T-Helper 22 Cell Type Responses to Nickel in Contact Allergic Subjects Are Associated with T-Helper 1, T-Helper 2, and T-Helper 17 Cell Cytokine Profile Responses and Patch Test Reactivity

Nickel (Ni) allergy, a T-cell-mediated delayed-type hypersensitivity reaction, is the most common type of allergic contact dermatitis (ACD). The T cells react with the hapten Ni (Ni ions) that have penetrated the skin and bound to endogenous proteins. Based on high IFN-γ and interleukin (IL)-2 production, in parallel with a low T-helper cell (Th) 2 cytokine production, displayed by Ni-specific CD4+ T-cell clones from allergic subjects, the response was initially considered to be Th1 reactivity [1]. Later studies, however, demonstrated the production of Th2 cytokines also, IL-4, -5, and -13, when peripheral blood mononuclear cells (PBMCs) from allergic subjects were in vitro stimulated with Ni [2, 3]. IL-31, another cytokine produced by Th2 cells, has been shown to be involved in promoting dermatitis [4], and skin biopsies from patients with ACD contain correlated IL-4, IL-13, and IL-31 mRNA [5]. Still, direct evidence that in vitro-stimulated Ni-specific PBMC produce IL-31 has not previously been shown.

In addition to the classification of CD4+ T cells into Th1 and Th2, based on a mutually exclusive production of IFN-γ and IL-4, respectively [6], other Th phenotypes have been described and shown to be of relevance for beneficial as well as detrimental immune reactions in the skin. Th17 cells are classified by the production of IL-17A [7] but can also produce IL-22 and other cytokines [8]. IL-22-producing Th that do not produce the hallmark cytokines of Th1, Th2, and Th17 are in turn defined as Th22 [9]. Both Th17 and Th22 cell types, as well as Th1, are increased in blood of psoriatic individuals [10] and have also been implicated in other skin disorders [11].

Several studies have implicated the involvement of Ni-specific Th17 in ACD. IL-17A is secreted by T cells from Ni-allergic subjects stimulated with Ni-pulsed dendritic cells [12]. In addition, IL-17A was produced at higher levels by Ni-stimulated PBMC from Ni-allergic subjects suffering from complications with implants compared to other Ni-allergic subjects [13]. However, in some studies where PBMCs from allergic and healthy subjects were stimulated with Ni, no difference in the IL-17A levels was found [14, 15]. It has also been shown that Ni-challenged skin from allergic subjects contains infiltrating CD4+ T cells expressing IL-22 in addition to IFN-γ- and IL-17A-expressing T cells. However, cytokine co-expression patterns were not established nor was the antigen specificity of the IL-22-secreting T cells [12, 16].

The hallmark cytokines of Th1, Th2, Th17, and Th22 are produced by distinct cell populations but also by intermediate cell populations displaying mixed cytokine profiles. All these Th phenotypes have been shown in different ways to be involved in detrimental or protective inflammatory reactions. In particular, Th17 and Th22 have been implicated in inflammatory reactions in the skin, where Th17 cells protect against Candida albicans and staphylococcal infections but are also involved in pathogenic reactions in autoimmune diseases and various skin disorders [17‒20]. Similarly, Th22 are believed to be recruited to the skin for tissue repair and for protection against pathogens but are also implicated in psoriasis and atopic and allergic dermatitis [10, 11, 16, 21]. The importance of the distribution of distinct and intermediate Th subtypes for pro-inflammatory versus down-regulatory inflammatory reactions in ACD remains to be elucidated.

In the present study, PBMCs from fifteen Ni-allergic women as well as five non-allergic controls were analyzed for Ni-induced cytokine responses using a panel of enzyme-linked immunosorbent assays (ELISAs) measuring cytokines representing cell populations of Th1, Th2, Th17, and Th22 type. In addition, cytokines like IL-2 and IL-3 representing more general T-cell activation markers were analyzed. IL-6, which can be produced by T cells during chronic inflammation, was analyzed as well [22]. FluoroSpot, where the footprint of single cells secreting one or several cytokines is detected, was used to assess if single cells responding to Ni produced single Th signature cytokines or if they displayed intermediate Th population characteristics with secretion of hallmark cytokines from different Th populations. FluoroSpot is based on the same assay principle as ELISpot, but by utilizing fluorescent detection systems, it enables analysis of cells secreting one or multiple cytokines in one well simultaneously.

Fifteen female subjects (22–46 years; median age 40 years) with previously established patch test reactivity to Ni, ranging from +1 (weak) to +3 (strong), were included in the study. Six subjects were +3 patch test positive, 5 subjects +2, and 4 subjects were +1. Details on their patch test reactivity and collection of blood samples are outlined in a previous study [23]. Blood was also collected from five non-allergic individuals (4 female and 1 male; 35–61 years; median age 45 years) included in the study based on no history or manifestation of Ni allergy. The study was approved by the Ethics Committee of the Medical Faculty at Lund University (Lund, Sweden), and all subjects gave their written consent prior to participation. PBMCs were prepared from the blood by Ficoll separation and stored in liquid nitrogen [23].

PBMCs were thawed and transferred to cell culture medium: RPMI-1640 supplemented with HEPES (10 mM), l-glutamine (2 mM), penicillin G sodium-streptomycin sulfate (diluted 1:100) (all from Gibco BRL, Life Technologies Ltd., Paisley, UK), and 10% fetal bovine serum (HyClone Laboratories Inc., Logan, UT, USA). PBMCs at a concentration of 3 × 10⁶ cells/mL in a final volume of 1.5 mL were incubated with either 50 μM NiCl₂ × 6H₂O (Merck, Darmstadt, Germany), 20 μg/mL of tetanus toxoid (TT; SBL Vaccines, Solna, Sweden), phytohemagglutinin (Thermo Fisher Oxoid, Göteborg, Sweden), or without stimulus. Incubations were done for 40 h at 37°C with 5% CO₂ whereafter the supernatant was centrifuged and stored at −20°C. ELISA analysis of supernatants was performed using cytokine ELISA kits for human IFN-γ, IL-2, IL-3, IL-5, IL-6, IL-13, IL-17A, IL-22, and IL-31 (Mabtech, Nacka Strand, Sweden) according to the instructions of the manufacturer. For the analysis, supernatants were serially diluted from 1:2 in four consecutive three-fold dilutions to obtain absorbance values in the linear range. Samples with higher cytokine levels were further diluted and retested. Supernatants with non-detectable cytokine levels were set to <5 pg/mL for all ELISAs except for IL-3 and IL-6 where 20 pg/mL was used as the lowest detection level based on the lowest detection limits in the different assays.

Low-fluorescent 96-well PVDF plates (Millipore Life Science, Solna, Sweden) were pre-wetted with 15 μL 35% ethanol for 1 min, immediately followed by washing with sterile water five times (200 μL/well). Capture mAbs at 15 μg/mL (for each mAb) in sterile PBS (100 μL) were added to the wells and incubated for 20 h at 4°C. The capture mAbs for different cytokines were 1-D1K (IFN-γ), TRFK5 (IL-5), MT44.6 (IL-17A), MT12A3 (IL-22), IL-13-I (IL-13) (Mabtech). The following day, plates were washed five times with sterile PBS (200 μL/well) and blocked with 200 μL/well of cell culture medium (as above) for 1 h. The medium was removed and fresh medium containing PBMC, without or with stimulus, was added in a volume of 100 μL/well. PBMC from three strongly reactive Ni-allergic donors were thawed, rested for 1 h at 37°C with 5% CO₂, counted, and added to each well with or without stimulus (500,000 cells/well except for PHA 50,000 cells/well). Each sample was analyzed in duplicates or triplicates (based on the availability of PBMC). The stimuli used were 50 μM of NiCl₂, 20 μg/mL of Candida albicans extract (Greer, Lenoir, NC, USA), or 2 μg/mL PHA. PBMCs were incubated for 40 h at 37°C with 5% CO₂ and cells were removed by washing five times with PBS (200 μL/well) in an automated ELISA washer (Bio-Tek Instruments Inc., Winooski, VT, USA). Mixed detection mAbs were added and incubated for 2 h at room temperature followed by washing as above and incubation with mixed secondary fluorophore-labeled detection reagents for 2 h. Fluorescence enhancer (Mabtech) was added (50 μL/well) and plates were dried before analysis in an ELISpot/FluoroSpot reader (iSpot Spectrum, AID, Strassberg, Germany) with software version 7.0, build 14790. Fluorescent spots were counted utilizing separate filters for fluorophores absorbing and emitting light at 490/520, 550/570, and 640/660 nm. Camera settings (exposure and gain) were adapted for each filter to obtain high-quality spot images, preventing over- or under-exposure. Four different combinations of cytokines were analyzed: IFN-γ/IL-5/IL-17A; IFN-γ/IL-22/IL-17A; IFN-γ/IL-13/IL-22; IL-5/IL-13. Detection reagents used for the different cytokines were: for IFN-γ, detection mAb 7-B6-1 conjugated with FITC followed by mAb anti-FITC conjugated with 490 fluorophore; for IL-13, the biotinylated detection mAb IL-13-3 (3 μg/mL) followed by streptavidin (SA) conjugated with fluorophore 550; for IL-22, the biotinylated detection mAb MT7B27 (0.5 μg/mL) followed by SA conjugated with fluorophore 550 in the IFN-γ/IL-22/IL-17A FluoroSpot. In the IFN-γ/IL-13/IL-22 FluoroSpot, IL-22 was detected by mAb MT7B27 conjugated to the peptide tag BAM (0.5 μg/mL) followed by mAb anti-BAM conjugated to 640. For IL-17A, detection was made using mAb MT504-BAM (0.5 μg/mL) followed by a-BAM-640. For IL-5, detection was achieved with mAb 5A10-biotin (2 μg/mL) and SA-550 in the IFN-γ/IL-5/IL-17A FluoroSpot and with 5A10-BAM (1 μg/mL) and a-BAM-490 in the IL-5/IL-13 FluoroSpot. All detection reagents were from Mabtech. Detection reagents, not specified by concentrations, were diluted 1:200 according to the manufacturer’s instructions.

The Mann-Whitney U test was used to compare the responses of the allergic and the control groups using GraphPad Prism software version 5.04; 2010, (GraphPad, San Diego, CA, USA). The correlation between the levels of different cytokines was analyzed using Spearman’s rank order correlation coefficient (rs). Spearman rank test was performed using Analyse-it for Microsoft Excel version 1.65 (Analyse-it Software Ltd., Leeds, England, UK). Due to zero or low values in the non-allergic control groups, only allergic subjects were included in the Spearman analysis. The background values (PBMC in medium without stimulus) were subtracted from the stimulus-induced values prior to statistical analyses. If values below the detection limit were obtained, the values were set to 5 pg/mL or to 20 pg/mL (for IL-3 and IL-6), which was considered the lowest detection limit of the systems. A probability of p < 0.05 was considered statistically significant.

PBMC from Ni-allergic (n = 15) and non-allergic (n = 5) subjects were stimulated with NiCl₂ for 40 h followed by ELISA analysis of cytokine levels in the supernatants (shown in Fig. 1). In Figure 1, it is also shown which differences between groups were significantly significant. The group of allergic subjects comprised subjects with +3 (n = 6), +2 (n = 5), and +1 (n = 4) patch test reactivity. IFN-γ, IL-2, IL-3, IL-5, IL-6, IL-13, IL-17A, IL-22, and IL-31 were all induced by Ni at higher levels in PBMC from the +3 and +2 groups versus the control group. The levels of the +3 group were also statistically higher than the +1 group for all cytokines except IL-3; the +2 group was higher than the +1 group for IFN-γ, IL-2, IL-5, and IL-6. The +1 group displayed higher levels than the controls for IFN-γ, IL-2, IL-5, IL-13, IL-22, and IL-31.

Tetanus toxoid (TT), a protein antigen included for comparison, induced measurable levels of all cytokines but with little difference between groups. The exceptions were that the +2 group had higher IL-2, IL-5, and IL-13 levels than the controls and IL-5 levels were also higher than the +1 group. Also, IFN-γ differed with higher levels in controls versus the +1 group. PBMC from all subjects responded with all cytokines to the polyclonal activator PHA (data not shown).

The level of different cytokines induced by Ni displayed a statistically positive correlation between most cytokines (shown in Fig. 2a). IL-2, sometimes classified as a Th1 marker but more often considered a general Th activation marker [24], displayed the highest degree of correlation with all other cytokines, whereas IL-6 displayed the lowest correlation and only correlated significantly with IFN-γ, IL-2, and IL-22. Cytokine responses to TT displayed a more limited correlation (shown in Fig. 2b). Similar to the Ni-induced responses, however, TT yielded a high correlation between IL-2 and IL-22 and also between the Th2 cytokines IL-5, IL-13, and IL-31.

After having established that Ni elicits all the hallmark cytokines of Th1, Th2, Th17, and Th22 when PBMCs from allergic subjects were restimulated with Ni, the question was asked if these cytokines are produced by distinct populations with distinct Th profiles and/or intermediate Th profile populations capable of secreting mixed Th cytokines.

PBMC from three +3 subjects were stimulated with Ni and cytokine responses were assessed using FluoroSpot assays for various combinations of cytokines. A majority of cells identified in FluoroSpot assays for IFN-γ/IL-13/IL-22 (shown in Fig. 3a), IFN-γ/IL-17A/IL-22 (shown in Fig. 3b), and IFN-γ/IL-5/IL-17A (shown in Fig. 3c) were single positive, i.e., the cells only produced one of the cytokines assessed. Images from one PBMC preparation analyzed for IFN-γ/IL-5/IL-17A are shown in Figure 4 as an example. In line with the ELISA data, spots positive for cytokines from all four Th populations were found.

The largest population of Ni-reactive double-positive cells was identified when IL-5 and IL-13, both Th2 cytokines, were analyzed in the same assay; 10% of the spots were double positive (shown in Fig. 3d). In the IFN-γ/IL-17A/IL-22 assay, 3% of the Ni-induced spots were double positive for IL-17A/IL-22; these spots could potentially derive from Th17 cells co-secreting IL-22 (shown in Fig. 3b). Double-positive spots representing hallmark cytokines from different Th populations were rare for other combinations of Th1, Th2, Th17, and Th22 cytokine combinations. Triple-positive spots were rare.

T-cell responses to Candida albicans extract were assessed in parallel for comparison since it has been shown by FluoroSpot that C. albicans-derived antigens elicit Th1, Th17, and Th22 responses in most healthy donors [25]. PBMC from the three Ni-allergic donors displayed C. albicans-induced secretion of single cytokines representing Th1, Th17, and Th22 and Th2 profiles (shown in Fig. 3e–g). The largest proportion of double-positive spots was IL-17A/IL-22 which represented 9% of the spots in the IFN-γ/IL-22/IL-17A analysis. Proportions of spots double-positive for cytokines from different Th profiles like IFN-γ/IL-17A and IFN-γ/IL-22 were higher after stimulation with C. albicans compared to Ni (shown in Fig. 3a, c, e, g), whereas double-positive IL-5/IL-13 spots were comparable (shown in Fig. 3h).

A difference was observed when production of IL-13 was analyzed simultaneously with IFN-γ and IL-22 (shown in Fig. 3c, g) versus with IL-5 (shown in Fig. 3d, h). When analyzed together with IL-5, the number of IL-13 spots obtained were 232 spots/well (Ni; shown in Fig. 3d) and 180 spots/well (C. albicans; shown in Fig. 3h). In contrast, when analyzed together with IFN-γ and IL-22, the number of IL-13 spots were 87 spots/well (Ni; shown in Fig. 3c), and 38 spots/well (C. albicans; shown in Fig. 3g). To assess if the simultaneous analysis of IFN-γ and IL-22 affected the IL-13 production negatively, analysis of IL-13 was performed alone in wells or together with IFN-γ and/or IL-22. The results showed that analysis of IL-13 in parallel with IFN-γ and IL-22 reduced the IL-13 response compared to analysis of IL-13 alone (data not shown). The explanation for this phenomenon is the fact that capture of IFN-γ and IL-22, when secreted by cells in response to a stimulus, impacts the production of IL-13 by other stimulus-specific cells. Similar evaluations were made with the IFN-γ/IL-5/IL-17A and the IFN-γ/IL-22/IL-17A systems, but here the simultaneous analysis of cytokines had little impact on the production of any of the cytokines (data not shown). PBMC from non-allergic subjects did not respond to Ni but displayed responses to C. albicans in a manner comparable to the allergic subjects (data not shown).

The presence of Ni-specific Th1, Th2, and Th17 in PBMC from Ni-allergic subjects has been shown previously. Also, potential Th22 cells have been implicated in ACD based on the finding of infiltrating IL-22-producing T cells in Ni-provoked skin lesions [12]. We found that Ni-stimulated PBMC from Ni-allergic subjects produced all the hallmark cytokines of Th1, Th2, Th17, and Th22 type cells in a correlated manner. Cytokines levels were also positively associated with disease severity as defined by patch test reactivity. By FluoroSpot, it was shown that these cytokines to a large extent are produced by cells with a distinct Th profile with few cells co-secreting cytokines associated with different Th populations.

The cytokines investigated herein represent signature cytokines from different Th populations: IFN-γ for Th1; IL-5, IL-13, and IL-31 for Th2; IL-17A for Th17; IL-22 for Th22. IL-2 and IL-3 that are more general markers of Th activation were also measured, as was IL-6, shown to be secreted by potassium dichromate-stimulated PBMC from allergic subjects [26]. Using ELISpot, production of the Th2 cytokines IL-4 and IL-13 by Ni-stimulated PBMC from the same allergic subjects as in this study was previously shown to be strongly correlated with patch test reactivity [23]. IL-4 was not included in the ELISA analysis herein since IL-4 is notoriously difficult to measure in supernatants from antigen-stimulated PBMC due to IL-4 consumption by cellular receptors [27, 28]. IFN-γ, IL-2, and IL-5 have previously been shown to be produced by Ni-reactive PBMC and to be associated with patch test reactivity, as has IL-17A, albeit not in all studies which may relate to the quality of the immunoassays used [3, 12, 14, 29]. IL-3, IL-6, IL-22, and IL-31 have, to our knowledge, not previously been directly shown to be produced by Ni-specific cells after in vitro stimulation of PBMC from allergic subjects. Ni-reactive PBMC, from Ni-allergic subjects, that produce IL-13 and IFN-γ in ELISpot assays have previously been phenotyped as CD4+ cells [30]. Phenotyping of cells was not performed in this study, and hence the cells are referred to as displaying Th-type cytokine profiles.

When analyzed by FluoroSpot, a majority of the Ni-reactive PBMC were found to produce cytokines representing distinct Th profiles, i.e., most spots were single positive for either IFN-γ (Th1), IL-17 (Th17), IL-22 (Th22) or IL-5/IL-13 (Th2). In contrast to these distinct Ni-reactive cell populations found in PBMC, analysis of Ni-reactive T-cell clones has previously shown a more promiscuous cytokine expression pattern. Ni-reactive T-cell clones derived from the skin and blood of Ni-allergic subjects have been shown to display a more mixed cytokine expression pattern with a majority of clones secreting mixtures of Th1 (IFN-γ), Th2 (IL-4), and/or Th17 (IL-17A) cytokines [31]. Similarly, when T-cell clones against beta-lactam antibiotics, derived from hypersensitive individuals, were analyzed for IFN-γ, IL-13, and IL-22 production, most clones displayed a mixed cytokine production pattern [32]. The difference between ex vivo-stimulated T cells in PBMC and T-cell lines/clones may be explained by the fact that T-cell lines/clones undergo DNA modifications and obtain chromosomal abnormalities with increased numbers of cell cycles which in turn may result in atypical cytokine expression patterns [33]. Thus, data defined using T-cell clones/lines may have to be interpreted with caution.

The balance between the antagonizing Th1 and Th2 responses was the first Th dichotomy described to have an impact on both protective and detrimental immune response. Pathological reactivity in atopic allergies, for example, is associated with an enhanced Th2 response [34]. In atopic allergy, increased Th17 and Th22 have later been found together with Th2 in allergic subjects [35]. Systemic sclerosis, a chronic autoimmune disease, has also been associated with increased Th2, Th17, and Th22 [36]. Other examples where Th1 reactivity is found together with Th17 and Th22 reactivity have been described. For example, in individuals with psoriasis, Th1, Th17, and Th22 reactivity is found and has been suggested to contribute to cutaneous inflammation [10]. In contrast, when analyzing PBMC from subjects hypersensitive to the b-lactam antibiotic piperacillin, antigen-specific Th1, Th2, Th22 were found but not Th17 [32]. In ACD to Ni, as found herein, antigen-specific T cells representing all four Th populations were found to be increased; this was also the case with T cells specific for C. albicans antigens. This study shows the involvement of distinct Th1, Th2, Th17, and Th22 population in ACD to Ni, but their respective contribution to detrimental and/or down-regulatory immune responses remains to be further elucidated.

The authors would like to thank Jens Gertow and Peter Jahnmatz for critical reading of the manuscript.

The study was approved by the Ethics Committee of the Medical Faculty at Lund University (Lund, Sweden). Ethical application number: Dnr 457/2004. Written informed consent was obtained from participants to participate in the study.

Khosro Masjedi and Niklas Ahlborg are employed by Mabtech AB, Sweden. Several of the reagents used in this study are manufactured by Mabtech. Mabtech had no influence on the content of this study or interpretation of results. Magnus Bruze has no conflict of interest to declare.

This study was funded by Mabtech AB, Sweden. The company Mabtech had no influence on the content of this study or interpretation of results.

Design of the study, analysis of data, and writing of manuscript: K.M., M.B., and N.A.; patient contact and sample collection: MB; and sample preparation and experiments: K.M.

All data generated or analyzed during this study are included in the article except some additional control data mentioned as “data not shown” which is available from the corresponding author upon request. Further inquiries can be directed to the corresponding author.