CCR8⁺FOXp3⁺ T_reg cells as master drivers of immune regulation

Significance

The current study identifies CCR8⁺ regulatory T cells (T_reg cells) as drivers of immunosuppression and provides compelling evidence of a self-feeding mechanism by which, at an autoimmune site, CCL1 produced by FOXp3⁺ T_reg cells upregulates the expression of its own receptor, CCR8, on these cells, and potentiates their in vivo proliferation and suppressive activities as driver T_reg cells. The suppression of ongoing autoimmunity by a stabilized version of the chemokine (CCL1–Ig) highlights the translational potential of these findings.

Abstract

The current study identifies CCR8⁺ regulatory T cells (T_reg cells) as drivers of immunosuppression. We show that in human peripheral blood cells, more than 30% of T_reg up-regulate CCR8 following activation in the presence of CCL1. This interaction induces STAT3-dependent up-regulation of FOXp3, CD39, IL-10, and granzyme B, resulting in enhanced suppressive activity of these cells. Of the four human CCR8 ligands, CCL1 is unique in potentiating T_reg cells. The relevance of these observations has been extended using an experimental model of multiple sclerosis [experimental autoimmune encephalomyelitis, (EAE)] and a stabilized version of mouse CCL1 (CCL1–Ig). First, we identified a self-feeding mechanism by which CCL1 produced by T_reg cells at an autoimmune site up-regulates the expression of its own receptor, CCR8, on these cells. Administration of CCL1–Ig during EAE enhanced the in vivo proliferation of these CCR8⁺ regulatory cells while inducing the expression of CD39, granzyme B, and IL-10, resulting in the efficacious suppression of ongoing EAE. The critical role of the CCL1–CCR8 axis in T_reg cells was further dissected through adoptive transfer studies using CCR8^−/− mice. Collectively, we demonstrate the pivotal role of CCR8⁺ T_reg cells in restraining immunity and highlight the potential clinical implications of this discovery.

Two major populations of CD4⁺ regulatory T cells (T_reg cells), defined by whether they express the forkhead box protein 3 transcription factor (FOXp3), are thought to play a key role in the maintenance of self-tolerance (1–8). Both FOXp3⁺ and FOXp3⁻ subtypes participate in the regulation of inflammatory autoimmunity and in the maintenance of self-tolerance by various mechanisms, including regulating the biological function of effector TH1 and TH17 CD4⁺ T cells (6, 7, 9–12). CD4⁺ FOXp3⁻ regulatory T cells can be categorized as T regulatory-1 cells (Tr1), which primarily produce IL-10 (12–15), and Th3, which express high levels of TGFβ (16).

Chemokines are small (∼8–14 kDa) secreted proteins, structurally similar to cytokines, that regulate cell trafficking through interactions with a subset of seven transmembrane G protein-coupled receptors (GPCRs) (17), and many of them are associated with chemotaxis of leukocytes to inflammatory sites (18–20). Almost 15 y ago we showed that, in addition to their role in chemoattraction, chemokines are also involved in directing effector CD4⁺ T-cell (T_eff) polarization, by showing that the CXCR3 ligand CXCL10 directs the lineage development of TH1 cells (21, 22). More recently we identified two different chemokines that are involved in the lineage development of Tr1 cells (23, 24). The current study focuses on FOXp3⁺ T_reg cells and on the interplay between CCR8 and its ligands.

In mouse, the chemokine receptor CCR8 is expressed principally on T_reg cells and also notably on small fractions of TH2 cells, monocytic cells, and NK cells (25–28), but not TH1 cells (29, 30). A similar expression pattern is seen in humans, in which CCR8 is additionally found on ∼2% of CD8⁺ cells (30). CCR8 is known to be critical for T_reg function. For example, Coghill et al. recently showed in a graft versus host disease (GVHD) model, donor T_reg cells lacking CCR8 were severely impaired in their ability to prevent lethal GVHD (31). However, the underlying mechanisms of such observations remained unclear.

The current study uncovers the mechanistic basis by which the CCR8–CCL1 axis potentiates T_reg cells, its relevance to human biology, and explores the clinical implications of these findings using an experimental autoimmune disease of the central nervous system (CNS) (32).

Results

Of the Four CCR8 Ligands, CCL1 Is Unique in Potentiating the Suppressive Function of Human T_reg Cells.

Human CCR8 has four known ligands: CCL1, CCL8, CCL16, and CCL18 (33). First we examined whether one or more of the four human CCR8 ligands may enhance the suppressive activity of human T_reg cells. Fig. 1 summarizes data obtained from 10 different healthy donors, indicating that of the four human CCR8 ligands, CCL1 was unique in potentiating the suppressive function of human T_reg cells, detected as an enhanced suppression of T_eff proliferation (Fig. 1 A–C). Because CCR8 is considered to be the exclusive receptor for CCL1 (26), we sought to confirm that these effects were achieved via CCR8 signaling. As shown in Fig. 1D, antihuman CCR8 blocking mAb abolished CCL1-mediated effects in this assay.

Fig. 1.

CCL1 selectively potentiates suppressive function in CD4⁺CD25⁺CD127^low T cells. (A–C) CCL1 potentiates the suppressive function of human CD4⁺CD25⁺CD127^low T cells. Freshly isolated human T_reg cells were activated in the presence of each of the known CCR8 ligands used in suppression assay (SI Methods). Proliferation of T_eff cells was detected by incorporation of [³H] thymidine incorporation (results in A are shown as mean of triplicates ± SE), or by CFSE staining of effector T cells (B and C). B shows the results of a representative experiment. C summarizes CFSE results of all 10 healthy donors as a scattered plot (P < 0.001 unpaired Student’s t test). (D) CCL1 directs its function via CCR8. Suppression assays were conducted as described for A above, with or without addition of an anti-CCR8 blocking mAb. Results of one of three independent experiments are shown as mean of triplicates ± SE. Significance was determined by two-tailed unpaired Student’s t test (*P < 0.01).

Independently, each human CCR8 ligand was tested in a fluorometric imaging plate reader (FLIPR) assay (34) configured to detect the induction of intracellular Ca²⁺ flux in response to CCR8 activation in CHO-K1 cells overexpressing human CCR8. As shown in Fig. S1, CCL1 induced a dose-dependent up-regulation of Ca²⁺ flux in response to CCR8 activation, whereas no effect was seen in the presence of CCL8, CCL16, or CCL18. We note that these data differ from a publication showing that CCL18 may also induce Ca²⁺ flux via CCR8 (33).

Fig. S1.

CCL1 induces Ca²⁺ flux in CHO-K1 cells overexpressing human CCR8. Each of the four known human CCR8 ligands was tested for its ability to induce intracellular Ca²⁺ flux in CHO-K1 cells overexpressing human CCR8 by fluorometric imaging plate reader (FLIPR) assay. Results of one of three independent experiments are shown as mean of triplicates ± SE. Significance was determined by two-tailed unpaired Student’s t test (*P < 0.001).

Collectively, these data show that of the four CCR8 ligands, only CCL1 induces Ca²⁺ flux and potentiates the suppressive activity of these cells.

CCL1 Potentiates Human T_reg Cells by Inducing CCR8, FOXp3, CD39, Granzyme B, and IL-10 Expression.

At 36 h postactivation of cultured human T_reg cells that were, or were not, supplemented with CCL1, they were examined for the transcription of various genes known to be associated with the T_reg phenotype (7) by real-time PCR. We observed between 4- and 5-fold increases in the transcription of FOXp3 and CCR8 (P < 0.0001), a 3.7-fold increase in the transcription of CD39 and a 2.5-fold increase in granzyme B and IL-10 (P < 0.01) (Fig. 2A). Results were then confirmed at the protein level by flow cytometry (Fig. 2 B–D). We observed that during in vitro activation in the presence of CCL1, the expression of CCR8 on CD4⁺CD25⁺C127^low T_reg increases from 2.65% to 36% of the cells (Fig. 2B), whereas its expression on CD4⁺CD25⁻C127^high remained at a low level. This finding suggests that T_reg cells (CD4⁺CD25⁺C127^low) preferentially respond to CCL1 by a profound increase in the expression of CCR8.

Fig. 2.

CCL1 potentiates human T_reg cells by inducing CCR8, FOXp3, CD39, granzyme B, and IL-10 expression. (A) CCL1 enhances the transcription of FOXp3, CCR8, CD39, granzyme B, and IL-10 in CD4⁺CD25⁺CD127^low T cells. CCL1 was added to cultured CD4⁺CD25⁺CD127^low T cells as described above, and 36 h later, the relative transcription of various genes was detected. Results are shown as mean of triplicates ± SE. Significance was determined by two-tailed unpaired Student’s t test (*P < 0.001). The results represent one of three different experiments with similar observations. (B) CC1 preferentially induces the expression of CCR8 on CD4⁺CD25⁺CD127^low T-cells CCL1 was added to cultured CD4⁺CD25⁺CD127^low and CD4⁺CD25⁻CD127^low T cells as described above and expression of CCR8 was detected by flow cytometry. A representative plot is shown together with a scatterplot summarizing five different samples of healthy donors. Significance was determined by unpaired Student’s t test (C and D) CCL1 enhances the expression of FOXp3 (C), CD39, granzyme B, and IL-10 (D) in human CD4⁺CD25⁺CD127^low T cells. CCL1 was added to cultured CD4⁺CD25⁺CD127^low T cells as described above, and 36 h later the expression of various gene products was determined by flow cytometry. A representative plot is shown together with a scatterplot summarizing five different samples of healthy donors. Significance was determined by unpaired Student’s t test (B) P < 0.0001, (C) P < 0.01, and (D) P < 0.0001. (E) CCL1 does not convert FOXp3⁻ T cells into T_reg cells: CCL1 was added to cultured CD4⁺CD25⁻ (FOXp3⁻) T cells undergoing anti–CD3-induced activation. TGF-β was used as a positive control for induction of iT_reg cells.

An increased expression of FOXp3 (from 89.6% to 95.7%, P < 0.01) was also observed in these CD4⁺CD25⁺C127^low T cells (Fig. 2C). Further analysis of FOXp3⁺ T cells (Fig. 2D) revealed a significant increase in CD39 (from 4.43% to 16.1%, P < 0.0001), granzyme B (from 3.25% to 18.3%, P < 0.0001), and IL-10 (from 6.26% to 15.4%, P < 0.0001). Further dissection of the differential transcription of these molecules has been conducted in the murine setup and showed preferential early transcription of CCR8 and FOXp3, as discussed later (Fig. S2).

Fig. S2.

CCL1 induces murine T_reg cells via CCR8 in a STAT3-dependent manner. (A) CCL1 selectively induces the suppressive activity of murine FOXp3⁺ T cells: The three murine CCR8 ligands CCL1, CCL16, and CCL18 were examined for the induction of enhanced suppression of T_eff proliferation by FOXp3⁺ T_reg cells (isolated from FOXp3^GFP mice as described above). Results of one of three independent experiments with similar data are presented as mean triplicates ± SE. Significance was determined by two-tailed unpaired Student’s t test (*P < 0.01). (B) The suppressive activity of CCL1 is CCR8 dependent. CCL1 was examined for the induction of enhanced suppression of FOXp3⁺ T_reg cells that were obtained from wild-type or CCR8^−/− mice. In this experiment, cells were separated based on CD4⁺CD25⁺. Results of one of three independent experiments with similar data are presented as mean triplicates ± SE. Significance was determined by two-tailed unpaired Student’s t test (*P < 0.001). (C) CCL1 enhances the transcription of FOXp3, CCR8, CD39, granzyme B, and IL-10 in CD4⁺FOXp3⁺ T cells. CCL1 was added to cultured CD4⁺FOXp3⁺ T cells (isolated from FOXp3–GFP reporter mice) as described in legend to Fig. 1, 16 h (Left) and 36 h (Right); later the relative transcription of various genes was detected. Results are shown as mean of triplicates ± SE. Significance was determined by two-tailed unpaired Student’s t test (*P < 0.001). The results represent one of three different experiments with similar observations. (D) A STAT3 inhibitor reverses the induction of IL-10 and granzyme B by CCL1. Cultured CD4⁺FOXp3⁺ T cells undergoing anti–CD3/anti–CD28-induced activation were supplemented with CCL1 with or without a STAT3 inhibitor. After 72 h, levels of granzyme B and IL-10 were recorded by ELISA. Results of one of three independent experiments with similar data are presented as mean triplicates ± SE. Significance was determined by two-tailed unpaired Student’s t test (*P < 0.001). (E) STAT3 inhibitor reverses the induction of CD39 by CCL1: CD39 expression by cultured mouse CD4⁺FOXp3⁺ T cells was detected by flow cytometry following anti–CD3/anti–CD28-induced activation in the presence of CCL1, with or without addition of a STAT3 inhibitor. A representative plot is shown together with scatterplot summarizing samples of five different experiments. Significance was determined by unpaired Student’s t test (P < 0.001). (F) CCL1 does not convert FOXp3⁻ T cells into FOXp3⁺. FOXP3⁺ T_reg cells were separated from FOXp3⁻ cells from spleen cells of FOXp3^GFP reporter mice and subjected to anti–CD3/anti–CD28-induced activation in the presence of CCL1 and analyzed for FOXp3 expression on CD4⁺ T cells. CCL1 failed in transforming FOXp3⁻ T cells into FOXp3⁺.

A subsequent set of experiments was conducted to determine whether CCL1 might also increase FOXp3 expression in CD4⁺CD25⁻CD127^high T cells (FOXp3⁻) to convert them into FOXp3⁺. We could not find compelling evidence to show that CCL1 may directly convert FOXp3⁻ T cells into FOXp3⁺ (Fig. 2E).

Taken together, these data imply that CCL1 potentiates the suppressive activity of CCR8⁺ T_reg cells by inducing FOXp3, CD39, granzyme B, and IL-10 in these cells without converting FOXp3⁻ cells into FOXp3⁺.

CCL1-Induced Potentiation of T_reg Cells Is STAT3 Dependent.

Fig. 3A shows that CCL1 induces the phosphorylation of STAT3 but none of the other STAT proteins, as determined by phospho-specific detection using flow cytometry (35), and that this phosphorylation is selective to CCR8⁺ cells. To further validate this observation, we assessed CCL1-mediated effects in the presence or absence of a STAT3 inhibitor. As shown in Fig. 3B, flow cytometry analysis confirmed that inhibiting STAT3 prevented the up-regulation of CD39 in FOXp3⁺ T_reg cells (change from 7.25% to 19.8% (CCL1 alone) vs. 13% (CCL1 with inhibitor). Similarly, ELISA detection confirmed abolition of CCL1-induced granzyme B (Fig. 3C) and IL-10 (Fig. 3D) following STAT3 inhibition.

Fig. 3.

CCL1 potentiates human T_reg cells in a STAT3-dependent manner. CCL1 induces STAT3 phosphorylation. (A) STAT phosphorylation was determined in human CD4⁺CD25⁺CD127^low T cells by flow cytometry analyses. CCL1 was added (100 µg/mL) 24 h after CD3- and anti–CD28-induced activation. The data shown represent one of three independent experiments. (B) STAT3 inhibitor reverses the induction of CD39 by CCL1 in FOXp3⁺ T_reg cells. Flow cytometry analyses of the expression of CD39 in cultured CD4⁺CD25⁺CD127^low T cells undergoing anti–CD3- and anti–CD28-induced activation in the presence of CCL1 with or without addition of a STAT3 inhibitor (CP 690550, Santa Cruz Biotechnology, sc-202818, 20 µM). A representative plot is shown together with a scatterplot summarizing five different samples of healthy donors. Significance was determined by unpaired Student’s t test (P < 0.0001). (C and D) STAT3 inhibitor reverses the induction of granzyme B and IL-10 by CCL1. Cultured CD4⁺CD25⁺CD127^low T cells undergoing anti–CD3/anti–CD28-induced activation were supplemented with CCL1 with or without a STAT3 inhibitor. After 72 h, granzyme B and IL-10 levels were recorded by ELISA. Results of one of three independent experiments with similar data are presented as mean triplicates ± SE. Significance was determined by two-tailed unpaired Student’s t test (*P < 0.001).

Collectively, these results show that in human cells, CCL1 potentiates T_reg cells first by inducing the expression of its target receptor on CD4⁺CD25⁺C127^low T cells followed by the induction of STAT3-dependent increase of CD39, granzyme B, and IL-10, which are key drivers of the suppressive function of these cells.

We then sought to determine whether similar CCL1-drived pathways exist in mouse, using murine CD4⁺ T cells. As shown in Fig. S2A, of the three known murine ligands for mouse CCR8 (CCL1, CCL8, and CCL16), CCL1 exclusively enhanced the suppressive activities of mouse T_reg cells (41% increase, P < 0.01) in an ex vivo T_eff suppression assay. The dependence on CCR8 of these effects was confirmed using T cells isolated from CCR8-deficient mice. As shown in Fig. S2B, murine CCL1 enhanced the suppressive activity of T_reg cells from wild-type mice (72% increase, P < 0.001) but had no effect on T_reg cells obtained from CCR8^−/− mice. In experiments corresponding to those in the human system described in Fig. 2, we have similarly shown that mouse CCL1 enhances the transcription of CCR8, FOXp3, granzyme B, CD39, and IL-10 in murine T_reg cells (Fig. S2C). Here we also compared the relative increase of the transcription of these genes after 16 and 36 h, showing that at the earliest time point, the increased transcription of FOXp3 and CCR8 is mostly dominant over granzyme B, CD39, and IL-10. We also verified these results at the protein level and confirmed that blockade of STAT3 abrogates this increase (Fig. S2 D and E).

Finally, we examined whether CCL1 can polarize murine FOXp3⁻ T cells into FOXp3⁺ [i.e., induction of induced T_reg (iT_reg)], as reported for TGFβ (36). Fig. S2F shows that under in vitro conditions, CCL1 does not convert FOXp3⁻ T CD4⁺ cells into FOXp3⁺.

Taken together, these data confirm very close similarity in the mechanistic basis of CCL1-mediated effects on T_reg cells between human and mouse.

An Autocrine Role for the CCL1–CCR8 Axis in Potentiating T_reg Cells at the Autoimmune Site.

To better understand the potential role of CCL1 in autoimmune disease, we sought to investigate the relative transcription of CCL1 in the CNS as a function of disease status in the mouse EAE model. Fig. 4A shows a marked elevation of CCL1 levels after the peak of disease (up to 14-fold increase on day 22). At this time CD4⁺ T cells were isolated from the spinal cord and separated into FOXp3⁺CD4⁺ T cells and FOXp3⁻CD4⁺ T cells. Each subtype was then analyzed by real-time PCR for the relative transcription of CCL1 (normalized to β2M). CCL1 was up-regulated by 13.8-fold in FOXp3⁺ T_reg cells (Fig. 4B), suggesting that these cells were potentially the principal source of CCL1. Comparative analyses showed that within the inflamed CNS, CCL1 is largely transcribed by FOXp3⁺ T_reg cells but not by microglia cells (Fig. 4C).

Fig. 4.

An autocrine role for the CCL1–CCR8 axis in potentiating T_reg cells at the autoimmune site. (A) The kinetics of CCL1 expression at the CNS with EAE disease course: C57BL/6 mice (FOXp3^GFP transgenic) were subjected to active induction of disease. Representative mice (n = 3) were killed at different time points and the relative transcription of CCL1 at the lumbar spinal cord samples was quantitated and normalized to β2M by real-time PCR. (B) At the peak of disease in FOXp3–GFP reporter mice FOXp3⁺CD4⁺ and FOXp3⁻CD4⁺ populations were separated from the lumbar spinal cord. Each subtype was then analyzed by real-time PCR for the relative transcription of CCL1 (normalized to β2M), showing a 13.8-fold increase (P < 0.001) in its transcription by FOXp3⁺ T_reg cells. The results of one of five experiments is shown (black bars) and a summary of all five experiments is shown at Right (P < 0.001 unpaired Student’s t test). (C) At the peak of disease in FOXp3–GFP reporter mice, FOXp3⁺CD4⁺ T cells and microglia cells were separated from the lumbar spinal cord and subjected to PCR analyses of CCL1 normalized by GAPDH. (D) Analyses of the differential expression of CD39, granzyme B, and IL-10 on CCR8⁺ and CCR8⁻ T_reg cells in CD4⁺ T cells isolated from the lumbar spinal cord of EAE mice at the peak of disease. Representative flow cytometry plot is accompanied by scatterplot of five different experiments (unpaired Student’s t test, P < 0.0001).

Next, we examined the relative number of CCR8⁺ cells within the CD4⁺ FOXp3 subset at the inflamed CNS and compared the expression of granzyme B, CD39, and IL-10 in CCR8⁺ FOXP3⁺ and CCR8⁻ FOXP3⁺ CD4⁺ T cells within the CNS (Fig. 4D). We found that at the peak of disease, about 9% of CD4⁺FOXp3⁺ T cells are CCR8⁺ and that these cells preferentially express granzyme B (32.5% vs. 8.9% in CCR8⁻), CD39 (44% vs. 11.5% in CCR8⁻), and IL-10 (11.3% vs. 5.3% in CCR8⁻). Collectively, these data suggest an autocrine role for the CCL1–CCR8 axis in potentiating T_reg cells at the autoimmune site.

CCL1–Ig Suppress Ongoing EAE by Inducing CCR8 and Potentiating T_reg Cells via This Receptor.

We sought to investigate whether administration of CCL1 could affect ongoing EAE. The short in vivo half-life of chemokines limits their exposure following systemic administration. Therefore, using similar strategies for extending half-life to those that we have previously used for CXCL12 (24), CXCL10, and CXCL11 (23), we generated a mouse CCL1–Ig fusion protein (Fig. S3 A and B). This construct retains the biological activities of the native chemokine, including chemoattraction (Fig. S3C) and ERK1/2 phosphorylation of the BW5147 CCR8⁺ thyoma cell line (37) (Fig. S3D). Preliminary pharmacokinetic (PK) analysis in mice confirmed that CCL1 formatted as a Fc-fusion protein had an acceptable exposure and half-life to enable its use in disease models (Fig. S3E). Pharmacodynamic (PD) analysis confirmed that it retained biological activity in vivo, achieving a durable induction of FOXP3⁺CD39⁺ T_reg cells both in the spleen and in the spinal cord (Fig. S4). We were therefore able to investigate the potential therapeutic effect of CCL1–Ig on ongoing EAE. Fig. 5A summarizes data from one of three independent experiments with similar observations, showing that administration of CCL1–Ig during ongoing EAE rapidly suppressed the development and progression of disease (day 21 mean maximal score of 1 ± 0.13 compared with 2.5 ± 0.23, P < 0.01). Clinical observations were confirmed histologically (Fig. 5B, mean histological score 0.5 ± 0.1 compared with 2.6 ± 0.3 in and 2.5 ± 0.3 in control groups, P < 0.01).

Fig. S3.

Generation of CCL1–Ig and its in vivo pharmacokinetics. (A) Schematic view of the CCL1–Ig construct. (B) Western blot analysis of CCL1–Ig fusion protein. (C) CCL1–Ig maintains its ability to attract CCR8-expressing cells. Chemotaxis assay using Transwell containing CCL1 and CCL1–Ig showing that CCL1–Ig preserves the chemotactic properties of CCL1. (D) CCL1–Ig phosphorylates ERK1/2 in bw5147 thymoma cells. The 5 × 10⁶ bw5147 cells were incubated in DMEM 5% with 250 ng CCL1–Ig for different durations and tested for the phosphorylation of erk1/2 in Western blot. (E) PK of CCL1–Ig: C57BL/6 mice were injected with mouse CCL1–Ig (i.p. 100 µg). Blood was drawn at the indicated time points. Plasma extracted and the presence of CCL1–Ig was detected by Western blot using anti-His HRP-conjugated antibodies. IgG light chain served as a loading control.

Fig. S4.

Pharmacodynamics of CCL1–Ig. At the onset of MOG_35–55 induced EAE in FOXp3–GFP C57BL/6 mice that were administered at the onset of disease with CCL1–Ig (i.p. 100 µg/mouse) or control IgG. Spleen and lumbar spinal cord CD4⁺ T cells were subjected to FACS analysis for CD39 and FOXp3 at different time points. Results of five mice per group are shown.

Fig. 5.

CCL1–Ig suppress ongoing EAE. (A) C57BL/6 female mice were injected with MOG_p35-55 to induce active EAE and at the onset of disease (day 12) were separated into groups with comparable disease scores (n = 9 mice per group, from which three were killed at the peak of disease). On days 13, 15, 17, and 19 after the induction of disease mice were injected (i.p.) with either PBS, 300 µg per mouse of mCCL1–Ig or IgG isotype control. An observer blind to the experimental protocol monitored the development and progression of disease. The results (n = 9 mice per each group until day 17 and n = 6 from day 17 onward) are shown as the mean maximal score ± SE. The results show one of three independent experiments with similar data. One-way ANOVA for paired data was used to determine the significance of the time–response curves (*P < 0.01). The arrows indicate the days of mCCL1–Ig or IgG administration. (B) Histopathological evaluation: At the peak of disease (day 17), three representative mice per group were killed and lumbar spinal cord was subjected to histological analysis (18 sections per spinal cord) using a score of 0–3 as described in ref. 63 (see also SI Methods). The mean histological score ± SE was calculated for each treatment group. Representative histological sections are shown, and a statistical analysis of all sections is also given. Significance was determined by two-tailed unpaired Student’s’s t test (*P < 0.001).

Protective administration of CCL1–Ig led to a significant increase in the relative number of FOXp3⁺ T_reg cells both in the periphery (spleen) and the CNS, and of the relative expression of CCR8 on these cells (Fig. S5A). As the relative increase in FOXp3⁺ T cells is systemic, it is not likely that it is due to differential migration of cells, but rather increased expression of FOXp3 and/or increased in vivo proliferation of FOXp3⁺ T cells. We also tested 5-bromo-2′-deoxyuridine (BrdU) uptake in this in vivo model (Fig. S5B) showing a significant increase in the proliferative response of FOXp3⁺ T cells at the CNS and spleen of CCL1–Ig-treated mice (P < 0.01) combined with reduced proliferation of FOXp3⁻ CD4⁺ T cells (P < 0.01), which could be due to the increased suppressive effect of FOXp3⁺ T_reg cells. Furthermore, the relative expression of CD39, granzyme B, and IL-10 significantly increased in these CCR8⁺ T cells following CCL1–Ig administration (Fig. S5C, P < 0.001). Collectively, these data suggest that CCL1–Ig suppresses ongoing EAE, in part, by inducing the proliferative response of T_reg cells and potentiating their activity via CCR8, which is induced by itself via CCL1.

Fig. S5.

Therapy with CCL1–Ig increases the relative level of T_reg cells expressing high levels of FOXp3, CD39, IL-10, and granzyme B and also their proliferative rate (BrdU uptake). (A) Comparative flow cytometry analyses of T cells isolated from spleen and spinal cords of control and CCL1–Ig-treated mice (day 17). (Left) Staining for FOXp3 vs. CD4. (Right) Gating on CD4⁺ analysis of CCR8 expression on FOXp3⁺ CD4⁺ T cells. In each experiment a representative plot is shown as well as analyses of three independent experiments (unpaired Student’s t test, P < 0.001). (B) BrdU uptake of FOXp3⁺ T cells. T cells isolated from the spinal cords of control and CCL1–Ig-treated mice by flow cytometry (gated on CD4⁻). In each experiment a representative plot is shown as well as analyses of three independent experiments (unpaired Student’s t test, P < 0.001). (C) Analyses of the expression of CD39, granzyme B, and IL-10 by T cells isolated from the spinal cords of control and CCL1–Ig-treated mice by flow cytometry (gated on CD4⁺CCR8⁺). In each experiment a representative plot is shown as well as analyses of three independent experiments (unpaired Student’s t test, P < 0.001).

The in Vivo Suppressive Activity of CCL1 on T_reg Cells Is CCR8 Dependent.

We conducted a set of experiments to further examine the relevance of the interplay between CCL1 and its receptor CCR8 in restraining EAE. The experimental system included CCR8^−/− mice reconstituted with CD4⁺ T_reg cells from either CCR8^−/− or WT mice followed by administration of CCL1–Ig. To optimize the methodology of the adoptive transfer protocol, we performed pilot studies in which mice were administered 5 × 10⁵, 5 × 10⁶ , or 5 × 10⁷ T_reg cells (WT) and monitored for the development and progression of EAE. Administration of cells in the range of 5 × 10⁶–5 × 10⁷ significantly suppressed the disease without the need to inject CCL1–Ig. Therefore, administration of 5 × 10⁵ cells was selected for further studies.

First, we confirmed that CCR8⁺ T_reg cells from FOXp3–GFP reporter mice (FOXp3^GFP) injected i.p. entered the CNS within 48 h (Fig. S6), and that administration of CCL1–Ig to mice lacking CCR8 had no effect on the severity of disease (Fig. S7A), in contrast to its effect on WT mice (Fig. S7B). Importantly, we observed that in CCR8^−/− mice reconstituted with CCR8⁺ T_reg cells (Fig. S7C) administration of CCL1–Ig rapidly suppressed disease (day 21, 1.5 ± 0.166 compared with 3 ± 0.23, P < 0.01), whereas reconstitution with T_reg cells from CCR8^−/− mice was without effect (Fig. S7D). This finding further indicates the pivotal role of the CCL1–CCR8 axis on T_reg cells in the regulation of EAE.

Fig. S6.

CCR8⁺ T cells injected in the periphery enter the CNS. CCR8^−/− EAE mice injected with FOXp3–GFP CD4⁺ T cells (CCR8⁺) at a dose of 5 × 10⁵ per mouse. These cells could then be recorded in the CNS 48 h later (FACS).

Fig. S7.

The in vivo suppressive activity of CCL1 on T_reg cells is CCR8 dependent. C57BL/6 CCR8^−/− female mice (n = 54) were subjected to active induction of MOG_p35-55-induced EAE. One day postinduction, EAE-induced mice were separated into three groups, with or without injection with 5 × 10⁵ CD4⁺CD25⁺ cells from a wild-type donor. At the onset of symptomatic disease (day 11) the mice of each group were separated into three groups based on disease score (n = 6 mice per group). On days 12, 14, 16, and 18 (arrows) after the induction of disease, mice were injected i.p. with either PBS control (squares), 200 μg per mouse of mCCL1–Ig (circles), or IgG isotype control (rectangles). An observer blind to the experimental protocol monitored the development and progression of disease. The results are shown as the mean maximal score ± SE. (A) CCR8^−/− mice, (B) WT mice, (C) CCR8^−/− mice reconstituted with T_reg cells from CCR8^+/+ mice, and (D) CCR8^−/− mice reconstituted with T_reg cells from CCR8^−/− mice. The results show one of three independent experiments with similar data. One-way ANOVA for paired data was used to determine the significance of the time–response curves (*P < 0.01).

Because the mechanism of T_reg potentiation includes up-regulation of IL-10, we sought to address whether increased production of IL-10 in T_reg cells is the principal mechanism of CCL1-induced potentiation of these cells. An in vitro suppression assay comparing T_reg and effector CD4⁺ T cells from WT or IL-10 KO mice revealed that CCL1-mediated potentiation of T_reg cells was achieved even in the absence of IL-10 (Fig. S8A). Subsequently, we used the adoptive transfer model to show that CCL1–Ig effectively suppresses EAE when acting on T_reg cells from IL-10 KO mice (Fig. S8B). Collectively, these data imply CCL1 may potentiate T_reg cells even in the absence of IL-10.

Fig. S8.

CCL1 potentiates T_reg cells from IL-10 deficient mice. (A) CCL1 enhance the in vitro suppressive activity of T_reg cells from IL-10 KO mice. In vitro suppression assay was established as described in the legend to Fig. 4 A and B. CD25⁺ T_reg cells were isolated from either WT or IL-10 KO mice. We show that CCL1 enhances the suppressive activity of T_reg cells from both WT and IL-10 KO mice. (B) CCL1–Ig enhances the in vivo suppressive activity of T_reg cells from IL-10 KO mice. CCR8^−/− mice were administered with 5 × 10⁵ CD4⁺CD25⁺ cells from WT or IL-10^−/− donors. After the onset of disease (days 13, 15, 17, and 18) the mice were injected i.p. with 200 µg of mCCL1–Ig and monitored for the development and progression of disease. Results are shown as mean EAE score ± SE.

Our adoptive transfer experiments (Fig. S7) show that the interaction between CCL1 and CCR8 on T_reg cells is essential for suppressing EAE. However, this observation still does not exclude the possibility that CCL1 may also affect the development and progression of disease via its interaction on other CCR8⁺ cells, such as CCR8⁺ macrophages or natural killer (NK) cells, or other CD4⁺ T cell subsets. Other than T_reg cells, Th2 cells are the only CD4⁺ population that express high levels of CCR8 (38). Th2 cells produce IL-4 to mediate their biological function, so we examined whether CCL1 potentiates IL-4 production in these cells but found no evidence of such an effect (Fig. S9), consistent with the regulatory effect of CCL1 on CD4⁺ T cells being predominantly T_reg dependent.

Fig. S9.

CCL1 does not enhance IL-4 expression by CD4⁺ Th2 T cells. Primary CD4⁺ T cells (CD44^loCD62L^hi) were purified from spleen of BALB/C mice. T cells were stimulated for 2 d by anti-CD3 (10 µg/mL; BioLegend) and anti-CD28 (1 µg/mL; BioLegend) in the presence of IL-2 (2.5 ng/mL), IL-4 (10 ng/mL; PeproTech), and mAb to IFN-γ (5 µg/mL; R4-6A2, BioLegend). Cells were cultured for an additional 3 d without stimulation of the TCR with or without CCL1 (200 ng/mL). Intracellular staining was done for IFN-γ and IL-4 (BioLegend). We show that CCL1 has no effect on IL-4 production by Th2 cells.

Discussion

The current study focuses on the interplay between the chemokine receptor CCR8 and its ligands, particularly CCL1, and its role in the generation and maintenance of active tolerance. We show that of the known CCR8 ligands, CCL1 is unique in its ability to induce Ca²⁺ flux via this receptor and to potentiate the suppressive activities of T_reg cells. We also show that this chemokine up-regulates the expression of its target receptor on these cells (in vitro and in vivo) and by so doing, further induces their suppressive activities in an autocrine loop. Evidence from the EAE model suggests that at the autoimmune site, CCL1 is largely produced by T_reg cells to potentiate their suppressive function in an autocrine loop, further emphasizing the key role of this interaction in the regulation of autoimmunity.

We suggest two complementary pathways by which T_reg cells, via the CCR8–CCL1 interaction, function as driver suppressor cells in vivo: (i) the induction of their proliferation to increase their relative number and (ii) an increased expression of key mediators of suppressive immune regulation. These include: the suppressor cytokine IL-10, which suppresses both innate and specific immune activities (39); granzyme B, which is capable of specifically killing antigen-presenting B cells, thus limiting autoimmunity (40); and CD39, which decreases the extracellular concentration of ATP, and has recently been tightly associated with control of autoimmunity within the CNS (41). Interestingly, we showed that CCL1 could effectively potentiate CCR8⁺ T_reg cells lacking IL-10 (Fig. S7). This result implies that IL-10 could be important, but not critical, for CCL1-induced potentiation of T_reg cells. Future complementary studies using IL-10R KO mice are required to investigate whether there is an alternative cytokine that induces IL-10R–dependent activation of STAT3 in these mice.

Even though CCR8 is expressed by “driver” T_reg cells, its expression is not exclusive to these cells (28, 30, 42–45). Our results (Fig. S7) showing that CCL1–Ig rapidly suppresses disease in mice reconstituted with T_reg cells from wild-type but not CCR8-deficient mice suggest that T_reg cells are the dominant cell type by which CCL1 achieves its beneficial effects. Nevertheless, we do not exclude the possibility that the interaction of CCL1 with other CCR8⁺ cells may also affect the dynamics of disease.

A further question relates to the role of STAT3 in T_reg potentiation. The involvement of STAT3 in the polarization of both Th17 and FOXp3⁺ T_reg cells (46–49) might explain, in part, the plasticity of the Th17/T_reg pathways (10, 50) and its implications in the regulation of immunity, particularly within the gut (51–53). Moreover, in Th17 cells STAT3 phosphorylation is induced by the proinflammatory cytokines IL-6 and IL-23 (48). Taken together, these data imply that STAT3 phosphorylation is induced by different cytokines/chemokines in different cell types and that the biological consequences of this phosphorylation may be cell type specific and/or that alternative signaling pathways, yet to be identified, are required for the CCL1–CCR8-induced potentiation of T_reg cells.

We note, and draw some parallels with, previous reports that low dose IL-2 rapidly induces the in vivo expansion of T_reg cells to suppress type I diabetes in nonobese diabetic mice, even though the in vitro effect of low-dose IL-2 on T_reg expansion was very limited (54, 55). These low-dose IL-2 studies were later extended to clinical trials, e.g., for hepatitis C virus-induced vasculitis (56), and GVHD (57). Indeed, the induction of in vivo expansion of T_reg cells as a therapeutic strategy for autoimmune disease remains an area of intense interest. However, our data suggest that intervention with a CCL1-based therapy may be preferred to IL-2, because the latter has the potential, even at low dose, to activate effector T cells and NK cells and potentially aggravate disease (58).

From the translational perspective, CCL1–Ig could be a preferred candidate for therapy of autoimmune diseases, because CCR8⁺ T_reg insufficiency (functionally and/or numerically) is likely to be an important contributor to a range of diseases (31). Conversely, blockade of the CCR8–CCL1 axis may be used as a strategy to enhance anticancer immunity (59).

Methods

Animals.

The 6-wk-old female C57BL/6 mice were purchased from Harlan and maintained under specific pathogen-free conditions in our animal facility. FOXp3^GFP mice (internal ribosome entry site–GFP knocked in to the FOXp3 locus, on the C57BL/6 background) were kindly provided by Vijay Kuchroo, Harvard Medical School, Boston, MA. The generation of CCR8^−/− mice has been previously described (by S.A.L.) (29).

The use of animals and experimental protocols were approved by the Animal Care and Use Committee of the Technion.

Antibodies, Cytokines, and Chemokines.

Anti-mouse CCL1 neutralizing antibody (AF845), hTGF-β and hIL-2, and all recombinant chemokines (human and mouse) were purchased from R&D Systems.

Human Samples.

All human samples were purchased from the Israel Blood Banks. All human biological samples were sourced ethically and their research use was in accordance with the terms of the informed consents.

In Vitro Proliferation Assays.

T-cell proliferation was determined either by thymidine incorporation or by carboxyfluorescein succinimidyl ester (CFSE) staining. CFSE labeling studies used a CFSE Cell Division Tracker Kit (423801 BioLegend) according to the manufacturer’s protocol and only CD4⁺ T_eff cells were labeled.

Cell Separation and Suppression Assays.

The basic protocol for the mixed lymphocyte suppression assay was conducted according to Collison and Vignali (60). The detailed protocols for murine and human are specified in SI Methods.

Measurement of Intracellular Calcium Mobilization.

A Fluorometric Imaging Plate Reader (FLIPR, Molecular Devices) was used to detect calcium flux. Data were analyzed using GraphPad Prism (v5) as specified in detail in SI Methods.

Phospho-Specific Flow Cytometry and STAT Inhibitors.

Phospho-specific flow cytometry was conducted according to ref. 35. The biological relevance was verified by using STAT-specific inhibitors. All protocols are specified in detail in SI Methods.

Cytokine Measurement by ELISA.

Methods of cytokine measurement by ELISA are specified in SI Methods.

Real-Time PCR Primers.

Construction of mIgG plasmid and CCL1–Ig cloning are specified in detail in SI Methods.

Expression and Purification of Fusion Proteins.

Fusion proteins were expressed and purified using CHO dhfr^−/− (DG44) cells (provided by L. Chasin, Columbia University, New York, NY) according to the method described in detail in ref. 61. The protocols are specified in SI Methods.

Induction of Acute and Semichronic EAE.

Studies were conducted according to ref. 62 and are further detailed in SI Methods.

BrdU Uptake.

BrdU (Sigma) was added to the drinking water (1 mg/mL) for 14 d according to the manufacturer’s protocol. BrdU uptake was conducted by flow cytometry using anti-Brdu mAb (BioLegend, clone Bu20a).

Histopathology.

Histopathology was conducted as we previously described (63) and is explained in details in SI Methods.

Statistical Analysis.

Statistical analysis was done according to the recommendations provided by Nature for reporting life sciences research and are specified in SI Methods.

SI Methods

Animals.

The 6-wk-old female C57BL/6 mice were purchased from Harlan and maintained under specific pathogen-free conditions in our animal facility. These mice are developed from breeders obtained from The Jackson Laboratory. FOXp3^gfp mice (internal ribosome entry site–GFP knocked in to the FOXp3 locus, on the C57BL/6 background) were kindly provided by Vijay Kuchroo (Harvard Medical School, Boston, MA). CCR8^−/− mice were previously generated by one of us (S.A.L.) (29).

Antibodies, Cytokines, and Chemokines.

Anti-mouse CCL1 neutralizing antibody (AF845), human (h)TGF-β and hIL-2, and all recombinant chemokines (human and mouse) were purchased from R&D Systems.

Peptides.

Myelin oligodendrocyte glycoprotein (MOG)_(p35–55) was constructed by the Beckman Center, Stanford University. After purification by HPLC, the sequence was confirmed by amino acid analysis and the correct mass was confirmed by mass spectroscopy. Purity of the peptide used in the current study was >95%.

In Vitro Proliferation Assays.

T-cell proliferation was determined either by thymidine incorporation or by carboxyfluorescein succinimidyl ester (CFSE) staining. For thymidine incorporation, each well was pulsed with 2 µCi of thymidine (specific activity 10 Ci/mmol) for the final 16 h of the assay followed by harvesting onto fiberglass filters. Results are shown as the mean counts per minute of six replicates ± SE from three independent experiments, divided by the mean counts per minute of control cells. CFSE labeling studies used a CFSE Cell Division Tracker Kit (423801 BioLegend) according to the manufacturer’s protocol and only CD4⁺ T_eff cells were labeled. Experiments were conducted in three replicates and a flow cytometry diagram of representative replicate is presented.

Cell Separation and Suppression Assays.

The basic protocol for the mixed lymphocyte suppression assay was conducted according to Collison et al. (60) as follows:

Human.

CD4⁺ T cells were separated from blood samples using CD4⁺ T-cell separation kit (Stem Cell). T_eff (CD4⁺CD127⁺CD25⁻) and T_reg cells (CD4⁺CD127⁻CD25⁺) were separated using a BD FACSAria cell sorter after staining with anti-CD25 and anti-CD127 mAbs, (BioLegend). Each cell subtype was subjected to anti-CD3 and anti-CD28 activation (1 μg/mL, BioLegend) supplemented with IL-2 (500 units/mL, R&D Systems) and APC in a 1:1 ratio (5 × 10⁴ of each cell type). Recombinant chemokines (CCL1, CCL8, CCL16, or CCL18) were added to CD4⁺CD127⁻CD25⁺ T_reg cells (200 ng/mL, R&D Systems). After 24 h, T_reg cells were collected, washed, and added to the cultured T_eff cells (ratio of 1:0.5) for an additional 72 h. At this stage cultures were also supplement with IL-2 (500 units/mL, R&D Systems). Effector cells were labeled with CFSE (5 μM). To reciprocal wells, tritiated thymidine was added for the last 16 h of a 96-h activation and the samples were counted in a beta scintillation counter.

Murine.

For purification of antigen presenting cells (APCs) mouse spleens were digested for 30 min at 37 °C in the presence of collagenase IV (Sigma) and DNase (2 μg/mL) (Sigma) in complete medium (modified RPMI 1640 supplemented by 10% FBS (HyClone), 50 μM 2-ME (Sigma-Aldrich), 1% sodium pyruvate, 1% nonessential amino acids, 1% Hepes, 100 units/mL penicillin, 100 μg/mL streptomycin, and 2 mM L-glutamine (all from Invitrogen). CD4⁺ T cells were isolated from whole spleens and lymph nodes using Easysep CD4⁺ enrichment magnetic beads (Stemcell). For the separation of CD4⁺FOXp3⁺ cells from spleens of FOXp3^GFP reporter mice, CD4⁺-enriched T cells were sorted using BD FACSAria cell sorter based on the expression of GFP (purity > 96%). This method was used in in vitro and ex vivo suppression assays, adoptive transfer studies, and transcription analysis of T_reg cells. Mouse suppression assays were conducted as described for human.

Measurement of Intracellular Calcium Mobilization.

CHO-K1 cells were stably transfected with expression plasmids encoding human CCR8 and Gqi5 such that CCR8 agonism could be detected by intracellular Ca2⁺ flux. Cells were cultured in Ham’s F-12 medium supplemented with 10% heat-inactivated FBS, geneticin, glutamax, and hygromycin. Twenty-four hours before testing, cells were seeded in black 96-well tissue culture microplates (Greiner Bio-one) at 4.0 × 105 cells per well.

FLIPR Calcium 3 Dye (Express Kit, Molecular Devices) was diluted in FLIPR assay buffer [Hank’s Balanced Salt Solution (HBSS, Sigma) supplemented with Hepes and sodium bicarbonate solution 7.5%] to generate the dye loading buffer. Culture medium was dispensed with care from preseeded cells and replaced with dye loading buffer and plates were incubated for 1 h. Test molecules were prepared in FLIPR assay buffer supplemented with 0.1% BSA and 2.5 mM probenecid, adjusted to pH 7.4, and added to cells in triplicate wells to achieve a concentration gradient (maximum final concentration 300 nM). Intracellular Ca2⁺ flux in response to receptor activation was quantitated using a Fluorometric Imaging Plate Reader (FLIPR, Molecular Devices). Data were analyzed using GraphPad Prism (v5).

Phospho-Specific Flow Cytometry and STAT Inhibitors.

Human CD4⁺CD25⁺CD127^low T_reg cells or murine FOXp3⁺(GFP⁺) cells were sorted using flow cytometry as described above and activated for 24 h with anti-CD3/ and anti-CD28 antibodies. Medium was replaced and cells were rested for 2 h and restimulated using CCL1 (200 ng/mL) for several time points. Each time point was tested for the phosphorylation of STAT1, 3, 4, 5, and STAT6. Cells were also stimulated with appropriate cytokines as a positive control (for example STAT3 phosphorylation was tested following treatment with IL-6 for 20 min). Cells were subsequently fixed using PFA 1% for 10 min and permeabilized using cold methanol for 10 min. Cells were washed and stained with phospho-specific antibodies. To examine the biological significance STAT3 phosphorylation, T_reg cells were incubated for 3 d and treated with CCL1 and/or the STAT3 inhibitor Stattic (100 nM) as appropriate, before testing for the protein expression of CD39, granzyme B, and IL-10.

Single-cell suspensions were stained using the following conjugated antibodies according to the manufacturer’s protocol: anti-mouse CD4 (GK1.5); anti-mouse IFN-γ (XMG1.2); anti-mouse IL-17 (TC11-18H10.1); anti-mouse CD39(5F2); anti-mouse IL-10(JES5-16E3); anti-mouse CTLA4(UC10-4B9) (all from Biolegend); anti-mouse granzyme B (NGZB, eBioscience); and anti-mouse CCR8 (polyclonal, Novus Biologicals).

Human cells were stained using: anti-human CD39 (A1); anti-human IL-10 (JES3-9D7); anti-human CTLA4 (L3D10) (all from Biolegend); anti-mouse granzyme B (B18.1, NOVUS); and anti-human CCR8 (L264G8).

Real-Time PCR Primers Used for Quantitating Gene Transcription in T_reg Cells.

The real-time PCR primers used for quantitating gene transcription in T_reg cells are as follows:

• GAPDH: F: 5′-ACCACAGTCCATGCCATCAC-3′

• R: 5′-TCCACCACCCTGTTGCTGTA-3′;

• FOXp3: F: 5′-GCACCTTCCCAAATCCCAGT-3′

• R: 5′-GCAGGCAAGACAGTGGAAACC-3′

• CTLA-4: F: 5′-CAGTGGAAATCAAGTGAACCTCAC-3′

• R: 5′-GCACGGTTCTGGATCAATTACA-3′

• CD25: F: 5′-CGCAGAATAAAAAGCGGGTCA-3′

• R: 5′-ACTTGTTTCGTTGTGTTCCGA-3′

• CCL1: F: 5′-CTCATTTGCGGAGCAAGAGAT-3′

• R: 5′-GCCTCTGAACCCATCCAACTG-3′

• CCR8: F: 5′-GTGTGACAACAGTGACCGACT-3′

• R: 5′-CTTCTTGCAGACCACAAGGAC-3′

• granzyme B: F: 5′-CCCTGGGAAAACACTCACACA-3′

• R: 5′-GCACAACTCAATGGTACTGTCG-3′

• CD39: F: 5′-AGGTGCCTATGGCTGGATTAC-3′

• R: 5′-CCAAAGCTCCAAAGGTTTCCT-3′

• GALECTIN-1: F: 5′-TCGCCAGCAACCTGAATCTC-3′

• R: 5′-GCACGAAGCTCTTAGCGTCA-3′

• IL-10: F: 5′-ACCTGCCTAACATGCTTCGAG-3′

• R: 5′-CC AGCTGATCCTTCATTTGAAAG-3′

• TGF-β 1: F: 5′-ATTGCTTCAGCTCCACGGA-3′

• R: 5′-CCCGGGTTATGCTGGTTGTAC-3′.

Construction of mIgG Plasmid and CCL1–Ig Cloning.

cDNA encoding the constant region (Hinge-CH2-CH3) of mouse IgG1 Fc was generated by RT-PCR using RNA extracted from mouse spleen cells that were cultured for 4 d with LPS and IL-4. The primers used for this reaction were: sense 5′-ctcgagGTGCCCAGGGATTGTGGTTG-3′ and antisense 5′-gggcccTTTACCA GGAGAGTGGGAGA-3′.

PCR products were digested with XhoI and ApaI and ligated into mammalian expression/secretion vector pSecTag2/Hygro B (Invitrogen Life Technologies). The following sets of primers were used to generate cDNA encoding mouse CCL1–Ig from RNA extracted from mouse Th2 cells induced by IL-4: sense 5′-CTAGCTAGCatgaaacccactgccatggca-3′ and antisense 5′-CCGCTCGAGgcaggggttcaccttcttcag-3′.

PCR products were digested with NheI and XhoI and subcloned into the vector containing the mouse IgG1 fragment. Because alterations in the amino acid sequence at the N terminus of chemokines might change their properties, NheI was selected for the cloning procedure, and the original murine kappa chain leader sequence found in pSecTag2/Hygro B was replaced by mouse CCL1 signal peptide sequence. The fused fragments were sequenced by dideoxynucleotide sequencing in our facility (Sequins version 2; Upstate Biotechnology).

Expression and Purification of Fusion Proteins.

Fusion proteins were expressed and purified using CHO dhfr^−/− (DG44) cells (provided by L. Chasin, Columbia University, New York, NY) according to the method described in detail in ref. 61. The fusion protein was expressed as a disulfide-linked homodimer similar to IgG1 and was purified from the culture medium by protein A affinity chromatography (High-Trap, GE Healthcare, BD Biosciences).

Cytokine Measurement by ELISA.

Cytokines were detected using commercial ELISA kits (mIL-10 (R&D Systems), mTGF-β (BD Biosciences) mGranzyme B (eBioscience), hIL-10 (eBioscience), hTGF-β (BD Biosciences), and hGranzyme B (eBioscience), according to the manufacturer’s protocols.

Induction of Acute and Semichronic EAE.

Experimental autoimmune encephalomyelitis (EAE) was induced by immunizing mice with MOG(p35–55)/CFA, as described by Tompkins et al. (62). Seven-to-eight week-old female C57BL/6 mice were immunized s.c. with 200 μL of an emulsion containing 400 μg of Mycobacterium tuberculosis H37Ra (Difco) and 200 μg MOGp35–55. 200 μL of Bordetella pertussis toxin (Sigma) (1 μg/mL) was also administered i.p. on day 0 and day 2 postimmunization. In all experiments at the onset of disease, before therapy was initiated, mice were regrouped so each group included equally sick mice. At this time, mice with a clinical score of less than +1 were excluded.

Mice were then monitored daily for clinical signs by an observer blind to the treatment protocol. EAE scores are as follows: 0, normal; 1, flaccid tail; 2, hind limb paralysis; 3, total hind limb paralysis, accompanied by an apparent front limb paralysis; 4, total hind limb and front limb paralysis; and 5, death. Histological scores: Lumbar spinal cord sections were stained (H&E) and each section was evaluated for tissue damage and mononuclear infiltration using the following scale: 0, no mononuclear cell infiltration; 1, 1–5 perivascular lesions per section with minimal parenchymal infiltration; 2, 5–10 perivascular lesions per section with parenchymal infiltration; and 3, >10 perivascular lesions per section with extensive parenchymal infiltration (Fig. 5B).

BrdU Uptake.

BrdU (Sigma) was added to the drinking water (1 mg/mL) for 14 d according to the manufacturer’s protocol. BrdU uptake was conducted by flow cytometery using anti-BrdU mAb (BioLegend, clone Bu20a).

Histopathology.

The lumbar spinal cord was dissected, fixed in 4% paraformaldehyde, dehydrated, and embedded in paraffin. The 5-µm thick sections were stained with H&E. Each section was evaluated for tissue damage and mononuclear infiltration using the following scale: 0, no mononuclear cell infiltration; 1, 1–5 perivascular lesions per section with minimal parenchymal infiltration; 2, 5–10 perivascular lesions per section with parenchymal infiltration; and 3, >10 perivascular lesions per section with extensive parenchymal infiltration. The mean histological score ± SE was calculated for each treatment group.

CCL1 Transcription in Microglia vs. FOXp3⁺ T_reg Cells.

Entire brain was extracted following perfusion and placed with PBS on ice until transferred to 7 mL RPMI-1640 at room temperature. The brain was thoroughly dissociated using an homogenizer and 3 mL of stock isotonic percoll (SIP) was add (for a final 30% Percoll). The mixture was gradually added on top of 2 mL 70% SIP in PBS and centrifuged at 500 × g, 30 min, 18 °C). The top myelin layer was removed and discarded and 2 mL of the 30–70% intermediate layer (the microglial cells) was taken and washed with PBS. Microglial was sorted as CD45⁺CX₃CR1⁺ cells. RNA was extracted and subjected to RT-PCR analysis (using the following primers: 5′-TTCCCCTGAAGTTTATCCAGTGTT-3′ and 5′-TGAACCCACGTTTTGTTAGTTGAG-3′.

Statistical Analysis.

Statistical analysis was done according to the recommendations of Nature for reporting life sciences research, as specified in the figure legends. For comparison of two groups, linear regression with 95% confidence interval and unpaired two-tailed Student’s t test were used. One-way ANOVA for paired data was used to determine the significance of the time–response curves (for example EAE scores). P values of <0.05 were considered statistically significant. For adjustment of the significance value for multiple comparisons, a Bonferroni correction was applied with a corrected significance value of 0.017.