TRAV7-2*02 Expressing CD8+ T Cells Are Responsible for Palladium Allergy

While metallic biomaterials have led to an improvement in the quality of life, metal allergies, especially to palladium (Pd), has caused a recent increase in allergic patients. Metal allergy is known to be a T cell-mediated delayed-type hypersensitivity (DTH); however, the pathogenic T cell subsets and the specific T cell receptor (TCR) have not been identified. Therefore, we attempted to identify the pathogenic T cells responsible for Pd allergy. We found that activating CD8+ T cells significantly increased and that the TRAV (TCRα variable) 7-2*02 chain skewed in Pd allergic mice. Furthermore, adoptive transfer experiments revealed that in vitro-cultured Pd-stimulated antigen presenting cells (APCs) function as memory APCs with recipient mice developing Pd allergy and that the frequency of TRAV7-2*02 increases the same as conventional Pd allergic mice. In contrast, neither proliferation of CD8+ T cells nor increasing of TRAV7-2*02 was observed in major histocompatibility complex I (MHC I)-deficient Pd-APCs transferred to mice. Taken together, we revealed that TRAV7-2*02-expressing CD8+ T cells are the pathogenic T cells for the development of Pd allergy. We also identified the CDR3 consensus motif of pathogenic TCRs as CAAXSGSWQLIF in TRAV7-2*02/TRAJ (TCRα junction)22*01 positive cells. These results suggest that the specific TCRs represent novel targets for the development of diagnostics and treatments for metal allergy.


Introduction
Metal is superior to other materials in hardness, strength, durability, and workability and, therefore, is broadly useful in many fields. Palladium (Pd), in particular, has a wide array of uses, including as an extremely important metallic biomaterial for dentistry in the reconstruction of occlusions during dental restorations. However, metallic biomaterials sometimes cause metal allergy [1], and the incidence of Pd allergy, in particular, has increased recently [2]. However, the pathological mechanism of metal allergy has not been clarified. Clinical symptoms of metal allergy include intractable diseases, such as dermatitis, inflammation of oral mucosa, oral lichen planus, and stomatitis.
Metal allergy is categorized as a delayed-type hypersensitivity (DTH) reaction, which is a T cell-mediated disease with onset at 24 hours after exposure to the causal metal [3,4]. The pathogenic mechanism of metal allergy includes the function of metal ions as haptens that exert antigenicity and induce pathogenic T cells, which cause metal allergy. T cells are classified into two broad subsets, CD4 + T cells and CD8 + T cells. CD4 + T cells binding major histocompatibility complex II (MHC II) on presenting cells and play roles as helper and regulatory T cells by producing cytokines. CD8 + T cells recognize MHC I and function as proinflammatory T cells that directly kill infected or transformed cells. It has been reported that both CD8 + and CD4 + T cells are required for DTH [3]. Thus, both types of cells [5,6] were generally thought to be responsible for Pd allergy; however, controversy remains regarding the role of pathogenic T cell subsets in DTH.
Various cytokines control the activation or inhibition of immune responses that comprise the host biological defense system against foreign antigens. IFN-γ, the most important cytokine in metal allergy, is an inflammatory cytokine that enhances immune responses, including Th1 responses and cytotoxic activity [6][7][8]. Previously, we reported that there was a decrease in ear swelling in IFN-γ-deficient mice (IFN-γ −/− mice) compared to wild-type (WT) mice after Pd challenge [6]; however, in a previous study the distribution of T cell subsets was not examined in IFN-γ −/− mice after Pd challenge.
In order to establish acquired immunity, the TCR recognizes antigenic peptides present in the context of MHC, which leads to T cell activation [9,10]. To obtain vast repertoire diversity, the TCR α and β genes undergo complex rearrangement of variable (V), diverse (D), and junction (J) region gene segments resulting in the generation of a specific TCR from 1 × 10 18 possible clonotypes. Examination of peripheral blood mononuclear cells obtained from patients with metal allergy showed that metal ions can induce proliferation of human T cells in vitro and limited TCRs are utilized in human T cells from metal allergy patients [11]. However, the specific TCR present on pathogenic T cells of patients with Pd allergy has not previously been identified. Therefore, we attempted to identify the T cell subsets responsible for Pd allergy and investigated whether a specific TCR exists on these pathogenic T cells.

Activated T Cells Accumulate in Regional Lymph Nodes during Pd Allergy
To examine the accumulation of activated T cells in regional lymph nodes, we analyzed the T cell populations present in submandibular lymph node (SLN) cells of Pd-allergic mice by flow cytometry. To this end, Pd + lipopolysaccharide (LPS)-sensitized WT mice (sensitized-WT mice) and phosphate buffered saline (PBS)-injected WT mice (unsensitized-WT mice) were challenged by injection of Pd solution into ear auricles and ear swelling was assessed ( Figure S1A). Following Pd challenge, ear swelling of sensitized-WT mice was significantly increased compared to that of unsensitized-WT mice (** p < 0.01). Moreover, WT mice sensitized with Pd alone or LPS alone and then challenged with Pd showed no evidence of ear swelling [6]. Both CD8 + and CD4 + T cells accumulated in ear auricles of sensitized WT mice ( Figure S1B). Since the ears were most swollen at 24 h post Pd challenge, we assessed the T cell population in the SLNs at this time and found no difference in CD3 + T cell numbers between unsensitized-WT mice and naïve mice. In contrast, there was a significant increase in CD3 + T cells in the SLNs of sensitized-WT mice (** p < 0.01) ( Figure 1A). Although CD4 + T cells appeared to be slightly increased in the SLN of sensitized WT mice, this difference was not statistically significant ( Figure 1A). In contrast, CD8 + T cells were significantly increased in the SLN of sensitized WT mice compared to unsensitized WT (* p < 0.05) or naïve mice (** p < 0.01) ( Figure 1A). We next measured IFN-γ production by SLN T cells. Previously, we showed that IFN-γ producing CD8 + T cells are responsible for Pd allergy [6]. However, in the previous study we analyzed IFN-γ production in the limited condition of T cells enriched by sequential adoptive transfer in Pd allergic mice. Therefore, we investigated which T cells are activated in a conventional model of Pd allergy. Interestingly, not only CD8 + T cells, but also CD4 + T cells produced IFN-γ in sensitized-WT mice (** p < 0.01) ( Figure 1B,C).
Since ear auricles were not swollen in response to Pd challenge of IFN-γ −/− mice ( Figure 1D), we assessed CD8 + and CD4 + T cells in the SLN cells of sensitized-IFN-γ −/− mice 24 h after Pd challenge and found no increase in CD8 + T cells (* p < 0.05) compared to that seen in sensitized-WT mice ( Figure 1E). These results suggest that activation of CD8 + T cells, rather than CD4 + T cells, enhances the development of Pd allergy.
sensitized-WT mice (** p < 0.01) ( Figure 1B,C). Since ear auricles were not swollen in response to Pd challenge of IFN-γ −/− mice ( Figure 1D), we assessed CD8 + and CD4 + T cells in the SLN cells of sensitized-IFN-γ −/− mice 24 h after Pd challenge and found no increase in CD8 + T cells (* p < 0.05) compared to that seen in sensitized-WT mice ( Figure 1E). These results suggest that activation of CD8 + T cells, rather than CD4 + T cells, enhances the development of Pd allergy.

MHC I and Activated CD8 + T Cells Are Critical for Pd Allergy
To investigate whether MHC I or II plays a more important role in Pd allergy, we induced Pd allergy in C57BL/6-deficient β2m-microgloblin (B2m) mice (B2m −/− mice) and C57BL/6-deficient MHC II H2-Ab1 (I-A b ) mice (I-A b−/− mice). Since B2m constitutes a part of the MHC I molecule, MHC I is not expressed on the cell surface in these mice and, thus, CD8 + T cells are impaired, while in I-A b−/− mice CD4 + T cells are impaired. Compared to unsensitized-WT mice, ear auricles were significantly more swollen in sensitized I-A b−/− mice at 24 h after Pd challenge (** p < 0.01). However, ear swelling was not observed in B2m −/− mice following Pd challenge ( Figure 2E). In addition, CD4 + T cells were not significantly increased in SLNs of B2m −/− mice ( Figure 2F), while the number of CD8 + T cells in SLNs of I-A b−/− mice was greatly increased (** p < 0.01) ( Figure 2F). Furthermore, TRAV7-2*02 was highly expressed on SLN cells from Pd-sensitized I-A b−/− and WT mice (data not shown). Therefore, MHC I is essential and activated CD8 + T cells are important for the development of Pd allergy.

Adoptive Transfer of In Vitro Prepared Pd-Treated APCs Induces Pd Allergy
We examined whether Pd allergy could develop in mice with adoptively-transferred Pdstimulated APCs. To this end, APCs were prepared from bone marrow cells and cultured in mouse macrophage colony stimulating factor (mM-CSF) containing medium. After seven days in culture,

MHC I and Activated CD8 + T Cells Are Critical for Pd Allergy
To investigate whether MHC I or II plays a more important role in Pd allergy, we induced Pd allergy in C57BL/6-deficient β2m-microgloblin (B2m) mice (B2m −/− mice) and C57BL/6-deficient MHC II H2-Ab1 (I-A b ) mice (I-A b−/− mice). Since B2m constitutes a part of the MHC I molecule, MHC I is not expressed on the cell surface in these mice and, thus, CD8 + T cells are impaired, while in I-A b−/− mice CD4 + T cells are impaired. Compared to unsensitized-WT mice, ear auricles were significantly more swollen in sensitized I-A b−/− mice at 24 h after Pd challenge (** p < 0.01). However, ear swelling was not observed in B2m −/− mice following Pd challenge ( Figure 2E). In addition, CD4 + T cells were not significantly increased in SLNs of B2m −/− mice ( Figure 2F), while the number of CD8 + T cells in SLNs of I-A b−/− mice was greatly increased (** p < 0.01) ( Figure 2F). Furthermore, TRAV7-2*02 was highly expressed on SLN cells from Pd-sensitized I-A b−/− and WT mice (data not shown). Therefore, MHC I is essential and activated CD8 + T cells are important for the development of Pd allergy.

Adoptive Transfer of In Vitro Prepared Pd-Treated APCs Induces Pd Allergy
We examined whether Pd allergy could develop in mice with adoptively-transferred Pd-stimulated APCs. To this end, APCs were prepared from bone marrow cells and cultured in mouse macrophage colony stimulating factor (mM-CSF) containing medium. After seven days in culture, CD11b + F4/80 + APCs were sorted and stimulated with Pd + LPS (Pd-APCs) or LPS (LPS-APCs) for 24 h. Flow cytometric analysis revealed that these APCs also express the co-stimulatory molecules CD80, CD86, and CD40 ( Figure S3A). To assess their effects on Pd sensitization in mice, these APCs were adoptively transferred into naïve WT mice. Seven days after the transfer, recipient mice were challenged with Pd injection into ear auricles. Ear swelling in the mice given Pd-APCs was significantly increased (** p < 0.01) compared to that of mice given LPS-APCs ( Figure 3A). Further, CD8 + and CD4 + T cells were significantly increased in SLNs of mice that received Pd-APCs ( Figure 3B). These results indicate that APCs cultured with Pd + LPS in vitro function as the driving force in the development of Pd allergy. In addition, the TCR repertoire in mice given Pd-APCs is skewed in a manner similar to that seen in Pd-sensitized WT mice, with significant increases in TRAV7-2*02 expression, but not TRBV expression, starting 24 h after Pd challenge ( Figure 3C,D). Therefore, Pd-memory APCs lead to Pd allergy, and the TCRα chain, rather than TCRβ chain, is involved in the development of Pd allergy. CD11b + F4/80 + APCs were sorted and stimulated with Pd + LPS (Pd-APCs) or LPS (LPS-APCs) for 24 h. Flow cytometric analysis revealed that these APCs also express the co-stimulatory molecules CD80, CD86, and CD40 ( Figure S3A). To assess their effects on Pd sensitization in mice, these APCs were adoptively transferred into naïve WT mice. Seven days after the transfer, recipient mice were challenged with Pd injection into ear auricles. Ear swelling in the mice given Pd-APCs was significantly increased (** p < 0.01) compared to that of mice given LPS-APCs ( Figure 3A). Further, CD8 + and CD4 + T cells were significantly increased in SLNs of mice that received Pd-APCs ( Figure  3B). These results indicate that APCs cultured with Pd + LPS in vitro function as the driving force in the development of Pd allergy. In addition, the TCR repertoire in mice given Pd-APCs is skewed in a manner similar to that seen in Pd-sensitized WT mice, with significant increases in TRAV7-2*02 expression, but not TRBV expression, starting 24 h after Pd challenge ( Figure 3C,D). Therefore, Pdmemory APCs lead to Pd allergy, and the TCRα chain, rather than TCRβ chain, is involved in the development of Pd allergy.

MHC I-Dependent CD8 + T Cell Activation Is Essential for the Onset of Pd Allergy
Neither MHC I molecules or CD8 + T cells function in B2m −/− mice as development of CD8 + T cells is impaired in the absence of this molecule. Thus, we investigated whether MHC I-deficient Pd-APCs derived from B2m −/− mice are able to induce Pd allergy. To this end, we prepared WT Pd-APCs and B2m −/− Pd-APCs and adoptively transferred them into naïve WT mice. Seven days later, these mice were challenged with Pd injection and ear thickness was measured from 24 h after the challenge. Ear swelling of B2m −/− Pd-APC recipient mice was significantly less (** p < 0.01) than that of mice given WT Pd-APCs ( Figure 4A). Neither CD4 + nor CD8 + T cells increased significantly in mice given B2m −/− Pd-APCs compared to those given WT Pd-APCs (** p < 0.01) ( Figure 4B). Moreover, the frequency of TRAV7-2*02 was not increased in mice given B2m −/− Pd-APCs compared to mice given WT Pd-APCs (** p < 0.01) ( Figure 4C). These results indicate that TRAV7-2*02 expressing CD8 + T cells are activated and proliferate in a manner dependent on MHC I, which leads to Pd allergy.

MHC I-Dependent CD8 + T Cell Activation Is Essential for the Onset of Pd Allergy
Neither MHC I molecules or CD8 + T cells function in B2m −/− mice as development of CD8 + T cells is impaired in the absence of this molecule. Thus, we investigated whether MHC I-deficient Pd-APCs derived from B2m −/− mice are able to induce Pd allergy. To this end, we prepared WT Pd-APCs and B2m −/− Pd-APCs and adoptively transferred them into naïve WT mice. Seven days later, these mice were challenged with Pd injection and ear thickness was measured from 24 h after the challenge. Ear swelling of B2m −/− Pd-APC recipient mice was significantly less (** p < 0.01) than that of mice given WT Pd-APCs ( Figure 4A). Neither CD4 + nor CD8 + T cells increased significantly in mice given B2m −/− Pd-APCs compared to those given WT Pd-APCs (** p < 0.01) ( Figure 4B). Moreover, the frequency of TRAV7-2*02 was not increased in mice given B2m −/− Pd-APCs compared to mice given WT Pd-APCs (** p < 0.01) ( Figure 4C). These results indicate that TRAV7-2*02 expressing CD8 + T cells are activated and proliferate in a manner dependent on MHC I, which leads to Pd allergy.

Discussion
In this study we demonstrate the skewing of specific TCRs during the development of Pd allergy and the contribution of CD8 + T cells, rather than CD4 + T cells, to the development of this allergy in mice. Moreover, we identify that TRAV7-2*02 expressing CD8 + T cells mainly function as pathogenic T cells. Furthermore, using adoptive transfer experiments we show that APCs stimulated in vitro with Pd function as metal memory APCs in vivo and induce sensitization for metal allergy when transferred. Indeed, skewing of TRAV7-2*02 was observed in the Pd allergic WT mice that received in vitro derived Pd-APCs. These results indicate that Pd memory APCs were generated in vitro and induce the pathogenic T cell activation seen during the Pd allergic response.
Although there has previously been controversy regarding this, we found that activated CD8 + T cells, rather than CD4 + T cells, play a key role as proinflammatory T cells in Pd allergy (Figure 1).
Twenty-four hours after Pd challenge, both CD8 + and CD4 + T cells became activated and produced IFN-γ ( Figure 1B,C). Thus, we assessed CD8 + and CD4 + T cells in the SLNs of sensitized-IFN-γ −/− mice 24 hours after Pd challenge and found no increase in CD8 + T cells (* p < 0.05) over that seen in sensitized-WT mice ( Figure 1E). Therefore, CD4 + T cells are likely involved only in exacerbating Pd allergy, because IFN-γ production by CD4 + T cells increased in sensitized-WT mice. However, while the ears of sensitized-I-A b−/− mice and sensitized-WT mice were swollen, those of sensitized-B2m −/− mice were not ( Figure 2E). In addition, CD8 + T cells were significantly increased in sensitized-I-A b−/− mice without the cooperation of CD4 + T cell activation ( Figure 2F). These results indicate that Pd allergy is caused by antigen presentation by MHC I to CD8 + T cells; however, Pd allergy does not occur following the MHC II-CD4 + T cell interaction. Thus, both MHC I expression and activation of CD8 + T cells are essential for the development of Pd allergy. Consistent with our results, CD8 + T cells have been shown to function as effector cells in DTH responses [14].
It has been reported that PdCl 2 is sufficiently reliable reagent for development of Pd allergy and that solution of PdCl 2 injection induces T cell dependent allergic diseases and in mice [6,15,16]. Therefore, we used PdCl 2 for these experiments and referred the appropriate dose of PdCl 2 [6,15,16]. Interestingly, since Na 2 [PdCl 4 ] has been shown to be effective in human patch testing [17], Na 2 [PdCl 4 ] may be more useful reagent than PdCl 2 in Pd allergic mouse model. Future experiments will be required for applying animal experiments.
LPS is a component of microbes that, when recognized by immune cells, induces MHC expression on APCs ( Figure S3B). Thus, it has been suggested that this microbial component plays a role in the immunological response leading to metal allergy [18]. Moreover, it is likely that the resident flora contributes to the progression of metal allergy.
The predominant TCRα chain expressed during Pd allergy is TRAV7-2*02/TRAJ22*01. While TRAV7-2*02 was expressed on both subsets of T cells ( Figure 2C,D), increased frequencies of TRAV7-2*02 expressing CD8 + T cells were observed in sensitized-I-A b−/− allergic mice (data not shown). Therefore, adoptive transfer experiments were performed to assess whether TRAV7-2*02 expressing CD8 + T cells are important for Pd allergy. Interestingly, we found ear swelling and increases in CD8 + T cells in mice given Pd-APCs at a similar level to conventional Pd allergic mice ( Figure 3A,B). Furthermore, the TRAV7-2*02 repertoire was found to be significantly represented in mice given Pd-APCs after Pd challenge ( Figure 3C). In contrast, the frequency of TRAV7-2*02 was not increased in mice given B2m −/− Pd-APCs compared to mice given WT Pd-APCs ( Figure 4C). Consistent with the lack of a response by TRAV7-2*02 expressing CD8 + T cells, Pd allergy did not develop in mice given B2m −/− Pd-APCs ( Figure 4). Therefore, these results indicate that CD8 + T cells play a role as pathogenic T cells for the development of Pd allergy, and that TRAV7-2*02 expressing CD8 + T cells are activated by Pd-stimulated APCs expressing MHC I. Accordingly, our findings indicate the presence of a specific TCR in metal allergy. Furthermore, the adoptive transfer experiments indicate that APCs stimulated in vitro with Pd function to sensitize recipient naïve mice. Therefore, it is possible that adoptive transfer experiments may be useful for further study of antigen presentation in metal allergy.
The skewed TCRs identified in Pd allergy were TRAV7-2*02/TRAJ22*01. As for the TCRβ chain, the TRBV13 family has been reported to be frequently expressed in various mouse disease models including in NOD mice [19] and in an induced arthritis model [20]. These results imply that TCRβ chains exhibit flexible binding capacities against pathogenic antigens. However, in Pd allergy there was actually more frequent skewing of the TCRα chain than of the TCRβ chain, and only Pd-specific TCRα chains were detected. Consistent with our results, TRAV has a high specificity for recognizing nickel bound to peptide in humans [21,22]. Therefore, the Pd binding peptide may be restricted to the TCRα chain. Moreover, we found that TRAJ22*01 was the most commonly-expressed J segment (Table 1), and we identified the CDR3 consensus frame, CAAXSGSWQLIF, in TRAV7-2*02/TRAJ22*01 ( Table 2). Taken together, it can be concluded that TRAV7-2*02/TRAJ22*01 expressing CD8 + T cells function as pathogenic T cells in Pd allergy. Based on our results, transgenic mice expressing TRAV7-2*02/TRAJ22*01 on CD8 + T cells should be highly-sensitive Pd allergy animal models and would be a valuable tool to explore the molecular mechanisms underlying the pathogenesis of metal allergy. Furthermore, the response of TRAV7-2*02 expressing CD8 + T cells could be assessed as an indicator of metal safety and, thus, our findings may be helpful for development of innovative biomaterials.

Ethics Statement
All mice were maintained under specific pathogen-free conditions, and used according to the guidelines of the Institutional Animal Care and Use Committee established at Tohoku University. The project identification code is 2016AcA-016 (date of approval: 25 March 2016).

Induction of Pd Allergy
Induction of Pd allergy was previously described [6]. Briefly, for sensitization mice were injected twice intraperitoneally with 250 µL of 10 mM PdCl 2 containing 10 µg/mL LPS in PBS at an interval of seven days. LPS was used as an adjuvant [15]. As a control, mice were injected with PBS only (unsensitized). Seven days after the second sensitization, mice were challenged by intradermal injection of 20 µL of 0.5 mM PdCl 2 into each ear. Ear swelling was measured before and after the challenge using a Peacock dial thickness gauge (Ozaki MFG Co., Ltd., Tokyo, Japan) [6,15,23,24].

Histological Analysis
To identify the T cells present in ear auricles of WT mice with Pd allergy, frozen ear auricles were sliced and immunostained with anti-mouse CD4 mAb (H129. 19), and anti-mouse CD8a mAb (53-6.7). CD4 + and CD8 + T cells were visualized by staining with 3,3 -diaminobenzidine (DAB) chromogen, and these sections were counter-stained with hematoxylin. The DAB signals were detected with an Olympus IX81 microscope, an Olympus DP71 CCD camera (Olympus, Tokyo, Japan), and Lumina Vision software (Mitani Corporation, Fukui, Japan). The scale bar indicates 100 µm.

Adoptive Transfer of Bone Marrow-Derived Antigen-Presenting Cells
To differentiate APCs, mouse bone marrow cells were cultured in RPMI1640 complete medium with 10% FBS and 10% CMG14-12 culture supernatant containing mM-CSF [25,26]. On day 7, the collected APCs were treated with 0.2 mM PdCl 2 and 10 ng/mL LPS (Pd-APCs) or 10 ng/mL LPS (LPS-APCs, as a control), and were cultured for 24 h. On day 8, these APCs were washed twice in PBS to completely remove the supernatant, and the harvested Pd-APCs or LPS-APCs (1 × 10 6 cells per mouse) were adoptively transferred via intravenous injection into naïve C57BL/6 mice for sensitization ( Figure S4). Seven days after the adoptive transfer, recipient mice were challenged by intradermal injection using the same procedure described above.

T Cell Repertoire Sequence Analysis Using a Next-Generation Sequencer
To investigate the skewing of TCRs in Pd allergy, we analyzed the TCR repertoire using a next generation sequencer. Briefly, total RNA was prepared from the SLN 24 hours after Pd challenge [16]. Complementary DNA was synthesized from total RNA and TCR chains were amplified using adaptor ligation-mediated PCR. The specific PCR primers were 5 -AGG TGA AGC TTG TCT GGT TGC TC-3 (TCRα) and 5 -TGC AAT CTC TGC TTT TGA TGG CTC-3 (TCRβ). Then, using the PCR products as templates, TCR sequences were analyzed using the 454 GS junior+ system (Roche Applied Science, Indianapolis, IN, USA) according to the manufacturer's protocol. Alignments among approximately 100,000 sequences/run were performed with IMGT/V-QUEST (http://www.imgt.org).

Statistics
Student's t-test was used for analysis of differences and values of p < 0.05 or 0.01 were considered statistically significant [6,27].

Conclusions
In conclusion, we show that MHC I-dependent CD8 + T cells activate and proliferate as pathogenic T cells and we identified TRAV7-2*02/TRAJ22*01 as the specific TCR in Pd allergic mice. Furthermore, Pd-APCs prepared in vitro can function as memory APCs and activate pathogenic T cells in recipient mice. In addition, the CDR3 consensus frame of pathogenic TCRs is CAAXSGSWQLIF in TRAV7-2*02/TRAJ22*01. Thus, these specific TCRs are promising, novel targets for generating improved diagnostics and treatments for Pd allergy.