Lupus-prone female New Zealand Black × New Zealand White F1 (NZB/W F1) and non-lupus-prone female Balb/c x New Zealand White F1 (CW F1) mice were obtained from the Federal Institute for Risk Assessment, Berlin, Germany, and the Research Facility for Experimental Medicine of the Charité – University Medicine, Berlin, Germany. The mice were kept in a specific pathogen-free (SPF) environment at the German Rheumatism Research Centre (DRFZ), Berlin, Germany. All experiments were performed according to institutional and federal guidelines (State Office for Health and Social Affairs, Berlin, Germany). NZB/W F1 mice of different disease stages were used: young, clinically healthy mice (8-12 weeks of age, in average 10.4 weeks, no proteinuria); mice at disease onset (5-7 months of age, in average 6.4 months, proteinuria = 30-100 mg/dL); mice with established disease (7-9 months of age, in average 8.8 months, proteinuria = 300-2000 mg/dL). Each mouse was analyzed individually.

The SmD1(83-119) peptide (=SmD1p) (VEPKVKSKKREAVA GRGRGRGRGRGRGRGRGRGGPRR) and a control peptide (= control) with amino acids identical to the SmD1(83–119) peptide, however, with a random sequence of the same amino acids (CREKGRVGRGRPAVGRRGVGRPGRRGSRARGEGKGRK), were synthesized according to a standard procedure. For CD4⁺ T cell epitope mapping 7 overlapping 15-mer peptides derived from SmD1(83-119) [epitope 1 = SmD1(83–97), -epitope 2 = SmD1(85–99), -epitope 3 = SmD1(87–101), -epitope 4 = SmD1(90–104), -epitope 5 = SmD1(93–107), -epitope 6 = SmD1(97–111), and -epitope 7 = SmD1(105–119)] were synthesized by JPT Peptide Technologies GmbH, Berlin/Germany. An 11-mer peptide derived from hen egg lysozyme (HEL (106-116) = NAWVAWRNRCK) was used as control for the 15-mer SmD1 peptides. All peptides were solved in PBS, sterile filtered by 0,2µm filters, and tested negative for the presence of endotoxins.

For the in vivo generation of SmD1p-specific CD4⁺ T cells, 8-10 weeks old female NZB/W F1 mice or 8-10 weeks old female CW F1 mice were immunized intravenously in the tail vein with 25µg of SmD1(83-119) together with 10µg of lipopolysaccharide (LPS, from E. coli 026:B6*C – Sigma-Aldrich, Taufenkirchen/Germany) as adjuvant diluted in 100µl PBS. After 8 days, a booster immunization was performed subcutaneously in the tail base with 25µg SmD1(83-119) diluted in PBS and emulsified with incomplete Freund’s adjuvant (IFA) (1:1) in a total volume of 50 µl. Another 7 days later, mice were sacrificed and spleen cells were harvested.

For the in vivo depletion of CD25⁺ Treg, 8-10 weeks old female NZB/W F1 or CW F1 mice were injected i.v. with 200µg of an anti-CD25-antibody (rat IgG1, clone PC61.5, DFRZ, Berlin/Germany) one week before the initial immunization. The success of depletion was confirmed by flow cytometry in peripheral blood cells using a non-cross-reactive anti-CD25 antibody (clone 7D4, DRFZ, Berlin/Germany).

Spleens from NZB/W F1 or CW F1 mice were mashed individually through a 70µm cell strainer (BD Biosciences, Heidelberg/Germany), and the cell suspension was washed once with PBS/0,2% BSA. Erythrocytes were eliminated by lysis using a solution containing 0,01M KHCO3, 0,155M NH4Cl, and 0.1mM EDTA at pH 7.5 followed by washing with PBS/0,2% BSA/2mM EDTA.

Erythrocyte-depleted splenocytes were incubated with monoclonal PE-conjugated anti-CD25 antibody (7D4) followed by anti-PE MicroBeads (both Miltenyi Biotec, Bergisch Gladbach, Germany) according to the manufacturer’s instructions. CD25⁺ cells were depleted using the AutoMACS (Miltenyi Biotec, Bergisch Gladbach, Germany). The CD25^- fraction showed a purity of over 99% as analyzed by flow cytometry.

For the detection of antigen-specific CD4⁺ T cells and CD4⁺ T cell epitope mapping, splenocytes from un-depleted, in vivo CD25⁺ depleted, in vitro CD25⁺ depleted and both in vivo and in vitro CD25⁺ depleted mice were used. Cells were re-stimulated in vitro with SmD1p, 15-mers derived from SmD1p, or with control peptides (randomized peptide, HEL) at a concentration of 20µg/ml for 6h in AIM-V medium (Invitrogen, Darmstadt/Germany) in 96-well plates at a cellular concentration of 1x10⁷ cells/ml. For the comparison between surface and intracellular staining of CD154, different concentrations ranging between 5 µg/ml and 100 µg/ml of SmD1p were used for in vitro re-stimulation.

For intracellular staining of CD154⁺ cells, 5µg/ml brefeldin A (Sigma-Aldrich, Hamburg/Germany) was added 2 hours after the addition of antigens to the cell culture. After additional 4 hours of stimulation, cells were fixed in 2% paraformaldehyde for 20 min at room temperature, washed three times, and kept in PBS containing 0,2% BSA and 0,01% sodium azide at 4°C until FACS analyses were performed. Fixed cells were permeabilized in PBS/0,2% BSA containing 0.5% saponin and stained intracellularly using allophycocyanin (APC)-conjugated anti-CD154 antibodies (clone MR1, Miltenyi Biotec, Bergisch Gladbach/Germany). Anti-Fc-γ-receptor antibody (clone 2.4G2, 40µg/ml, DRFZ, Berlin/Germany) was added during the staining procedure to avoid unspecific binding. For surface staining of CD154, unconjugated anti-CD40 antibody (clone 1C10, 10µg/ml, BioLegend, San Diego/USA), anti-Fc-γ-receptor antibody (clone 2.4G2, 40µg/ml, DRFZ, Berlin/Germany) and APC-conjugated anti-CD154 antibody (clone MR1, Miltenyi Biotec, Bergisch Gladbach/Germany) were added to the culture from the beginning. After intracellular staining of CD154 or after the in vitro re-stimulation period (6h) for the surface staining of CD154, cells were washed with PBS/0,2% BSA and stained with FITC-conjugated anti-CD4 antibody (clone RM4-5, 5 µg/ml, BD Pharmingen, Heidelberg/Germany) on ice for 15 minutes. Stained cells were analyzed with a FACSCalibur 4-color flow cytometer (BD Biosciences, Heidelberg/Germany).

Fixed cells were permeabilized in PBS/0,2% BSA containing 0.5% saponin and stained for intracellular cytokines and surface markers. Anti-Fc-γ-receptor antibody (clone 2.4G2, 40µg/ml, DRFZ, Berlin/Germany) was used to avoid unspecific binding. The following monoclonal antibodies were used for intracellular and surface staining: APC-conjugated anti-CD154 (clone MR1, Miltenyi Biotec, Bergisch Gladbach/Germany), FITC-conjugated anti-IFN-γ (clone XMG1.2, DRFZ, Berlin/Germany), FITC-conjugated anti-TNF-α (clone MP6-XT22, DRFZ, Berlin/Germany), PE-conjugated anti-IL-2 (clone JES6-5H4, DRFZ, Berlin/Germany), PE-conjugated anti-IL-10 (clone JES5-16E3, DRFZ, Berlin/Germany), PE-conjugated anti-IL-17 (clone TC11-18H10, DRFZ, Berlin/Germany), Streptavidin-PerCP (BD Pharmingen, Heidelberg/Germany) and biotin-conjugated anti-CD4 (clone GK1.5, DRFZ, Berlin/Germany). FoxP3-staining was performed using FITC- or PE-conjugated anti-FoxP3 (clone FJK-16s, eBioscience, San Diego/USA) according to the manufacturer’s instructions. Stained cells were analyzed with a FACSCalibur 4-color flow cytometer (BD Biosciences, Heidelberg/Germany).

GraphPad Prism 5 software (Windows version, GraphPad Software Inc., La Jolla/USA) was used for statistical analyses. The two-tailed paired t-test was used to compare differences between SmD1p and control peptide-stimulated samples. Mann-Whitney test was used for the comparison of frequencies of SmD1p-specific CD4⁺ T cells between CW F1 and NZB/W F1 mice and to compare differences in the percentages of FoxP3⁺ and FoxP3^- cells among CD4⁺CD25⁺ splenocytes at different disease stages in NZB/W F1 mice. Two-way ANOVA was used to compare differences in the frequencies of SmD1p-specific CD4⁺ T cells between undepleted, in vitro depleted, in vivo depleted and in vivo and in vitro CD25⁺ depleted samples and between controls. Two-way ANOVA was also used for the comparisons between surface and intracellular staining of CD154. The statistical tests used for each experiment are indicated in the respective figure legends. P values of < 0.05 were considered statistically significant.

Our previous works suggested that the highest frequencies of SmD1p-specific CD4⁺ T cells could be generated by immunization of NZB/W F1 mice against the autoantigen (3). Thus, we primarily used splenocytes from immunized mice for the detection of SmD1p-specific CD4⁺ T cells after antigenic in vitro re-stimulation. Firstly, we aimed to compare the yield of detectable SmD1p-specific CD4⁺ T cells obtained by conventional intracellular staining for the Th1 lineage cytokine IFN-γ, the most abundant pro-inflammatory CD4⁺ T cell cytokine in murine lupus (11, 19), with intracellular staining for CD154. Intracellular staining for IFN-γ yielded a mean frequency of 0.32% (range: 0.18-0.40%) of SmD1p-specific IFN-γ⁺ cells among CD4⁺ T cells which were significantly higher than the percentages of CD4⁺IFN-γ⁺ T cells obtained by stimulation with the control peptide (mean 0.16%; p=0.02, SmD1p vs control; Figure 1A ). However, the unspecific background response obtained by re-stimulation with the control peptide was relatively high representing app. 50% of IFN-γ⁺ cells obtained by re-stimulation with SmD1p. In addition, the staining was diffuse with a high variability in fluorescence intensity and high standard deviations (app. 0.1%) among the individually analyzed mice. Using intracellular staining for CD154 yielded a mean frequency of 0.12% (range: 0.10-0.13%) of SmD1p-specific CD154⁺ cells among CD4⁺ T cells. Although the frequencies of detectable antigen-specific CD4⁺ T cells were lower compared to the IFN-γ⁺ staining, the unspecific background activation was nearly absent when using the CD154 staining approach (mean 0.02% for re-stimulation with control peptide) and standard deviations were low, providing also a higher level of statistical significance for the difference between SmD1p and control peptide stimulated samples (p=0.01, Figure 1B ). In addition, the discrimination of positively stained CD4⁺ T cells from negatively stained CD4⁺ T cells was superior by using the CD154 staining approach in comparison to conventional cytokine staining. Because of the more reliable detection and superior background discrimination of autoantigen-specific CD4⁺ T cells, we used the CD154 intracellular staining approach for the detection of autoantigen-specific CD4⁺ T cells in our further experiments.

Treg are known for their capability to suppress T cell responses, in particular against autoantigens, in order to efficiently counteract autoimmunity (8). To investigate whether Treg are capable to suppress SmD1p-specific CD4⁺ T cell responses we first depleted CD25⁺ cells in vitro from splenocytes of immunized NZB/W F1 mice. Efficient depletion of CD25⁺ cells, of which app. 90% expressed the Treg lineage marker FoxP3 and co-expressed CD4, among the CD4⁺FoxP3⁺ Treg population was confirmed before antigenic re-stimulation ( Figure 2A , left). Upon re-stimulation with SmD1p, samples depleted of CD25⁺ Treg showed an average increase in the frequency of detectable CD4⁺CD154⁺ T cells by 0.04% compared to un-depleted samples (mean 0.16% vs 0.12%, p=0.0044), whereas the response to the control peptide remained low (mean 0.03%, p<0.0001, SmD1p vs control peptide) comparable to this of un-depleted samples ( Figures 2A, C right).

To further substantiate this putative inhibitory effect of Treg on antigen-specific T cell responses, we now depleted CD25⁺ Treg in vivo by injecting depleting anti-CD25 antibodies before immunization of NZB/W F1 mice against SmD1p ( Figure 2B , left). In vivo depletion of CD25⁺ Treg resulted in an average increase in the frequency of detectable CD4⁺CD154⁺ T cells by 0.05% upon in vitro re-stimulation with SmD1p compared to un-depleted samples (mean 0.17% vs 0.12%, p=0.0017) ( Figures 2B, C , right). This increase in the frequency of SmD1p-specific CD4⁺ T cells was comparable to the increase obtained by the in vitro depletion approach (difference not significant). Similar to the in vitro depletion experiments, the response to the control peptide remained almost absent (mean 0.02%, p<0.0001; SmD1p vs control peptide; Figure 2C , right).

In light of these results, we asked whether the combination of both approaches, the in vivo and in vitro depletion of CD25⁺ Treg, could have an additive effect on autoantigen-specific CD4⁺ T cell responses. Indeed, the combined depletion approach resulted in the detection of the highest frequencies of CD154⁺ cells among CD4⁺ T cells with a mean frequency of 0.20% ( Figure 2C ), representing an increase of SmD1p-specific CD4⁺ T cells by app. 0.08% compared to un-depleted samples (p=0.0002) and by app. 0.3% and 0.4% compared to in vitro (p=0.009) and in vivo (p=0.0293) depleted samples, respectively. Again, there was no relevant response to the control peptide (mean 0.03%, p<0.0001; SmD1p vs control peptide; Figure 2C ) allowing for the reliable discrimination between positive and negative cells.

To address whether our observations are specifically associated with susceptibility for lupus, we analyzed for the presence of SmD1p-specific CD4⁺ T cells in the genetically related mouse strain (BALB/c x NZW) F1 (CW F1) which bears the same MHC haplotypes as the NZB/W F1 strain but does not develop lupus-like disease. Immunized CW F1 mice at the same age as NZB/W F1 were subjected to a combined in vivo (before immunization) and in vitro depletion of CD25⁺ Treg, as described above for the NZB/W F1 mice. Although, SmD1p-specific CD4⁺ T cells according to CD154 expression were detectable above background in CW F1 mice upon in vitro re-stimulation (mean 0.09% vs 0.02%, p=0.002, SmD1p vs control peptide; Figure 3A ) their frequency was more than 2-fold lower compared to the responses observed in Treg depleted NZB/W F1 lupus-prone mice (p=0.002; CW F1 vs NZB/W mice; Figure 3A ).

To obtain insights into the relevance of SmD1p-specific CD4⁺ T cells and the inhibitory potential of Treg on autoantigen-specific CD4⁺ T cell responses in disease pathogenesis, we examined whether SmD1p-specific CD4⁺ T cells can be also detected and their frequency increased by depletion of CD25⁺ Treg during spontaneous disease development. Accordingly, splenocytes from non-immunized NZB/W F1 mice at different disease stages were in vitro depleted from CD25⁺ Treg or left un-depleted before stimulation with SmD1p. In young, clinically healthy NZB/W F1 mice no relevant frequencies of SmD1p-specific CD4⁺CD154⁺ T cells above background levels could be observed, neither in un-depleted nor in depleted samples ( Figure 3B ). Although, frequencies of SmD1p-specific CD4⁺CD154⁺T cells were higher in un-depleted samples of NZB/W F1 mice at the onset of the disease compared to young NZB/W F1 mice, there was no significant difference compared to the control peptide re-stimulation in both un-depleted and depleted samples ( Figure 3B ). However, in mice with established disease there was a significantly higher frequency of SmD1p-specific CD4⁺ T cells in un-depleted samples compared to the stimulation with the control peptide (p<0.05) and the frequency reached similar levels as observed in young mice after immunization with SmD1p, although a much higher level of unspecific background staining for CD154 was evident in the controls suggesting a higher state of constitutive T cell activation in mice with active disease ( Figure 3B ). Unexpectedly, this difference was no more detectable after in vitro depletion of CD25⁺ cells as the frequency of SmD1p-specific CD4⁺CD154⁺ T cells dropped to the same level obtained by stimulation with the control peptide. Since CD25 is not exclusively expressed on Treg and is known to be expressed also on activated conventional CD4⁺ T cells (Tcon), and frequencies of CD25⁺ cells among CD4⁺FoxP3^- Tcon were shown to be increased in NZB/W F1 mice with active disease (11), the observed loss of SmD1p-specific CD4⁺ T cells after depletion of CD25⁺ cells in diseased mice could be explained by the depletion of the CD4⁺CD25⁺ Tcon population containing also activated SmD1p-specific CD4⁺ T cells. Thus, we analyzed the percentages of FoxP3⁺ and FoxP3^- cells within the CD4⁺CD25⁺ T cell population, respectively, at different disease stages. In line with our assumption, the percentages of FoxP3⁺ Treg among CD4⁺CD25⁺ T cells were significantly reduced in diseased NZB/W F1 mice compared to young (p=0.0019) and onset (p=0.04) mice, whereas the percentages of FoxP3^- Tcon among CD4⁺CD25⁺ T cells were significantly higher compared to young (p=0.0019) and onset (p=0.04) mice ( Figure 3C ).

To exclude that the observed increase in the frequency of auto-antigen specific CD4⁺ T cells by depletion of CD25⁺ Treg is biased due to changes in the cellular composition induced by the removal of CD25⁺ cells, we tested whether the in vitro re-addition of CD25⁺ Treg after Treg depletion is capable to restore Treg-mediated suppression of autoantigen-specific T cell responses by decreasing the percentage of detectable SmD1p-specific CD4⁺ T cells again. Accordingly, we used splenocytes from in vivo and in vitro CD25⁺ depleted and immunized NZB/W F1 mice and re-added either none, 5%, or 10% of purified CD25⁺ Treg from unmanipulated, 8-10 weeks old NZB/W F1 mice (total cell number in the cell culture remained the same in all settings) before in vitro re-stimulation with SmD1p. Indeed, the re-addition of 10% of purified CD25⁺ Treg resulted in a decrease in the percentages of detectable SmD1p-specific CD4⁺CD154⁺ T cells from a mean of 0.15% to a mean of 0.09% (p=0.01, Figure 4 ), which corresponds to frequencies obtained by stimulation without prior Treg depletion. The re-addition of 5% of purified CD25⁺ Treg, however, was not sufficient to reduce the percentages of SmD1p-specific CD4⁺CD154⁺ T cells.

Next, we aimed to delineate the lineage differentiation by determining the cytokine profile of SmD1p-specific CD4⁺CD154⁺ T cells in splenocytes from immunized NZB/W F1 mice that were subjected to the combined Treg depletion approach. We found that SmD1p-specific CD4⁺CD154⁺ T cells produced mainly TNF-α (mean frequency 53%) and IFN-γ (mean frequency 48%), and to a much lesser extent IL-10 and IL-17 (mean frequency < 4% for both), indicating that the majority of SmD1p-specific CD4⁺ T cells belonged to the Th1 lineage ( Figure 5A ).

We also performed fine epitope mapping with 7 different overlapping 15-mer peptide sequences derived from the SmD1p to determine the pre-dominant CD4⁺ T cell epitope. We compared the frequencies of antigen-specific CD4⁺CD154⁺ T cells obtained from immunized NZB/W F1 mice without depletion of CD25⁺ Treg to those from immunized NZB/W F1 mice with combined in vivo and in vitro depletion. In both experimental settings, re-stimulation with the SmD1p sequence 3 (amino acids 87-101) provided the highest frequencies of epitope-specific CD4⁺CD154⁺ T cells (mean 0.04%, p=0.03 for un-depleted vs control; mean 0.11%, p=0.001 for CD25⁺ depleted vs control; Figure 5B ). Notably, in the CD25⁺ depleted samples the frequencies of epitope 3-specific CD4⁺CD154⁺ T cells were significantly higher than in the un-depleted samples (p=0.007; Figure 5B ). Apart from the epitope sequences 2 and 4, which elicited a moderate increase in the frequency of CD4⁺CD154⁺ T cells in CD25⁺ depleted samples, no relevant T cell responses to other peptide sequences were observable ( Figure 5B ). Thus, in NZB/W F1 mice epitope sequence 3 was identified as the dominant CD4⁺ T cell epitope within the entire SmD1p sequence.

One advantage in the detection of (auto)antigen-specific CD4⁺ T cells by CD154 staining and flow cytometry is the possibility to stain CD154 on the surface of living cells, in order to isolate antigen-specific CD4⁺ T cells alive with only little manipulations of these cells. Thus, we compared the frequencies of SmD1p specific CD4+ T cells obtained by surface staining of CD154 with those obtained by the intracellular staining approach by using increasing concentrations of the SmD1p antigen. The surface staining of CD154 on living SmD1p-specific CD4⁺ T cells in general revealed lower frequencies of these cells compared to the intracellular CD154 staining approach for all antigen concentrations tested ( Figure 5C ). However, significantly higher frequencies of SmD1p-specific CD4+ T cells compared to the re-stimulation with the control peptide were evident for all antigen concentrations (p=0.01) and highest frequencies were observed by stimulation with concentrations of 20 µg/ml and 50 µg/ml. The intracellular CD154 staining approach yielded significantly higher frequencies of CD4⁺CD154⁺ T cells than the surface CD154 staining in nearly all concentrations of SmD1p used for re-stimulation (p<0.05 and p<0.01), except for the re-stimulation experiment with the concentration of 20 µg/ml (ns, p=0.06). Similar to what was observed with the surface staining, there was a tendency toward a dose dependency of the SmD1p-specific CD4+ T cell responses ( Figure 5C ).