Equine CD4⁺ CD25^high T cells exhibit regulatory activity by close contact and cytokine-dependent mechanisms in vitro

Abstract

Horses are particularly prone to allergic and autoimmune diseases, but little information about equine regulatory T cells (Treg) is currently available. The aim of this study therefore was to investigate the existence of CD4⁺ Treg cells in horses, determine their suppressive function as well as their mechanism of action. Freshly isolated peripheral blood mononuclear cells (PBMC) from healthy horses were examined for CD4, CD25 and forkhead box P3 (FoxP3) expression. We show that equine FoxP3 is expressed constitutively by a population of CD4⁺ CD25⁺ T cells, mainly in the CD4⁺ CD25^high subpopulation. Proliferation of CD4⁺ CD25⁻ sorted cells stimulated with irradiated allogenic PBMC was significantly suppressed in co-culture with CD4⁺ CD25^high sorted cells in a dose-dependent manner. The mechanism of suppression by the CD4⁺ CD25^high cell population is mediated by close contact as well as interleukin (IL)-10 and transforming growth factor-β1 (TGF-β1) and probably other factors. In addition, we studied the in vitro induction of CD4⁺ Treg and their characteristics compared to those of freshly isolated CD4⁺ Treg cells. Upon stimulation with a combination of concanavalin A, TGF-β1 and IL-2, CD4⁺ CD25⁺ T cells which express FoxP3 and have suppressive capability were induced from CD4⁺ CD25⁻ cells. The induced CD4⁺ CD25^high express higher levels of IL-10 and TGF-β1 mRNA compared to the freshly isolated ones. Thus, in horses as in man, the circulating CD4⁺ CD25^high subpopulation contains natural Treg cells and functional Treg can be induced in vitro upon appropriate stimulation. Our study provides the first evidence of the regulatory function of CD4⁺ CD25⁺ cells in horses and offers insights into ex vivo manipulation of Treg cells.

Introduction

Regulatory T cells (Treg) were first described as a subpopulation of circulating CD4⁺ T cells that express the interleukin-2 (IL-2) receptor α chain (CD25) and play a crucial role in maintaining self-tolerance.¹ Over the years, several additional functions of Treg cells in human and murine systems have been suggested, such as control of allergy,^2–5 limitation of chronic inflammation in various diseases,^6–8 induction of transplantation tolerance,^9–11 but also inhibition of the immune responses to tumours.^12–14 Horses, unlike other livestock animals, are particularly prone to allergic and autoimmune diseases and have infectious diseases and tumours related to those occurring in humans.¹⁵

A number of different regulatory T cell (Treg) populations have been described, the best-defined to date being CD4⁺ CD25⁺ Treg. A major CD4⁺ CD25⁺ Treg population is derived from the thymus, and is referred to as natural Treg (nTreg). In humans and mice, these represent the majority of circulating Treg under steady state conditions.^2,16,17 They are characterized by the constitutive expression of the transcription factor forkhead box P3 (FoxP3), which programmes their development in the thymus.^18–21 In addition to the naturally occurring Treg cells, CD4⁺ Treg can also be induced in the periphery (iTreg), a mechanism through which peripheral tolerance or immunosuppression can be adoptively modulated.^22,23 iTreg cells may be FoxP3⁺ or FoxP3⁻ and generally secrete the inhibitory cytokines such as IL-10 and transforming growth factor (TGF)-β1.^24–26In vitro studies in several species have demonstrated that the critical requirements for FoxP3⁺ iTreg induction are T cell receptor (TCR) stimulation and the cytokines IL-2 and TGF-β1.^27–29

Generally, Treg cells respond poorly to T cell receptor (TCR) stimulation in terms of cytokine production and proliferation, but possess the ability to suppress the immune response of other effector cells.³⁰ Various non-exclusive mechanisms of how Treg cells mediate their suppression have been proposed.^31,32 These include direct cell-to-cell contact^33,34 or secretion of the inhibitory cytokines IL-10 and TGF-β1.^17,35–37 An additional proposed mechanism is the consumption of IL-2.^38–40

So far, little information is available about equine Treg cells, but essential tools such as anti-CD25 and anti-FoxP3 antibodies have been described recently.⁴¹ The aim of this study was to examine the existence of equine Treg cells in healthy horses and analyse some of their specific mechanisms of inhibition.

This study shows the existence of equine circulating Treg cells, which constitutively express FoxP3. This population is comparable to the nTregs described in humans, but different from those described in other species such as mice, pigs⁴² or dogs.²⁷ These cells exhibit a suppressive capability and can be expanded in vitro while maintaining their function. Moreover, we demonstrate that equine Treg cells with a suppressive capability can be induced in vitro.

Materials and methods

Study group

Blood samples were obtained by jugular venipuncture from 12 clinically healthy adult horses and collected into sterile sodium heparin-containing Vacuette tubes (Greiner Bio-One Vacuette GmbH, St-Gallen, Switzerland). The horses (three mares and nine geldings) belonged to various breeds and were all born and living in Switzerland. These horses were regularly dewormed and vaccinated.

Examination of circulating Tregs

Peripheral blood mononuclear cells (PBMC) were isolated from healthy horses (n = 12) by a Ficoll-Hypaque procedure as described;⁴³ 4 × 10⁶ freshly isolated PBMC were examined for expression of CD4, CD25 and FoxP3 by intracellular staining using flow cytometry as described below.

Measurement of FoxP3-expressing CD4⁺ CD25⁺ T cells by flow cytometry

Freshly isolated or 4-day-cultured PBMC were labelled with 5 μl mouse anti-equine CD4 (CVS4; Serotec, Düsseldorf, Germany) at 4° for 30 min, followed by donkey anti-mouse immunoglobulin G-fluorescein isothiocyanate (IgG-FITC) (Jackson Immunoresearch Europe Ltd, Suffolk, UK). This was followed by staining with goat anti-human CD25 (R&D Systems, London, UK),⁴¹ or its relevant isotype control antibody (goat IgG; Santa Cruz Biotechnology, Inc., Heidelberg, Germany) at 4° for 30 min. As secondary antibody, donkey anti-goat IgG-allophycocyanin (APC) (Jackson Immunoresearch Europe Ltd) was used. Phosphate-buffered saline (PBS) containing ethylenediamine tetraacetic acid (EDTA) (13·4 mm), gelatine (1%) and sodium azide (0·02%) buffer was used for labelling and washing steps for cell surface staining. Thereafter, cells were fixed and permeabilized using Fix-perm buffer (eBioscience, Hatfield, UK) at 4° for 15 min, followed by washing twice with PBS containing 0·5% Saponin (Sigma-Aldrich, St Louis, MO) and 0·5% BSA (PPA Laboratories GmbH, Pasching, Austria). Whole mouse-IgG molecules (10 μg/ml) were added for 15 min to block any remaining binding sites of the secondary antibody conjugate from CD4 to avoid cross-reaction with anti-mouse FoxP3. This was followed by staining with rat anti-mouse FoxP3 phycoerythrin (PE) (FJK-16s; eBioscience) for 30 min on ice, followed by two further washes in PBS containing 0·5% Saponin and 0·5% bovine serum albumin (BSA). Isotype control used was rat anti-mouse IgG-PE (eBioscience). Washed cells were then resuspended in PBS and measured by flow cytometry using an LSRII fluorescence activated cell sorter (FACS) (Becton-Dickinson, Franklin Lakes, NJ). Analysis was performed using flowjo software (TreeStar Inc., Ashland, OR). A gate was set around lymphocytes and gated cells were analysed for CD4 expression, where another gate was positioned. CD4⁺ cells were subsequently analysed for the expression of CD25 following an approach established in humans and pigs.^44–46 As in those species, CD25⁺ cells do not form a discrete population, but rather tail out of the CD25⁻ cells. Thus, CD25⁻ cells were first distinguished from those with a dim fluorescence signal (CD25^dim), and additionally from those with a distinct (> 10-fold) brighter fluorescence signal (CD25^high). As described,⁴⁴ three gates with small gaps between them were placed, defining CD4⁺ CD25⁻, CD4⁺ CD25^dim and CD4⁺ CD25^high cells. The percentage of FoxP3-expressing cells was determined within each of the three gates (Fig. 1a).

Figure 1.

Determination of forkhead box P3 (FoxP3)-expressing cells in freshly isolated peripheral blood mononuclear cells (PBMC). (a) Gating strategy used for the determination of FoxP3-expressing CD4⁺ CD25⁺ T cells in PBMC from healthy horses. A gate was positioned around lymphocytes and cells within this gate were used for identifying CD4⁺ T cells. Based on the fluorescence signal of the CD25 staining, CD25⁺ cells with the bright fluorescence signal were distinguished as CD25^high from those with a dim fluorescence signal (CD25^dim), whereby the difference between both populations was approximately 10-fold in CD25 expression. Accordingly, three non-overlapping gates were placed for CD4⁺ CD25⁻, CD4⁺ CD25^dim and CD4⁺ CD25^high cells. Keeping the gates wide but separate ensured the inclusion of all relevant cells for analysis, but without minute shifts in fluorescence influencing the results unduly. FoxP3 histogram plots overlaid with the isotype control plot were used to determine the number of FoxP3⁺ cells. (b) Proportion of CD4⁺ CD25⁻, CD4⁺ CD25^dim and CD4⁺ CD25^high T cells and (c) expression of FoxP3 within these three subpopulations in freshly isolated PBMC from healthy horses determined using flow cytometry. Results from all horses examined (n = 12) are presented as box-plots. The centre horizontal line of the box plot marks the median of the sample. The edges of the box mark the first and third quartiles. The whiskers define the upper adjacent value which is the largest observation that is less than or equal to the 75th percentile plus 1·5 times the interquartile range (IQR) and the lower adjacent value, which is the smallest observation that is greater than or equal to the 25th percentile minus 1·5 times IQR. Comparisons were performed using the Kruskal–Wallis multiple-comparison Z-value test with Bonferroni correction. An asterisk indicates statistical significance difference of P ≤ 0·05.

Magnetic cell separation and sorting

PBMC were isolated from healthy horses (n = 5), as described above. CD4⁺ cells were enriched by positive selection using magnetic antibody cell sorting (MACS). PBMC were incubated with mouse anti-equine CD4 (as described above) for 10 min on ice, washed once in ice-cold MACS buffer (PBS containing 2 mm EDTA and 0·5% FCS) and then incubated with rat anti-mouse IgG1 microbeads (Miltenyi Biotec, Bergisch Gladbach, Germany) for 15 min on ice. After a further washing step, cells were resuspended in ice-cold MACS buffer and magnetic cell separation was performed in LS columns using MidiMACS separators (Miltenyi Biotec). After a washing step, the CD4 positively selected cell fractions were stained with donkey anti-mouse IgG-FITC (Jackson) and then stained for CD25 as described above.

Using a BD FACSAria sorter, CD4⁺ CD25⁻, CD4⁺ CD25^dim and CD4⁺ CD25^high lymphocytes were sorted. For this experiment, CD4⁺ CD25^high cells were identified using a more stringent gating (incorporating only the upper 3% of all CD4⁺ T cells with the brightest fluorescence signal for CD25). Part of these sorted cell subtypes were kept at −80° for mRNA analysis. The other part was used further in the experiments to measure the ability of CD4⁺ CD25⁺ cells to suppress proliferation.

T cell proliferation and suppression assay

To study the suppressive function of Treg we used an allogenic mixed leucocyte reaction (MLR). PBMC from a different horse than any included in our study were used as stimulators of an allogenic reaction. Allogenic PBMC were isolated and irradiated (44 Gy) using a Gammacell 40 research irradiator (Department of Clinical Research, University of Bern, Switzerland). In a 96-well plate (Sarstedt, Nümbrecht, Germany), 5 × 10⁴ sorted CD4⁺ CD25⁻ responder cells were cultured alone or with different ratios of autologous CD4⁺ CD25⁺ putative suppressor cells (CD25⁻:CD25^dim or ^high; 1:1, 1:0·5, 1:0·25, 1:0·125) in the presence of 5 × 10⁴ allogenic irradiated PBMC. As a control, the same experiment was performed with CD4⁺ CD25⁻ cells (CD25⁻:CD25⁻; 1:1, 1:0·5, 1:0·25, 1:0·125). T cell proliferation was measured by [3H]thymidine incorporation assay. At day 5, 5 μCi/well [3H]thymidine (Amersham Biosciences, Little Chalfont, UK) was added, followed by harvesting 18 hr later. The amount of incorporated [3H]thymidine was determined by liquid scintillation spectroscopy (Topcount NXT Scintillation, Packard Instrument Company, Ramsey, MN).

Mechanism of Treg suppression

To assess whether CD4⁺ CD25^dim or CD4⁺ CD25^high cells exert their regulatory function through direct cell-to-cell contact or through soluble factors, we performed a series of transwell experiments. Once purified, CD4⁺ CD25^dim or CD4⁺ CD25^high cells were added at a ratio of 0·25:1 to autologous CD4⁺ CD25⁻ cells seeded at 8 × 10⁵ in the lower chamber of a 24-well plate (Corning Incorporated, New York, NY) in the presence of 8 × 10⁵ allogenic irradiated PBMC. CD4⁺ CD25^dim or CD4⁺ CD25^high cells were either added to the lower chambers (in contact) or to the upper chambers (separated by a 0·4-μm pore membrane). Control cultures using CD4⁺ CD25⁻ cells (instead of CD4⁺ CD25^high) cells were performed under identical conditions.

To investigate whether the suppressor function of Treg cells is related to the release of regulatory cytokines, anti-human TGF-β1 (5 μg/ml; R&D Systems), cross-reacting with equine TGF-β1⁴⁷ and anti-equine IL-10 (5 μg/ml; R&D Systems), were added separately or in combination. Respective isotype control antibodies were also used. Antibodies were added to both settings of the transwell experiment.

On day 6, cells from the lower chambers of each well were collected, resuspended in 600 μl fresh medium and transferred into four wells of a 96-well plate (150 μl/well). The cell proliferation was measured by adding 5 μCi/well [3H]thymidine for 18 hr.

In vitro expansion and induction of Treg cells

PBMC (4 × 10⁶) were cultured with a set of stimuli. Concanavalin A (ConA, 5 μg/ml; Sigma-Aldrich), recombinant human IL-2 (rh.IL-2, 100 U/ml; Peprotech, London, UK) and recombinant human TGF-β1 (rh.TGF-β1, 2 ng/ml; Peprotech EC) were used separately or in a combination of all three (cocktail). At day 4, cells were harvested for flow cytometric measurement of FoxP3 expression by different subsets of CD4⁺ cells, as described above.

Furthermore, as shown in Fig. 2, freshly isolated CD4⁺ T cells were sorted into CD4⁺ CD25⁻ (n/0), circulating CD4⁺ CD25^dim (d/0) and circulating CD4⁺ CD25^high (h/0), as described above (magnetic cell separation and sorting). These three populations were cultured separately with cocktail or left unstimulated. Four days later, cultured CD4⁺ CD25⁻ (n/0) cells were stained for CD4 and CD25 and resorted. The proportion of CD4⁺ CD25⁻ (n), induced CD4⁺ CD25^dim (_Id) and induced CD4⁺ CD25^high (_Ih) T cells were determined within each sorted population.

Figure 2.

Strategy of in-vitro induction and expansion of CD4⁺ CD25⁺ T cells. Freshly isolated peripheral blood mononuclear cells (PBMC) from healthy horses were sorted according to their double expression of CD4 and CD25. To enrich maximally for forkhead box P3 (FoxP3)⁺ cells, the CD25^high gate was set to incorporate only the top 3% of all CD4⁺ T cells with the brightest fluorescence signal for CD25. The sorted CD4⁺ CD25⁻ (n/0), CD4⁺ CD25^dim (d/0) and CD4⁺ CD25^high (h/0) cells were cultured with a combination of recombinant human interleukin-2 (rh.IL-2), rh.TGF-β1 and concanavalin A (ConA) (cocktail). Four days later, resorting of the stimulated CD4⁺ CD25⁻ cells was performed and percentages of induced CD4⁺ CD25^dim (_id), CD4⁺ CD25^high (_ih) cells as well as remaining CD25⁻ (n) cells were measured.

Additionally, the day 0 sorted CD4⁺ CD25^dim and ^high cells were stimulated via the cocktail and expanded CD4⁺ CD25^dim (_Ed) and CD4⁺ CD25^high (_Eh) cells were obtained. All sorted and stimulated or expanded populations were used to determine the expression of FoxP3 by flow cytometry or the mRNA expression of IL-10 or TGF-β1 by quantitative reverse transcription–polymerase chain reaction (qRT–PCR). Unstimulated CD4⁺ CD25⁻ (n/–), CD4⁺ CD25^dim (d/–) and CD4⁺ CD25^high (h/–) T cells were used as controls.

The ability of induced CD4⁺ CD25^dim (_Id) and CD4⁺ CD25^high (_Ih) as well as of expanded CD4⁺ CD25^dim (_Ed) and CD4⁺ CD25^high (_Eh) cells to suppress proliferation of freshly isolated CD4⁺ CD25⁻ T cells was determined in a co-culture assay using a ratio of 1:0·25 (CD4⁺ CD25⁻:_Id or _Ih), respectively. The suppression by freshly isolated CD4⁺ CD25^dim or CD4⁺ CD25^high cells was used for comparison.

RNA extraction, reverse transcription of RNA and qRT–PCR

Total RNA was isolated from flow-sorted cells using the RNeasy Plus Mini kit (Qiagen, Hombrechtikon, Switzerland), according to the manufacture's protocol. Sorted cells (5 × 10⁵) were lysed using 350 μl guanidine–isothiocyanate-containing buffer to inactivate RNases to ensure isolation of intact RNA. Lysed cells were then stored at −80° until they were processed for total RNA isolation. Cell lysates were homogenized using Qiashreder spin column and passed through a gDNA eliminator spin column, which allows efficient removal of genomic DNA. Ethanol was added to the flow-through to provide appropriate binding conditions for RNA and the sample was then applied to an RNeasy spin column, where total RNA binds to the membrane and contaminants are efficiently washed away. RNA was then eluted in 50 μl RNase-free water. The RNA concentration was measured by a spectrophotometer (Nanodrop, Wilmington, DE). Total RNA (500 ng) was incubated for 50 min at 42° with the reverse transcription (RT) enzyme Superscript II (RNase H-reverse transcriptase; Invitrogen, Paisley, UK) generating the complementary DNA (cDNA) used for qRT–PCR. Subsequently, mRNA coding for FoxP3, IL-10 and TGF-β1 was quantified by TaqMan® assay using primers and probes previously described.⁴⁸ Duplicate cDNA samples were amplified for 45 cycles (15 seconds at 95° and 60 seconds at 60°) in an AB7300 real-time PCR instrument (Applied Biosystems, Rotkreuz, Switzerland) using the TaqMan® Universal PCR Master Mix (Applied Biosystems). Control samples of RNA were incubated as described above, but in the absence of the RT enzyme in order to assure that genomic DNA was not co-amplified. FoxP3 and or cytokine gene expression were normalized against the 18s rRNA content (TaqMan® ribosomal RNA control reagents; Applied Biosystems) and 1/2^DCt values were calculated as described.⁴³ The 1/2^DCt expresses the relative quantity of specific cytokine mRNA molecules present in the samples compared to the number of the 18s rRNA molecules in the sample. Assay optimization and validation was carried out using dilutions of positive controls. Cycle threshold (C_t) values were determined and the linearity of detection confirmed to achieve a correlation coefficient of > 0·99 over the detection range.

Statistical analyses

The NCSS 2007 software program (NCSS, Kaysville, Utah 84037) was used for statistical analyses. When the data were not normally distributed, non-parametric tests were chosen. This was the case for the data presented in Figs 1 and 5, where the non-parametric analysis of variance (anova) Kruskal–Wallis multiple-comparison Z-value test was used to analyse the frequency of the three CD4⁺ T cell subsets (CD4⁺ CD25⁻, CD4⁺ CD25^dim, CD4⁺ CD25^high) (Figs 1b and 5a) and the expression of FoxP3 by different CD4⁺ T cell subsets (Figs 1c and 5b). To compare the different culture conditions to mock conditions, the non-parametric Wilcoxon signed-rank paired t-test was used (Fig. 5a,b).

Figure 5.

Proportion of CD4⁺ CD25⁻ (white box-plots), CD4⁺ CD25^dim (grey box-plots) and CD4⁺ CD25^high (black box-plots) T cells (a) and expression of forkhead box P3 (FoxP3) within these three subpopulations (b) in peripheral blood mononuclear cells (PBMC) from healthy horses after stimulation as measured by flow cytometry. PBMC from healthy horses (n = 12) were either left unstimulated (mock) or stimulated with recombinant human interleukin-2 (rh.IL-2) or rh.transforming growth factor-β1 (TGF-β1) separately or combined or concanavalin A (ConA) alone or the three combined (cocktail). Except for the cocktail, the culture conditions did not significantly alter the composition of CD4⁺ subsets. However, while all culture conditions changed the expression of FoxP3, only the cocktail significantly increased FoxP3 expression in all populations. Box-plots show the data of the 12 horses (for definition of the box-plots and whiskers, see legend of Fig. 1). Comparison between the three different subsets was performed using the Kruskal–Wallis multiple-comparison Z-value test with Bonferroni correction. *n: denotes a significant difference from CD4⁺ CD25⁻ cells within the respective culture conditions; *d indicates a significant difference from CD4⁺ CD25^dim; *mock indicates a significant difference with reference to unstimulated cells (mock). Comparisons were performed using the non-parametric Wilcoxon signed-rank test. P ≤ 0·05.

One-way anova was used to compare proliferation of the different CD4⁺ subpopulations (Fig. 3b) and to compare counts per minute (CPM) values between the different cultures (Fig. 4). This test was also used to compare the differences in FoxP3-expressing cells (Fig. 6) and to compare levels of IL-10 and TGF-β1 mRNA (Fig. 8) between all T cell populations. The Bonferroni test was used to correct for multiple comparisons.

Figure 3.

(a) Suppression of proliferation of CD4⁺ CD25⁻ cells by CD4⁺ CD25^high or CD4⁺ CD25^dim T cells from healthy horses. CD4⁺ CD25⁻ sorted lymphocytes from healthy horses (n = 4) were cultured with irradiated allogenic peripheral blood mononuclear cells (PBMC) alone (CD25n) or in the presence of CD4⁺ CD25^dim (CD25n + CD25d) or CD4⁺ CD25^high (CD25n + CD25h) cells in different ratios (CD25n: CD25^dim or ^high), ranging from 1:1 to 1:0.125. As a control, the same experiment was performed using CD4⁺ CD25⁻ cells (CD25n + CD25n) in different ratios (CD25n: CD25n) ranging from 1:1 to 1:0·125. Proliferation was measured using [3H]thymidine incorporation assay. Results are shown as counts per minute (CPM). Bars represent mean values of the four horses examined + standard error (SE). *P ≤ 0·05 in the paired t-test indicate significant differences from CD4⁺ CD25⁻ (CD25n). (b) Proliferation of the three subsets of sorted CD4⁺ T cells in mixed lymphocyte reaction (MLR); 5 × 10⁴ CD4⁺ CD25⁻, CD4⁺ CD25^dim or CD4⁺ CD25^high sorted lymphocytes from healthy horses (n = 4) were cultured with 5 × 10⁴ irradiated allogenic PBMC. Proliferation was measured using [3H]thymidine incorporation assay. Results are shown as CPM. Bars represent mean values of the four horses examined + SE. Statistical analysis was performed using one-way analysis of variance (anova) with Bonferroni correction, *P ≤ 0·05 was considered statistically significant.

Figure 4.

Mechanism of suppression by CD4⁺ CD25^high T cells. CD4⁺ CD25⁻ sorted lymphocytes from healthy horses (n = 4) were cultured with irradiated allogenic peripheral blood mononuclear cells (PBMC) alone (CD25n) or in the presence of CD4⁺ CD25^high cells (co-culture) in a ratio of (1:0·25). To test the involvement of cell–cell contact in the mechanism of suppression by CD4⁺ CD25^high cells, the co-cultures were cultivated in the bottom of the well (CD25n + CD25h) or separated by a Transwell insert (CD25n/CD25h). Contribution of interleukin (IL)-10 and transforming growth factor (TGF-β1) was examined by the addition of their neutralizing antibodies to the co-culture (CD25n + CD25h + antibodies) or to the upper Transwell chamber (CD25n/CD25h + antibodies). Proliferation was measured using [3H]thymidine incorporation assay. Results are shown as count per minute (CPM). Bars represent mean values of the four horses + standard error (SE). *n, *c indicate a significant difference from CD25n or to (CD25n + CD25h), respectively. Statistical analysis was performed using one-way analysis of variance (anova) with Bonferroni correction. *P ≤ 0·05 was considered significant.

Figure 6.

Percentage of forkhead box P3 (FoxP3)-expressing cells measured by flow cytometry, in circulating and induced or expanded CD4⁺ CD25⁺ T cell populations. Cells were obtained from three healthy horses according to the protocol outlined in Fig. 2. Freshly isolated sorted CD4⁺ CD25⁻ (n/0) cells could be converted into CD4⁺ CD25^{dim or high} (_id and _ih) cells expressing FoxP3 upon stimulation by cocktail. Furthermore, CD4⁺ CD25⁺ FoxP3⁺ cells could be expanded (_Ed and _Eh) in both CD4⁺ CD25^dim (d/0) and CD4⁺ CD25^high (h/0) populations. (n/–) refers to cultured unstimulated CD4⁺ CD25. Bars represent mean values of the examined three horses + standard error (SE). Statistical analysis was performed using one-way analysis of variance (anova) with Bonferroni correction. *P ≤ 0·05 was considered significant.

Figure 8.

Expression of interleukin-10 (IL-10) (a) and transforming growth factor (TGF-β1) (b) mRNA by circulating, induced and expanded equine CD4⁺ CD25⁺ T cells as measured by quantitative reverse transcription–polymerase chain reaction (qRT-PCR). Flow-sorted cells obtained as described in Fig. 2 from healthy horses (n = 3) were subjected to relative quantification of cytokine mRNA levels normalized against 18s. While the expression of both IL-10 and TGF-β1 was significantly increased in induced CD25^high T cells (_ih, significantly higher than in all other populations), only IL-10 expression was significantly increased in expanded CD25^high (_Eh) compared to circulating (h/0) or cultured but not stimulated (h/–) CD25^high. Bars represent mean values of the examined three horses + standard error (SE). Statistical analysis was performed using one-way analysis of variance (anova) with Bonferroni correction. *P ≤ 0·05 was considered significant.

The paired t-test was used to compare CPM values (Figs 3a and 7) between CD4⁺ CD25⁻ T cells cultured in the absence and the presence of CD4⁺ CD25^dim or CD4⁺ CD25^high cells. On each occasion, P values ≤ 0·05 were regarded as significant.

Figure 7.

Ability of induced and expanded CD4⁺ CD25^dim and CD4⁺ CD25^high T cells to suppress proliferation of freshly isolated CD4⁺ CD25⁻ (CD25n) T cells. CD4⁺ CD25⁻ T cells sorted from freshly isolated peripheral blood mononuclear cells (PBMC) from three healthy horses were cultured with irradiated allogenic PBMC alone (CD25n) or in the presence of different cell populations. The ratio used (CD25n:CD25^{dim or high}) was 1:0·25. Freshly isolated CD4⁺ CD25⁻ cells were added as a control (CD25n + CD25n); CD4⁺ CD25^dim, freshly isolated (CD25n + CD25d), induced (CD25n + CD25_id) or expanded (CD25n + CD25_Ed) or CD4⁺ CD25^high, freshly isolated (CD25n + CD25h), induced (CD25n + CD25_ih) or expanded (CD25n + CD25_Eh) were added as a putative suppressor cells. Proliferation was measured using [3H]thymidine incorporation assay. Results are shown as count per minute (CPM). In accordance with Fig. 3a, CD4⁺ CD25⁻ cells have an increased proliferation and freshly isolated CD4⁺ CD25^dim at a ratio of 1:0·25 did not suffice to suppress proliferation. Bars represent mean values of the examined three horses + standard error (SE). *P ≤ 0·05 in the paired t-test indicate differences from CD4⁺ CD25⁻ (CD25n).

Results

Presence of CD4⁺ CD25⁺ cells expressing FoxP3 in equine PBMC

Three-colour flow cytometry allowed the detection of circulating equine FoxP3⁺ within CD4⁺ CD25⁺ T cells in freshly isolated PBMC from healthy horses (n = 12), as exemplified for one horse in Fig. 1a. The CD4⁺ CD25⁻ cells (median, range; 71, 58–78%) represent the major subpopulation of T cells compared to CD4⁺ CD25^dim (14, 8–16%) and CD4⁺ CD25^high (7, 4–9%) T cell subpopulations (Fig. 1b). Conversely, Fig. 1c shows that the significantly highest percentage of FoxP3-expressing cells was detected in the CD4⁺ CD25^high (8, 3–16%) followed by a small proportion of FoxP3⁺ cells in the CD4⁺ CD25^dim (2, 0·7–4%) T cell subpopulations. In contrast to the CD4⁺ CD25⁺ cells, CD4⁺ CD25⁻ (0·5, 0·1–1·3%) T cells showed only a marginal expression of FoxP3.

Circulating CD4⁺ CD25⁺ T cells have a suppressive capability

Because we have shown that a population of peripheral blood CD4⁺ CD25⁺ cells co-express FoxP3, it was of interest to determine whether these cells can suppress proliferation of CD4⁺ CD25⁻ T cells. Figure 3a displays the results of allogenic MLRs without or with the addition of sorted CD4⁺ CD25^dim or CD4⁺ CD25^high cells. The CD4⁺ CD25⁻ cells proliferated in response to allogenic PBMC [mean CPM ± standard error (SE), n = 4; 12 034 ± 1962] and the addition of further CD4⁺ CD25⁻ cells resulted in a linear increase in proliferation (1:1, 21 157 ± 2358; 1:0·5, 17 677 ± 891; 1:0·25, 13 980 ± 1952; 1:0·125, 11 801 ± 2677), reflecting the increased numbers of cells. Conversely, the MLR was inhibited by the addition of either CD4⁺ CD25^high (1:1, 714 ± 235; 1:0·5, 1628 ± 694; 1:0·25, 2967 ± 1139; 1:0·125, 4392 ± 1176) or CD4⁺ CD25^dim (1:1, 4189 ± 1446; 1:0·5, 6812 ± 1764; 1:0·25, 10 132 ± 1982; 1:0·125, 10 711 ± 3348) T cells in a dose-dependent manner. However, CD4⁺ CD25^high cells suppressed the proliferation far more efficiently than CD4⁺ CD25^dim cells, which were only able to convey a statistically significant suppression at the highest concentration (1:1). Another difference between CD4⁺ CD25^dim and CD4⁺ CD25^high cells was their ability to proliferate in an MLR. While the CD4⁺ CD25^dim (4666 ± 2400) cells proliferated, although to a lower extent than the CD4⁺ CD25⁻ (12 000 ± 2000) cells, the CD4⁺ CD25^high (343 ± 80) cells were anergic (Fig. 3b).

Mechanism of suppression by equine CD4⁺ CD25⁺ T cells

Several non-exclusive mechanisms have been proposed in the human and mouse systems so far.^33,34,37,49 For a functional analysis, we first determined the relevance of contact-dependent mechanism by co-culturing sorted CD4⁺ CD25⁻ together with sorted CD4⁺ CD25^high T cells (CD25n + CD25h) or separated through transwell chambers (CD25n/CD25h). Figure 4 shows that proliferation of CD4⁺ CD25⁻ cells (mean CPM ± SE, n = 4; 13 262 ± 2442) was significantly decreased when they were cultured together with CD4⁺ CD25^high T cells (CD25n + CD25h) (6108 ± 1775). In the presence of the Transwell insert (CD25n/CD25h), the proliferation was still decreased (8879 ± 2105), but less than under close contact.

Based on our previous results⁵⁰ we also investigated a role for cytokines, in particular IL-10 and TGF-β1, in the mechanism of suppression. A pilot experiment had shown increased levels of IL-10 and TGF-β1 mRNA in the MLR co-culture of CD4⁺ CD25⁻ with CD4⁺ CD25^high cells compared to the MLR of CD4⁺ CD25⁻ cells as well as freshly isolated (CD4⁺ CD25⁻ or CD4⁺ CD25^high) cells (data not shown). Indeed, the addition of anti-IL-10 and anti-TGF-β1 to the MLR of CD4⁺ CD25⁻ with CD4⁺ CD25^high T cells reduced the suppressive activity (7978 ± 1419), but did not abrogate it (Fig. 4). It should be noted that the use of individual anti-cytokine antibodies was not evaluated further, after a preliminary experiment showed no appreciable effects when the antibodies were used separately (Fig. S1). Addition of isotype control antibodies made no difference in this assay system (Fig. S1.)

The notion that close contact as well as inhibitory cytokines contribute to the activity of CD4⁺ CD25^high T cells was reflected by the results of using Transwell chambers and neutralizing antibodies in combination (CD25n/CD25h + antibodies) (11 358 ± 1993). Under these conditions, the suppressive activity of the regulatory cells was significantly decreased.

Modulation of FoxP3 expression in equine CD4⁺ lymphocytes

It has been shown that FoxP3 expression can be induced in vitro via TCR activation in combination with IL-2 and TGF-β1.^27,51–53 Based on these findings, we have examined the proportion of FoxP3⁺ cells within both CD4⁺ CD25⁺ and CD4⁺ CD25⁻ populations in PBMC after in vitro stimulation. PBMC were cultured for 4 days with rh.IL-2, rh.TGF-β1 separately, combined or with ConA alone or the three stimuli in combination (cocktail) or left unstimulated (mock) (Fig. 5).

Figure 5a demonstrates a general increase in the proportion of CD4⁺ CD25⁺ T cells upon stimulation of PBMC with IL-2 or ConA separately. This increase occurred specifically in the population of CD4⁺ CD25^dim T cells. Moreover, the cocktail induced a particularly significant increase in CD4⁺ CD25^dim cells. In contrast, there was a decrease in the CD4⁺ CD25^dim upon stimulation by TGF-β1 alone. Furthermore, stimulation by IL-2 alone, in combination with TGF-β1, ConA alone, or the cocktail, significantly increased the percentage of FoxP3-expressing cells by CD4⁺ CD25^dim (Fig. 5b). Notably, stimulation by the cocktail but not by any other stimulus significantly increased the expression of FoxP3 in all T cell populations (Fig. 5b).

No significant change was detected in the proportion of CD4⁺ CD25^high T cells with any of the stimuli (Fig. 5a). However, the frequency of FoxP3-expressing cells in CD4⁺ CD25^high T cell population was significantly increased by all stimuli compared to mock (Fig. 5b). Accordingly, the CD4⁺ CD25^high T cell subpopulation showed the highest expression of FoxP3 regardless of stimulation (Fig. 5b).

CD4⁺ CD25⁻ T cells can be converted into FoxP3⁺ CD4⁺ CD25⁺ iTreg

Our results have shown an increased percentage of FoxP3 expressing CD4⁺ CD25⁺ T cells after stimulation with the cocktail. This raised the question of whether this culture condition expanded the existing FoxP3⁺ CD25⁺ cells only or could also induce them from CD25⁻ cells.

We have therefore examined the ability of sorted CD4⁺ CD25⁻ T cells to be converted into CD4⁺ CD25⁺ T cells expressing FoxP3. Stimulation of CD4⁺ CD25⁻ T cells by the cocktail resulted in the induction of CD4⁺ CD25^dim (mean ± SE, n = 3; _Id, 18 ± 2%) and CD4⁺ CD25^high (_Ih, 5·7 ± 1·5%) T cells (data not shown). As shown in Fig. 6, a significant percentage of the induced CD4⁺ CD25⁺ (mean ± SE, n = 3; _Id, 13 ± 4% and _Ih, 51 ± 1·6%) population also express FoxP3 compared to CD4⁺ CD25⁻ [freshly isolated (n/0, 0·3 ± 0·05%), unstimulated (n/–, 0·6 ± 0·2%) and stimulated (n, 1·6 ± 0·8%)] cells, where FoxP3 expression was hardly detected.

Additionally, we have determined whether this cocktail can expand the number of FoxP3-expressing cells in the freshly isolated CD4⁺ CD25^dim and CD4⁺ CD25^high populations (Fig. 6). Stimulation of these two populations by the cocktail indeed significantly increased the percentage of FoxP3-expressing cells (d/0, 1 ± 0·4% versus _Ed, 18 ± 6% and h/0, 3 ± 0·9% versus _Eh, 31 ± 4%).

Induced CD4⁺ CD25⁺ T cells exhibit a suppressive function

We have also measured the ability of the induced (_Id and _Ih) and expanded (_Ed and _Eh) CD4⁺ CD25⁺ T cells to suppress the proliferation of freshly isolated CD4⁺ CD25⁻ T cells. As controls, freshly isolated CD4⁺ CD25^dim (CD25d) and CD4⁺ CD25^high (CD25h) were included. Figure 7 shows clearly that the addition of the induced (CD25_Ih) as well as the expanded (CD25_Eh) CD4⁺ CD25^high T cells significantly inhibited the proliferation of CD4⁺ CD25⁻ T cells [mean CPM ± SE, n = 3; (CD25n, 36 851 ± 6880); (CD25n + CD25_Ih, 8836 ± 3975); (CD25n + CD25_Eh, 2563 ± 891)]. However, the capacity of suppression was more efficient through freshly isolated CD4⁺ CD25^high (CD25n + CD25h, 766 ± 70) or CD25_Eh T cells than by CD25_Ih. Notably, while freshly isolated CD4⁺ CD25^dim T cells (CD25d) had no suppressive effect (CD25n + CD25d, 38 033 ± 4514), both the induced (_Id) (CD25n + CD25_Id, 15 386 ± 285) and expanded (_Ed) (CD25n + CD25_Ed, 15 740 ± 3062) CD4⁺ CD25^dim T cells were able to efficiently suppress proliferation of CD4⁺ CD25⁻ T cells (Fig. 7).

Increased expression of IL-10 and TGF-β1 mRNA by induced and expanded CD4⁺ CD25⁺ T cells

It has been suggested in other species that induced but not natural Treg cells produce IL-10 and TGF-β1.^26,54 We therefore intended to determine whether there are differences in the mRNA expression levels of these two cytokines between circulating, induced and expanded equine CD4⁺ CD25⁺ T cells. As shown in Fig. 8a, IL-10 mRNA is expressed at similar levels by circulating CD4⁺ CD25^{dim or high} (mean ± SE, n = 3; d/0, 1·6E−03 ± 4·8E−05; h/0, 1·2E−03 ± 1·4E−05) and CD4⁺ CD25⁻ (n/0, 1·3E−03 ± 4·9E−05) T cells.

Expansion of natural CD4⁺ CD25^high (h/0) T cells by cocktail (_Eh) significantly increased the level of IL-10 (h/0, 1·2E−03 ± 1·4E−05 versus _Eh, 2·2E−03 ± 1·4E−05). The highest levels of IL-10 mRNA were found to be expressed by induced CD4⁺ CD25^high (_Ih, 3·5E−03 ± 4·0E−04). In contrast, there were no changes in the expression of IL-10 mRNA between circulating (d/0, 1·6E−03 ± 4·8E−05), induced (_Id, 2·0E−03 ± 8·4E−05) and expanded (_Ed, 1·7E−03 ± 9·1E−06) CD4⁺ CD25^dim T cells.

Similarly to IL-10 mRNA expression, no differences were detected in TGF-β1 mRNA levels (Fig. 8b) between circulating CD4⁺ CD25^{dim or high} (d/0, 6·7E−03 ± 5·5E−04; h/0, 4·5E−03 ± 2·2E−04) and CD4⁺ CD25⁻ (5·1E−03 ± 4·6E−04) cells. Conversely to IL-10 mRNA, expansion of CD4⁺ CD25^high T cells did not increase TGF-β1 mRNA expression (h/0, 4·5E−03 ± 2·2E−04 versus _Eh, 6·7E−03 ± 5·5E−04). The highest levels of TGF-β1 mRNA were expressed by induced CD4⁺ CD25^high (_Ih, 2·1E−02 ± 2·1E−04). As for IL-10 mRNA, there were no differences in the expression of TGF-β1 mRNA between circulating (d/0), induced (_Id, 8·2E−03 ± 2·1E−03) and expanded (_Ed, 6·7E−03 ± 1·3E−03) CD4⁺ CD25^dim T cells.

Discussion

Horses are particularly good models for the study of allergic and autoimmune diseases because they develop spontaneous diseases which share many features with those of humans, such as heaves/asthma, insect bite hypersensitivity/atopic eczema or equine recurrent uveitis/autoimmune uveitis.^43,55,56 The discovery of Treg cells in humans highlights the involvement of these cells in controlling the immune response in various autoimmune diseases and allergy.^57–59 So far, studies related to equine Treg cells have been lacking. Therefore, we investigated the presence of circulating equine CD4⁺ CD25⁺ FoxP3⁺ that can be addressed as natural Treg (nTreg) cells and whether it was possible to expand those as well as to induce Treg cells in vitro de novo (iTreg).

One important marker for defining Treg cells in human and other species is FoxP3.^27,42,45,60 Published studies have consistently reported that FoxP3 is expressed in human,^61–66 murine,^21,67–69 porcine⁴⁵ and canine⁷⁰ CD4⁺ CD25⁺ Treg cells. However, these studies also revealed differences between the species,⁷¹ foremost in the amount of FoxP3-positive nTreg within CD4⁺ CD25^high cells. While almost all CD4⁺ CD25⁺ cells in mouse are FoxP3⁺, only CD4⁺ CD25^high cells are FoxP3⁺ in humans⁷² or pigs.^42,45

In this study, we first examined freshly isolated PBMC from healthy horses for the presence of circulating CD4⁺ CD25⁺ FoxP3⁺ T cells which were predominantly detected as a subpopulation of CD4⁺ CD25^high cells (Fig. 1c), as in human or pigs. However, unlike in those species, equine CD4⁺ CD25^high cells are heterogeneous, as only a minority of cells co-express FoxP3⁺. Even under the stringent gating conditions including the highest 3% of CD25-expressing cells, only 40% (median) co-expressed FoxP3 (data not shown). A similar heterogeneity of CD4⁺ CD25^high cells has been described in cattle recently, but a suppressive function has not yet been described.⁷³

Although only a minority of the equine CD4⁺ CD25^high cells express FoxP3, we found that the constitutive expression of FoxP3 correlated with their suppressor function. CD4⁺ CD25^high cells had a strong suppressive activity (Fig. 3a). In contrast, CD4⁺ CD25^dim T cells exhibited a significant effect only at the highest concentration (Fig. 4a), and it should be taken into account that a small proportion of those were also FoxP3⁺. Importantly, CD4⁺ CD25^high cells did not proliferate in an MLR, whereas, CD4⁺ CD25^dim did (Fig. 3b). This is in agreement with the situation in humans and mice, where Treg cells themselves are anergic.^74–76 Taken together, our results demonstrate that equine CD4⁺ CD25^high cells contain nTreg cells similar to those described in human or other species, whereas the CD4⁺ CD25^dim are primarily activated T cells which might contain some regulatory cells, possibly precursors of the suppressor CD25^high cells.

The mechanism of suppression mediated by equine circulating CD4⁺ CD25^high T cells was also investigated. The functional assay showed that the suppressor activity of equine Treg cells (CD4⁺ CD25^high) was significantly decreased when both physical separations of CD4⁺ CD25⁻ target cells from CD4⁺ CD25^high cells and addition of neutralizing anti-IL-10 and anti-TGF-β1 antibodies were combined (Figs 4 and S1).

Nevertheless, our data do not exclude the involvement of further cytokines that need to be addressed in additional studies. The results indicate that the circulating equine CD4⁺ CD25^high population contain both nTreg, which have been implicated to suppress by cell-to-cell contact,^36,77 and iTregs, suppressing through regulatory cytokines.^17,35–37 Similar to the situation in humans, the ontogeny of circulating FoxP3⁺ cells, the plurality of mechanisms used and the relevance of suppressive cytokines need to be further investigated.⁷²

Because the relative contribution of nTreg versus iTreg is still a matter of debate, we considered it important to investigate the possibility of inducing FoxP3⁺ iTreg cells in vitro. Studies in other species have shown that IL-2 and TGF-β1 play crucial roles in the differentiation of FoxP3⁺ regulatory T cells.^27,29,45,52 Starting from unsorted PBMC, IL-2 alone significantly augmented the frequency of FoxP3⁺ cells in both CD4⁺ CD25^high and CD4⁺ CD25^dim T cell subpopulations but not within CD4⁺ CD25⁻ cells (Fig. 5). This, however, was paralleled by an increase in the number of CD4⁺ CD25^dim and ^high T cells, suggesting that IL-2 (alone or in combination with endogenous factors) acted via the conversion of CD25⁻ cells into CD25⁺ cells expressing FoxP3. This is in accordance with data from humans that primary stimulation of T cells can induce Foxp3.

TGF-β1 alone significantly increased FoxP3⁺ cells within CD4⁺ CD25^high T cells, without an increase in the number of CD25⁺ T cells. This is consistent with studies in other models, which demonstrated that TGF-β1 is required for expression of FoxP3.^28,78–80 However, as this cytokine alone reduces proliferation of CD4⁺ T cells⁸¹ and renders them anergic,^28,82 IL-2^29,83,84 and TCR stimulation^53,85–87 are needed to sustain FoxP3 expression and ensure proliferation of CD4⁺ T cells. Consequently, using a cocktail of TGF-β1, IL-2 and ConA, we were able to induce the expression of FoxP3 by CD4⁺ CD25⁺ T cells as well as increase the number of CD4 ⁺ CD25⁺ T cells. Interestingly, we have also detected a significant increase in the expression of FoxP3 by CD4⁺ CD25⁻ T cells compared to unstimulated cells.

Several studies in the human system have reported that CD4⁺ effector T cells express FoxP3 upon activation without becoming suppressive.^85,88,89 Therefore, it was necessary to verify in our system that the increase in equine FoxP3⁺ T cells via the cocktail was associated with a suppressive activity. The induction of CD25^high FoxP3⁺ T cells was accompanied by suppressor capacity similar to that of the natural and expanded CD4⁺ CD25⁻ cells, thus resembling FoxP3⁺ iTreg. Despite the limited ability of circulating equine CD25^dim T cells to suppress proliferation, the induced and expanded CD25^dim T cells were found to be suppressive. This might be due to a higher proportion of FoxP3⁺ cells within induced and expanded CD25^dim cells compared to freshly isolated ones (Fig. 6). Taken together, these findings show the induction of equine Treg cells in vitro and suggest that the increased expression of FoxP3 is not just a consequence of activation.

Finally, we addressed the question of how the expression of regulatory cytokines is affected by the induction or expansion of CD4⁺ CD25⁺ cells from sorted populations. Our results show that freshly isolated CD4⁺ CD25^high T cells express similar levels of IL-10 and TGF-β1 mRNA to CD4⁺ CD25⁻ T cells (Fig. 8). This emphasizes that these cells contain cells resembling nTreg. It is important to note that CD4⁺ CD25^high T cells increased the expression of IL-10 and TGF-β1 mRNA during the functional MLR to test for inhibition (data not shown), which might explain why these cytokines contribute to suppression. Furthermore, expanded CD4⁺ CD25^high cells had an increased expression of IL-10 mRNA. Interestingly, induced CD4⁺ CD25^high cells showed the highest levels of IL-10 and TGF-β1 mRNA (Fig. 8). Triple staining of FoxP3, IL-10 and TGF-β1 in the future could reveal whether these cytokines are produced by FoxP3⁻ or FoxP3⁺ cells and to which extent any of these populations resemble Tr1.

In summary, our results provide conclusive evidence for the existence of CD4⁺ Treg cells in the circulation of horses. As in humans, mice or pigs, freshly isolated equine CD4⁺ CD25^high cells are anergic, constitutively express FoxP3 and have the ability to suppress T cell proliferation. However, equine CD4⁺ CD25⁺ cells underline the heterogeneity between species, as only a minority of cells were FoxP3⁺. Functionally, in the non-clonal, outbred situation of the equine immune system several of the known mechanisms⁷² seem to exist in parallel and further studies are necessary to dissect them at a clonal level.

Next to circulating Treg cells we show for the first time that peripheral equine CD4⁺ CD25⁻FoxP3⁻ T cells can be converted in vitro into a CD4⁺ CD25^dim or ^high FoxP3⁺ iTreg phenotype. The converted cells are anergic themselves and suppress proliferation of freshly isolated CD4⁺ CD25⁻ T cells. Circulating Treg cells can also be expanded in vitro. The ability to induce and expand regulatory cells ex vivo provides the opportunity to design treatment strategies for allergic and autoimmune diseases in horses as a model species.

Disclosures

The authors have no conflicts of interest to disclose.

Supporting Information

Additional Supporting Information may be found in the online version of this article:

Figure S1. Effect of anti-interleukin (IL)-10 or anti-transforming growth factor (TGF)-β1 alone or combined on the suppression of proliferation.

Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.