Oxidation States Sensitization Capacity of Nickel at Higher Nickel

Nickel Allergy in Mice: Enhanced Sensitization Capacity of Nickel at Higher Oxidation States1

Suzan Artik,2,3*† Christian von Vulte´e,3* Ernst Gleichmann,* Thomas Schwarz,‡ and Peter Griem*

Attempts to induce contact hypersensitivity to nickel in mice using, e.g., Ni(II)Cl2 often failed. Here, we report that sensitization was achieved by injecting Ni(II)Cl2 in combination with either CFA or an irritant, such as SDS and PMA, or IL-12, or by administering nickel at higher oxidation states, i.e., Ni(III) and Ni(IV). Although Ni(II), given alone, was ineffective in T cell priming, it sufficed for eliciting recall responses in vivo and in vitro, suggesting that Ni(II) is able to provide an effective signal 1 for T cell activation, but is unable to provide an adequate signal 2 for priming. Immunization of mice with nickel-binding proteins pretreated with Ni(IV), but not with Ni(II), allowed them to generate nickel-specific CD4؉ T cell hybridomas. Ni(II) sufficed for restimulation of T cell hybridomas; in this and other aspects as well, the hybridomas resembled the nickel-specific human T cell clones reported in the literature. Interestingly, restimulation of hybridomas did not require the original Ni(IV)-protein complex used for priming, suggesting either that the nickel ions underwent ligand exchange toward unknown self proteins or peptides or that nickel recognition by the TCR is carrier-independent. In conclusion, we found that Ni(III) and Ni(IV), but not Ni(II) alone, were able to sensitize naive T cells. Since both Ni(III) and Ni(IV) can be generated from Ni(II) by reactive oxygen species, released during inflammation, our findings might explain why in humans nickel contact dermatitis develops much more readily in irritated than in normal skin. The Journal of Immunology, 1999, 163: 1143–1152.

ickel is a metal that is used in many different alloys. It inflamed rather than normal skin. Here, we report that, when skin occurs, for example, in coins, buttons, costume jewelry, inflammation is mimicked in the mouse by coadministration of and items made out of stainless steel or aluminum. Its Ni(II)Cl with either an irritant, CFA, or IL-12, contact hypersen- N 2 widespread use contributes to the fact that nickel is one of the most sitivity to Ni(II), indeed, ensues. In inflammation, reactive oxygen common contact allergens (1). Nickel-specific T cells have been species, such as hydrogen peroxide (H2O2) (4) and hypochlorite detected in sensitized humans and animals. These T cells could be (OClϪ)4, are produced by phagocytes. These powerful oxidants restimulated by Ni(II) in vitro (2–4) and, upon adoptive transfer, in can oxidize Ni(II), i.e., nickel in the oxidation state ϩ2, to the vivo (5). Paradoxically, however, the mechanism underlying de higher oxidation states Ni(III) and Ni(IV), respectively (12, 13). novo sensitization to this common contact allergen is obscure. In The higher nickel oxidation states possess a far greater chemical the two studies reporting on successful sensitization of mice to reactivity than Ni(II). Up until now, however, no study has inves- nickel, special measures had to be taken, such as breeding and tigated the possible significance of higher oxidation states in al- rearing the animals in a nickel-free environment (5) or adminis- lergic contact hypersensitivity to nickel. tering exceedingly high concentrations of nickel salts (6). In con- When assessing the sensitization potential of different gold com- trast, most previous attempts to experimentally sensitize mice to pounds, our group has shown that gold(I) has a poor sensitizing nickel failed (7–11). capacity whereas the higher oxidation state, gold(III), due to its In trying to induce sensitization to nickel in mice, we took into superior chemical reactivity, proved to be a potent sensitizer (14– consideration the clinical experience that, in humans, allergic 17). In this context, it is worth noticing that OClϪ can oxidize contact hypersensitivity to nickel develops much more readily in gold(I) to gold(III) (15, 18). By analogy to gold, here we asked whether the higher oxidation states of nickel would be more potent than Ni(II) in inducing de novo sensitization to this heavy metal. *Division of Immunology, Medical Institute of Environmental Hygiene and †Derma- Two independent experimental approaches were made to assess tology Clinic, Heinrich Heine University, Du¨sseldorf, Germany; ‡Ludwig Boltzmann Institute for Cell Biology and Immunobiology of the Skin, Department of Dermatol- the sensitizing capacity of higher nickel oxidation states. In the ogy, University of Mu¨nster, Mu¨nster, Germany first approach, Ni(III) and Ni(IV), respectively, were used for Received for publication December 28, 1998. Accepted for publication May 11, 1999. priming mice, and specific T cell responses were assessed in vivo The costs of publication of this article were defrayed in part by the payment of page using the mouse ear swelling test (MEST) and the popliteal lymph charges. This article must therefore be hereby marked advertisement in accordance node (PLN) assay. In the second approach, Ni(IV) complexed with with 18 U.S.C. Section 1734 solely to indicate this fact. protein was used for priming, and nickel-specific CD4ϩ T cell 1 This work is part of C. von Vulteé’s doctoral thesis, which was supported by Deut- sche Forschungsgemeinschaft, Bonn, Germany, through a fellowship from the Post- hybridoma clones were established and characterized. graduate Training Program in “Toxicology and Environmental Hygiene” granted to Du¨sseldorf University. 2 Address correspondence and reprint requests to Dr. Suzan Artik, Division of Im- munology, Medical Institute of Environmental Hygiene, Auf’m Hennekamp 50, Ϫ 4 Abbreviations used in this paper: OCl , hypochlorous acid; HSA, human serum al- D-40225 Du¨sseldorf, Germany. E-mail address: [email protected] bumin; MEST, mouse ear-swelling test; MSA, mouse serum albumin; Ni(IV)(OH)2O, 3 S.A. and C.v.V. contributed equally to this paper. nickel oxide hydroxide; PLN, popliteal lymph node; DC, dendritic cell.

Materials and Methods ness compared with the prechallenge value; measurements were performed Mice by using a micrometer (Oditest D 1000 gauge; Dyer, Lancaster, PA). Data shown represent the mean ear-swelling response of groups comprising 5 to Ϫ Specific pathogen-free female C57BL/6J (H-2b) mice obtained from Har- 6 mice, expressed in units of mm ϫ 10 2 Ϯ SD. lan Olac (Bicester, U.K.) were used throughout. Animals were used at 9–20 wk of age at the onset of the experiments. They had free access to Proliferation of auricular lymph node cells drinking water and standard rodent lab chow (No. 1324, Altromin, Lage/ Lippe, Germany). No measures were taken to protect the animals from In addition to determining ear swelling after recall with Ni(II)Cl2, prolif- exposure to nickel: cages (made from plastic) were covered by stainless eration in vitro of auricular lymph node cells was measured. Lymph nodes ϫ 5 steel covers, and drinking water was provided by glass bottles covered with were obtained 48 h after challenge, and 3 10 cells were incubated for ␮ water outlets made from stainless steel. 16 h in the presence or absence of 100 M Ni(II)Cl2 in RPMI 1640 medium (Life Technologies, Paisley, U.K.) supplemented with gentamicin, Cell lines penicillin, streptomycin, essential and nonessential amino acids, 10% FCS, and 50 ␮M 2-ME. Cell proliferation was determined by adding 18.5 kBq ␣Ϫ␤Ϫ Thymoma line BW5147 (TCR ) (19) was kindly provided by Dr. [3H]thymidine at the beginning of the culture, harvesting the cells onto H.-G. Burgert (Max Planck Institute for Immunobiology, Freiburg, nitrocellulose filter, and measuring incorporated radioactivity in a scintil- Germany). lation counter. Reagents Murine popliteal lymph node assay Au(I)TM (disodium aurothiomalate) was kindly provided by E. Tosse (Hamburg, Germany). In addition, the following metal salts were used: The popliteal lymph node (PLN) assay was performed as described (24). ⅐ Briefly, 50 ␮l of saline containing H O and Ni(II)Cl at the concentrations ZnCl2, CoCl2 6H2O, and K2Cr2O7 were purchased from Sigma-Aldrich 2 2 2 indicated, or of either substance alone, were injected s.c. into the left hind- Chemie (Deisenhofen, Germany); Na2PtCl6,Na2PdCl6,Na2PdCl4, and foot pad on day 0. For determination of primary responses, mice (five to six Pt(NH3)4Cl2 were gifts from Degussa (Frankfurt/Main, Germany); ⅐ 3 ⅐ ⅐ ⅐ animals per group) were killed on day 6, and both PLNs were removed. HAuCl4 [ H]2O, CuSO4 5H2O, FeCl3 6H2O, and Ni(II)Cl2 6H20 were purchased from E. Merck (Darmstadt, Germany). Sterile, pyrogen-free 0.9% Cell numbers of each PLN were determined, and the PLN cell count index saline was purchased from Fresenius (Bad Homburg, Germany). Nitroso- of each animal was calculated by dividing the value obtained from the benzene was purchased from Sigma-Aldrich; before use, it was dissolved draining (ipsilateral) PLN through that of the control (contralateral) PLN. in absolute ethanol (Merck) and diluted in 0.9% saline to the concentration For induction of secondary responses, on day Ϫ10, mice were primed by indicated. H O (30%) was purchased from Merck. SDS was obtained from s.c. injection into the left hindfoot pad 50 ␮l of either a mixture of 7.35 2 2 ␮ ␮ Sigma-Aldrich. Recombinant murine IL-12 was kindly provided by S. mol H2O2 and 0.25 mol Ni(II)Cl2 in saline, each substance alone, or Wolf (Genetics Institute, Cambridge, MA). For in vivo injection, IL-12 saline alone. On day 0, mice were challenged for recall by s.c. injection ␮ ␮ was diluted in sterile endotoxin-free saline and injected i.p. at a dose of into the right hindfoot pad of either 0.25 mol Ni(II)Cl2 in 50 l saline, or 500 ng. saline alone. On day 6, mice were killed, and their PLN cell count indices Ni(III) and Ni(IV) were freshly prepared for each experiment because were determined. In control experiments with nitrosobenzene, the doses they are rather unstable and not available commercially. Ni(III) was pre- used for priming and challenge were 50 nmol and 3 nmol nitrosobenzene, respectively. The latter dose was termed “suboptimal” because it was too pared by mixing Ni(II)Cl2 with a 30-fold molar excess of H2O2 (10 mM and 300 mM final concentration, respectively) (12, 20). In one experiment, small for priming, but sufficed for eliciting a secondary PLN response.

catalase (Sigma-Aldrich Chemie) was used to degrade excess H2O2 that had not reacted with Ni(II). After incubating Ni(II) with H2O2 for1h, Ags used for generation of T cell hybridomas catalase (0.68 U/ml) was added to the mixture. The amount of H O pos- 2 2 Mouse serum albumin (MSA), human serum albumin (HSA), and bovine sibly remaining after the addition of catalase was determined by fluores- ribonuclease A (RNase A) were purchased from Sigma. For complex for- cence measurement using a Perkin-Elmer LS-5-Luminescence Spectrom- mation with Ni(II), MSA and HSA were incubated for 1 h with a 100-fold eter (Perkin-Elmer, Norwalk, CT). In the presence of HRP, the fluorescent molar excess (2.9 mM) of Ni(II)Cl in 0.9% saline; the final protein con- scopoletin is known to be oxidized by H O , thereby losing its fluorescence 2 2 2 centration was 2 mg/ml. Complex formation of the proteins with (21, 22). Nickel oxide hydroxide (Ni(IV)(OH) O) was freshly prepared by 2 Ni(IV)(OH) O was conducted as follows: to a solution of the respective oxidation of Ni(II) with NaOCl (12.5% active chlorine; Riedel-de-Haen, 2 protein in 0.9% saline (concentration 2 mg/ml), a 100-fold molar excess Seelze, Germany), as described elsewhere (13, 23); at the alkaline pH (2.9 mM in the case of MSA and HSA) or 55-fold molar excess (7.9 mM occurring after this reaction, Ni(IV)(OH) O forms a black suspension. The 2 in the case of RNase) of Ni(IV)(OH) O was added. Diluted HCl was added suspension was spun down and washed twice with saline to remove excess 2 Ϫ up to complete dissolution of the Ni(IV) compound. After1hofincuba- OCl . The Ni(IV)(OH) O was resuspended in saline by stirring and used 2 tion, the solution was neutralized with diluted NaOH. in the experiments in vivo and in vitro. Although we did not directly dem- onstrate the formation of higher nickel oxidation states by physico-chemical techniques, for the sake of brevity, we shall refer to the mixture of T cell hybridomas

Ni(II) and H2O2 as Ni(III); correspondingly, (Ni(IV)(OH)2O), which re- ␮ Ϫ Mice were immunized by s.c. injection at the base of tail with 100 gof sults from oxidation of Ni(II) through OCl (13, 23), will be referred to as the indicated Ag emulsified in CFA (1:1). After 8–10 days, spleens were Ni(IV). removed, and single-cell suspensions were prepared in PBS. After lysis of MEST erythrocytes in hypotonic buffer (17 mM Tris-HCl (pH 7.2) and 160 mM NH4Cl) and enrichment of splenic T cells by nylon wool separation, T cells A modification of the MEST was used, as described by van Hoogstraten et (2 ϫ 106/ml) were cultured in supplemented RPMI 1640 medium and al. (5). For induction of sensitization, groups of mice were injected intra- restimulated with Ag (0.1 mg/ml) in the presence of irradiated (20 Gy) dermally with 50 ␮l into both flanks, each injection containing either 1) 1 syngeneic spleen cells (2 ϫ 106/ml) as APC. After 3 days, living cells were ␮ ␮ mol Ni(II)Cl2 in 0.9% saline, 2) a mixture of 1 mol Ni(II)Cl2 and 30 isolated by gradient centrifugation on Ficoll-Paque (Pharmacia, Freiburg, ␮ ␮ ␮ mol H2O2, i.e., Ni(III), in saline, 3) 30 mol H2O2 in saline, or 4) 1 mol Germany) and expanded in culture medium for 3 days in the presence of 20 Ni(II)Cl2 in saline mixed 1:1 with CFA. In the case of irritants, the intra- U/ml rIL-2. T cell blasts isolated by Ficoll-gradient centrifugation were dermal injections, 50 ␮l each, into both flanks were administered subse- fused with BW5147 tumor cells using polyethylene glycol 1500 (Boehr- ␮ quently, as follows: 2% SDS in saline given 4 h before 1 mol Ni(II)Cl2, inger, Mannheim, Germany), as described by the manufacturer. After se- ␮ or 600 ng PMA given 1 h before 1 mol Ni(II)Cl2. In the case of IL-12, lection for hybrid cells in hypoxanthine/aminopterin/thymidine (HAT) me- i.p. injections of 500 ng rIL-12 in saline were given at 24 h and 3 h before dium for 2 wk, cells were cultured in medium containing hypoxanthine and ␮ the intradermal injections of 1 mol Ni(II)Cl2 into each flank. Ten days thymidine. T cell hybridomas reacting to Ag in the stimulation assay de- ␮ later, mice were challenged for recall by injecting 0.2 mol Ni(II)Cl2 in 20 scribed below were subcloned by limiting dilution. ␮l saline, or saline alone, into the pinnae of each ear, as described (5). Mice Ag-induced stimulation of T cell hybridomas was tested in standard ␮ ϫ 5 ␮ were sensitized to chromate by injecting an emulsion of 1 mol K2Cr2O7 IL-2 release assays. For this purpose, hybridomas (1 10 /200 l sup- in 50 ␮l saline plus 50 ␮l of CFA into each flank (5). Ten days later, they plemented RPMI medium) were stimulated with either native protein ␮ were challenged for recall by injecting 70 nmol K2Cr2O7,or0.2 mol (MSA, HSA, or RNase), Ni(IV)(OH)2O-pretreated protein, Ni(II)Cl2-pre- ␮ Ni(II)Cl2,in20 l saline into the pinnae of each ear. Forty-eight hours (in treated protein, or Ni(II)Cl2 alone. For screening of hybridomas, a protein one experiment also 24 and 72 h) after challenge, delayed-type hypersen- concentration of 0.1 mg/ml and a nickel concentration of 100 ␮M were sitivity reactions were determined by measuring the increment in ear thick- used. All assays were performed in the presence of syngeneic spleen cells The Journal of Immunology 1145

FIGURE 1. Coadministration of irritants with Ni(II) or administration of the higher nickel oxidation states Ni(III) and Ni(IV) lead to the induction of ␮ contact hypersensitivity to Ni(II). Groups of mice were injected intradermally into each ﬂank with (A)1 mol Ni(II)Cl2 in saline after injection of 2% SDS ␮ ␮ ␮ 4 h earlier or with either of these substances alone, 1 mol Ni(II)Cl2 in CFA, or 1 mol Ni(II)Cl2 after injection of 600 ng PMA 1 h earlier, (B)1 mol ϫ ␮ ␮ ␮ Ni(II)Cl2 in saline, or saline alone, after pretreatment of mice with 2 500 ng rIL-12 or saline, or (C)1 mol Ni(II)Cl2,1 mol Ni(II)Cl2 plus 30 mol ␮ H2O2, i.e., Ni(III), or 1 mol Ni(IV)(OH)2O suspended in saline. Ten days after priming, animals were challenged for recall by injection into each ear of Յ ء ␮ ϫ Ϫ2 ϩ either 0.2 mol Ni(II)Cl2 in saline, or saline alone. Data are expressed as mean ear-swelling response ( 10 mm) SD at 48 h after challenge ( , p .(p Յ 0.001, compared with the groups denoted by bars ,ءءء p Յ 0.01; and ,ءء ;0.05

(5 ϫ 105) as APC. Hybridomas and APC left without Ag served as neg- PBS) were stained for 20 min at 4°C in 96-well plates. Fluorescence was ative controls. Hybridoma stimulation lasted 24 h, thereafter culture su- measured using a FACScan flow cytometer and the CellQuest analysis pernatants (50 ␮l) were removed, frozen at Ϫ80°C, and, after thawing, program (Becton Dickinson, Mountain View, CA). tested for IL-2 content by adding spleen cells (1.5 ϫ 104) that had been cultured for 24 h with Con A (2.5 ␮g/ml) and for another 24 h without mitogen. Cell proliferation was measured by adding 18.5 kBq [3H]thymi- Statistical analysis dine for the last6hof24h,asdescribed above. Statistical significance of results was determined by ANOVA tests. The In some experiments, APC were pulsed with Ni(II)Cl for 24 h and then 2 level of significance was set at p Յ 0.05. All experiments were performed washed twice before addition of T cell hybridoma clones. In the assay, ␮ at least twice to assure reproducibility. either no Ni(II)Cl2 or Ni(II)Cl2 (100 M) was added. In some other experiments, fixed APC were used (25). Briefly, spleen cells were washed in PBS and resuspended in 0.5 ml 0.05% glutaraldehyde/106 cells for 60 s at room temperature. The reaction was stopped by adding an equal volume of Results 0.2 M L-lysine to the cells and washing them twice in PBS. Then, Ni(II) at Secondary responses to nickel in the MEST the concentrations indicated, or medium alone, was added to these fixed APC and, finally, T hybridoma cells were added. To mimic nickel contact with inflamed skin, mice were injected into the flanks with a 2% solution of the irritant SDS, while control ␮ MHC-blocking experiments mice received saline only. Two hours later, 1 mol of Ni(II)Cl2 To assess the MHC dependence of nickel-specific T cell hybridoma clones, was injected into the same skin areas on both flanks. Upon chal- ␮ blocking experiments were performed. T cell hybridomas were cocultured lenge with 0.2 mol Ni(II)Cl2 into the ears, only mice primed with ␮ with syngeneic spleen cells as APC and 100 M Ni(II) either without mAb Ni(II)Cl2 plus SDS, but not Ni(II)Cl2 or SDS alone, showed a or in the presence of 10 ␮g/ml anti-MHC class II mAb (clone M5/114; significant increase in ear thickness (Fig. 1A). When other inflam- Boehringer), or isotype-matched IgG2b, ␬ mAb (clone R35-95; PharMin- gen, Hamburg, Germany) for control. mation-inducing substances were employed, we observed that administration of PMA or CFA also was able to induce sensitization Analysis of TCR by flow cytometry to Ni(II) (Fig. 1A). The following anti-mouse FITC-conjugated mAb, all purchased from IL-12 has been identified as a critical mediator in the induction PharMingen, were used to determine the TCR V␣ and V␤ repertoire of of contact hypersensitivity (26, 27). To study whether IL-12 en- nickel-specific T cell hybridoma clones: pan ␣␤ (H57-597), V␣2 (B20.1), ables sensitization to Ni(II), mice received two i.p. injections of V␣3.2b (RR3-16), V␣8 (B21.14), V␣11b,d (RR8-1), V␤2 (B20.6), V␤4 500 ng rIL-12 before administration of Ni(II). As shown in Fig. (KT4), V␤5.1,5.2 (MR9-4), V␤6 (RR4-7), V␤7 (TR310), V␤8.1,8.2 (MR5-2), V␤8.3 (1B3.3), V␤9 (MR10-2), V␤10b (B21.5), V␤11(RR3-15), 1B, pretreatment of mice with rIL-12 before injection of Ni(II) V␤12 (MR11-1), V␤13 (MR12-3), V␤14 (14-2), and V␤17a (KJ23); mAb resulted in the priming of T cells for a nickel-specific secondary V␤3 (KJ25) was PE conjugated. T cell hybridoma cells (105 cells/20 ␮l reaction. 1146 ENHANCED SENSITIZATION VIA NICKEL AT HIGHER OXIDATION STATES

Sensitization of mice by nickel at higher oxidation states Since the application of irritants like SDS induces inflammatory infiltrates consisting of activated neutrophils and other phagocytes (28, 29) and, hence, release of reactive oxygen species, such as Ϫ H2O2 and OCl , we examined whether H2O2 itself can enhance sensitization to Ni(II)Cl2. Mice were injected at the flanks with a mixture of Ni(II)Cl2 and H2O2 that is known to result in the formation of Ni(III) (12, 20). After priming with Ni(III), a secondary response upon challenge with Ni(II) was observed (Fig. 1C),

whereas injection of Ni(II) or H2O2 alone did not result in successful T cell priming (Fig. 2A). Signiﬁcantly increased ear-swelling responses were obtained at 48 and 72 h after challenge, but not after 24 h (Fig. 2A). Similar results as with Ni(III) were found with Ni(IV); mice that had been primed with the higher nickel oxidation

state Ni(IV) by injection of a suspension of Ni(IV)(OH)2O showed a secondary response to Ni(II), as determined by MEST (Fig. 1C). Speciﬁcity of the ear-swelling response was demonstrated by

criss-cross immunization with K2Cr2O7 (Fig. 2B), a common sensitizer that has also been studied in the MEST (5, 9). A signiﬁcant

ear-swelling response was seen in mice primed with K2Cr2O7 and challenged for recall with K2Cr2O7. By analogy, mice primed with Ni(III) responded to challenge with Ni(II). No cross-reaction be-

tween K2Cr2O7 and Ni(II) was observed.

Proliferation of auricular lymph node cells in vitro In some experiments, the proliferative response of cells from the draining lymph nodes also was determined. Auricular lymph node cells from mice that had been primed at the ﬂanks with Ni(III) and challenged at the ears with Ni(II) showed enhanced spontaneous proliferation in vitro, indicating an ongoing secondary response. This effect became even more prominent when the lymph node ␮ cells were incubated in the presence of 100 M Ni(II)Cl2 (Fig. 2C). Thus, the proliferative response of cells from the draining lymph nodes corresponded with the results of the MEST performed in the donors of these lymph nodes.

Primary and secondary responses to nickel in the PLN assay Unlike the MEST, which detects only secondary immune responses, the PLN assay, another well-established test for identiﬁcation of sensitizing chemicals (24), allows detection of primary immune responses as well. Primary PLN responses were studied ﬁrst. Mice were

injected into one hind footpad with a mixture of H2O2 and Ni(II)Cl2 FIGURE 2. Coadministration of H2O2 enables induction of contact hy- to examine whether the combination would induce a stronger primary persensitivity to Ni(II). A, Groups of mice were injected intradermally into ␮ ␮ PLN response than each compound alone. As shown in Fig. 3A,10 each flank with either 1 mol Ni(II)Cl2 plus 30 mol H2O2, i.e., Ni(III), in saline, or with either of these substances alone. Ten days after priming, mM Ni(II)Cl2 completely failed to induce PLN enlargement. When animals were challenged for recall by injection into each ear of either 0.2 H2O2 was admixed to Ni(II)Cl2, however, distinct PLN enlargement ␮ mol Ni(II)Cl2 in saline. B, Groups of mice were primed by intradermal was observed. With all three H2O2 concentrations used, the PLN re- ␮ ␮ injection into each flank of either 1 mol Ni(II)Cl2 plus 30 mol H2O2, action to the Ni(II)Cl2/H2O2 mixture, i.e., Ni(III) (12, 20), was sig- i.e., Ni(III), either of these substances alone, or an emulsion of 1 ␮mol nificantly higher than the reaction caused by H O alone. 2 2 K2Cr2O7 and CFA. Ten days after sensitization, animals were challenged Secondary PLN responses were studied by challenging mice by ␮ for recall by injection into each ear of either 0.2 mol Ni(II)Cl2 or 70 nmol s.c. injection of Ni(II)Cl2 into the right foot pad. A secondary K2Cr2O7 in saline. C, Recall response in vitro to Ni(II) by auricular lymph response to nickel could be elicited only in mice that had been node cells obtained from groups of mice used in MEST. Lymph node cells primed with Ni(III) (Fig. 3B). Notably, Ni(II) proved to be suffi- from each group were pooled and cultured with 18.5 kBq [3H]thymidine cient for eliciting specific secondary responses. Specificity of the for 16 h in the presence or absence Ni(II)Cl2, as indicated. Data in A and response to Ni(II) was demonstrated in criss-cross immunization B were obtained at 48 h after challenge; in C, mean cpm values of tripli- Յ ءءء Յ ءء Յ ء experiments using nitrosobenzene (Fig. 3C). Nitrosobenzene is a cates are shown. , p 0.05; , p 0.01; , p 0.001. In A, asterisks denote differences in comparison with each other group; in B, asterisks protein-reactive chemical known to elicit primary and specific denote differences between groups compared by brackets. secondary PLN responses.5 Only nitrosobenzene-primed mice

5 M. Wulferink, J. Gonzalez, C. Goebel, and E. Gleichmann. T cells ignore ani- exhibited a positive PLN response upon recall with nitrosoben- line, a phohapten, but respond to its reactive metabolites generated by phagocytes: possible implications for the pathogenesis of toxic oil syndrome (TOS). (Submit- zene, and only mice primed with Ni(III) responded to Ni(II) upon ted for publication.) recall; there was no cross-reaction. The Journal of Immunology 1147

Chemical evidence for the use of nickel at higher oxidation states and for its sensitizing capacity The results described above (Figs. 1–3) raised two important questions. First, how certain can we be to have employed nickel at higher oxidation states? Second, how sure can we be that the successful priming observed with oxidized Ni(II) was indeed due to Ni(III) and Ni(IV), respectively, and not to an adjuvant effect of Ϫ the oxidants used, i.e., H2O2 and OCl ? With regard to the ﬁrst question, it should be realized that direct, physico-chemical detection of nickel at higher oxidation states is technically very de- manding and has not been achieved up to now in a cell culture system, not to mention an animal; it was therefore not attempted in the present investigation. Instead, we took indirect approaches to ascertain that, indeed, higher oxidation states of nickel were used. On inspection, we noted disappearance of the characteristic green color of the Ni(II)-water complex after addition of the oxidants Ϫ H2O2 and OCl , respectively, and this was conﬁrmed by UV-Vis spectroscopy. Addition of H2O2 to Ni(II)Cl2, as used for the generation of Ni(III), resulted in partial disappearance, and addition of OClϪ, as used for the generation of Ni(IV), resulted in complete disappearance of the green color. The green color, however, reap- peared if thiol-containing reducing compounds, such as glutathi- Ϫ one, were added (data not shown). Addition of OCl to Ni(II)Cl2, followed by centrifugation, resulted in formation of a black precipitate that has been shown to contain Ni(IV), by means of in- frared spectroscopy, x-ray photoelectron spectroscopy, and electron spin resonance (13). Hence, the combined evidence indicates that indeed we employed nickel at higher oxidation states. With regard to the second question, the following experiments were undertaken to remove the oxidants added to Ni(II) for syn- thesis of the higher oxidation states of nickel and, thus, exclude a

possible adjuvant effect of residual oxidant. In the case of H2O2, catalase was added to the mixture of Ni(II) and H2O2 to degrade excess H2O2, and the remaining H2O2 was determined (21, 22). We calculated that, after catalase treatment, a maximum of 1 nmol ␮ H2O2 could have been present in the volume (100 l) injected for induction of sensitization. With the Ni(III) remaining after the deg-

radation of H2O2, we performed the MEST. The ear-swelling response of mice primed with Ni/H2O2 plus catalase was comparable with that of mice primed with Ni/H2O2 (Fig. 4). In the case of addition of OClϪ, the resulting suspension of Ni(IV) was precip- itated by centrifugation so that excess OClϪ in the supernatant was FIGURE 3. Injection of Ni(III) evokes a primary response and allows removed, as described under Materials and Methods. The possi- elicitation of Nickel-specific recall responses in the PLN assay. A, On day bility remained, however, that some OClϪ was occluded in the 0, groups of mice received a single s.c. injection into the left hindfoot pad black precipitate. For this purpose, in one test, instead of resus- of either saline (column at the very left), Ni(II)Cl2 alone (second column pending the precipitate in saline, it was dissolved in diluted HNO3, from the left), the indicated concentration of H2O2 alone (hatched bars), or 10 mM of Ni(II)Cl plus the indicated concentration of H O (black bars). and 0.3 M AgNO3 was added. This did not lead to the character- 2 2 2 istic white precipitate indicative of AgCl that would have been On day 6, PLN cell counts were determined. B, Groups of mice were Ϫ ␮ formed had OCl been present. Taken together, these results in- injected into the left hindfoot pad with a mixture of 7.35 mol H2O2 and ␮ 0.25 mol Ni(II)Cl2 (i.e., Ni(III)), with either substance alone in saline, or dicate that the induction of sensitization obtained with Ni(III) and with saline alone. Ten days after immunization, mice were challenged for Ni(IV), respectively, was due to the higher nickel oxidation states recall by s.c. injection into the right hindfoot pad of either 0.25 ␮mol per se; adjuvanticity of the oxidants used played no, or at least no Ni(II)Cl2 in saline, or saline only, as indicated. Six days later, mice were significant, role in this process. killed, and the PLN cell count index was determined. C, Groups of mice were primed by injection into the left hindfoot pad of either a mixture of Generation of CD4ϩ nickel-specific hybridoma clones Ni(II)Cl2 plus H2O2 in saline (same doses as in B), either of these substances alone, or 50 nmol nitrosobenzene (NB). Two months later, when To prime mice to nickel, we decided to use nickel and carrier PLN indices of nitrosobenzene-treated mice had reverted to background proteins, such as albumin, known to have nickel-binding capacity level, mice were challenged for recall by injection into the right hindfoot (30). In the first experiment, mice were immunized against pad of either 0.25 ␮mol Ni(II)Cl or a suboptimal dose of nitrosobenzene 2 Ni(II)Cl -preincubated MSA. After generation of T cell hybridoma p Յ 2 ,ءء ;p Յ 0.05 ,ء .nmol). The PLN index was determined 6 days later 3) p Յ 0.001. In A and C, asterisks denote differences between clones from splenic T cells, only one clone of 136 clones tested ,ءءء ;0.01 groups compared by brackets; in B, in comparison with each other group. exhibited specificity for Ni(II); i.e., it recognized Ni(II) in the presence of syngeneic spleen cells, irrespective of whether MSA was present or not (Table I). Thus, this clone was not specific for a 1148 ENHANCED SENSITIZATION VIA NICKEL AT HIGHER OXIDATION STATES

FIGURE 4. Sensitization by Ni(II) plus H2O2 is caused by Ni(III) per se and not by adjuvanticity of the H2O2. Groups of mice were injected intra- ␮ ␮ dermally into each ﬂank with either 1 mol Ni(II)Cl2 plus 30 mol H2O2, i.e., Ni(III), in saline, or with either of these substances alone. In one group

of mice, after incubating Ni(II) with H2O2 for 1 h, catalase (3000 U/mg, i.e., 0.68 U/ml) was added to the mixture to delete excessive H2O2 that had not reacted with Ni(II). With the Ni(III) after removal of H2O2 we then performed the MEST. Ten days after priming, animals were challenged for ␮ recall by injection into each ear of either 0.2 mol Ni(II)Cl2 in saline. Data are expressed as mean ear-swelling response (ϫ 10Ϫ2 mm) ϩ SD at 48 h p Յ 0.001; compared with ,ءءء ;p Յ 0.01 ,ءء ;p Յ 0.05 ,ء) after challenge the groups denoted by bars).

Ni(II)-MSA complex (data not shown). Next, we tested HSA as nickel-complexing agent because of its inherent antigenicity in mice: 59 of 360 hybridomas proved to be HSA specific, but no nickel-specific T cell hybridoma was identified upon immunization

with Ni(II)Cl2-preincubated HSA. We then chose Ni(IV)(OH)2O for generation of T cell hybridomas. Following immunization

against MSA preincubated with Ni(IV)(OH)2O, 61 Ni(II)-speciﬁc T cell hybridomas were identiﬁed among 350 hybridomas screened; in addition, two T cell clones reacted against MSA alone and thus were autoreactive. In another immunization, we used bo-

vine RNase A preincubated with Ni(IV)(OH)2O; this protein, too, is known to bind Ni(II) (31). This time, 35 Ni(II)-specific T cell hybridomas were found among 384 hybridomas tested (Table I). Again, the nickel-specific T cell clones did not require the presence of the carrier protein used for immunization (Table I). Hence, for successful generation of Ni(II)-specific clones, not the kind of pro- FIGURE 5. Characterization of nickel-specific CD4ϩ T cell hybrid- tein, but the oxidation state of the nickel compound used for im- omas in the presence of nonfixed spleen cells as APC. After 24-h cell munization seems to be the crucial factor. Once they had been culture, supernatants were removed and assayed for IL-2 release. Means of generated, however, our nickel-specific T cell hybridomas reacted triplicate values are shown. A, Reactivity of nickel-specific T cell hybrid- equally well to Ni(II) and Ni(IV) (data not shown). It is noteworthy oma clones to increasing concentrations of Ni(II)Cl2. Background prolif- Ϯ that the T cell hybridomas recognized nickel irrespective of eration of IL-2-dependent cells was 465 95 cpm. B, Cross-reaction of nickel-specific clone F4 13A6 with CoCl . Reactivity to increasing con- whether the original protein used for priming was present during 2 centrations of metal [Me(II)] salts, namely Ni(II)Cl and CoCl , is shown. recall or not (Table I). 2 2 Background proliferation of IL-2-dependent cells was 1102 Ϯ 677 cpm. C, Nickel-specific T cell hybridomas are MHC class II restricted. Five clones ϩ were cocultured with 100 ␮M Ni(II)Cl , either in the absence of mAb Table I. Survey of the reactivity of CD4 T cell hybridomas obtained 2 after immunization against proteins complexed with either Ni(II) or (open columns), or in the presence of anti-MHC class II mAb (black col- Ni(IV) umns) or isotype-matched control mAb (hatched columns). Background proliferation of IL-2-dependent cells was 2490 Ϯ 623 cpm. Number of T Cell Hybridomas Reacting Against Ag Used for Immunization Protein alone Ni Total No. Tested Dose-response relationships in responses of nickel-specific T cell hybridomas MSA/Ni(II) 0 1a 136 HSA/Ni(II) 59 0 360 Responses of five different, randomly selected nickel-specific T a MSA/Ni(IV) 2 61 350 cell hybridoma clones to increasing concentrations of Ni(II)Cl are RNase/Ni(IV) 162 35a,b 384 2 shown in Fig. 5A. Four of five clones reached their maximum IL-2 a Clones reacted to Ni(II) and Ni(IV) irrespective of whether the protein used for response at about 200 ␮M Ni(II), a concentration that is already immunization was present or not. b In addition to the clones listed here, three clones were found exclusively to react toxic for Con A-stimulated T cells (4, 17) and reduced the prolif- against Ni(IV)-pretreated RNase (34). erative capacity of T cell hybridoma clones (data not shown). The Journal of Immunology 1149

Table II. Staining of Ni-specific T cell hybridoma clones with FITC- of these epitopes could, for instance, depend on the way nickel conjugated, TCR-specific mAb interacts with self protein, on how nickel-protein complexes are processed, or on the stability of nickel-peptide-MHC complexes T Cell Hybridoma V␣ V␤ Pan ␣/␤ that are presented by APC and subjected to repeated washing. The F4 8A2 8 8.1, 8.2 ϩa most straightforward possibility would be that T cells recognize F4 11C4 —b 4 ϩ Ni(II) complexed with dominant self peptides that are naturally F4 11C2 — — ϩ presented by MHC class II molecules (32, 33). In this case, it F4 4C5 — — ϩ might be expected that T cells also recognize fixed APC to which ϩ F4 13A6 — 14 Ni(II) was added after the fixation. Alternatively, before peptide F4 9D2 8 — ϩ F4 3A6 — 13 ϩ presentation by MHC II molecules Ni(II) could form chelate com- F4 14C4 — — ϩ plexes with soluble or membrane-bound self proteins; the changes F4 17C3 — — ϩ in protein conformation thus caused would alter processing of the ϩ F4 8B5 — — respective proteins and lead to presentation of different Ni(II)-pep- F4 1D1 — — ϩ tide complexes or of cryptic self peptides (17, 34). In this case, the a ϩ denotes positive staining. T cells would recognize nickel-induced, processing-dependent b The gene segment could not be identified with the set of mAbs used. epitopes that could not be formed by APC that were fixed before the addition of Ni(II). Limited cross-reactivity of T cell hybridoma clones with other To address these questions, we compared nonfixed and glutar- metal ions aldehyde-fixed APC in hybridoma stimulation assays. Eighteen It has been shown that part of the human nickel-specific T cell nickel-specific T cell hybridoma clones were cocultured with non- clones cross-react with copper, palladium, and/or cobalt (2, 4). To fixed or fixed APC in the absence or presence of Ni(II). One of answer the question, whether some of the murine nickel-specific T these clones, F4 9D2, still responded when Ni(II) was added to cell hybridoma clones also show cross-reactivity, several transition previously fixed APC (Fig. 6A), whereas all other clones failed to metals, namely Au(I), Au(III), Co(II), Cu(II), Pd(II), Pd(IV), do so (clone F4 14C4 shown as example). Consistent with this Pt(II), and Pt(IV), were tested in specificity assays. Only one clone result, clone F4 D92 also was the only hybridoma clone of those of 23 tested, F4 13A6, responded to one of these metal salts, tested that reacted with APC that were first pulsed with Ni(II) and namely Co(II) (Fig. 5B). The stimulatory concentration range for thereafter fixed and washed (data not shown). Apparently, clone Ni(II) and Co(II) was between 100 ␮M and 10 ␮M, with a max- F4 9D2 recognizes a Ni(II)-peptide-MHC class II complex that is imum response at 100 ␮M. Even though the maximum response to processing independent and involves a dominant self peptide that Ni(II) exceeds that to Co(II) in the experiment shown, a preferen- is not destroyed or deformed by glutaraldehyde fixation. For the tial response of clone F4 13A6 to Ni(II) in comparison to Co(II) other seventeen clones, exemplified by F4 14C4, it cannot be con- cannot be deduced from these results, because in other experiments cluded with certainty, however, that they recognize processing- the response to Co(II) was sometimes higher than that to Ni(II) dependent epitopes induced by nickel. The reason for this is that (data not shown). the glutaraldehyde fixation could have destroyed or deformed the naturally presented self peptides with which Ni(II) prefers to com- MHC dependence and TCR elements of nickel-specific T cell plex. To decide between these possibilities, we stimulated 10 T hybridoma clones cell hybridoma clones with nonfixed APC that had been pulsed for Human nickel-specific T cell clones presumably recognize Ni(II) 24 h with Ni(II), or medium alone, and had then been washed. In chelate complexes formed with MHC molecules and unknown self the stimulation assay, either no Ni(II) or 100 ␮M Ni(II) was added. peptides bound to them (4, 32). To test the MHC dependence of Data of three representative clones are shown in Fig. 6B. When ϩ nickel-specific CD4 T cell hybridoma clones of mice, blocking Ni(II) was added during the stimulation assay, all clones re- experiments with anti-MHC class II mAb were performed. Fig. 5C sponded, irrespective of whether Ni(II)-pulsed washed APC or me- shows the results obtained with five nickel-specific T cell hybrid- dium-pulsed washed APC were used. This was expected because ␮ oma clones that were cocultured with APC and 100 M Ni(II)Cl2 the APC were not fixed and were not subjected to washing after in the presence or absence of anti-MHC class II mAb or isotype- addition of Ni(II) in the assay. Only one clone, F4 9D2, also re- matched control mAb. In the presence of anti-MHC class II mAb, sponded to Ni(II)-pulsed washed APC when no additional Ni(II) the response of all five T cell hybridoma clones to Ni(II)-treated was provided in the assay; another clone, F4 13A6, showed a low, APC was completely inhibited; in contrast, the IgG2b control mAb but reproducible, response under these conditions. The other eight failed to exert an inhibitory effect. clones, exemplified by F4 1D1, failed to respond to Ni(II)-pulsed Staining of nickel-specific T cell hybridoma clones with a panel washed APC and reacted only if the provision of Ni(II) was re- of TCR V␣␤-specific mAb was performed to assess TCR diversity newed in the stimulation assay. Apparently, this lack of response of nickel-specific T cells. Eleven T cell hybridoma clones were was due to the loss of Ni(II) from the recognized Ni(II)-peptide- stained with FITC-conjugated, TCR-specific anti-mouse mAb. MHC complexes during washing of Ni(II)-pulsed APC. Interest- Four clones of 11 tested showed positive staining, with one of the ingly, clone F4 9D2, which also reacted to fixed APC treated with anti-V␤ mAb tested, and for two clones a V␣-chain was identified Ni(II) (Fig. 6A), recognized Ni(II)-pulsed washed APC without (Table II). All four clones express different V␤-chains, whereas the addition of Ni(II) in the assay (Fig. 6B); together with data shown two V␣-chains detected were identical. This result does not sug- in Fig. 5C, these findings suggest that clone F4 9D2 recognizes a gest a preferential V␣␤-chain repertoire of murine nickel-specific stable complex formed by Ni(II) and a dominant self peptide pre- T cells, but the number of TCR elements is too small to draw any sented by MHC II molecules. definitive conclusions.

Response of nickel-speciﬁc T cell hybridomas to pretreated APC Discussion It is conceivable that different nickel-speciﬁc CD4ϩ T cell hybrid- Here we demonstrated that mice, albeit reared and kept under con- oma clones recognize different nickel-induced epitopes. Formation ventional conditions, were readily sensitized to Ni(II) if one of the 1150 ENHANCED SENSITIZATION VIA NICKEL AT HIGHER OXIDATION STATES

poor sensitizer Ni(II) to Ni(III) and Ni(IV), where the latter two succeeded in priming the animals. This possibility is supported by the results obtained under the second and third condition speciﬁed above: Ni(III) was found capable of priming mice for nickel-spe- ciﬁc recall responses in vivo and in vitro, whereas Ni(II) alone, or

H2O2 alone, were not. As far as the third condition, administration

of Ni(IV), is concerned, Ni(IV)(OH)2O proved to be effective in two different experimental approaches. First, Ni(IV), but not Ni(II), primed mice for recall responses in the MEST. Second, when complexed with protein, Ni(IV), but not Ni(II), effectively primed CD4ϩ T cells so that we succeeded in generating nickel- specific hybridomas. It should be realized, however, that, once the T cells had been primed, the sole provision of Ni(II) sufficed for elicitation of secondary responses in all test systems used in vivo and in vitro. Although we cannot exclude distinct mechanisms of adjuvanticity operating in each of the conditions specified, it is conceivable that there is a final common pathway for their ability to induce hypersensitivity. This could be the activation and mobilization of nickel-laden dendritic cells (DC) to emigrate to the draining lymph node and there provide signal 1, i.e., nickel-induced neoantigen, together with signal 2 to naive T cells. The activation of resident DCs is known to be inducible by nonspecific tissue lesion in their environment (36, 37). The superior capacity of Ni(III) and Ni(IV) for induction of sensitization presumably is due to their higher chemical reactivity and, hence, greater toxicity when compared with Ni(II). Thus, unlike Ni(II), Ni(III) and Ni(IV) could trigger FIGURE 6. Comparison of T cell hybridoma responses to Ni(II) pre- the activation and mobilization of DCs by exerting a nonspecific sented by fixed and nonfixed APC, respectively. A, Only clone F4 9D2 local toxicity, a “danger” signal (38). Although CFA, PMA, SDS, responded when Ni(II) was added to already fixed APC. Cells of F4 9D2 and a representative nonresponding clone (F4 14C4), respectively, were and H2O2 per se are able to activate DC, it is conceivable that the cocultured with unfixed (open columns) or glutaraldehyde-fixed (black col- mobilizing effect of all these agents is enhanced when combined

umns) APC in the presence or absence of Ni(II)Cl2. After 24 h of coculture, with the greater chemical reactivity of nickel at higher oxidation supernatants were removed and assayed for IL-2 release. Means of tripli- states. This is supported by our results obtained from experiments cate values are shown. Background proliferation of IL-2-dependent cells for induction of nickel-specific T cell hybridomas. Here, CFA plus 6 was 946 Ϯ 175 cpm. B, Nonfixed APC (2.5 ϫ 10 /ml) were cultured for 0.15 ␮mol Ni(IV) proved to be more effective in sensitizing than ␮ 24 h in the absence (open bars) or presence of 100 M Ni(II) (gray bars) CFA plus the same dose of Ni(II). Other than for the induction of and were then washed twice before they were added to T cell hybridoma T cell hybridomas, priming with Ni(II) admixed to CFA was suc- clones; no Ni(II) was added during the stimulation assay. Alternatively, APC were cultured 24 h in the absence (hatched bars) or presence of Ni(II) cessful in the MEST; there, however, the nickel dose used (1.0 ␮ (closed bars) and then washed. The Ni(II)-pulsed washed APC as well as mol/mouse) was 6.5-fold higher than that used for induction of 100 ␮M Ni(II) were then added to T cell hybridoma clones. Background nickel-specific T cell hybridomas. proliferation of IL-2-dependent cells was 1373 Ϯ 246 cpm. Obviously, Ni(II) can provide signal 1 but, when given alone, lacks the capacity for evoking a sufficient signal 2 for the priming of naive T cells. As evident from the fourth condition specified following conditions was used for priming: 1) coadministration of above, the lack of signal 2 could also be compensated for by sys- Ni(II) with either CFA or an irritant, such as SDS or PMA; 2) temic treatment with rIL-12. IL-12 appears to be critically in- volved in the induction of contact hypersensitivity, since blocking administration of Ni(III), provided by mixing Ni(II)Cl2 and H2O2 of IL-12 in vivo by injection of neutralizing Abs directed against (12, 20); 3) administration of Ni(IV)(OH)2O; or 4) systemic pretreatment of mice with rIL-12, followed by administration of IL-12 inhibits contact sensitization against haptens (26, 27). The current concept of the induction phase of sensitization implies that Ni(II). The s.c. injection of Ni(II)Cl2 alone was insufficient for the induction of hypersensitivity. Sensitization was monitored both in Langerhans cells migrate out of the skin to the regional lymph vivo, using the PLN assay and the MEST, and in vitro, by mea- nodes. During this migration, they mature into effective APCs suring the proliferation of cells from the draining lymph node and with all the costimulatory molecules required for T cell priming. by IL-2 production by nickel-specific CD4ϩ T cell hybridomas. Macatonia et al. showed that DCs are a potent source of IL-12 What makes the four conditions specified above special that and thereby direct the development of Th1 cells from naive ϩ they allow for induction of sensitization to nickel? At the present CD4 T cells (39). The dominant role of IL-12 during sensitiza- time, we can only speculate on this. With regard to the first con- tion is also indicated by the fact that IL-12 was identified as the dition, it is not known which of the many different effects of CFA first cytokine that is able to prevent and even break established and irritants was responsible for the adjuvant effect seen in priming hapten tolerance (27, 40, 41). to Ni(II). A possible explanation could invoke the fact that SDS Theoretically, another possibility that could explain the greater and CFA induce massive dermal infiltration by, and activation of, immunogenicity of higher nickel oxidation states is that these phagocytic cells (28, 29, 35), and this yields, among other, a high might bind with greater affinity than Ni(II) and thus form more

local production of reactive oxygen species, such as H2O2 and stable complexes with proteins and peptides; a quantitative differ- OClϪ. Conceivably, these reactive oxygen species oxidized the ence in stability of protein complexes could result in increased The Journal of Immunology 1151 immunogenicity. Consistent with this possibility, the binding con- tides. If so, this would imply that the recognition of Ni(II) by the stants for Ni(III) to terpyridine are greater than for Ni(II) (42). In TCR is carrier independent to a certain degree, similarly as has any event, however, a supposed greater stability of protein or pep- been observed with some of the trinitrophenyl-specific T cell tide complexes with nickel at higher oxidation states did not result clones (47, 48) and some of the human nickel-specific T cell in formation of neoantigens that were so stable that in the elicita- clones (49). tion of recall responses they would have required presence of nick- With respect to their MHC class II dependence, TCR diversity, el-protein complexes that were identical with those used for prim- limited cross-reactivity, and recognition of processing-dependent ing, e.g., bovine RNase complexed with Ni(IV). and -independent nickel epitopes, the murine CD4ϩ nickel-specific What is the chemical basis of the oxidation of Ni(II) to the T cell hybridomas reported here closely resemble the human nick- higher oxidation states of nickel, which, as we hypothesize, may el-specific T cell clones described in the literature (2, 4). As sev- take place in vivo? With regard to the formation of Ni(III), it has eral of their nickel-specific CD4ϩ clones also reacted to glutaral- to be taken into account that, due to charge delocalization favored dehyde-fixed APC to which Ni(II) was added after the fixation, by ligands with high electron density, the standard electrode po- Weltzien and coworkers (4) concluded that these clones are pro- tential of Ni(II)/Ni(III) can be dramatically lowered when Ni(II) is cessing independent, suggesting that they recognized nickel com- complexed by peptides or proteins (43). This allows Ni(II)-peptide plexed with naturally presented self peptides. Likewise, one of our complexes to be oxidized to Ni(III)-peptide complexes by rather nickel-specific T cell clones, F4 9D2, recognized fixed APC to mild oxidants, such as molecular oxygen or H2O2 (20), in a Fen- which Ni(II) was added after the fixation and, thus, presumably ϩ 3 ϩ Ϫ ϩ . ton-like reaction: Ni(II) H2O2 Ni(III) OH OH (44). recognized naturally presented self peptides complexed with Both Ni(III) and the hydroxyl radical are chemically more re- Ni(II). Moreover, this particular epitope proved to be relatively active than their precursor molecules, Ni(II) and H2O2, respec- stable because it was not lost when nonfixed Ni(II)-pulsed APC tively and, hence, should be more powerful elicitors of the danger presenting the epitope were subjected to repeated washing. In con- signal (38). As far as production of Ni(IV) is concerned, this can Ϫ trast, the epitope(s) recognized by most other T cell hybridomas be generated from Ni(II) by reaction with OCl (13), a powerful tested was unstable by this criterion. The other part of the CD4ϩ oxidant that is formed from H2O2 by myeloperoxidase; both H2O2 nickel-specific clones described by Weltzien et al. (4) was classi- and myeloperoxidase are released from activated neutrophils and fied as processing dependent because they failed to respond to monocytes into the microenvironment. fixed APC plus Ni(II). Likewise, the majority of our T cell hy- The present paper is the first to report the generation of nickel- bridomas failed to recognize fixed APC plus Ni(II) and, therefore, specific murine T cell hybridomas. The nickel-specific T cell hy- might be interpreted to be processing dependent. Processing de- bridoma clones could be restimulated not only by the higher nickel pendence of Ni-induced epitopes implies that nickel interacts with oxidation state used for priming, but also by Ni(II). In the latter soluble or membrane-bound proteins and leads to altered process- respect, the results obtained by the study of hybridomas conform ing of the self proteins thus complexed and, hence, to presentation with those obtained from the MEST, the PLN assay, and the in of either Ni(II)-complexed or cryptic self peptides (17, 34, 50). vitro testing of human T cell clones specific for nickel (3, 4). The This conclusion should be drawn with caution, however, since loss crucial difference between the higher oxidation states of nickel on of T cell responsiveness in the presence of fixed APC and Ni(II) the one hand side and Ni(II) on the other hand side seems to be does not necessarily prove processing dependence. Fixation of their differential capacity for induction of costimulatory signals APC with glutaraldehyde cross-links their surface proteins and rather than a differential capacity for provision of signal 1. Since the requirements for costimulation, in particular DC mobilization, could thereby alter the conformation of peptide-MHC complexes are lower in secondary than in primary immune responses, the such that the APC, upon addition of nickel, can no longer form the defective capacity of Ni(II) for induction of costimulation needs same metal-peptide-MHC complexes as unfixed APC and, hence, not be disclosed when studying recall responses. Ni(II) on the need no longer be recognized by a given T cell clone. other hand does induce ICAM and E-selectin on endothelial cells Although it is known that the oxidation state of heavy metals (45, 46) so that primed T cells are enabled to get to the site of may be subject to change in vivo and that such changes may pro- challenge with Ni(II) in the mouse ear. foundly alter their biological function, little attention has been paid Unexpectedly, the presence of the proteins MSA and RNase, to this in allergology. An exception here is the oxidative conver- respectively, that was part of the Ni(IV)-protein complex used for sion of chemically less reactive gold(I) into the highly reactive gold(III) intermediate (15, 18). As mentioned before, the biooxi- priming was not required for eliciting recall responses by the nick- Ϫ el-specific T cell hybridomas. In all likelihood, the clones reacted dation of Ni(II) by H2O2 and OCl , respectively, parallels that of to MHC-embedded peptides cleaved from unidentified self pro- gold(I) in that in both cases the higher oxidation states of heavy teins that were altered by Ni(II) and Ni(IV), respectively, in one metal proved to be much stronger inducers of sensitization than the way or the other. Hence, none of the nickel-binding proteins used lower ones (Refs. 14, 16, and 34, and results of the present paper). for priming acted like the classical protein carriers of covalently However, there also seems to be a difference between the two bound organic haptens, such as trinitrophenyl derivatives (47, 48), metals, and this concerns the reason why their lower oxidation that provide the peptide part of the MHC-embedded hapten-pep- states, gold(I) and Ni(II), are inadequate agents for T cell priming. tide conjugate. Two different explanations may account for this In the case of gold, the difference between the lower and the higher lack of requirement of the original protein used for priming. First, oxidation state lies in its differential ability to generate neoantigen, the proteins used for priming to nickel merely served as vehicles or signal 1; neoantigen formation by gold(III) is due to oxidation conserving higher oxidation states of the metal; this was followed and, hence, irreversible denaturation of protein (17, 34), something by ligand exchange and complex formation with unknown self the lower oxidation state, gold(I), is unable to do. In contrast, the proteins or self peptides, and it was these nickel complexes that inability of the lower oxidation state of Ni, Ni(II), for induction of actually primed naive T cells. Alternatively, some of the epitopes sensitization should be sought not in inadequate provision of signal resulting from nickel complexing with the exogenous proteins 1, but in an inability to elicit effective signal 2 for T cell priming. used for priming were identical with those that resulted from the The latter can be efficiently elicited by higher oxidation states of complexing of nickel ions with unidentified self proteins or pep- nickel, which presumably are generated under conditions such as 1152 ENHANCED SENSITIZATION VIA NICKEL AT HIGHER OXIDATION STATES injury, inflammatory skin disease, or concomitant exposure to ir- 23. Nakagawa, K., R. Konaka, and T. Nakata. 1961. Oxidation with nickel peroxide. ritating chemicals. I. Oxidation of alcohols. J. Org. Chem. 27:1597. 24. Bloksma, N., M. Kubicka-Muranyi, H.-C. Schuppe, E. Gleichmann, and H. Gleichmann. 1995. A predictive immunotoxicological test system: suitability Acknowledgments of the popliteal lymph node assay in rats and mice. Crit. Rev. Toxicol. 25:369. 25. Shimonkevitz, R., J. Kappler, P. Marrack, and H. Grey. 1983. Antigen recog- We thank Dr. S. Wolf (Genetics Institute, Cambridge, MA) for generously nition by H-2-restricted T cells. I. Cell-free antigen processing. J. Exp. Med. providing rIL-12, as well as Professor C. Frank Shaw III (University of 158:303. 26. Riemann, H., A. Schwarz, S. Grabbe, Y. Aragane, T. A. Luger, M. Wysocka, Wisconsin-Milwaukee, WI), Professor Peter J. Sadler (University of Ed- M. Kubin, G. Trinchieri, and T. Schwarz. 1996. Neutralization of IL-12 in vivo inburgh, U.K.), and the faculty and participants of the AAI Advanced prevents induction of contact hypersensitivity and induces hapten-specific toler- Course in Immunology (July 1998, Berkeley, CA) for suggestions and ance. J. Immunol. 156:1799. discussions. 27. Mu¨ller, G., J. Saloga, T. Germann, G. Schuler, J. Knop, and A. H. Enk. 1995. IL-12 as mediator and adjuvant for the induction of contact sensitivity in vivo. J. Immunol. 155:4661. References 28. Anderson, C. 1988. Dermal cell infiltrates: allergic, toxic, irritant, and type I reactions. Acta Derm Venereol (Stockh) 68(Suppl. 135):24. 1. Basketter, D. A., G. Briatico Vangosa, W. Kaestner, C. Lally, and W. J. Bontinck. 29. Lever, W. F., and G. Schaumburg-Lever. 1990. Histopathology of the Skin. Lip- 1993. Nickel, cobalt and chromium in consumer products: a role in allergic con- pincott, Philadelphia, p. 106. tact dermatitis? Contact Dermatitis. 28:15. 30. Yongqia, Z., and H. Xuying. 1994. The novel behaviour of interaction between 2. Pistoor, F. H., M. L. Kapsenberg, J. D. Bos, M. M. Meinardi, Ni(II) ion and human or bovine serum albumin. Biochem. J. 304:23. B. M. von Blomberg, and R. J. Scheper. 1995. Cross-reactivity of human nickel- 31. Gill, G., A. A. Richter-Rusli, M. Ghosh, C. J. Burrows, and S. E. Rokita. 1997. reactive T-lymphocyte clones with copper and palladium. J. Invest. Dermatol. Nickel-dependent oxidative cross-linking of a protein. Chem. Res. Toxicol. 105:92. 10:302. 3. Cavani, A., D. Mei, E. Guerra, S. Corinti, M. Giani, L. Pirrotta, P. Puddu, and 32. Romagnoli, P., A. M. Labhardt, and F. Sinigaglia. 1991. Selective interaction on G. Girolomoni. 1998. Patients with allergic contact dermatitis to nickel and non- Ni with an MHC-bound peptide. EMBO J. 10:1303. allergic individuals display different nickel-specific T cell responses: evidence for 33. Sinigaglia, F. 1994. The molecular basis of metal recognition by T cells. J. Invest. the presence of effector CD8ϩ and regulatory CD4ϩ T cells. J. Invest. Dermatol. Dermatol. 102:398. 111:621. 34. Griem, P., C. von Vulteé, K. Panthel, S. L. Best, P. Sadler, and C. F. Shaw. 1998. 4. Moulon, C., J. Vollmer, and H. U. Weltzien. 1995. Characterization of processing T cell cross-reactivity to heavy metals: identical cryptic peptides may be pre- requirements and metal cross-reactivities in T cell clones from patients with sented from protein exposed to different metals. Eur. J. Immunol. 28:1941. allergic contact dermatitis to nickel. Eur. J. Immunol. 25:3308. 35. Weltfriend, S., M. Bason, K. Lammintausta, and H. I. Maibach. 1996. Irritant 5. Van Hoogstraten, I. M., C. Boos, D. Boden, M. E. von Blomberg, R. J. Scheper, dermatitis. In Dermatoxicology, 5th Ed. F. N. Marzulli and H. I. Maibach, eds. and G. Kraal. 1993. Oral induction of tolerance to nickel sensitization in mice. Taylor & Francis, Washington, p. 87. J. Invest. Dermatol. 101:26. 36. Ibrahim, M. A. A., B. M. Chain, and D. R. Katz. 1995. The injured cell: the role 6. Ishii, N., N. Moriguchi, H. Nakajima, S. Tanaka, and F. Amemiya. 1993. Nickel of the dendritic cell system as a sentinel receptor pathway. Immunol. Today sulfate-specific suppressor T cells induced by nickel sulfate in drinking water. 16:181. J. Dermatol. Sci. 6:159. 37. Cumberbatch, M., R. C. Scott, D. A. Basketter, E. W. Scholes, J. Hilton, 7. Mo¨ller, H. 1984. Attempts to induce contact allergy to nickel in the mouse. R. J. Dearman, and I. Kimber. 1993. Influence of sodium lauryl sulphate on Contact Dermatitis 10:65. 2,4-dinitrochlorobenzene-induced lymph node activation. Toxicology 77:181. 8. Cornacoff, J. B., R. V. House, and J. H. Dean. 1984. Comparison of a radioiso- 38. Matzinger, P. 1994. Tolerance, danger, and the extended family. Annu. Rev. topic incorporation method and the mouse ear-swelling test (MEST) for contact Immunol. 12:991. sensitivity to weak sensitizers. Fundam. Appl. Toxicol. 10:40. 39. Macatonia, S. E., N. A. Hosken, M. Litton, P. Vieira, C. S. Hsieh, 9. Vreeburg, K. J., K. De Groot, I. M. Van Hoogstraten, B. M. von Blomberg, and J. A. Culpepper, M. Wysocka, G. Trinchieri, K. M. Murphy, and A. Ogarra. 1995. R. J. Scheper. 1991. Successful induction of allergic contact dermatitis to mer- Dendritic cells produce IL-12 and direct the development of Th1 cells from naive cury and chromium in mice. Int. Arch. Allergy Appl. Immunol. 96:179. CD4ϩ T cells. J. Immunol. 154:5071. 10. Mandervelt, C., F. L. Clottens, M. Demedts, and B. Nemery. 1997. Assessment 40. Schwarz, A., S. Grabbe, Y. Aragane, K. Sandkuhl, H. Riemann, T. Luger, of the sensitization potential of five metal salts in the murine local lymph node M. Kubin, G. Trinchieri, and T. Schwarz. 1996. Interleukin-12 prevents ultravi- assay. Toxicology 120:65. olet B-induced local immunosuppression and overcomes UVB-induced tolerance. 11. Kimber, I. 1996. The local lymph node assay. In Dermatoxicology, 5th Ed. J. Invest. Dermatol. 1187. F. N. Marzulli and H. I. Maibach, eds. Taylor & Francis, Washington, p. 469. 41. Schmitt, D. A., L. Owenschaub, and S. E. Ullrich. 1995. Effect of IL-12 on 12. Oller, A. R., M. Costa, and G. Oberdo¨rster. 1997. Carcinogenicity assessment of immune suppression and suppressor cell induction by ultraviolet radiation. J. Im- selected nickel compounds. Toxicol. Appl. Pharmacol. 143:152. munol. 154:5114. 13. Christoskova, S. G., N. Danova, M. Georgieva, O. K. Argirov, and 42. Subak, E. J., V. M. Loyola, and D. W. Margerum. 1985. Substitution and rear- D. Mehandzhiev. 1995. Investigation of a nickel oxide system for heterogeneous rangement reactions of Ni(III) peptide complexes in acid. Inorg. Chem. 24:4350. oxidation of organic compounds. Appl. Catalysis A: General 128:219. 43. Salnikow, K., M. Gao, V. Voitkun, X. Huang, and M. Costa. 1994. Altered 14. Verwilghen, J., G. H. Kingsley, L. Gambling, and G. S. Panayi. 1992. Activation oxidative stress responses in nickel-resistant mammalian cells. Cancer Res. 54: of gold-reactive T lymphocytes in rheumatoid arthritis patients treated with gold. 6407. Arthritis Rheum. 35:1413. 44. Stohs, S. J., and D. Bagchi. 1995. Oxidative mechanisms in the toxicity of heavy 15. Goebel, C., M. Kubicka-Muranyi, T. Tonn, J. Gonzalez, and E. Gleichmann. metal ions. Free Radical Biol. Med. 18:321. 1995. Phagocytes render chemicals immunogenic: oxidation of gold(I) to the T 45. Goebeler, M., G. Meinardus-Hager, J. Roth, S. Goerdt, and C. Sorg. 1993. Nickel cell sensitizing gold(III) metabolite generated by mononuclear phagocytes. Arch. chloride and cobalt chloride, two common contact sensitizers, directly induce Toxicol. 69:450. expression of intercellular adhesion molecule-1 (ICAM-1), vascular cell adhesion 16. Schuhmann, D., M. Kubicka-Muranyi, J. Mirtscheva, J. Gu¨nther, P. Kind, and molecule-1 (VCAM-1), and endothelial leukocyte adhesion molecule (ELAM-1) E. Gleichmann. 1990. Adverse immune reactions to gold. I. Chronic treatment by endothelial cells. J. Invest. Dermatol. 100:759. with an Au(I) drug sensitizes mouse T cells not to Au(I), but to Au(III), and 46. Wagner, M., C. L. Klein, H. Kleinert, C. Euchenhofer, U. Fo¨rstermann, and induces autoantibody formation. J. Immunol. 145:2132. C. J. Kirkpatrick. 1997. Heavy metal ion induction of adhesion molecules and 17. Griem, P., K. Panthel, H. Kalbacher, and E. Gleichmann. 1996. Alteration of a cytokines in human endothelial cells: the role of NF-␬B, I␬B-␣ and AP-1. Patho- model antigen by Au(III) leads to T cell sensitization to cryptic peptides. Eur. biology 65:241. J. Immunol. 26:279. 47. Cavani, A., C. J. Hackett, K. J. Wilson, J. B. Rothbard, and S. I. Katz. 1995. 18. Shaw, C. F., S. Schraa, E. Gleichmann, Y. P. Grover, L. Dunemann, and Characterization of epitopes recognized by hapten-specific CD4ϩ T cells. J. Im- Ϫ A. Jagarlamudi. 1994. Redox chemistry and [Au(CN)2] in the formation of gold munol. 154:1232. metabolites. Metal-Based Drugs 1:351. 48. Kohler, J., S. Martin, U. Pflugfelder, H. Ruh, J. Vollmer, and H. U. Weltzien. 19. White, J., M. Blackman, J. Bill, J. Kappler, P. Marrack, D. P. Gold, and W. Born. 1995. Cross-reactive trinitrophenylated peptides as antigens for class II major 1989. Two better cell lines for making hybridomas expressing specific T cell histocompatibility complex-restricted T cells and inducers of contact sensitivity receptors. J. Immunol. 143:1822. in mice: limited T cell receptor repertoire. Eur. J. Immunol. 25:92. 20. Cheng, C. C., S. E. Rokita, and C. J. Burrows. 1993. Nickel(III)-promoted DNA 49. Vollmer, J., M. Fritz, H. U. Weltzien, and C. Moulon. 1997. Dominance of the cleavage with ambient oxygen. Angew. Chem. 32:277. BV17 element in nickel-specific human T cell receptors relates to severity of 21. Root, R. K., J. Metcalf, N. Oshino, and B. Chance. 1975. H2O2 release from contact sensitivity. Eur. J. Immunol. 27:1865. human granulocytes during phagocytosis: documentation, quantitation, and some 50. Kubicka-Muranyi, M., J. Kremer, N. Rottmann, B. Lu¨bben, R. Albers, regulating factors. J. Clin. Invest. 55: 945. N. Bloksma, R. Lu¨hrmann, and E. Gleichmann. 1996. Murine systemic autoim- 22. Clifford, D. P., and J. E. Repine. 1984. Measurement of oxidizing radicals by mune disease induced by mercuric chloride: T helper cells reacting to self pro- polymorphonuclear leukocytes. Methods Enzymol. 105:393. teins. Int. Arch. Allergy Immunol. 109:11.