AhR activation increases IL-2 production by alloresponding CD4⁺ T cells initiating the differentiation of mucosal-homing Tim3⁺Lag3⁺ Tr1 cells

Abstract

Activation of the aryl hydrocarbon receptor (AhR) by immunosuppressive ligands promotes the development of regulatory T (Treg) cells. Although AhR-induced Foxp3⁺ Treg cells have been well studied, much less is known about the development and fate of AhR-induced Type 1 Treg (AhR-Tr1) cells. In the current study, we identified the unique transcriptional and functional changes in murine CD4⁺ T cells that accompany the differentiation of AhR-Tr1 cells during the CD4⁺ T-cell-dependent phase of an allospecific cytotoxic T lymphocyte (allo-CTL) response. AhR activation increased the expression of genes involved in T-cell activation, immune regulation and chemotaxis, as well as a global downregulation of genes involved in cell cycling. Increased IL-2 production was responsible for the early AhR-Tr1 activation phenotype previously characterized as CD25⁺CTLA4⁺GITR⁺ on day 2. The AhR-Tr1 phenotype was further defined by the coexpression of the immunoregulatory receptors Lag3 and Tim3 and non-overlapping expression of CCR4 and CCR9. Consistent with the increased expression of CCR9, real-time imaging showed enhanced migration of AhR-Tr1 cells to the lamina propria of the small intestine and colon. The discovery of mucosal imprinting of AhR-Tr1 cells provides an additional mechanism by which therapeutic AhR ligands can control immunopathology.

Introduction

The aryl hydrocarbon receptor (AhR) is an important physiological regulator of many aspects of the immune system [1, 2]. As a heterodimeric transcription factor, AhR interacts with several different signaling pathways to impact the activation, differentiation and survival of cells that are involved in generating immune responses [3, 4]. CD4⁺ T cells are particularly sensitive targets for AhR regulation, and activation of AhR by exogenous ligands results in suppression of Th1-, Th2- and Th17-dependent immunity [5, 6]. Accordingly, treatment of mice with AhR ligands has been shown to ameliorate the development of several T-cell dependent autoimmune diseases, such as type 1 diabetes [7, 8], inflammatory bowel disease [9–11], and experimental autoimmune encephalomyelitis [12–15], as well as to suppress transplant rejection and inhibit graft-versus-host (GVH) disease [16–18]. The potent efficacy of AhR ligands in these preclinical models supports the development of AhR ligands for therapeutic use in prevention and treatment of immune-mediated diseases [19]. However, the clinical development of AhR ligands will be facilitated by a better understanding of the mechanisms by which AhR signaling induces such profound immunosuppressive effects.

The importance of CD4+ cells as direct targets of AhR-mediated immune suppression was first established in a parent-into-F1 GVH model, wherein AhR activation intrinsic to the CD4⁺ T cell was required for suppression of an allospecific cytotoxic T lymphocyte (allo-CTL) response [20]. The phenotype of the donor CD4⁺ T cells was altered within 48 hr following their transfer, resulting in a significant increase in the percentage of cells expressing high levels of CD25, CTLA-4 and GITR, along with the downregulation of CD62L [21], suggesting that AhR signaling was promoting the activation of the alloresponding T cells. Subsequent studies showed that the CD25⁺ population of donor CD4⁺ T cells was capable of suppressing the proliferation of T effector cells ex vivo, leading to the conclusion that AhR was driving the differentiation of regulatory T cells [21]. The absence of Foxp3 expression but increased IL-10 production supported the concept that AhR activation was inducing type 1 regulatory T cell (Tr1)-like cells that suppressed the development of Th1 cells needed to generate the allo-CTL response [21, 22]. A role for AhR in the generation of Tr1 cells was further described by Apetoh et al. [23]. Under TGFβ/IL-27 polarizing conditions, an AhR-c-Maf complex was shown to enhance IL-10 and IL-21 transcription resulting in Tr1 differentiation. Production of IL-10 and IL-21 was enhanced by the addition of AhR ligands. Furthermore, in the absence of exogenous AhR ligand, the induction of IL-10⁺ Tr1 cells by anti-CD3 in vivo was impaired in mice that expressed the low affinity AhR allele [23].

Apart from these studies, very little is known about how exogenous AhR ligands alter the differentiation of CD4⁺ T cells in vivo. In the current study, we utilized the non-irradiated parent-into-F1 alloresponse model as a defined in vivo system to track CD4⁺ T cell activation [24]. By focusing on the first four days of the alloresponse, we identified the unique transcriptional and functional changes in alloresponding CD4⁺ T cells that accompany the generation of AhR-induced Tr1 cells (AhR-Tr1). Increased expression of several genes were validated at the protein level, including Tim3 and Lag3 as well as the mutually-exclusive expression of CCR4 or CCR9. Consistent with the increased expression of CCR9, real time imaging showed enhanced migration of AhR-Tr1 cells to mucosal tissues, and specifically to the small intestine and colon. These findings suggest that AhR activation by exogenous AhR ligands leads to intestinal mucosa imprinting of AhR-Tr1s. The ability of AhR-Tr1 cells to rapidly disseminate could enhance their ability to control immunopathology at mucosal surfaces.

Results

Sustained AhR activation after day 3 is not required to suppress the allo-CTL response

The first three days of the alloresponse represent the CD4-dependent phase of CTL priming. Previous studies have shown that AhR activation by the prototypic ligand TCDD had to be initiated during this window of time in order to suppress the development of CTL [25]. However, because TCDD is resistant to metabolic breakdown (half-life of approximately 11 days [26]), it induces sustained activation of AhR throughout the experimental time period. Thus, it was not clear if AhR signaling during the CD4-dependent phase of the alloresponse (days 0-3) would be sufficient to suppress the CTL response. To address this question, we used Cl-BBQ, a high-affinity but rapidly metabolized AhR ligand (half-life of 2 hr) that has been shown to suppress the allo-CTL response when given daily at a dose (10 mg/kg) that maintains comparable AhR activation as a single dose of TCDD (15 μg/kg) [17]. Suppression of the allo-CTL response by either Cl-BBQ or TCDD is AhR-dependent [17, 20]. In the present study, host mice were treated daily with Cl-BBQ on days 0-3 or once with TCDD on day 0 relative to donor cell injection, and the allo-CTL response was measured by CD44^hiCD45^low expression on donor CD8 cells [25, 27] on day 10 (Figure 1A). Treatment with Cl-BBQ for three days was sufficient to prevent the loss of body weight associated with the alloresponse (Figure 1B) and to inhibit the development of donor-derived CTL to the same degree as TCDD (Figure 1C and Supplemental Figure 1). In accordance with the suppression of the CTL response, mice treated with Cl-BBQ or TCDD showed less destruction of host cells as measured by host B cell depletion on day 10 (Figure 1D and Supplemental Figure 1). These results indicate that AhR activation during the CD4⁺ T cell-dependent phase of the alloresponse is sufficient to suppress allo-CTL development. This finding is consistent with prior studies showing that TCDD does not suppress a CD4-independent CTL response [28] nor directly impair influenza-specific CD8⁺ T cell expansion [29].

Figure 1. AhR activation on days 0-3 is sufficient to suppress allo-CTL on day 10.

(A) Experimental timeline. Donor cells from B6 mice were injected i.v. into F1 host mice followed by i.p. injection with 10mg/kg Cl-BBQ on days 0-3 or 15μg/kg TCDD on day 0. Grey bars represent the period of AhR activation. Control mice were treated with vehicle on days 0-3. Donor cells were injected into B6 mice as a syngeneic control. Donor (H2Dd⁻) and host (H2Dd⁺) splenocytes were analyzed on day 10. (B) Body weight change due to the alloresponse relative to Day 6. (C) % CTL (CD8⁺CD44^hiCD45RB^low) gated on donor cells. (D) % CD19⁺ gated on host cells. Dotted line represents syngeneic control (C and D). Data are presented as mean + SEM. Data in B-D are from a single experiment representative of two independent experiments with 5 mice per group per experiment. *p<0.05, **p<0.01, ***p<0.001 compared to vehicle-treated mice determined by one-way ANOVA with Tukey’s test for multiple comparisons.

High concordance in CD4⁺ T cell differential gene expression following treatment with TCDD or Cl-BBQ

To gain insight into how AhR activation alters the early response of allospecific CD4⁺ T cells, we performed genome-wide analysis of changes in gene expression on days 2 and 3 of the alloresponse in cells isolated from mice treated with TCDD, Cl-BBQ, or vehicle. Highly enriched populations of alloantigen specific CD4+T cells were obtained by sorting donor CD4⁺ T cells that had undergone two or more cell divisions based on CFSE dilution (Supplemental Figure 2).

The expression profiles of differentially regulated genes were highly concordant (p<0.0001) between TCDD- and Cl-BBQ-treated mice on both day 2 and day 3 of the alloresponse (Figure 2A,B). The increased expression of known AhR-regulated genes such as Cyp1a1, Cyp1b1, Ahrr, and Tiparp validated AhR activation. Differentially expressed genes that are known to play a role in T cell function are annotated in Figure 2B and Table I. Foxp3 expression was not altered, consistent with our previous data showing that Foxp3⁺ cells are not induced by AhR ligands during the early stages of T cell activation [17, 22]. Canonical pathway analysis based on the differentially regulated genes identified several “Immune response”-related pathways that were positively associated with AhR activation. These pathways included IL-10, IL-12 and IL-27 signaling as well as regulatory T cell-mediated pathways involving modulation of antigen-presenting cells and the function of effector T cells and NK cells (Figure 2C). Taken together, results of the pathway analysis provide evidence for AhR promoting the activation of CD4⁺ T cells and also driving the expression of genes that promote the development of Foxp3⁻ regulatory T cells.

Figure 2. AhR-induced transcriptional profile in alloresponding CD4⁺ donor cells highlights upregulation of T-cell activation and regulatory genes and downregulation of cell cycle genes.

Gene expression was analyzed in alloresponding donor (dividing CFSE⁺) CD4⁺ cells. For the day 2 gene expression analysis, F1 mice were treated i.p. with 15μg/kg TCDD on day 0 or 2mg/kg Cl-BBQ on days 0 and 1. For the day 3 gene expression analysis F1 mice were treated with 15μg/kg TCDD on day 0 or 10mg/kg Cl-BBQ on days 0-2. (A) Heatmap displays differentially expressed genes compared to vehicle treated mice (p<0.05). (B) Highly differentially regulated genes following TCDD and Cl-BBQ treatment on day 3. Genes that were also differentially expressed on day 2 are depicted in orange. Cell cycle relate genes are depicted in blue. (C) Enriched canonical pathways associated with AhR activation on day 3. Pathways that were also enriched based on the day 2 gene expression data are underlined. Data in A-C are combined from two separate experiments, with 48 hour and 72 hour data performed separately. Data represent 4 mice per treatment group and time point. (D-F) F1 mice were treated i.p. with 15μg/kg TCDD on day 0 or 10mg/kg Cl-BBQ on days 0-3. BrdU was injected i.p. on day 3 and mice were sacrificed on day 4. (D) CFSE dilution in donor CD4⁺ cells. (E) Division index measured by CFSE dilution. (F) % BrdU⁺ cells gated on donor CD4⁺ cells. Data are from a single experiment representative of 2 experiments with 5-6 mice per group per experiment. Bar graphs represent mean + SEM. *p<0.05, **p<0.01, ***p<0.001 determined by one-way ANOVA with Tukey’s test for multiple comparisons.

Table I.

Genes related to immune function that are differentially expressed in allospecific CD4+ T cells from mice treated with TCDD or Cl-BBQ on days 2 and 3 of the GVH response

	Gene Name	Gene Symbol	Day 2 TCDD	Day 2 BBQ	Day 3 TCDD	Day 3 BBQ	Adhesion/Chemotaxis	Cytokine Receptor	Immune Regulation	Cytotoxicity
Upregulated	Interleukin 12 receptor subunit beta 2	Il12rb2a	3.1	2.2	16.5	28.3		+
	Granzyme A	Gzmaa	5.3	2.9	9.4	27.9				+
	G protein-coupled receptor 15	Gpr15	5.7	2.8	17.2	25.9	+		+
	Granzyme B	Gzmba	3.3	2.5	9.3	19.0				+
	Ankyrin repeat SOCS box containing 2	Asb2	1.9	1.4	7.4	10.7	+
	Integrin subunit alpha E	Itgae)	1.8	1.3	6.4	9.8	+
	Leukocyte Ig like receptor B4	Lilrb4	ns	ns	5.6	9.6			+
	C-X-C motif chemokine recpetor 6	Cxcr6	2.5	1.8	5.8	8.3	+
	Synaptotagmin like 2	Sytl2	2.3	1.7	5.0	7.6				+
	C-C motif chemokine recpetor 5	Ccr5	3.4	2.4	5.1	7.4	+
	Myosin IF	Myo1f	3.7	2.6	5.0	7.0			+
	Integrin subunit alpha 2	Itga2	2.6	1.9	3.2	6.3			+
	IFN induced transmembrane protein 2	Ifitm2	1.6	1.3	5.1	6.2				+
	Interleukin 7 receptor	Il7r	1.6	1.6	2.6	6.2		+
	Interleukin 18 receptor accessory protein	Il18rap	1.6	1.3	2.9	5.3		+
	Perforin 1	Prf1	1.4	1.3	3.2	5.1				+
	Natural killer cell granule protein 7	Nkg7	ns	ns	3.2	4.8				+
	Granzyme C	Gzmc	ns	ns	2.3	4.8				+
	C-C motif chemokine receptor 9	Ccr9a	6.8	3.7	2.7	4.5	+
	Hepatitis A virus cellular receptor 2	Havcr2	2.9	1.6	2.3	3.7			+
	C-C motif chemokine receptor 4	Ccr4	2.9	1.7	3.0	3.6	+
	Galectin 3	Lgals3	1.6	1.3	2.1	3.6			+
	Interleukin 10	Il10a	2.1	1.3	3.2	3.5			+
	C-C motif chemokine receptor 2	Ccr2	2.5	1.6	2.4	3.4	+
	SLAM family member 7	Slamf7	ns	ns	2.7	3.4			+
	Interleukin 18 receptor	Il18r	ns	ns	1.7	3.1		+
	CD69 molecule	Cd69	1.2	ns	2.1	2.6			+
	Fas ligand	Fasl	ns	ns	2.1	2.6				+
	L1 cell adhesion molecule	L1cam	1.7	1.5	1.8	2.5			+
	Tranforming growth factor beta 3	Tgfb3a	2.2	1.6	2.1	2.4			+
	PR/SET domain 1	Prdm1a	2.4	1.6	1.9	2.3			+
	Interleukin receptor 4 subunit alpha	Il4ra	1.3	ns	1.7	2.2		+
	Interleukin 12 receptor subunit beta 1	Il12rb1	ns	ns	1.5	2.2		+
	Interferon gamma receptor 1	Ifngr1	1.3	1.2	1.7	2.2		+
	Ikaros family zinc finger 3	Ikzf3	1.3	1.2	1.9	2.2			+
	Selectin P ligand	Selplg	1.3	1.3	1.7	2.2			+
	Runt related transcription factor 2	Runx2	1.5	1.2	1.6	2.2			+
	TNF receptor superfamily member 9	Tnfrsf9	ns	ns	2.0	2.1			+
	Suppressor of cytokine signaling 3	Socs3	1.3	1.3	1.3	1.8			+
	Interleukin 27 receptor subunit alpha	Il27ra	ns	ns	1.5	1.7		+
	Interleukin 10 receptor subunit alpha	Il10ra	1.4	ns	1.4	1.7		+
	Interleukin 21 receptor	Il21r	1.1	ns	1.3	1.6		+
	CD27 molecule	Cd27	ns	ns	1.4	1.6			+
	B-cell CLL/lymphoma 3	Bcl3	1.2	1.2	1.5	1.6			+
	NFKB inhibitor zeta	Nfkbiz	ns	ns	1.3	1.6			+
	Lymphocyte activating 3	Lag3	ns	ns	1.5	1.5			+
	Retinoic acid receptor alpha	Rara	1.3	ns	1.5	1.5			+
	Interleukin 2	Il2	1.4	1.3	1.6c	ns			+
Downregulated	CD86 molecule	CD86	−1.8	−1.7	ns	ns
	Integrin subunit alpha L	Itgal	−1.2	ns	−1.3	−1.4	+
	Lymphocyte antigen 6 complex C1	Ly6c	ns	ns	−2.8	−2.0	+
	LIM domain only 4	Lmo4	ns	ns	−1.5	−2.0			+
	Integrin subunit alpha 4	Itga4)	ns	ns	−2.0	−2.0	+
	TNF superfamily member 4	Tnfsf4	−1.8	−1.4	−1.6	−2.5
	Semaphorin 4A	Sema4a	2.0	−1.6	−2.9	−3.4
	CD22 molecule	Cd22	−2.0	−1.4	−4.1	−4.2	+
	Neuropilin 1	Nrp1	−1.4b	−1.6b	−6.6	−6.8			+
	Integrin subunit beta 1	Itgb1	−1.6	−1.6	−8.0	−9.2	+
	Selectin L	Sell	−1.6	−1.4	−8.9	−19.5	+
	Stefin A3	Stfa3	−2.4	−3.2	−7.4	−28.3

Differential Expression based on fold change (p<0.05); ns=not significant

^aGene previously associated with AhR activation in CD4+ cells (Marshall, et al., 2008);

^bp=0.06

In contrast to the upregulated genes, few immune-related genes were downregulated following AhR activation, and these genes were primarily related to cell adhesion (e.g., Itgal, Ly6c, Itga4, Itgb1, Cd22, Sell) (Table I). In addition, 11 canonical cell cycle pathways were negatively associated with AhR activation on day 3 of the alloresponse, which reflected the downregulation of more than 100 genes involved in cell proliferation and cell death (Figure 2B,C). This finding was unexpected since we had sorted on dividing alloresponding CD4⁺ T cells and there was no difference in the degree of proliferation between AhR ligand-treated and control groups at that point in time. To determine if the down-regulation in cell cycle genes on day 3 portended impaired proliferation, the proliferative status of CFSE-labeled donor cells was analyzed on day 4. Alloresponding CD4⁺ T cells from TCDD- and Cl-BBQ-treated mice showed a significantly reduced division index, a measure of the average number of cells that were generated by a single dividing cell over the four day period (Figure 2D, E). To confirm that proliferation was impaired between day 3 and day 4 by AhR activation, mice were injected with BrdU on day 3 and incorporation into dividing cells was assessed on day 4. Indeed, BrdU incorporation was significantly reduced in donor CD4⁺ T cells from TCDD- and Cl-BBQ-treated mice (Figure 2F). This delayed response could reflect suppression of effector T cell proliferation by AhR-Tr1 cells.

AhR activation enhances early IL-2 production to induce the activated AhR-Tr1 cell phenotype

To gain greater insight into early AhR-mediated transcriptional changes in CD4⁺ T cells prior to day 2 of the alloresponse, we performed an analysis of transcription factors (TFs) predicted to function upstream of the genes that were significantly upregulated by AhR activation on days 2 and 3 (Figure 3A). Even though the changes in gene expression between TCDD- and Cl-BBQ-treatment groups were highly concordant on day 2 (Figure 2A), the dose of Cl-BBQ used (2 mg/kg) did not activate AhR to a level equivalent to the 15ug/kg dose of TCDD. Therefore, data from the day 2 study were not included in the prediction analysis. Predicted TFs included AhR itself, as well as its binding partner ARNT, and Bach2, a direct target of AhR [30]. The other TFs are known to be involved in T cell activation and synergistic induction of IL-2 production and signaling, including NFAT, Fra1, Fra2, CREB3, and STAT5b [31–33].

Figure 3. AhR-induced excess IL-2 is responsible for driving the CD4⁺CD25⁺CTLA4⁺GITR⁺ donor cell activation phenotype on day 2.

(A)Transcription factors that are predicted to function upstream of TCDD-induced upregulated genes on day 2 and 3 and Cl-BBQ-induced upregulated genes on day 3. On day 2, the 2mg/kg Cl-BBQ dose did not lead to predicted TFs and was therefore not included in the analysis. Data are combined from two separate experiments, with 48 hour and 72 hour data collected separately. Data represent 4 mice per treatment group and time point. (B) Donor cells from B6 mice were injected i.v. into F1 host mice followed by i.p. injection with 15μg/kg TCDD on Day 0. IL-2 production on day 1 was measured using a capture flow cytometry assay. (C) CFSE labeled donor cells from B6 mice were injected i.v. into F1 hosts and treated i.p. with anti-IL-2 or isotype control and 15μg/kg TCDD or vehicle. Alloresponding donor (dividing CFSE⁺) CD4⁺ cells were analyzed on day 2 for % CD25⁺CTLA4⁺GITR⁺. (D, E) Donor cells from B6 mice were injected i.v. into F1 hosts followed by i.p. injection with 15μg/kg TCDD on Day 0. At 36 hours, the percentage (D) and mean channel fluorescence (MCF) (E) of pSTAT5 and CD25 expressing donor CD25⁺ cells gated on CD4⁺ cells. Data in B-E represent mean + SEM and are from single experiments with 5 mice per group. *p<0.05, ***p<0.001. Significant was determined by Student’s t-test (B) or by one-way ANOVA with Tukey’s test for multiple comparisons (C).

Since IL-2 is known to play a major role in regulating the induction of regulatory T cells [34, 35], and the IL-2 promoter contains AhR/ARNT binding sites [36], we hypothesized that AhR activation increases the production of IL-2 to induce the activated AhR-Tr1 phenotype (CD25⁺CTLA4⁺GITR⁺) on day 2 of the alloresponse [21]. IL-2⁺ donor CD4⁺ T cells were identified by flow cytometry at 24 hr post donor cell injection. As shown in Figure 3B, a small but significant increase in the percentage of IL-2⁺ cells within the donor CD4⁺ T cell population was observed in TCDD-treated mice. Based on the fact that only 10% of the donor CD4⁺ T cell pool is allospecific [37], this small increase in the fraction of IL-2 producing donor cells could be biologically significant. To determine if excess IL-2 was indeed responsible for inducing the activated AhR-Tr1 phenotype on day 2, IL-2 signaling was blocked with antibodies targeting two different epitopes of IL-2 [35, 38]. AhR activation induced a robust increase in alloresponding, but not host (Supplemental Figure 3), CD4⁺ cells that expressed CD25, CTLA4, and GITR from 8 ± 2% to 26 ± 7%, and this increase was completely normalized by anti-IL-2 treatment (Figure 3C). Furthermore, these CD25⁺ cells responded to IL-2 ex vivo with enhanced STAT5 phosphorylation (Figure 3D), and there was a direct correlation between CD25 and pSTAT5 expression levels (Figure 3E). These results show that AhR activation enhances IL-2 production and induces functional CD25 expression on alloresponding donor CD4⁺ T cells, leading to the activated AhR-Tr1 phenotype.

Protein validation of differential gene expression associated with activated AhR-Tr1 cells

To further characterize the phenotype AhR-Tr1s, we evaluated the protein expression levels of 26 differentially expressed genes (annotated in Figure 2B) that have been associated with different types of regulatory T cells. As expected, AhR activation resulted in increased expression of CD25, CTLA4, GITR, and ICOS, as well as decreased expression of CD62L in alloresponding donor CD4⁺ T cells on day 2 (Figure 4C and Supplemental Table 1). Additionally, AhR activation increased the expression of IL-10Rα, FasL, Tim3, Lag3, CCR4, and CCR9, along with decreased expression of Nrp1. Altered expression of the other genes could not be verified at the protein level which could be related to timing (day 2 versus day 3) or reagents (Supplemental Table 1).

Figure 4. Lag3, Tim3, CCR4 and CCR9 are preferentially expressed on AhR-induced CD25⁺ Tr1 cells.

Donor cells from B6 mice were injected i.v. into F1 host followed by i.p. injection with 15μg/kg TCDD on day 0 or 10mg/kg Cl-BBQ on days 0 and 1. On day 2 phenotypic analysis of splenic alloresponding donor CD4⁺ cells was conducted. (A) Gating strategy for alloresponding donor cells and AhR-induced CD25⁺ Tr1 cells. (B) Percentage and total number of alloresponding CD4⁺Foxp3⁻ cells that were CD25⁺CD62L⁻. (C) Representative FACS plots of Tim3, Lag3, CCR4, and CCR9 expression in alloresponding CD4⁺ cells and CD4⁺Foxp3⁻CD25⁺CD62L⁻ cells (left panels) and percentage of marker positive CD4⁺ T cells (right panels). (D) Coexpression of Lag3 and Tim3 on CD4⁺Foxp3⁻CD25⁺CD62L⁻ cells. (E) Expression of CCR4 and CCR9 on CD4⁺Foxp3⁻CD25⁺CD62L⁻Lag3⁺Tim3⁺ cells. Data in A-E are from a single experiment representative of 2 experiments with 4 mice per group per experiment. (F) F1 mice were treated i.p. with 15μg/kg TCDD on day 0. BrdU was injected i.p. on day 3 and mice were sacrificed on day 4. BrdU incorporation was evaluated on CD4⁺CD25⁺Lag3⁺Tim3⁺ cells and the non- CD4⁺CD25⁺Lag3⁺Tim3⁺ cells. Data are representative of 5 mice per group and present mean + SEM. V: Vehicle-treated, T: TCDD-treated, B: Cl-BBQ treated. *p<0.05, **p<0.01, ***p<0.001 in comparison to vehicle treated mice. ^#p<0.05, ^##p<0.01, ^###p<0.001 comparison between alloresponding CD4⁺ and CD4⁺Foxp3⁻CD25⁺CD62L⁻ cells. Significant was determined by one-way ANOVA with Tukey’s test for multiple comparisons (B, C, E) or by Student’s t-test (F).

To determine if any of the validated proteins were preferentially expressed on AhR-Tr1 cells, the frequency of positive cells for each marker in the CD4⁺CD25⁺CD62L^low subpopulation was compared to all alloresponding CD4⁺ cells. Foxp3⁺ cells, although low in number, were gated out to circumvent overlapping expression profiles in AhR-Tr1 cells and thymic-derived Treg cells (Figure 4A, B). As anticipated, expression of CTLA4, ICOS and GITR was increased in the AhR-Tr1 subpopulation with >94% of cells positive for these markers (Supplemental Table 1). In addition, the percentage of cells expressing Tim3, Lag3, CCR4 and CCR9 was increased in the AhR-Tr1 subpopulation compared to all alloresponding CD4⁺ cells (Figure 4C).

To determine if Tim3, Lag3, CCR4 and/or CCR9 could act as specific biomarkers for AhR-Tr1s, we compared the frequency of coexpression in CD25⁺CD62L^low cells that from vehicle and AhR ligand-treated mice. Notably, the percentage of alloresponding CD25⁺CD62L^low cells that coexpressed both Lag3 and Tim3 was significantly elevated from 9±1% in vehicle-treated mice to 32±2% and 34±4% in TCDD and Cl-BBQ treated mice, respectively (Figure 4D). This phenotype was specific to alloresponding cells, as less than 3% of host CD25⁺CD62L^low cells expressed both Lag3 and Tim3 in TCDD- or Cl-BBQ-treated mice (Supplemental Figure 4). CCR4 and CCR9 expression further defined nonoverlapping subsets of the alloresponding CD25⁺CD62L^lowLag3⁺Tim3⁺ population (Figure 4E). These results suggest that coexpression of Lag3 and Tim3 can be used as a biomarker of AhR-Tr1 cells.

AhR-Tr1 cells continue to proliferate after day 3 of the alloresponse

The suppression of proliferation of donor CD4⁺cells on day 4 of the alloresponse (see Figure 2F) may be a reflection of AhR-Tr1-mediated suppression of effector T cell proliferation. The identification of the Lag3⁺Tim3⁺ phenotype as a biomarker for AhR-Tr1s allowed us to address this possibility. We compared BrdU incorporation into the CD4⁺Foxp3⁻CD25⁺Lag3⁺Tim3⁺ (AhR-Tr1) subpopulation versus the remaining (non-(AhR-Tr1)) cells. Strikingly 10.1 ± 2.9% of non-(AhR-Tr1) cells incorporated BrdU in comparison over 35.6 ± 4.8% of AhR-Tr1 cells (Figure 4F). These results indicate that AhR-Tr1 cells were continuing to proliferate, while the proliferation of non-(AhR-Tr1) cells was highly suppressed. Together with prior data showing that AhR-Tr1 cells inhibited the proliferation of effector T cells ex vivo [21] these findings suggest that the reduction in cell cycle gene expression on day 3 occurs selectively in the T effector population.

AhR-Tr1 cells migrate to the intestines

The expression of CCR4 and CCR9 on discrete subsets of AhR-Tr1 cells suggested that AhR activation could influence their migration to the lung/skin (CCR4) [39] and/or small intestine/colon (CCR9) [40]. In order to evaluate the migratory capacity of T cells following AhR activation, we used donor cells from Tlux mice which express luciferase in both CD4 and CD8 T cells under the control of CD2 [41]. At the time of donor cell transfer, mice were treated with vehicle or TCDD. TCDD was used as the model for AhR activation based on the strong concordance in differential gene expression between TCDD- and Cl-BBQ-treated mice and the fact that TCDD only had to be injected once for sustained AhR activation. Following donor cell injection and treatment, host mice were imaged daily for four days to determine the kinetics of T cell distribution (Figure 5A). On day 1, the majority of cells in both vehicle- and TCDD-treated mice were confined to the cervical, thoracic, and abdominal regions, likely corresponding to cervical lymph nodes, lung, and spleen. In vehicle-treated mice, the intensity of luminescence increased through day 4 but the majority of signal was confined to the cervical region. In contrast, T cell migration in TCDD-treated mice was highly dynamic between days 1 and 4, with a significant increase in luminescence in the head, abdominal, and genital regions, in comparison to vehicle-treated mice (Figure 5B). The increase in signal in the head region was not associated with nasal-associated lymphoid tissues, tongue, nor salivary glands. The exact source of the signal in the head region remains to be identified.

Figure 5. AhR activation enhances the migration of alloresponding CD4⁺ T cells to mucosal regions.

Donor cells from Tlux mice were injected i.v. into F1 host followed by i.p. injection with 15 μg/kg TCDD on day 0. (A) Mice were live imaged for bioluminescence signal daily through day 4. Images are from a single mouse representative of 3 experiments with 3 mice per group per experiment. (B) Photons emitted/sec of regions of interest defined in the top left of panel A was measured each day. Data represent the mean + SEM from one of three experiments with 3 mice per group per experiment. (C-F) On day 4, PBMCs were isolated and CD25, Lag3, Tim3 and CCR9 expression was analyzed on alloresponding donor CD4⁺ cells (H2D^d-, CFSE diluted). (G,H) On day 4, the intestines were excised from host mice and the whole intestine was imaged for bioluminescence signal. (I-K) Lamina propria cells were isolated from the small intestine and colon and cells in addition to cells isolated from Peyer’s patches. (I) Viable donor CD45⁺ cells were measured. (J) CD4 and CD8 expression gated on viable donor CD45⁺ cells. (K) Total AhR-Tr1 cells in the small intestine, colon and Peyer’s patches. Data in C-K are from a single experiment representative of 2 experiments with 3 mice per group per experiment. Data represent mean + SEM. *p<0.05, **p<0.01, ***p<0.001 determined by Student’s t-test.

Since the bioluminescence signal reflects both CD4 and CD8 T cells, tissues were processed on day 4 to analyze specific cell subsets. Migrating donor T cells were detected in the blood of both vehicle- and TCDD-treated mice on day 4 with no difference in the percentages of donor CD4 and CD8 cells between treatment groups. However, there was a significant increase in the percentage of donor CD4⁺ cells that expressed CD25 in TCDD-treated mice in comparison to vehicle-treated mice (28.0 ± 8.1% versus 7.9 ± 2.5%) (Figure 5C,D). Within the CD25⁺ cell population, a greater proportion of the cells coexpressed Lag3 and Tim3 in TCDD-treated-mice in comparison to vehicle-treated mice (27.7 ± 2.7% versus 10.9 ± 2.1 %) (Figure 5C,E). Furthermore, in TCDD-treated mice, 12.4 ± 2.2% of CD25⁺Lag3⁺Tim3⁺ cells expressed CCR9 (Figure 5C,F). CCR4⁺ cells were not detected in the blood on day 4 (data not shown). These results confirm the enhanced migration of AhR-Tr1 cells.

Analysis of bioluminescence in the intestines revealed a significantly increased signal throughout the small and large intestines in TCDD-treated mice, whereas luciferase-expressing cells were localized to the Peyer’s patches in vehicle-treated mice (Figure 5G,H). Preliminary experiments showed that donor cells in the intestines of TCDD-treated mice were localized to the lamina propria (LP) and were not in the intraepithelial layer. Consistent with the bioluminescent signal, the percentage of donor cells in the LP increased from 7% to 23% in the small intestine and from 10% to 22% in the colon of TCDD-treated mice compared to vehicle-treated mice. In contrast, there was no difference in the percentage of donor lymphocytes in the Peyer’s patches between TCDD- and vehicle-treated mice (Figure 5I). Donor cells in the LP were further evaluated for CD4 and CD8 expression to determine which population accounted for the luminescent signal in the intestines. Both donor CD4⁺ and CD8⁺ T cells were identified in the Peyer’s patches, whereas only donor CD4⁺ cells were detected in LP of the small intestine and colon (Figure 5J).

Finally, AhR-Tr1 cell markers were examined in the LP of the small intestine and colon. While the percentage of CD4⁺ donor T cells that coexpressed CCR9, Lag3, and CD25 did not differ between vehicle- and TCDD-treated mice, the total number of cells coexpressing these markers was significantly increased in TCDD-treated mice (Figure 5K). Collectively, these data demonstrate a role for AhR activation in enhancing the migration of alloresponding AhR-Tr1 cells to the lamina propria.

Discussion

Ligand-induced activation of AhR alters the differentiation of CD4⁺ T cells, resulting in increased regulatory T cell activity and suppression of adaptive immune responses [5]. Based on the simplicity of ligand-dependent induction of AhR-Tr1 cells in vivo, the development of therapeutic AhR ligands for the treatment of immune-mediated diseases is a promising area of translational research. Therefore, understanding the underlying mechanisms by which AhR activation drives AhR-Tr1 cell development is an area of high priority. In the studies presented here, we used the acute parent-into-F1 alloresponse model to track activated CD4⁺ T cell responses in vivo following treatment with AhR ligands. We identified early transcriptional and functional events following AhR activation that drive the differentiation of CD4⁺ T cells into dynamic AhR-Tr1 cells, with upregulated expression of Tim-3 and Lag-3, and enhanced migratory capacity toward mucosal sites.

The earliest documented change in alloresponding CD4⁺ T cells following AhR activation was an increase in the percentage of IL-2⁺ cells. This increase in IL-2-producing cells could result from direct transcriptional regulation of the IL-2 gene by AhR, as previously reported [36] or from the ability of AhR ligands to induce a rapid increase in the concentration of free intracellular calcium that has been described in several different cell types [42, 43]. The AhR-mediated increase in calcium has been shown to induce a transient increase in CaMK activity leading to the activation NFAT as well as PKC activation and NFκB signaling could also contribute to the promotion of IL-2 gene expression [44]. The increase in IL-2 production was shown to be responsible for the early activation phenotype (CD25⁺CTLA4⁺GITR⁺) on day 2 of the alloresponse. This increase in IL-2 was transient, and IL-2 was no longer produced by CD4+ cells on day 2 [22], consistent with Tr1 cell differentiation [45].

Tr1 cells are a subset of induced regulatory T cells that produce the suppressive cytokine IL-10 and do not depend on the transcription factor Foxp3 [45]. AhR-Tr1 cells share these characteristics [21, 22], and also show increased expression of many genes associated with the Tr1 transcriptional signature, including Il10, Il10ra, Tgfb3, Lag3, Itga2 (Cd49b), Ctla4, Gzmb, Prf1, and Prdm1 [45]. The induction of Tr1 cells by AhR activation could occur through the IL-27 signaling pathway that leads to the production of IL-10 as well as IL-21, a promoter of Tr1 cell stability [46]. Increased expression of Il27ra and Il21r in alloresponding CD4⁺ T cells support a role for enhanced IL-27 signaling in the development of AhR-Tr1 cells. Likewise, preliminary data indicate that AhR promotes the expression of Il27, Il10, and Il21 in non-CD4⁺ lymph node cells, demonstrating that AhR activation can increase cytokine production as well as cytokine receptors to promote Tr1 differentiation. Interestingly, previous studies have also shown that IL-27 upregulates AhR expression [23], suggesting that AhR and IL-27 may interact in a feed forward loop to enhance Tr1 differentiation.

The primary function of Tr1 cells is to suppress the proliferation of effector T cells through a variety of suppressive mechanisms including production of cytotoxic mediators, expression of inhibitory signaling molecules, and secretion of suppressive cytokines [45]. AhR-Tr1 cells show increased expression of numerous genes associated with these mechanisms, as listed in Table I. Furthermore, suppression of effector T cell proliferation by AhR-Tr1 cells was suggested by the highly significant and global downregulation of genes related to cell cycling on day 3 of the alloresponse. This observation was initially surprising given that lymphocyte proliferation is generally not affected by AhR activation in vitro [47]. Furthermore, in the alloresponse model, donor CD4⁺ T cells are clearly capable of undergoing more than six rounds of cell division following AhR activation over the first 3 days, and the proliferation rate is the same as controls (Supplemental Figure 4 and [21]). Therefore, the reduction in cell cycle genes was likely an indirect effect rather than direct transcriptional regulation by AhR. An indirect effect was further supported by data showing that AhR-Tr1 cells continued to proliferate on day 4 whereas the proliferation of the remaining CD4⁺ T cells was reduced by 3.5-fold. Together with the known ability of AhR-Tr1 cells to suppress T cell proliferation ex vivo [21], the data suggest that the inhibition of cell cycle gene expression stems from an AhR-Tr1-mediated inhibition of Th1 cell proliferation.

In the present study, we uncovered a novel role for the AhR in enhancing migration of activated CD4⁺ T cells to mucosal tissues. AhR activation induced the expression of several genes encoding proteins involved in cell migration, including CCR2, CCR4, CCR5, CCR9, GPR15, CXCR6, CD103, and SELPLG. At the same time, AhR activation led to decreased expression of several integrins and adhesion molecules which could also influence T cell migration (e.g., CD11a, Ly6c, CD49d, CD22, CD29, CD62L). When T cell migration was tracked in real time, increased homing of CCR9-expressing AhR-Tr1 cells to the small intestine and colon was confirmed. In addition, AhR activation led to a significant upregulation of the gut-tropic orphan G-coupled protein receptor Gpr15, which plays a critical role in immune tolerance in the large intestine [48]. The preferential dissemination of AhR-Tr1 cells to the gut mucosa suggests that they are primed to regulate hyperinflammatory alloimmune responses in GVH target organs [49]. Our discovery of mucosal imprinting of AhR-Tr1 cells provides an additional mechanism by which AhR activation can promote intestinal immune tolerance [50–53]. Importantly, mucosal imprinting of AhR-Tr1 cells may enhance the protective effects of AhR ligands in controlling diseases that involve T cell-mediated immunopathology in barrier organs, such as inflammatory bowel diseases, Sjogren’s syndrome and allergic diseases.

Materials and Methods

Animals

C57BL/6J (B6; H-2^b/b) and B6D2F1 (F1; H-2^b/d) mice were purchased from The Jackson Laboratory. Tlux mice (C57BL/6 transgenic mice containing the firefly luciferase gene, luc, under control of a human CD2 mini-gene cassette) were a kind gift from Dr. Casey Weaver [41]. Mice were maintained in the specific pathogen-free animal facility at Oregon State University. All experiments used gender/age-matched mice between 9-12 weeks of age. All animal procedures were carried out under protocols approved by the Institutional Animal Care and Use Committee.

Donor Cell Transfer

Cells were pooled from the spleen and peripheral lymph nodes of donor B6 mice and the equivalent of 3-4×10⁷ donor T cells were injected i.v. into F1 host mice. In some studies, donor cells were labeled at a final concentration of 5 μM carboxyfluorescein succinimidyl ester (CFSE; Life Technologies) prior to donor cell injection.

AhR Ligand Treatments

F1 host mice were injected i.p. with 15 μg/kg TCDD diluted in 0.15% anisole in peanut oil, Cl-BBQ (a 40:60 mixture of 10- and 11-chloro-7H-benzimidazo[2,1-a]benzo[de]Iso-quinolin-7-one) diluted in 5% anisole in peanut oil, or vehicle alone. For experiments with both TCDD and Cl-BBQ treatments, the higher concentration of anisole was used as the vehicle control. Doses and timing of Cl-BBQ treatment are indicated in figure legends for each experiment.

Antibody neutralization of IL-2

At the time of donor cell transfer, F1 host mice were injected i.p. with 0.5mg anti-IL-2 antibody of a 1:1 mixture of S4B6 and JES6-1 antibodies (BioXCell) or isotype control (IgG2a). The JES6-1 antibody was used to prevent the formation of complexes between IL-2 and S4B6 [38].

BrdU incorporation

F1 host mice were injected i.v. with 2 mg of BrdU solution (BD) eighteen hours prior to sacrifice on day 4 of the alloresponse response.

Preparation of cells from spleen, Peyer’s patches, and intestines

Single cell suspensions of splenocytes were prepared and red blood cells were removed by hypotonic lysis. Cells from Peyer’s patches were isolated through mechanical dissociation between frosted glass slides. Isolation of LP cells from the colon and small intestine was performed as previously described [54]. Briefly, Peyer’s patches were excised and the epithelial layer of the intestine was removed through washes in a 5mM EDTA and 0.145mg/ml DTT solution while stirring at 37°C. The remaining intestinal tissue was diced and digested with 0.2U/ml Liberase TM and 0.05% DNase while shaking at 37°C. The digested tissue was washed three times with a 3% FBS solution and successively filtered after each wash through two 70μm filters and one 40μm filter in preparation for cell staining.

Flow Cytometry

Fc receptors were blocked with rat IgG (Jackson ImmunoResearch) and stained with the following antibodies: CD45 (30-F11), CCR9 (CW-1.2), CD45RB (C363-16A), CD44 (1M7), CD8 (53.6.7), CD19 (1D3), CD4 (RM4-5), CD25 (PC61.5), and GITR (DTA-1) from eBioscience, Lag3 (C9B7W) and CD62L (R1-2) from BD Bioscience and CCR4 (2G12), Tim3 (RMT3-23), H2D (34-2-12) from Biolegend). For intracellular staining, cells were fixed and permeabilized using the Foxp3 Fixation/Permeabilization buffer (eBioscience) and stained with CTLA-4 (UC10-4B9) and/or Foxp3 (FJK168) or fixed with Cytofix/Cytoperm (BD Biosciences) and stained with anti-BrdU according the manufacturer’s protocol (BrdU Flow Kit; BD Biosciences). Flow cytometric measurement of IL-2 secreting cells was conducted according to the manufacturer’s instructions (Mouse IL-2 Secretion Assay kit, Miltenyi Biotec).

Data were acquired on a FC-500 or Cytoflex flow cytometer (Beckman Coulter). Data were compensated and analyzed using WinList (Verity Software House) or FlowJo (Treestar) software. Division Index based on CFSE dilution was calculated using FCS Express 5 software (DeNovo) using the following equation:

$$\frac{{\displaystyle \sum _{i=1}^{P-1}{N}_i}}{{\displaystyle \sum _{i=1}^{P-1}\frac{N_i}{2^i}}}$$

P is the total number of peaks found and N is the number of cells in a generation. Fluorescence minus one controls were used for setting gates.

Gene expression analysis

Gene expression analysis was performed in two separate experiments on day 2 and day 3 of the alloresponse response using CFSE-labeled donor cells. Spleen and peripheral lymph nodes cells were pooled from each animal and CD4+ T cells were enriched through negative selection using the mouse CD4 T cell isolation kit (Miltenyi Biotec) and the autoMACS separator. Alloresponding donor cells were identified by gating on CD4⁺ CFSE⁺ cells that had divided at least twice, and then sorted to high purity (>96%) using a MoFlo XDP cell sorter (Supplemental Figure 4). RNA was isolated according to the manufacturer’s protocols (RNeasy, Qiagen). Labeled target cDNA was prepared using NuGen Pico WTA System v.2/Encore Biotin labeling protocol. Whole genome analysis was performed using the Affymetrix Mouse Gene 1.0 (day 2) and 2.0 (day 3) microarrays and uploaded to NCBI Gene Expression Omnibus (www.ncbi.nlm.nih.gov/geo/query/acc.cgi?token=yhshisgadnmrvuv&acc=GSE83260). Data were processed with Bioconductor using the ‘oligo’ and ‘limma’ packages. Raw intensity data were background corrected, quantile normalized by robust multi-array analysis summarization and statistically significant, differentially expressed transcripts were identified between treated and control at each time point by Moderated t-test with 5% false discovery rate calculation.

Unsupervised bidirectional hierarchical clustering of microarray data was performed using Euclidean distance metric and centroid linkage clustering to group treatments and gene expression patterns by similarity. The clustering algorithms, heat map visualizations and centroid calculations were performed with Multi-Experiment Viewer software based on log2 expression ratio values. Functional enrichment statistics were determined with MetaCore (GeneGo) to identify the most significant canonical pathways (p<0.01) affected by TCDD and Cl-BBQ. The statistical scores in MetaCore are calculated using a hypergeometric distribution where the p-value represents the probability of a particular mapping arising by chance for experimental data compared to the background, which were all genes on the Affymetrix platform. To identify major transcriptional regulators, the statistical Interactome tool in MetaCore was used to measure interconnected genes in the experimental dataset relative to all known interactions in the background dataset. Statistical significance of overconnected interactions was calculated using a hypergeometric distribution (p<0.005). Integration of data between the Affymetrix Mouse Gene 1.0 ST array studies and Gene 2.0 array studies was performed in Bioinformatics Resource Manager web tool (http://cbb.pnnl.gov/brm/) using the merge data workflow [55].

In vivo bioluminescence imaging

In vivo imaging of mice was conducted daily using the IVIS Lumina II (Perkin Elmer) as previously described [41] with the following changes. Ten minutes after luciferin injection, bioluminescent images were taken with 120 second exposures. For full body imaging on the final day of the study, mice were sacrificed 8 minutes post luciferin injection. Individual organs were then excised and imaged 20 minutes post luciferin injection. Regions of interest were drawn around anatomical compartments and total photons were normalized to image acquisition time in order to obtain photons emitted/second.

Statistics

All statistical analyses were performed using Graphpad Prism. All bar graphs represent the mean ± SEM. Students t-test was performed to compare two treatment groups and one-way ANOVA with Tukey’s test for multiple comparisons, with p < 0.05 considered statistically significant.

Abbreviations

AhR

aryl hydrocarbon receptor

Treg cells

regulatory T cells

Tr1 cells

Type 1 regulatory T cells

AhR-Tr1 cells

aryl hydrocarbon receptor-induced type 1 regulatory T cells

allo-CTL

allospecific cytotoxic T lymphocytes

TF

transcription factor

LP

lamina propria