer (NanoDrop Products, Wilmington, DE). One micro-gram of total RNA per sample was used to synthesize cDNA using aTranscriptor First Strand cDNA Synthesis Kit and anchored oli-go(dT)18 primers (Roche, Indianapolis, IN) according to the manu-facturer's instructions. Primers for murine CD70, CD40L, KirL1, b-actin, and Dnmt1 were obtained from Integrated DNA Technologies(Coralville, IA) and used in RT-PCR to measure mRNA geneexpression as previous described [8].

2.6. Anti-DNA antibody ELISA

Mouse IgG anti-dsDNA antibodies were measured by ELISA aspreviously described [11]. Briefly, Costar (Corning, NY) 96 well flatbottom microtiter plates were coated overnight at 4 �C with 10 mg/ml dsDNA in PBS, pH 7.2. Various dilutions of mouse sera or murinemonoclonal IgG anti-dsDNA antibody (Millipore, Billerica, MA)standard were added in PBS and incubated overnight at 4 �C. Boundanti-dsDNA antibodies were detected using HRP-goat anti-mouseIgG-Fc-specific (Bethyl Labs, Montgomery, TX) antibodies and OneStep Ultra TMB substrate (Thermo, Rockford, IL) and measured at450 nm [11].

2.7. Histologic analyses

Kidneys were fixed in 10% neutral buffered formalin, routinelyprocessed and paraffin embedded. Five micron sections weredeparaffinized, hydrated and rinsed for 5 min in tap water andstained with hematoxylin and eosin or treated with citrate antigenretrieval buffer according to the manufacturer's instructions (Vec-tor laboratories, Burlingame, CA). Endogenous peroxidase wasblocked by treating the tissue sections with 0.3% hydrogen peroxideinmethanol for 30min at ambient temperature followed by a 5minwash in PBS. The tissues were then incubated with 2.5% horseserum/PBS for 20 min and IgG detected using the IMPRESS Detec-tion system and NOVA-Red© substrate (Vector Laboratories, Inc).The tissue was counter stained with hematoxylin according to themanufacturer's instructions, mounted with Immu-Mount (ThermoFisher) and coverslipped.

2.8. Statistical analyses

The significance of differences between means was assessedusing Student's t-Test or ANOVA. The linear mixed model was usedto assess the effect of different treatments on antibody productionrelative to injections of Hanks balanced salt solution (HBSS) overtime, and the Chi square test to assess differences in proteinuria

Page 3: Strickland 2015

F.M. Strickland et al. / Journal of Autoimmunity 62 (2015) 75e80 77

between groups at 12 weeks.

Fig. 2. 5-azaC, H2O2 and ONOO- increase CD40L protein expression in CD4þ T cells.Splenocytes from female SJL mice were stimulated with Con-A then treated with H2O2

or ONOO� as in Fig. 1. 72 h later the cells were washed, stained with fluorochromeconjugated antibodies to CD4 and CD40L then analyzed by flow cytometry. *P ¼ 0.049,**P ¼ 0.004 by paired t-Test). Results are expressed as the mean ± SEM (n ¼ 5 mice)and are from one of two independent experiments.

3. Results

3.1. Murine CD4þ T cells treated with oxidizing agents overexpressmethylation sensitive genes

We previously reported that inhibiting the replication of humanCD4þ T cell DNA methylation patterns with the Dnmt inhibitor 5-azaC causes demethylation and overexpression of genes includingCD40L on the female inactive X chromosome in humans and mice[8,12] as well as the human KIR gene family, normally expressed byNK cells but not T cells [9,13]. Initial studies compared the effects ofH2O2 and ONOO� to 5-azaC on murine CD4þ T cell CD40L and Kirexpression. We recently reported that experimentally demethy-lated CD4þ T cells cause a more severe form of autoimmunity inlupus prone SJL mice relative to C57BL6mice, with higher titer anti-DNA antibodies and more severe kidney disease [8]. Therefore SJLmice were used for these studies. The spleen, thymus and axillarylymph nodes were removed from female SJL mice, then mono-nuclear cells were isolated and stimulated with Con-A in vitro.18e24 h later CD4þ T cells were purified and cultured with 5 mM 5-azaC, 20 mM H2O2 or 20 mM ONOO� for 72 h using protocols pre-viously described by our group [6]. Live cells were isolated on aficoll gradient followed by lysis, RNA isolation and cDNA prepara-tion. CD40L and KirL1 transcripts were measured relative to b-actinby RT-PCR. CD40L was selected because this gene is encoded on thesecond, inactive X chromosome in females, and the increase inexpression caused by DNA methylation inhibitors is due to deme-thylation of CD40LG on the inactive X in humans and mice [11,12].KirL1 was chosen because this gene is expressed by NK cells but notT cells, and is silenced in T cells primarily by DNA methylation [13].Fig. 1 shows that CD4þ T cells treated with 5-azaC, H2O2 or ONOO�

had significantly decreased Dnmt1 gene expression relative tountreated controls (Fig. 1A) and over-expressed CD70 (Fig. 1B),CD40L (Fig. 1C) and KirL1 (Fig. 1D) mRNA relative to untreated Tcells. These results are similar to our earlier reports using otherDNA methylation inhibitors [8,9].

The effects of 5-azaC, H2O2 and ONOO� on CD40L expression

Fig. 1. 5-azaC, H2O2 and ONOO� decrease Dnmt1 gene expression and increase the expressionthymus and axillary lymph nodes were removed from female SJL mice, then mononuclear cpurified and treated with 5 mM 5-azaC, 20 mM H2O2 or 20 mM ONOO� using previously descrfinal 6 h of culture then lysed and Dnmt1 (A), CD70 (B), CD40L (C) and KirL1 (D) transcripts mcontrol (100%) and represent the mean ± SEM of 3 experiments combined (*P � 0.05; **P

were confirmed at the protein level. Splenocytes from female SJLmice were stimulated with Con-A then treated with H2O2 orONOO� as in Fig. 1. Seventy-two hours later the cells were washed,stained with fluorochrome conjugated antibodies to CD4 andCD40L then analyzed by flow cytometry. Fig. 2 shows that 5-azaCtreatment increased CD40L expression as expected (p ¼ 0.004).Both H2O2 and ONOO� significantly increase CD40L expression onCD4þ T cells compared to the untreated control (p ¼ 0.049,p ¼ 0.004 respectively). ONOO� treatment induced more CD40Lexpression than 5-azaC (mean MFI 5819 ± 675 vs 2942 ± 147,p ¼ 0.01, n ¼ 5) but the difference between ONOO� and H2O2 wasnot statistically significant (mean MFI 3615 ± 475, p ¼ 0.06)(p ¼ 0.004).

3.2. CD4þ T cells treated with oxidizing agents cause lupus-likeautoimmunity in syngeneic mice

Functional significance of the changes in T cell gene expressioncaused by the oxidizers was tested using the adoptive transfer

of methylation sensitive genes CD40L, CD70 and KirL1 mRNA in CD4þ T cells. The spleen,ells were isolated and stimulated with Con-A in vitro. 18e24 h later CD4þ T cells wereibed protocols [6]. 72 h later the T cells were treated with PMA and ionomycin for theireasured relative to b-actin by RT-PCR. Results are expressed as percent of the untreated< 0.01; ***P � 0.001 by Student's t-Test.).

Page 4: Strickland 2015

Table 1CD4þ T cells treated with 5-azaC, H2O2 or ONOO� cause proteinuria in SJL mice.

Treatment Proteinuria P (vs HBSS)

HBSS 0/5Untreated T cells 2/9 NS5-azaC treated T cells 4/5 0.01H2O2 treated T cells 4/7 0.038ONOO� treated T cells 7/10 0.01

Female SJL mice received i.v. injections of HBSS or 5 � 106 CD4þ T cells treated withthe indicated agents every other week for a total of 7 injections. Urine protein wasmeasured serially using dipsticks. P values were calculated using Chi-square sta-tistics. Proteinuria: �30 mg/dl protein for �2 consecutive wk.

F.M. Strickland et al. / Journal of Autoimmunity 62 (2015) 75e8078

model previously described by our group for T cells treated pro-cainamide and hydralazine, which inhibit T cell DNA methylationand cause a lupus-like disease in genetically predisposed people[14]. Splenocytes, thymocytes and lymph node cells were isolatedfrom female SJL mice, stimulated with Con-A, then CD4þ T cellspurified, washed, and treated with 5 mM 5-azaC, 20 mM H2O2 or20 mM ONOO� for another 72 h as before. The cells were thenwashed, suspended in HBSS, and 5 � 106 cells injected through thetail vein into female SJL mice every 2 weeks for a total of 7 in-jections. Controls included untreated T cells and HBSS alone.

Fig. 3 shows that by 12 weeks CD4þ T cells treated with 5-azaC,H2O2 or ONOO� all stimulate higher levels of anti-DNA antibodiesrelative to untreated T cells in syngeneic recipient mice (p ¼ 0.037by ANOVA), resembling results from similar experiments reportedby our group for procainamide and hydralazine treated T cells [15].The ONOO� treated T cells caused the highest levels of anti-DNAantibodies at 12 weeks, and were significantly greater than thosecaused by HBSS (p ¼ 0.005) or untreated T cells (p ¼ 0.01).

Urinary protein levels weremeasured by dipstick 12 weeks afterthe first injection in mice receiving untreated CD4þ T cells or CD4þ

T cells treated with 5-azaC, H2O2 or ONOO�. Table 1 shows thatmice receiving 5-azaC, H2O2 or ONOO� treated CD4þ T cells, but notuntreated T cells, also developed significantly greater proteinuriarelative to mice receiving HBSS alone (p < 0.04). No hematuria wasobserved (not shown).

At 14 weeks the mice were sacrificed and kidneys removed forhistologic analysis. Fig. 4 shows hematoxylin and eosin (left) andanti-IgG (right) stained kidney sections from mice receiving un-treated, 5-azaC treated, H2O2 treated or ONOO� treated CD4þ Tcells. Panels A and B show representative glomeruli from micereceiving untreated T cells. The glomeruli are normal. Glomerulifrommice receiving 5-azaC treated CD4þ T cells (panel C) exhibitedhypercellularity, hyaline thrombi (arrow) and karyorrhexis (arrowhead). IgG deposits (panel D) were also detected in the glomerulifrom these animals. Similarly glomeruli from mice receiving H2O2treated T cells (Fig. 4E, F) and frommice receiving ONOO� treated Tcells (Fig 4G, H) exhibited hypercellularity, hyaline thrombi andkaryorrhexis. IgG staining was more intense in the glomeruli from

Fig. 3. CD4þ T cells treated with 5-azaC, H2O2 or ONOO- cause anti-DNA antibodies insyngeneic mice. CD4þ T cells treated as described for Fig. 1 were injected (5 � 106 cells/recipient) into the tail veins of female SJL mice every 2 weeks for a total of 7 injections.Controls included mice receiving untreated T cells and HBSS alone. Blood was obtained4, 8 and 12 weeks after the first injection and serum anti-dsDNA antibody levelsmeasured by ELISA. Results are presented as the mean ± SEM of groups of 5e8 mice.The linear mixed model of ANOVA was used to assess the statistical significance of theeffects of different treatments on antibody production relative to HBSS injectionswithout cells over time. P values shown are versus mice treated with HBSS (no cells).

these mice compared with animals that received 5-azaC treatedcells and is consistent with the higher levels of anti-dsDNA anti-body produced in mice receiving H2O2 or ONOO- treated cells.Together these studies demonstrate that CD4þ T cells treated withthe Dnmt1 inhibitor 5-azaC, or the oxidizing agents H2O2 or ONOO�

are sufficient to cause anti-DNA antibodies and an immune com-plex glomerulonephritis in genetically predisposed mice.

4. Discussion

A relatively extensive body of literature implicates CD4þ T cellDNA demethylation in lupus onset and flares. Early work demon-strated that inhibiting DNAmethylation in CD4þ T cells with 5-azaCalters gene expression and converts polyclonal as well as clonedantigen-specific CD4þ T cells in autoreactive cells that respond toself class II MHC determinants without the addition of specificantigen [16,17]. The autoreactivity was traced to overexpression ofthe adhesion molecule LFA-1 (CD11a/CD18), due to demethylationof the ITGAL (CD11a) gene [18], and causing LFA-1 overexpressionby transfection caused a similar MHC-restricted autoreactivity [19].Importantly, both the drug treated and the transfected T cells weresufficient to cause lupus-like autoimmunity when injected i.v. intosyngeneic mice [20]. The autoreactivity may be due to increasedcostimulatory signaling caused by increased LFA-1 expression,because causing LFA-1 overexpression by transfection resulted in asimilar autoreactivity [19]. Alternatively, LFA-1e ICAM 1 complexessurround the T cell receptor-class II MHCmolecule complex to formthe “immunologic synapse” [21]. The increased LFA-1 levels mayalso over stabilize the normally lowaffinity interaction between theT cell receptor and class II MHC molecules without the appropriatepeptide in the antigen binding cleft, causing T cell activation [19].Similarly, procainamide and hydralazine, which cause a lupus-likedisease in genetically predisposed people [22], also decrease totalCD4þ T cell DNA methylation, causing ITGAL overexpression and Tcell autoreactivity similar to that caused 5-azaC [20], providinganother example of exogenous agents causing lupus-like autoim-munity through T cell DNA demethylation. Subsequent studiesdemonstrated that the same changes in ITGAL methylation andexpression are found in T cells treated with procainamide and in Tcells from patients with active lupus. Further, the degree ofdemethylation and overexpression are directly related to diseaseactivity as measured by the SLEDAI [23], suggesting that similarmechanismsmay contribute to drug induced and idiopathic humanlupus flares. Later studies demonstrated that 5-azaC also deme-thylates and causes overexpression of genes encoding perforin,CD70, CD40L and the KIR gene family and that these genes are alsodemethylated and overexpressed by lupus T cells [14]. Perforincontributes to autoreactive macrophage killing [24], while CD70and CD40L contribute to antibody overstimulation [25,26], and KIRcontributes to interferon-g overproduction and autoreactivemacrophage killing [27].

Page 5: Strickland 2015

Fig. 4. CD4þ T cells treated with 5-azaC, H2O2 or ONOO� cause an immune complex glomerulonephritis in syngeneic mice. 14 weeks after the first injection the mice were sacrificed and

kidneys removed for histologic analysis. Hematoxylin and eosin stained kidney sections frommice receiving untreated (A), 5-azaC treated (C), H2O2 treated (E) or ONOO� treated (G)CD4þ T cells. Kidney sections from the same mice tested for IgG deposition using immunoperoxidase staining. Glomeruli from mice receiving 5-azaC (D). H2O2 (F) and ONOO� (H)D4þ T cells all show IgG deposition, identified by the red deposits while mice receiving untreated T cells (B) do not. Sections were counterstained with hematoxylin. (For inter-pretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

F.M. Strickland et al. / Journal of Autoimmunity 62 (2015) 75e80 79

Mechanistic studies revealed that the lupus T cell DNAmethylation defect is due to lowDnmt1 levels, caused by a failure toupregulate Dnmt1 during mitosis, and that decreased ERK pathwaysignaling contributes to the low Dnmt1 levels [28]. The lupus T cellERK pathway signaling defect was traced to PKCd, which fails toautophosphorylate when stimulated with PMA [3]. Interestingly,the lupus-inducing drug hydralazine also inhibits T cell PKCd acti-vation [3], and PKCdmutations cause lupus in people [29] and mice[30], suggesting an important role for PKCd inactivation in lupuspathogenesis.

Later studies demonstrated that PKCd is nitrated in T cells frompatients with active lupus, similar to the serum protein nitrationfound in patients with active lupus [4]. Further,

immunoprecipitations with anti-3-nitrotyrosine antibodiesrevealed that the nitrated PKCd fraction is refractory to PMA stim-ulation, while the unmodified fraction retained PMA-stimulatedautophosphorylation activity, indicating that the PKCd signalingfunction is sensitive to nitration [5]. More recent studies confirmedthat treating human CD4þ T cells with H2O2 or ONOO� similarlyinhibits PKCd activation, decreasing ERK pathway signaling,decreasing Dnmt1 levels and causing demethylation and over-expression of genes found to be demethylated and overexpressedin T cells from patients with active lupus, including CD70 [29] andthe Kir gene family [31]. These studies also revealed that ONOO� ismore potent than H2O2. This may reflect a requirement for H2O2 tocombine with endogenous NO to form ONOO� which then

Page 6: Strickland 2015

F.M. Strickland et al. / Journal of Autoimmunity 62 (2015) 75e8080

inactivates PKCd [5] as well as causing the protein nitration thatcharacterizes lupus flares [4]. Together these studies indicate thatoxidizing agents can epigenetically modify CD4þ T cells in vitro toresemble CD4þ T cells from patients with active lupus. However,whether the modified cells are capable of causing autoimmunityin vivo was unknown.

The studies described in the current report extend these ob-servations by demonstrating that ONOO� and H2O2 also causeoverexpression of mouse T cell genes normally suppressed by DNAmethylation, similar to their effects on human T cells [32]. Impor-tantly, the present studies also demonstrate that injecting theoxidized, epigenetically altered murine CD4þ T cells into syngeneicmice is sufficient to cause a lupus-like disease with anti-DNA an-tibodies and an immune complex glomerulonephritis similar tothat caused by other DNA methylation inhibitors such as procai-namide, hydralazine or 5-azaC [2], suggesting a mechanism for theassociation of oxidative stress with lupus flares. The autoimmunityinduced by the treated cells is not due to necrotic cells caused bytreatment with ONOO� or H2O2 because we have previously re-ported that heat killed, necrotic T cells do not cause lupus-likeautoimmunity in similar adoptive transfer experiments [2].

In summary, clinical observations indicate that agents whichcause oxidative stress, such as infections [33], sun exposure [34],silica [35] and smoking [36], are associated with lupus onset andflares. The studies described above indicate that CD4þ T cellsexposed to reactive oxygen species are sufficient to cause lupus-likeautoimmunity in mice. Together these reports suggest that envi-ronmentally induced oxidative stress may trigger lupus flares ingenetically predisposed people through effects on T cell signalingand DNA methylation. These studies also support the use of anti-oxidants to treat lupus flares, as reported by others [37].

Acknowledgments

We thank Erin Grant for her excellent technical assistance andMs. Stacy Fry for her excellent secretarial assistance. We also thankDr. Gabriela Gorelik for helpful discussions and advice on oxidativestress and PKC-d. This work was supported by PHS grants AR42525and ES017885, a Merit grant from the Department of Veterans Af-fairs and the Lupus Insight Prize from the Lupus Foundation ofAmerica, the Alliance for Lupus Research and the Lupus ResearchInstitute.

References

[1] E.C. Somers, B.C. Richardson, Environmental exposures, epigenetic changesand the risk of lupus, Lupus 23 (2014) 568e576.

[2] J. Quddus, K.J. Johnson, J. Gavalchin, E.P. Amento, C.E. Chrisp, R.L. Yung, et al.,Treating activated CD4þ T cells with either of two distinct DNA methyl-transferase inhibitors, 5-azacytidine or procainamide, is sufficient to cause alupus-like disease in syngeneic mice, J. Clin. Investigation 92 (1993) 38e53.

[3] G. Gorelik, J.Y. Fang, A. Wu, A.H. Sawalha, B. Richardson, Impaired T cellprotein kinase C delta activation decreases ERK pathway signaling in idio-pathic and hydralazine-induced lupus, J. Immunol. 179 (2007) 5553e5563.

[4] J.C. Oates, G.S. Gilkeson, The biology of nitric oxide and other reactive in-termediates in systemic lupus erythematosus, Clin. Immunol. 121 (2006)243e250.

[5] G.J. Gorelik, S. Yarlagadda, D.R. Patel, B.C. Richardson, Protein kinase Cdeltaoxidation contributes to ERK inactivation in lupus T cells, Arthritis Rheuma-tism 64 (2012) 2964e2974.

[6] Y. Li, G. Gorelik, F.M. Strickland, B.C. Richardson, Oxidative stress, T cell DNAmethylation, and lupus, Arthritis & Rheumatology 66 (2014) 1574e1582.

[7] A. Hewagama, G. Gorelik, D. Patel, P. Liyanarachchi, W.J. McCune, E. Somers, etal., Overexpression of X-linked genes in T cells from women with lupus,J. Autoimmun. 41 (2013) 60e71.

[8] F.M. Strickland, A. Hewagama, A. Wu, A.H. Sawalha, C. Delaney, M.F. Hoeltzel,et al., Diet influences expression of autoimmune-associated genes and diseaseseverity by epigenetic mechanisms in a transgenic mouse model of lupus,Arthritis Rheumatism 65 (2013) 1872e1881.

[9] Y. Liu, R. Kuick, S. Hanash, B. Richardson, DNA methylation inhibition in-creases T cell KIR expression through effects on both promoter methylation

and transcription factors, Clin. Immunol. 130 (2009) 213e224.[10] B. Richardson, A.H. Sawalha, D. Ray, R. Yung, Murine models of lupus induced

by hypomethylated T cells (DNA hypomethylation and lupus…), MethodsMol. Biol. 900 (2012) 169e180.

[11] F.M. Strickland, A. Hewagama, Q. Lu, A. Wu, R. Hinderer, R. Webb, et al.,Environmental exposure, estrogen and two X chromosomes are required fordisease development in an epigenetic model of lupus, J. Autoimmun. 38(2012) J135eJ143.

[12] Q. Lu, A. Wu, L. Tesmer, D. Ray, N. Yousif, B. Richardson, Demethylation ofCD40LG on the inactive X in T cells from women with lupus, J. Immunol. 179(2007) 6352e6358.

[13] A. Rouhi, L. Gagnier, F. Takei, D.L. Mager, Evidence for epigenetic maintenanceof Ly49a monoallelic gene expression, J. Immunol. 176 (2006) 2991e2999.

[14] B. Richardson, Primer: epigenetics of autoimmunity, Nat. Clin. Pract. Rheu-matol. 3 (2007) 521e527.

[15] R. Yung, S. Chang, N. Hemati, K. Johnson, B. Richardson, Mechanisms of drug-induced lupus. IV. Comparison of procainamide and hydralazine with analogsin vitro and in vivo, Arthritis Rheumatism 40 (1997) 1436e1443.

[16] R.L. Yung, J. Quddus, C.E. Chrisp, K.J. Johnson, B.C. Richardson, Mechanism ofdrug-induced lupus. I. Cloned Th2 cells modified with DNA methylation in-hibitors in vitro cause autoimmunity in vivo, J. Immunol. 154 (1995)3025e3035.

[17] E. Cornacchia, J. Golbus, J. Maybaum, J. Strahler, S. Hanash, B. Richardson,Hydralazine and procainamide inhibit T cell DNA methylation and induceautoreactivity, J. Immunol. 140 (1988) 2197e2200.

[18] Q. Lu, D. Ray, D. Gutsch, B. Richardson, Effect of DNA methylation and chro-matin structure on ITGAL expression, Blood 99 (2002) 4503e4508.

[19] B. Richardson, D. Powers, F. Hooper, R.L. Yung, K. O'Rourke, Lymphocytefunction-associated antigen 1 overexpression and T cell autoreactivity,Arthritis Rheumatism 37 (1994) 1363e1372.

[20] R. Yung, D. Powers, K. Johnson, E. Amento, D. Carr, T. Laing, et al., Mechanismsof drug-induced lupus. II. T cells overexpressing lymphocyte function-associated antigen 1 become autoreactive and cause a lupuslike disease insyngeneic mice, J. Clin. Investigation 97 (1996) 2866e2871.

[21] S.K. Bromley, W.R. Burack, K.G. Johnson, K. Somersalo, T.N. Sims, C. Sumen, etal., The immunological synapse, Annu. Rev. Immunol. 19 (2001) 375e396.

[22] R.L. Yung, B.C. Richardson, Drug-induced lupus, Rheumatic Dis. Clin. N. Am. 20(1994) 61e86.

[23] Q. Lu, M. Kaplan, D. Ray, D. Ray, S. Zacharek, D. Gutsch, et al., Demethylation ofITGAL (CD11a) regulatory sequences in systemic lupus erythematosus,Arthritis Rheumatism 46 (2002) 1282e1291.

[24] M.J. Kaplan, Q. Lu, A. Wu, J. Attwood, B. Richardson, Demethylation of pro-moter regulatory elements contributes to perforin overexpression in CD4þ

lupus T cells, J. Immunol. 172 (2004) 3652e3661.[25] K. Oelke, Q. Lu, D. Richardson, A. Wu, C. Deng, S. Hanash, et al., Overexpression

of CD70 and overstimulation of IgG synthesis by lupus T cells and T cellstreated with DNA methylation inhibitors, Arthritis Rheumatism 50 (2004)1850e1860.

[26] Y. Zhou, J. Yuan, Y. Pan, Y. Fei, X. Qiu, N. Hu, et al., T cell CD40LG geneexpression and the production of IgG by autologous B cells in systemic lupuserythematosus, Clin. Immunol. 132 (2009) 362e370.

[27] D. Basu, Y. Liu, A. Wu, S. Yarlagadda, G.J. Gorelik, M.J. Kaplan, et al., Stimulatoryand inhibitory killer Ig-like receptor molecules are expressed and functionalon lupus T cells, J. Immunol. 183 (2009) 3481e3487.

[28] C. Deng, M.J. Kaplan, J. Yang, D. Ray, Z. Zhang, W.J. McCune, et al., DecreasedRas-mitogen-activated protein kinase signaling may cause DNA hypo-methylation in T lymphocytes from lupus patients, Arthritis Rheumatism 44(2001) 397e407.

[29] A. Belot, P.R. Kasher, E.W. Trotter, A.P. Foray, A.L. Debaud, G.I. Rice, et al.,Protein kinase cdelta deficiency causes mendelian systemic lupus erythema-tosus with B cell-defective apoptosis and hyperproliferation, Arthritis Rheu-matism 65 (2013) 2161e2171.

[30] A. Miyamoto, K. Nakayama, H. Imaki, S. Hirose, Y. Jiang, M. Abe, et al.,Increased proliferation of B cells and auto-immunity in mice lacking proteinkinase Cdelta, Nature 416 (2002) 865e869.

[31] A.H. Sawalha, L. Wang, A. Nadig, E.C. Somers, W.J. McCune, C. Michigan Lupus,et al., Sex-specific differences in the relationship between genetic suscepti-bility, T cell DNA demethylation and lupus flare severity, J. Autoimmun. 38(2012) J216eJ222.

[32] D.R. Patel, B.C. Richardson, Epigenetic mechanisms in lupus, Curr. Opin.Rheumatology 22 (2010) 478e482.

[33] G. Zandman-Goddard, Y. Shoenfeld, Infections and SLE, Autoimmunity 38(2005) 473e485.

[34] A. Kim, B.F. Chong, Photosensitivity in cutaneous lupus erythematosus, Pho-todermatol. Photoimmunol. Photomed. 29 (2013) 4e11.

[35] S. Lee, H. Hayashi, M. Maeda, Y. Chen, H. Matsuzaki, N. Takei-Kumagai, et al.,Environmental factors producing autoimmune dysregulationechronic acti-vation of T cells caused by silica exposure, Immunobiology 217 (2012)743e748.

[36] K.H. Costenbader, E.W. Karlson, Cigarette smoking and systemic lupus ery-thematosus: a smoking gun? Autoimmunity 38 (2005) 541e547.

[37] A. Perl, Oxidative stress in the pathology and treatment of systemic lupuserythematosus, Nat. Rev. Rheumatol. 9 (2013) 674e686.