The aryl hydrocarbon receptor cell intrinsically promotes resident memory CD8⁺ T cell differentiation and function

SUMMARY

The Aryl hydrocarbon receptor (Ahr) regulates the differentiation and function of CD4⁺ T cells; however, its cell-intrinsic role in CD8⁺ T cells remains elusive. Herein we show that Ahr acts as a promoter of resident memory CD8⁺ T cell (T_RM) differentiation and function. Genetic ablation of Ahr in mouse CD8⁺ T cells leads to increased CD127⁻KLRG1⁺ short-lived effector cells and CD44⁺CD62L⁺ T central memory cells but reduced granzyme-B-producing CD69⁺CD103⁺ T_RM cells. Genome-wide analyses reveal that Ahr suppresses the circulating while promoting the resident memory core gene program. A tumor resident polyfunctional CD8⁺ T cell population, revealed by single-cell RNA-seq, is diminished upon Ahr deletion, compromising anti-tumor immunity. Human intestinal intraepithelial CD8⁺ T cells also highly express AHR that regulates in vitro T_RM differentiation and granzyme B production. Collectively, these data suggest that Ahr is an important cell-intrinsic factor for CD8⁺ T cell immunity.

In brief

Dean et al. show that Ahr, an environmental sensor, is important for the development and function of a specific immune cell type called tissue resident memory CD8⁺ T cells. These cells are critical in the response to local infections and tumors.

Graphical Abstract

INTRODUCTION

The continuous exposure to various commensals creates a unique “physiological inflammation state” in the gut. Recently, it has been shown that treating “steady-state” specific pathogen-free mice with an antiviral cocktail led to an intestinal intraepithelial (IEL)-specific reduction of most T cell subsets including CD8αβ T cells (hereinafter referred to as CD8⁺ T cells).¹ These IEL CD8⁺ T cells are CD69⁺CD103⁺ and known as T resident memory (T_RM) cells.² The lamina propria is home to a diverse population of other CD8⁺ T cell subsets including CD69⁺CD103⁺ CD8⁺ T cells.³ Following the recognition of a foreign commensal or pathogen, professional antigen-presenting cells migrate to draining lymph nodes and activate naive CD8⁺ T cells. Activation of naive CD8⁺ T cells results in clonal expansion and differentiation into diverse subsets of effector and memory cytotoxic T cell lymphocytes (CTLs).^4,5 This coordinated response helps to eliminate the pathogen and provide long-term protection.

CTLs vary in their phenotype, function, proliferative capacity, longevity, and ability to differentiate into cells conferring long-lived memory.⁶ Cells generated during the early phase of expansion likely have a spectrum of phenotypes arising via asymmetric cell division and/or effector cell de-differentiation.^7–10 Killer cell lectin-like receptor subfamily G, member 1 (KLRG1) and IL-7Rα (CD127), have been used as markers to distinguish effector CD8⁺ T cell subsets based on differences in migration, long-term survival, and memory differentiation potential. KLRG1⁺ cells, double-positive effector cells (CD127⁺KLRG1⁺, T_DPE), and short-lived effector cells (CD127⁻KLRG1⁺, T_SLE) cells, while being valuable functional effectors that produce cytolytic molecules such as perforin and granzyme B, at levels similar to other effector subsets,¹¹ have been described to exhibit limited memory cell differentiation potential.^12–14 In comparison, KLRG1⁻ cells, double-negative early effector cells (KLRG1⁻CD127⁻, T_EE) and memory precursor effector cells (KLRG1⁻CD127⁺, T_MPE) cells demonstrate better survival during the contraction phase and maintain stemness, as they are able to differentiate into multiple memory cell lineages.^13–15 This delineation is the canonical model for CD8⁺ T cell terminal effector vs. memory precursor differentiation; however, recent evidence suggests that there is more nuance and heterogeneity evident at this stage.^16–18

Memory T cell subsets that survive the contraction phase of infection vary based on their localization and function.¹⁹ T central memory (T_CM) cells express the lymph node homing receptors CCR7 and CD62L and have a high proliferative capacity but exhibit low cytotoxicity, while T effector memory (T_EM) cells are highly cytotoxic but are less proliferative, do not express CCR7 and CD62L, and therefore survey non-lymphoid tissues in the case of pathogen re-exposure.^20–22 In recent years, T_RM cells have gained significant attention because of their ability to remain in the tissue constantly surveilling in case of re-exposure to the same pathogen.^23–25 While the exact ontogeny of T_RM cells is not precisely known, the precursor cells poised to differentiate into T_RM have been suggested to resemble the classical T_MPE phenotype, and CD103 may function as a marker of this precursor population.^3,13,26 However, it has also been shown using a Klrg1-Cre fate mapping reporter mouse model that a number of cells that previously expressed KLRG1 and subsequently downregulate it (“ex-KLRG1” cells) give rise to all memory T cell lineages including T_RM cells.¹⁶ Runx3, Blimp1, and Hobbit are known transcriptional regulators of the T_RM lineage.²⁷ Runx3 is essential for the differentiation and long-term maintenance of CD8⁺ T_RM cells, while Blimp1 and Hobbit function to suppress Klf2, a transcriptional regulator important for T cell trafficking by promoting expression of S1P1 and CD62L and the T_CM cell lineage.^28–30 Upon secondary recognition of cognate antigen, memory cells are able to respond faster than a primary response, and it has recently been show that T_RM cells, after re-infection, can leave the tissue and contribute to the systemic immune responses by differentiating into a diverse pool of effector and memory cells.^31,32

The aryl hydrocarbon receptor (Ahr), a ligand-dependent environmental sensor and transcription factor, can be activated by xenobiotic compounds as well as natural ligands from diet and/or the microbiota.^33,34 Before its role in regulation of the immune system was appreciated, Ahr was shown to mediate the conversion and carcinogenic effects of environmental toxins like 2,3,7,8-tetrachloro-dibenzo-p-dioxin (TCDD). More recently, a physiological role for Ahr has emerged in the regulation of the immune system development and function via ligands from the microbiome, diet, and host cell metabolism. Compared with well-studied functions in CD4⁺ and CD8αα T cells,^33,35 the cell-intrinsic role and direct targets of Ahr in CD8⁺ T cell immune responses remain elusive.

RESULTS

Ahr is expressed by intestinal resident CD8αβ T cells

To more readily detect Ahr expression in different cell populations, we utilized Ahr^dCAIR mice,³⁶ in which a GFP reporter was knocked into the endogenous Ahr locus under the control of Ahr cis-acting regulatory elements. Comparing different tissues, we found that small intestine (SI) intraepithelial lymphocyte (IEL), as well as lamina propria lymphocyte (LPL) fractions in Ahr^dCAIR mice had the higher levels of GFP expression, indicating Ahr transcription, measured by flow cytometry in CD8⁺ T cells (Figures 1A and 1B). Examination of the different CD8⁺ T cells displaying markers of effector or memory subsets in the SI LPL (Figure S1A) showed that CD127⁻KLRG1⁺ short-lived effector CD8⁺ T cells (T_SLE) had low Ahr expression at a level similar to T_naive; however, CD127⁻KLRG1⁻ early effector CD8⁺ T cells (T_EE) and CD127⁺KLRG1⁻ memory precursor effector CD8⁺ T cells (T_MPE) had higher Ahr expression (Figure 1C). CD44⁺CD62L⁺ central memory CD8⁺ T cells (T_CM) had a low level of Ahr expression, while CD44⁺CD62L⁻ effector CD8⁺ T cells (T_EM) showed higher Ahr expression (Figure 1D). Furthermore, CD69⁺CD103⁺ CD8⁺ T cells (T_RM) exhibited the highest level of Ahr expression (Figure 1D). Of note, T_RM in the SI LPL and IEL had similar levels of Ahr expression (Figure S1B). However, in contrast to the LPL (Figure S1A), CD8⁺ T cells in IEL overwhelmingly consisted of the CD69⁺CD103⁺ T_RM population (>90%) (Figure S1G), consistent with higher expression of Ahr in IEL than in LPL (Figure 1B). Furthermore, we performed analysis of publicly available RNA-seq data of CD8⁺ T cell subsets following herpes simplex virus (HSV) infection in various tissues (GSE70813). Consistent with our data, the highest expression level of Ahr was detected in T_RM (CD69⁺CD103⁺), compared with those in T_CM (CD44⁺CD62L⁺), and T_EM (CD44⁺CD62L⁻) (Figure S1C). Together, these data suggest in CD8⁺ T cells Ahr is dynamically expressed along the T_RM differentiation pathway in the gut.

Figure 1. Ahr is expressed by intestinal resident CD8αβ T cells.

(A and B) Flow cytometry analysis of Ahr expression (GFP) in CD8⁺ T cells of Ahr^+/+ and Ahr^dCAIR/+ mice was performed. Histogram plot of GFP(Ahr) in CD8⁺ T cells isolated from spleen (Sp), peripheral lymph node (pLN), mesenteric lymph node (mLN), small intestine lamina propria (LPL), or intraepithelial lymphocytes (IEL) (A). Quantification of ΔgMFI in Ahr^dCAIR/+ compared with GFP-negative Ahr^+/+ mice (B).

(C and D) Flow cytometry quantification of ΔgMFI of Ahr-GFP in different CD8⁺ T cell populations isolated from LPL including CD44⁻CD62L⁺ (T_naive), CD127⁻KLRG1⁻ (T_EE), CD127⁻KLRG1⁺ (T_SLE), CD127⁺KLRG1⁻ (T_MPE) (C), and CD44⁺CD62L⁺ (T_CM), CD44⁺CD62L⁻ (T_EM), and CD69⁺CD103⁺(T_RM) (D). Data are shown as mean ± SEM (n = 3 mice per group). Data are representative of two independent experiments. See also Figure S1.

Ahr regulates intestinal CD8⁺ T cell compartment in a cell-intrinsic manner

To further investigate whether Ahr plays a role in the regulation of CD8⁺ T cell differentiation, we performed a bone marrow chimera experiment. Bone marrow from Ahr^+/+ CD45.1/.2 and Ahr^−/− CD45.2/.2 donor mice was transferred to half-lethally irradiated CD45.1/.1 mice and allowed to reconstitute for 2 months before flow cytometry analysis of spleen, SI LPL, and IEL (Figure S2A). The percentage of Ahr^−/−-derived cells was reduced in the SI IEL with no difference in the LPL or spleen (Figure S2B). This reduction in total lymphocytes was evidently caused by reduced CD8⁺ T cells, while CD4⁺ T cells showed no difference (Figures S2C and S2D), suggesting a role of Ahr in regulating CD8⁺ T cells in SI IEL. However, there was no difference in the protein level of Ki67 (Figure S2E) or annexin V (Figure S2F) expression in IEL Ahr^−/− CD8⁺ T cells examined directly ex vivo, suggesting that the observed decrease in these cells might be due to other mechanisms other than compromised cell proliferation or survival.

To further investigate the cell-intrinsic role of Ahr, we developed a genetic model (Ahr^f/fCd8^cre) to ablate Ahr expression specifically in CD8⁺ T cells but not in CD4⁺ T cells by Cre recombinase transgene under the control of the E8I CD8 enhancer region and the Cd8α promoter.³⁷ Ahr^f/fCd8^cre mice displayed efficient and specific deletion of Ahr in CD8β⁺ but not in CD8β⁻ cells (Figures S1D and S1E). Based on examination of different CD8⁺ T cell subsets in the LPL, we observed T_SLE (CD127⁻KLRG1⁺) were increased in the Ahr^f/fCd8^cre mice (Figure S1F, S2G, and S2H) and T_RM (CD69⁺CD103⁺) cells were modestly but significantly reduced in the SI IEL (Figures S1G, S2I, and S2J). These data suggest that although Ahr is most likely not a driver that directs the development of T_RM cells and/or their maintenance, it plays a cell-intrinsic role in promoting T_RM compartment at least at the steady state in the gut.

Ahr suppresses T_CM while promoting T_RM gene signature in intestinal CD8⁺ T cells

To determine the role of Ahr in regulation of SI epithelial resident CD8⁺ T cell compartment, we performed gene profiling analysis of CD8⁺ T cells from the IEL of Ahr^f/f and Ahr^f/fCd8^cre mice via RNA-seq. Principal component analysis and hierarchical clustering showed distinct separation between different genotypes (Figure S3A). 160 genes were differentially expressed (fragments per kilobase million (FPKM) ≥ 1, fold change ≥ 1.5, q ≤ 0.05) when comparing Ahr^f/f vs. Ahr^f/fCd8^cre mice with roughly equal upregulated (57%) and downregulated (43%) genes, with many being a regulator of or associated with CD8⁺ T cell memory subset fate (Figure 2A and S3B). Gene set enrichment analysis (GSEA) demonstrated a role for Ahr in promoting the expression of T_RM-associated genes while suppressing the expression of T_CM-associated genes (Figure 2B).²⁸ Notably, the magnitude in which the circulating core gene signature was enriched was greater than the enrichment for the resident core gene signature (Figures 2B and 2C). In the absence of Ahr, multiple transcription factors key to the regulation of CD8⁺ T cell memory development displayed a selectively differential expression pattern. The T_RM-promoting factors Prdm1 and Hic1^30,38,39 were downregulated, while the T_CM-promoting factors Eomes and Klf2^40,41 showed a trend of increase with no difference in Runx3, Zfp683 (Hobbit), and Tbx21 (T-bet) (Figure 2D). Consistently, at the protein level, in the CD69⁺CD103⁺ IEL resident CD8⁺ T cells of Ahr^f/fCd8^cre mice, Blimp-1 (encoded by Prdm1) was decreased (Figure S3C), while Eomes was increased (Figure S3D) compared with littermate Ahr^f/f control mice. Together, these transcriptomic changes are consistent with the observed T_RM defect and enhancement of T_CM cells in the gut of Ahr-deficient mice, suggesting a possibility that Ahr plays a role in CD8⁺ T cell fate decision.

Figure 2. Ahr suppresses T_CM while promoting T_RM gene signature in intestinal CD8⁺ T cells.

(A–E and H) RNA-seq analysis of IEL resident CD8⁺ T cells isolated from Ahr^f/f and Ahr^f/fCd8^cre mice and statistics (q values shown) were calculated via DESeq2 differential expression analysis, and FPKM values were quantified using RSEM. MA plot highlighting genes upregulated (red) and downregulated (blue) in Ahr^f/fCd8^cre compared with Ahr^f/f (A). Gene set enrichment analysis of resident and circulating core gene signatures (B). Heatmap depicting (fold change >1.5) genes enriched in respective signatures (C). FPKM expression values for transcriptional regulators (D) and secreted factors associated with cell function (E).

(F and G) Flow cytometry quantification of granzyme B protein levels of IEL resident CD8⁺ T cells isolated from Ahr^f/f and Ahr^f/fCd8^cre mice. The percentages (F) and total cell number (n = 5 mice per group) (G) are shown. Data are representative of two independent experiments.

(H) FPKM expression values for proliferation, cell cycle, and apoptosis genes.

(I–K) Naive CD8⁺ T cells were isolated from Ahr^+/+ and Ahr^−/− mice and then subjected to in vitro T_RM-like differentiation culture conditions. RNA was isolated for qRT-PCR expression analysis of Ahrr (I). Flow cytometry quantification of CD69⁺CD103⁺ in vitro T_RM-like cells (J) and CD44⁺CD62L⁺ in vitro T_CM-like population frequency (K). Data are shown as mean ± SEM (n = 3 mice per group). Data are representative of three independent experiments. See also Figures S1, S2, and S3.

TGFβ and IL-33 have been previously described to enhance in vitro CD69⁺CD103⁺ T_RM-like cell differentiation.^42,43 Of note, Ahr activation with 6-formylindolo(3.2b) carbazole (FICZ), confirmed by expression of the Ahr direct target gene Ahrr (Figure 2I), resulted in a significant enhancement of in vitro T_RM-like differentiation and reduced differentiation of the CD44⁺CD62L⁺ T_CM-like cells. These effects were abrogated in Ahr^−/− and Ahr^+/+ CH223191-treated (Ahr antagonist) CD8⁺ T cells (Figures 2J, 2K, S4B, and S4C), consistent with a role for Ahr in regulating T_RM and T_CM differentiation.

Granzymes, especially granzyme B, a well-described effector molecule secreted by cytotoxic lymphocytes to induce apoptosis in target cells,⁴⁴ were downregulated at the mRNA level in Ahr-deficient CD8⁺ cells, while other CD8⁺ T cell effector molecules including Tnf, Ifng, Prf1, and Ccl4, had no change (Figure 2E), suggesting a selective regulation of CD8⁺ T cell function in the intestine. Consistently, granzyme B, at the protein level on a per cell basis, was also reduced in Ahr-deficient IEL resident CD8⁺ T cells (Figures 2F and 2G) as well as in the bone marrow chimera model (Figure S4A).

Ahr directly binds DNA to regulate key T_RM genes

To elucidate the direct gene targets of Ahr in CD8⁺ T cells, we performed chromatin immunoprecipitation followed by sequencing (ChIP-seq) of Ahr using in vitro T_RM-like cells. 10,181 unique Ahr binding locations (peaks) were identified with the majority being located in the intron, intergenic, or promoter regions of genes (Figure 3A). Ahr binding at the promoter was enriched given that the relative abundance of promoter regions in the genome is estimated to be about 1%,⁴⁵ and 30% of Ahr peaks were located at the promoter (Figure 3A). Motif enrichment analysis (Figure 3B) showed the canonical Ahr:Arnt binding motif as the top hit. Other transcription factor motifs such as Runx2, ETS, Batf, IRF, and Stat5 were also found to be enriched, suggesting that Ahr might work together with these factors in a complex to bind DNA and regulate target gene transcription. Integration of the ChIP-seq and RNA-seq data by the software package Binding and Expression Target Analysis (BETA)⁴⁶ showed that Ahr functioned equally as a transcriptional activator and repressor (Figure 3C). This is different than in other cell types; for example, in ILC2s, Ahr mainly functions as a suppressor,⁴⁷ suggesting Ahr acts in a cell type-specific manner to regulate the expression of target genes. BETA direct target prediction identified Gzmb together with other T_RM-associated genes such as Itgae, Prdm1, and Hic1 as top Ahr-activated genes (Figure 3D), while T_CM-characteristic genes such as Klf2, Eomes, S1pr5, and Klrg1 were identified as Ahr-repressed genes (Figure 3D).

Figure 3. Ahr directly binds DNA to regulate key T_RM genes.

(A and B) ChIP-seq of Ahr was performed using in vitro T_RM-like CD8⁺ T cells. Analysis of Ahr binding was performed, and pie chart of peak annotation (A) as well as top seven enriched transcription factor motifs (B) are shown.

(C and D) Binding and Expression Target Analysis (BETA) was performed to integrate RNA-seq and ChIP-seq data. Visualization of transcription factor activating/repressive function prediction (C) and rank product volcano plot depicting top direct target candidates (D).

(E and F) Ahr^−/− in vitro T_RM-like CD8⁺ T cells transduced with retroviral constructs encoding MIG-EV (empty vector), MIG-Ahr, MIG-Y9A, or MIG-DbHLH. The cells were treated with DMSO or FICZ on day 3. On day 5, CD69 and CD103 expression was analyzed by flow cytometry (E), and RNA was isolated for qRT-PCR expression analysis of Ahr direct target gene Ahrr (F). Data are shown as mean ± SEM (n = 3 technical replicates per group). Data are representative of two independent experiments.

(G) Flow cytometry analysis of CD69 and CD103 expression in Ahr^−/− in vitro T_RM-like CD8⁺ T cells transduced with retroviral constructs encoding MIG-EV, MIG-Ahr, hCD2-EV, or hCD2-Blimp1. Data are representative of two independent experiments. See also Figure S5.

To determine whether Ahr DNA binding activity is required for in vitro T_RM-like differentiation, Ahr-deficient naive CD8⁺ T cells were transduced with empty vector (MIG-EV), wild-type Ahr (MIG-Ahr), Ahr DNA binding domain single amino acid mutant (MIG-Y9A),⁴⁸ or Ahr DNA binding domain deletion mutant (MIG-DbHLH)⁴⁹ and then differentiated to T_RM-like cells. Compared with MIG-EV control, MIG-Ahr transduction increased CD69⁺CD103⁺ in vitro T_RM-like differentiation, which was further enhanced by Ahr ligand FICZ treatment (Figure 3E) Both Ahr DNA binding mutants failed to upregulate Ahrr, an Ahr direct target gene^50,51 (Figure 3F), and exhibited markedly impaired CD69⁺CD103⁺ CD8⁺ T cell induction in T_RM-like skewing conditions (Figure 3E). These data suggest that Ahr acts in a DNA-binding-dependent manner to promote in vitro T_RM-like differentiation.

Given that Blimp1 (encoded by Prdm1), a well-known factor in promoting tissue residency,³⁰ was identified as an Ahr direct target (Figure 3D), we further transduced Ahr^−/− CD8⁺ T cells with hCD2-Blimp1. Notably, forced expression of Blimp1 resulted in a marked enhancement of in vitro T_RM-like differentiation similar to the MIG-Ahr transduced cells (Figures 3G and S5A). These data showed that restoration of Blimp1 levels was sufficient to rescue the observed in vitro T_RM-like differentiation defect in Ahr-deficient CD8⁺ T cells, suggesting that Blimp1 is one of the major downstream targets directly regulated by Ahr during T_RM differentiation. It is possible that the in vitro over-expression assay would not faithfully recapitulate the in vivo microenvironment and thus may not confer the whole T_RM gene expression program. Consistent with the regulation of CD69 and CD103 expression by Ahr, we analyzed Ahr ChIP-seq data and showed that Ahr directly bound to both the Cd69 and Itgae (CD103) gene loci (Figures S5B and S5C).

Ahr is a cell-intrinsic promoter of intestinal CD8 T_RM during infection

The reduction of CD69⁺CD103⁺ cells in the gut of Ahr-deficient mice under the steady state albeit significant is a modest reduction. Additionally, the results from the in vitro T_RM-like differentiation could be interpreted as Ahr regulating the expression of CD69 and CD103, not the T_RM population as a whole. Indeed, Ahr ChIP-seq analysis showed that Ahr directly bound to both the Cd69 and Itgae (CD103) gene loci (Figures S5B and S5C). Therefore, we sought to determine the role of Ahr in CD8⁺ T cell responses during infection to further validate these findings. To this end, we utilized a T_RM-dependent oral Listeria monocytogenes-ovalbumin (L.m.-OVA) infection model.³ The reasons for utilization of this model are 2-fold. It is an intestine-specific infection model that recapitulates human infection, allowing us to corroborate our findings on the role of Ahr in CD8⁺ T cells under the steady state of the gut; in addition, in combination with OVA-specific transgenic T cells (OTI) it allows for tracking mucosal CD8⁺ T cell responses in an antigen-specific manner. Host CD45.1/.1 mice were co-transferred with CD45.1/.2 Ahr^+/+ and CD45.2/.2 Ahr^−/− naive OTI CD8⁺ T cells before infection (Figure 4A). In the SI IEL, CD45.2/.2 Ahr^−/− OTI cells were significantly and gradually reduced (Figure 4B). At an effector time point (day 9 post infection), while there was no difference in their frequency in the IEL, Ahr^−/− OTI cells already exhibited an imbalanced memory precursor response. Specifically, CD127⁺ memory precursor populations were reduced in the Ahr^−/− OTI cells with a concomitant increase of CD127⁻KLRG1⁺ T_SLE (Figure 4C). Immature CD127⁻KLRG1⁻ T_EE was also increased in the Ahr^−/− OTI cells. The significance of the CD127/KLRG1 memory precursor vs. short-lived effector paradigm in the intestine is not well described. However, during the early stage of infection, CD69⁺CD103⁺ cells could be identified and are suggested to give rise to T_RM in the memory phase.³ Indeed, at this effector time point, Ahr^−/− cells had already shown a reduction in the CD69⁺CD103⁺ population with an increase in CD69⁻CD103⁻ cells (Figure 4D), consistent with defective CD8⁺ T_RM differentiation due to Ahr deficiency.

Figure 4. Ahr is a cell-intrinsic promoter of intestinal CD8⁺ T_RM during infection.

(A–F) Analysis of antigen-specific (OTI) CD8⁺ T cell response in the IEL during oral L.m.-OVA infection was performed as depicted in the schematic (A). Flow cytometry quantification of the percentage of Ahr^+/+ vs. Ahr^−/− OTI cells on day 9, 20, and 34 post infection (B). Flow cytometry quantification of memory precursor populations based on expression of CD127 and KLRG1 (C) as well as CD69 and CD103 (D) in Ahr^+/+ vs. Ahr^−/− OTI cells present in the IEL on day 9 post infection. Flow cytometry analysis of CD45.1 and GzmB gated on OTI cells depicting CD45.1⁺ Ahr^+/+(CD45.1/.2) and CD45.1⁻ Ahr^−/−(CD45.2/.2) OTI cells production of granzyme B on day 34 post infection (E). Quantification of granzyme B⁺ OTI T cells in the mice of indicated genotypes (F).

(G–J) L.m.-OVA re-infection was performed and Ahr^+/+ vs. Ahr^−/− OTI IEL resident CD8⁺ T cells analyzed on day 3 post re-infection. Flow cytometry analysis of CD45.1 and CD45.2 depicting percentage of Ahr^+/+(CD45.1/.2) and Ahr^−/−(CD45.2/.2) OTI IEL resident CD8⁺ T cells as well as (G) quantification of percentage are shown (H). Flow cytometry analysis (I) and quantification (J) of granzyme B production in analyzed cells. Data are compiled from two independent experiments and shown as mean ± SEM (n = 3–6 replicates per group). See also Figure S6.

In the spleen, we observed a similar frequency of OTI CD8⁺ T cells with or without Ahr; however, in MLNs there was an increase in Ahr-deficient OTI CD8⁺ T cells (Figures S6A and S6B), suggesting the priming of these cells is not impaired. Phenotypic analysis showed that in the absence of Ahr, T_CM-like (CD44⁺CD62L⁺) and T_MPE (CD127⁺KLRG1⁻) cells were increased in both the spleen and MLNs (Figures S6C–S6J), consistent with the notion that Ahr might function as a T_CM suppressor. Of note, in agreement with a previous report,³ we observed a higher frequency of T_MPE in the spleen in oral L.m.-OVA infection than those typically observed during systemic infection models. Additionally, while CD69⁺CD103⁺ cells were not present in the mesenteric lymph nodes (MLNs) at this time point, Ahr deficiency led to a decreased frequency of CD69⁺CD103⁻ cells that might be poised to become T_RM²⁶ (Figures S6K and S6L), suggesting that Ahr acts early to promote T_RM.

To further determine Ahr-mediated regulation of CD8⁺ T cell differentiation in the IEL at the early time point of infection, we performed RNA-seq analysis of Ahr^+/+ and Ahr^−/− OTI CD8⁺ T cells isolated from day 9 post L.m.-OVA infected mice. Consistent with our hypothesis that Ahr regulates the differentiation of T_RM, we observed, in the absence of Ahr, a suppression of T_RM signature gene expression (Figure S7A) with multiple genes overlapping with the steady state RNA-seq analysis (highlighted in red) including the Ahr direct target Tiparp and the T_RM master regulator Prdm1 (Figure S7B). Comparison of the two datasets, by principal component analysis, showed that upon addition of the day 9 post-infection data, the steady-state samples were no longer separated based on genotype, but for the infection dataset, Ahr^+/+ samples were clustered together and distinct from the Ahr^−/− samples, suggesting that infection magnifies the gene expression changes between Ahr^−/− and Ahr^+/+ CD8⁺ T cells that were observed under the steady state (Figure S7C). Furthermore, visualization of the transcriptome-wide fold change showed both in number and magnitude greater differential gene expression in the infection dataset compared with steady state (Figure S7D), suggesting that upon challenge with an infectious agent the impact of Ahr deficiency on CD8⁺ transcriptome is greater than that under the steady state. Consistently, the magnitude of the Residency Core Gene Signature enrichment was larger, and the number of differential genes was more compared with the steady state dataset (Figures 2B and 2C vs. S7A and S7B). Collectively, these data suggested that Ahr might play a more prominent role in shaping T_RM gene program during infection than it does at steady-state “physiological inflammation” in the gut.

At the memory phase of infection, bona fide T_RM is the sole CD8⁺ T cell subset that remains in the IEL by day 34 post infection.³ Given the reduction of T_RM precursor cells observed at day 9, we hypothesized that Ahr-deficient OTI cells would be reduced at this memory time point. Indeed, in the IEL, CD45.2/.2 Ahr^−/− OTI cells were significantly reduced at day 34 (Figure 4A). Additionally, CD45.2/.2 Ahr^−/− cells produced less granzyme B on a per cell basis compared with CD45.1/.2 Ahr^+/+ OTI cells (Figures 4E and 4F). Also, at this time point and consistent with the decreased expression of Prdm1 in Ahr-deficient CD8⁺ T cells analyzed by RNA-seq, IEL OTI cells of Ahr^−/− origin exhibited lower levels of Blimp1 staining compared with Ahr^+/+ cells (Figures S7E and S7F). Collectively these data suggest a critical role of Ahr in promoting T_RM cells and their function during infection potentially through regulation of Blimp1 expression.

To determine if the observed reduction of Ahr^−/− CD8⁺ T_RM cells also was present after a secondary recall challenge, following primary infection, mice were re-infected with L.m.-OVA, and then 3 days post re-infection, mice were sacrificed, and analysis of OTI-specific T cell responses was performed (Figure 4A). Consistent with results from the primary infection, in the IEL, CD45.2/.2 Ahr^−/− OTI cells were significantly reduced (Figures 4G and 4H) and produced less granzyme B (Figures 4I and 4J). However, in the spleen there was no difference in granzyme B production (Figures S7G and S7H), while Ahr^−/− OTI cells from the MLNs showed a trend of decreased granzyme B (Figures S7I and S7J), suggesting that Ahr regulation of CD8⁺ T cell function might be specific to the gut. Together, these results demonstrate that Ahr is a cell-intrinsic promoter of long-lived, functional T_RM cells during intestinal infection.

Single-cell RNA-seq identifies polyfunctional CD8⁺ T cell population dependent on Ahr

CD8⁺ tumor-infiltrating lymphocytes (TILs) displaying a T_RM-like phenotype have been well described to be critical to the orchestration of potent protective anti-tumor response.^52,53 Therefore, we next investigated the role of Ahr in CD8⁺ T cells in tumor immunity. Previous literature suggests that Ahr can function as either a tumor promoter or suppressor dependent of the context.^54,55 However, these studies did not address the cell-intrinsic role of Ahr in the tumor, in which perturbation of Ahr in different cell types such as cancer cells or immune cells would inevitably have different outcomes of tumorigenesis. To elucidate the role of Ahr in tumor immunity, B16F10 melanoma cells were injected subcutaneously into Ahr^−/− mice with Ahr germline deletion and into their littermate wild-type Ahr^+/+ mice. CD45⁺ TILs were purified and subjected to high-throughput single-cell RNA-seq using the 10X Genomics platform. 28 different clusters were identified in both groups with no marked differences between groups. (Figures S8A–S8C).

Given the crucial role of CD8⁺ T cells in anti-tumor immunity,^28,53 we decided to focus on the impact of Ahr on CD8⁺ T cells in the tumor microenvironment. Particularly of interest are CD8⁺ TILs labeled “polyfunctional” that simultaneously produce multiple effector molecules including cytokines, chemokines, and cytotoxic granules. Subsetting on the CD8⁺ T cell clusters and re-performing dimensionality reduction, clustering, marker gene finding, and visualization resulted in six unique clusters with visual differences between Ahr^+/+ and Ahr^−/− (Figures 5A and S8D). Ahr-deficient CD8⁺ T cells were decreased in the polyfunctional and activated (Ifit1⁺ and Il1b⁺) subsets while having an increase in the cluster with a circulating memory-like phenotype (Figure 5A right). Pseudotime analysis by Monocle3 showed that Ahr-deficient CD8⁺ T cells were less advanced in the pseudotime differentiation trajectory (Figures 5B and 5C), consistent with a role for Ahr in the regulation of CD8⁺ T cell differentiation. Differential gene expression analysis showed a reduction of genes critical for CD8⁺ T effector function (i.e., Gzmb, Prf1, Ifng, Ccl4, and Xcl1), residency-associated membrane receptors (i.e., Icos and Cd69), and transcriptional modulators (i.e., Smad7, Runx2, and Hif1a) (Figure 5D). Collectively, these results indicate that Ahr is critical to the development of highly activated and polyfunctional CD8⁺ TILs.

Figure 5. Single-cell RNA-seq identifies polyfunctional CD8⁺ T cell population dependent on Ahr.

(A–D) scRNA-seq analysis of Ahr^+/+ (n = 2) vs. Ahr^−/− (n = 2) TIL CD8⁺ T cells was performed. UMAP dimensionality reduction and cluster visualization (left) as well as pie chart frequency depiction (right) color-coded to represent cluster ID (A). Pseudotime visualization (B) and quantification (C). Differential gene expression depicted via color intensity as average expression and circle size as percent expressed in CD8⁺ T cells (D).

(E–H) Ahr^f/f and Ahr^f/fCd8^cre mice were inoculated subcutaneously with B16F10 mouse melanoma and tumor size monitored (E). At endpoint, tumor weight was quantified (F), and TILs were isolated for flow cytometry analysis. Pie chart visualization depicting polyfunctionality of TIL CD8⁺ T cells in mice with indicated genotypes (G). Polyfunctionality score quantification (triple-positive plus double-positive minus triple-negative divided by total cells) of TIL CD8⁺ T cells isolated from Ahr^f/f and Ahr^f/fCd8^cre tumor-bearing mice (H). Data are compiled from two independent experiments and are shown as mean ± SEM (n = 6 to 7 replicates per group). See also Figure S7.

To understand the CD8⁺ T cell-intrinsic role of Ahr in tumor immunity, Ahr^f/fCd8^cre mice were challenged with the aforementioned B16F10 melanoma model and exhibited significantly enhanced tumor growth (Figure 5E) and burden at endpoint (Figure 5F) compared with controls. While there were no differences in their frequency (Figure S8E), CD8⁺ TILs of Ahr^f/fCd8^cre mice were less polyfunctional and produced less granzyme B, IFNγ, and TNF, on a per cell basis. Triple-positive cells (granzyme B⁺IFNγ⁺TNF⁺) were reduced in the Ahr^f/fCd8^cre mice, while triple-negative cells (granzyme B⁻IFNγ⁻TNF⁻) were increased in the Ahr^f/fCd8^cre mice (Figures 5G, 5H, S8E, and S8F). Given the low immunogenicity of the B16F10 model, we sought to validate these findings using the MC38 murine colon carcinoma model. Consistently, Ahr^f/fCd8^cre mice exhibited significantly enhanced tumor growth (Figure S9A) and burden at endpoint (Figure S9B) compared with littermate controls. Furthermore, CD8⁺ TILs of Ahr^f/fCd8^cre mice produced less granzyme B, IFNγ, and TNF on a per cell basis (Figures S9C and S9D) and were less polyfunctional (Figure S9E) in the MC38 tumor model as well. Collectively, these results suggested that Ahr promotes polyfunctional CD8⁺ TILs in a cell-intrinsic manner.

AHR signaling promotes T_RM differentiation and function in human CD8⁺ T cells

To determine the role of AHR in human CD8⁺ T cells, we next examined the expression of AHR in human intestine resident CD8⁺ T cells. Compared with blood circulating CD45RA⁻CD45RO⁺ CD8⁺ T cells, virtually all IEL resident counterparts were CD69⁺CD103⁺ (Figure 6A). As expected, human IEL resident CD8⁺ T cells expressed less T-bet,⁵⁶ but more AHR (Figures 6B and 6C), consistent with the observation of enrichment of Ahr expression in mouse IEL CD8⁺ T cells. Next, we performed human in vitro T_RM-like differentiation using TGFβ and IL-33. Treatment with CH223191, a specific Ahr antagonist,⁵¹ caused a reduction in the frequency of CD103⁺ cells, whereas FICZ enhanced this population (Figure 6D). Although FICZ or CH223191 treatment did not significantly impact AHR expression, TGFβ and IL-33 treatment enhanced AHR expression in human CD8⁺ T cells (Figure 6E). As expected, FICZ treatment increased expression of the AHR target gene AHR, while CH223191 suppressed its expression. Notably, FICZ also significantly augmented the expression of the effector gene GZMB (Figures 6F and 6G). Together, these results demonstrate an evolutionarily conserved function of AHR in human CD8⁺ T cells.

Figure 6. AHR signaling promotes T_RM differentiation and function in human CD8 T cells.

(A–C) Human peripheral blood (PBMC) vs. IEL CD8⁺ T cells were analyzed via flow cytometry. Staining of CD45RA and CD45RO (top), CD103 and CD69 (bottom) (A), as well as T-BET and AHR (B) in tissue fractions as indicated in the figure. Quantification of AHR protein levels (gMFI) in human PBMC, LPL, and IEL CD8⁺ T cells (C). Data are compiled from three independent experiments and shown as mean ± SEM (n = 4 replicates per group).

(D–G) Human naive CD8⁺ T cells were isolated from PBMCs and then subjected to in vitro T_RM-like differentiation culture conditions. The cells were given differentiation cytokines and treated with DMSO (control), FICZ, or CH223191 on day 2. The assay was collected on day 4, and flow cytometry quantification of CD103 expression was performed (D). Data are compiled from three independent experiments and are shown as mean ± SEM (n = 4 replicates per group). RNA was isolated for qRT-PCR expression analysis of AHR (E), AHR direct target gene AHRR (F), and GZMB (G). Data are shown as mean ± SEM (n = 3 technical replicates per group). Data are representative of three independent experiments.

DISCUSSION

Proper regulation of CD8⁺ T cell memory development is critical to long-term protection against morbid infection. Multiple transcription factors have been suggested to contribute to the regulation of this process; however, data describing their cell-intrinsic effects are lacking. A role for Ahr supporting T_RM has been previously described.⁵⁷ Ahr expression was shown to be higher in skin compared with spleen CD8⁺ T cells, and Ahr functions to facilitate T_RM persistence in the epidermis.⁵⁷ Our study provided mechanistic insights into how Ahr regulates the transcriptional programs of CD8⁺ T cells and elucidated a role for Ahr in early T_RM differentiation, specifically in a gut-specific infection model.

Other studies, using an influenza infection model, demonstrated that early-life Ahr activation via TCDD impairs the priming of virus-specific CTLs; however, this phenotype is mediated by regulation of DNA methylation and found to be cell extrinsic.^58,59 Of additional importance, the cellular toxicity and long half-life of TCDD could lead to non-physiological activation of Ahr, thus complicating data interpretation. Given that Ahr is expressed in various cell types and plays a complex role, we ablated Ahr expression specifically in CD8⁺ T cells. We noted that differences in T_RM under the steady state between Ahr-deficient mice and their littermate controls were modest, presumably due to the lack of antigen-specific pathogen exposure and non-competitive environment. Although our data suggest that Ahr plays a cell-intrinsic role in CD8⁺ T cell differentiation and function, we opted to test this hypothesis in the context of gut infection as well. In the intestine, on day 9 post oral infection, we showed that Ahr-deficient CD8⁺ T cells exhibited a reduction of T_RM precursors with a concomitant increase in T_SLE and T_CM-like cells compared with wild-type counterparts. This decrease in precursors led to a marked reduction in the IEL resident T_RM cell compartment at the memory time point of day 34 post infection. Of note, for the gut infection model Ahr-deficient OTI CD8⁺ T cells responded to infection in the presence of an equal number of wild-type OTI CD8⁺ T cells. This allows for a well-controlled experiment because both are exposed to the same environment and stimuli; however, the magnitude of the observed phenotype could be due to competition. Furthermore, we showed that Ahr functions to promote in vitro T_RM-like cell differentiation while suppressing the T_CM phenotype. Collectively, these data prompted us to conclude that differentiation at least is one of the key mechanisms underlying the defective CD8⁺ T cell memory development that occurs in the absence of Ahr.

A previous report has shown Ahr is a critical regulator of monocyte differentiation acting to promote dendritic cell differentiation while suppressing macrophage differentiation.⁶⁰ In the current study, we demonstrate that in CD8⁺ T cells, Ahr acts as a cell fate decision regulator to suppress the differentiation of T_CM while promoting T_RM. Blimp1 is a well-described positive regulator of T_RM differentiation and function.²⁹ We showed that Prdm1 (Blimp1) is one of the top Ahr-regulated direct target genes in CD8⁺ T cells, consistent with the role of Ahr in promoting Prdm1 expression in monocytes.⁶⁰ We further showed that forced expression of Blimp1 was able to rescue in vitro T_RM-like differentiation defect in the absence of Ahr, suggesting that Blimp1 acts downstream of Ahr to regulate CD8⁺ T cell memory development. It is important to note that Ahr^−/− CD8⁺ T cells were still able to differentiate into CD69⁺CD103⁺ cells in vitro and in vivo; therefore, Ahr is likely acting as a promoter rather than a main driver of T_RM and in vitro T_RM-like differentiation.

Recent publications describe that Ahr promotes CD8⁺ T cell exhaustion in tumor immunity.^61,62 However, activated T cells as well as tissue-resident memory cells express many of the exhaustion-associated molecules (PD-1, Tim-3, CTLA-4, etc.).^28,56,63 It is also important to keep in mind that interpretation of the role of Ahr in tumor immunity with ligand treatment has to consider both cell-autonomous and non-autonomous effects, given the cell-specific role of Ahr in different immune cell types that infiltrate the tumor microenvironment, and in the tumor cells as well. It has been reported that when T_RM responses are diminished, polyfunctional CD8⁺ TILs are also reduced,^64,65 consistent with our findings. These polyfunctional tumor-resident memory-like cells have been positively associated with productive anti-tumor immunity.^66–68

Activation of the Ahr pathway in CD8⁺ T cells may be of therapeutic interest in the context of infection or cancer immunotherapy when a long-lived tissue-resident memory response is of utmost importance.

Limitations of the study

Our current study represents a serious effort to delineate the cell-intrinsic role of Ahr in CD8⁺ T cells, specifically in the gut. However, future efforts to understand the role of Ahr in CD8⁺ T cells in other infection models including skin infection with HSV and systemic lymphocytic choriomeningitis virus (LCMV) infection would help extend our understanding of this biology. Although no difference in the proliferation or cell death was observed in the intestinal Ahr-deficient T_RM cells ex vivo, whether Ahr also regulates T_RM maintenance in the intestine remains to be carefully determined. Future investigation is needed to elucidate the cell-intrinsic role of Ahr in regulation of T_RM maintenance, for example by development of CD8-specific inducible deletion of Ahr mouse model to examine T_RM and T_RM-like cells in various tissues under the steady state and during infection. Additionally, it is possible that Ahr deficiency could lead to decreased T_RM due to aberrant re-entering into circulation after having been in the tissue. In the future, more careful analysis of the recirculation of T_RM in the context of Ahr deficiency needs to be done through T_RM fate mapping mice as previously described.³²

STAR★METHODS

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead contact, Liang Zhou (liangzhou497@ufl.edu).

Materials availability

This study did not generate new unique reagents.

Data and code availability

All data supporting the findings of this study are available within the article and its supplementary information and from the corresponding author upon reasonable request. The accession number of the RNA-Seq, ChIP-Seq, and scRNA-Seq data reported in this paper have been uploaded to GEO at the accession number: GSE220944. This paper does not report original code. Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Human sample collection

Buffy coats from male healthy donors were purchased from Life South Community Blood Centers in accordance with the Institutional Review Boards at the University of Florida (UF IRB #201801563) and PBMCs isolated as described previously.⁵¹ Patients (both male and female) diagnosed with an IBD [Crohn disease (CD) and UC] were recruited for this study. Patients were at least 18 years old and recruited during clinical visits to the Inflammatory Bowel and Celiac Disease Program or before a previously scheduled colonoscopy. Tissue biopsies were taken from the colon. The mucosa was assessed based on frequently used scoring systems (Mayo Score for UC and Simple Endoscopic Score for Crohn Disease for CD). Biopsies collected were labeled as “healthy” based on normal vascularity and lack the presence of edema, erythema, friability, erosions, or ulcerations macroscopically. Intestinal tissue biopsies were collected in complete RPMI and then immediately processed to isolate lamina propria lymphocytes (LPLs) and intraepithelial lymphocytes (IELs) as described previously.⁶⁹

Mice

Mice used in this study were maintained in specific-pathogen-free (SPF) conditions at the University of Florida. All mouse studies were approved by the Institutional Animal Care and Use Committees of the University of Florida. Littermate controls as well as both male and female mice were used for experiments. Mice were used at 8 to 10-week-old age unless otherwise noted. Ahr^−/−,⁷⁰ Ahr^dCAIR 36, Ahr^f/f 51 were published previously. EI8Cre mice, and OTI TCR transgenic mice were purchased from Jackson Laboratory then crossed with Ahr^f/f and Ahr^−/− mice, respectively.

METHOD DETAILS

Lymphocyte isolation and flow cytometry

The isolation of lymphocytes from spleen, peripheral/mesenteric lymph nodes, small and large intestinal lamina propria, and small intestine intraepithelial lymphocytes were performed as previously described.⁷¹ After digestion, cells were further purified from the interphase of 37.5% and 75% Percoll gradient after 20 min spin at 2,500 rpm at room temperature. For flow cytometry analysis, cells were stained using Live and Dead violet viability kit (Invitrogen) or Zombie Aqua fixable viability kit (BioLegend). CD16/32 antibody (Thermo Fisher) was used to block the nonspecific binding followed by surface molecule staining on ice for 30 min. Cells were fixed and permeabilized with Foxp3 staining buffer Kit (eBioscience) for transcription factor staining. For cytokine staining, cells were stimulated with 50 ng/mL PMA and 500 ng/mL ionomycin for 3 h and Brefeldin A (2 μg/mL) was added 2 h before cells were collected. Sample acquisition was performed on BD FACSCantoII or Cytek Aurora flow cytometer and analyzed with FlowJo software (version 10.2).

Adoptive transfers and infection

The spleens of CD45.1/.2 Ahr^+/+ OT-I and CD45.2/.2 Ahr^−/− OT-I were isolated and single-cell suspensions were generated by mechanical disruption. Naive CD8 (Tcrb⁺CD8a⁺CD44^lo) OT-I T cells were first enriched using mouse naive CD8⁺ T cell isolation kit (Stemcell Technologies) then further purified using a Sony sorter SH800. Post-sort purity was routinely analyzed and higher than 97%. Purified cells were then mixed in a ratio of 1:1 and 5 × 10⁴ cells were co-transferred 1 day before infection. Mice were infected orally by feeding 2.5 × 10⁹ (primary infection) colony-forming units (CFUs) of L.m.-OVA InlAM (kindly provided by B. Sheridan, Stony Brook University) as described previously.³ At the indicated time points after infection, mice were sacrificed, and tissues were collected for analysis of OT-I T cell responses. For experiments investigating the recall response, following primary infection (>30 day later) mice were re-infected with 2.5 × 10¹⁰ CFU of L.m.-OVA InlAM. Then, 3 days later mice were sacrificed, and tissues were processed for analysis of OT-I T cell responses.

In vitro T_RM differentiation

CD8 T_RM cells were differentiated in vitro as described previously.⁴² Briefly, naive CD8⁺ T cells were purified from splenocytes using mouse naive CD8⁺ T cell isolation kit (Stemcell Technologies) then cultured in RPMI-1640 medium (plus β-mercaptoethanol) supplemented with 10% fetal bovine serum, 1% L-glutamine, 1% penicillin-streptomycin (cRPMI). 100 ng/mL IL-2 and anti-CD3/CD28 coated Dynabeads (Thermo Fisher) at 1:1 cell to bead ratio were added and cells incubated for 48 h. Then media volume was doubled, each well split into a second well, and 5 ng/mL TGF-β and 100 ng/mL IL-33 were added and incubated for another 48 h before harvest for flow cytometry and RT-qPCR analysis.

For human cells, naive CD8⁺ T cells were purified from PBMCs using human naive CD8⁺ T cell isolation kit (Stemcell Technologies) then cultured in cRPMI and treated as above with the human equivalent Dynabeads and cytokines. FICZ and CH223191 were added at a concentration of 200 nM and 1μM, respectively, as indicated in the text.

Quantitative PCR

Total RNA was isolated with Trizol reagent (Invitrogen). cDNA was synthesized by GoScript^™ Reverse Transcription Kit (Promega). Real-time PCR was performed using SYBR Green (Biorad) and different primer sets (Table S1). Reactions were run using the CFX Connect Real-Time PCR Detection System (Biorad). Each specific gene expression was normalized to β-actin expression.

RNA-seq and ChIP-Seq assay and analyses

For RNA-Seq analyses of IEL resident CD8αβ and CD8αα T cells, 2×10³ sorted CD8αβ and CD8αα T cells from the small intestine IEL of control or Ahr^f/fCd8^cre littermate mice were used. RNA was isolated by RNAeasy Micro Kit. cDNA generation was performed with SMART-Seq^® HT Kit (Takara). Sequencing libraries were generated with Nextera^® XT DNA Library Preparation Kit (Illumina). Libraries were sequenced on an Illumina HiSeq 4000 instrument to produce 50 bp single-end reads. The read mapping and mRNA quantification were performed as previously described.⁷² Briefly, FastQC was used to ensure high per-base sequence quality of reads. Sequenced reads were mapped and raw count values quantified with STAR⁷³ to the Mus musculus genome (GRCm38/mm10 assembly). RSEM⁷⁴ was used to quantify mRNA expression levels, FPKM aligned reads. Differentially expressed genes (max FPKM ≥1, fold change ≥1.5, q-value ≤0.05) were identified by DESeq2⁷⁵ analysis. Gene Set Enrichment Analysis (GSEA)^76,77 were performed using the circulating and resident gene signatures developed previously.²⁸ Log2 transformed FPKM values were used for principal component analysis in R⁷⁸ with the prcomp function and then visualized using the rgl⁷⁹ package. Heatmaps were created using the R package pheatmap.⁸⁰

For ChIP-Seq analyses of in vitro T_RM, cells were differentiated from naive CD8⁺ T cells as described above. Cells were then treated with FICZ (200 nM) for 4 h before harvest. Afterward, cells were collected and cross-linked with 1% formaldehyde for 15 min. Chromatin was sheared by sonication with Bioruptor Pico (30″ on and 30″ off for 25 cycles) and immunoprecipitated with anti-Ahr antibody (Enzo Life Science) using iDeal ChIP-Seq Kit for transcription factors or True MicroChIP Kit (Diagenode). Eluted DNA was used to generate an indexed library according to the manual of NEBNext Ultra II DNA Library Prep Kit for Illumina (NEB). The library was cleaned up with 1.2x SPRIselect beads (Beckman Coulter) before sequencing on an Illumina HiSeq 4000 instrument to produce 50 bp single-end reads. The read mapping and ChIP-Seq analysis were performed as previously described.⁵¹ Briefly, FastQC was used to ensure high per-base sequence quality of reads. Then ChIP-Seq reads were mapped to the mouse genome (GRCm38/mm10 assembly) with bowtie2 (v2.3.3)⁸¹ and further filtered using samtools (v1.7).⁸² The uniquely aligned reads were used to generate bedgraph files (scaled to 10 million reads) using bedtools (v2.25.0)⁸³ and then were uploaded to UCSC genome browser for visualization. ChIP-Seq peak finding, Motif enrichment analysis, and peak annotation were performed using Homer⁸⁴ with default parameters.

Retroviral transduction

Human embryonic kidney (HEK) 293T cells were transfected with retroviral plasmids and the packaging plasmid 10A1 using polyethyleneimine (PEI). Viral supernatant was collected after transfection. CD8⁺ T cells were isolated and cultured as described above. After 24 h of culture, retrovirus-containing supernatants supplemented with polybrene (8 μg/mL, Sigma-Aldrich) were added to the cells followed by centrifugation at 2500 rpm for 2 h at 32°C on days 1 and 2. The cells were further cultured under in vitro T_RM-like differentiation conditions before harvest as indicated in the text.

Plasmids

cDNA of mouse Ahr (1–805 amino acids) was cloned into MIG with hemagglutinin (HA) at the N terminus. For the Ahr Y9A (1–805 amino acids, Y9A), AhrΔbHLH (1–120 amino acids, were deleted) and subcloned into MIG with HA tag. Blimp1-hCD2 plasmid was kindly provided by Dr. Weishan Huang (Louisiana State University).

Tumor model

For the melanoma and MC38 tumor models, mice were injected in the right flank subcutaneously with B16-F10 or MC38 (5 × 10⁵) cells then tumor growth was monitored every other day with a digital caliper. At endpoint, tumors were excised and cut with scissors into <2mm pieces then incubated with collagenase and DNase I for 30 min at 37°C in a shaking incubator. Samples were passed through a 100-μm filter, tumor infiltrating lymphocytes (TILs) were isolated by interphase collection after a 40% and 80% Percoll gradient centrifugation and then analyzed via flow cytometry as described above.

Single cell RNA-Seq analyses

Live CD45⁺ total TILs were isolated and purified by FACS sorting from a total of 4 mice (2 Ahr^+/+ and 2 Ahr^−/−) before being loaded on the chromium controller aiming for a recovery of 10,000 cells. Single Cell 30 reagent kit v3.1 was used for reverse transcription, cDNA amplification, and library construction of gene expression libraries (10x Genomics) according to the manufacturer. For primary scRNA-Seq analysis: alignment, quantification, and quality control were performed using the Cell Ranger Software and default parameters. After quality control and removal of dead cells, doublets, and contaminating melanoma cells (Pmel mRNA count >0) a total of 9726 cells remained for downstream analysis. Normalization, cell clustering (resolution = 2), dimensionality reduction, differential expression, visualization, and pseudotime analysis were performed using the R packages Seurat⁸⁵ and Monocle.⁸⁶ The 3000 most variable features identified with the variance-stabilizing transformation (vst)-method were used for principal component analysis (PCA). Upon inspection of elbowplot and jackstraw plot, the first 20 principal components (PCs) were used for further analysis. Cell type annotation was performed manually by assessment of cluster marker genes.

QUANTIFICATION AND STATISTICAL ANALYSIS

All data are represented as mean ± SEM and have at least n = 3 per group from at least 2 independent experiments (refer to figure legend to detailed information). Unless otherwise noted, statistical analysis was performed with the unpaired Student’s t-test or one-way ANOVA with Dunnett’s correction for multiple comparisons. For analysis of tumor growth kinetics, linear regression was performed to test if the slopes are significantly different between groups. Statistical analyses were run using GraphPad Prism 8 software package. p values were indicated with asterisks (*p % 0.05; **p % 0.01; ***p % 0.001; ****p % 0.0001).

KEY RESOURCES TABLE.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Anti-mouse CD45.2 – FITC (clone 104)	TONBO biosciences	Cat# 35-0454-U500; RRID: AB_2621692
Anti-mouse CD45.1 – APC/Cyanine7 (clone A20)	BioLegend	Cat# 110716; RRID: AB_313505
Anti-mouse TCRβ – Brilliant Violet 650 (clone H57-597)	BioLegend	Cat# 109251; RRID: AB_2810348
Anti-mouse TCRβ – eFluor 450 (clone H57-597)	Invitrogen	Cat# 48-5961-82; RRID: AB_11039532
Anti-mouse CD3e – FITC (clone 145-2C11)	TONBO biosciences	Cat# 35-0031-U500; RRID: AB_2621659
Anti-mouse CD3 – APC-eFluor 780 (clone 17A2)	Invitrogen	Cat# 47-0032-82; RRID: AB_1272217
Anti-mouse CD8b – FITC (clone YTS156.7.7)	BioLegend	Cat# 126606; RRID: AB_961295
Anti-mouse CD8b – Brilliant Violet 711 (clone YTS156.7.7)	BioLegend	Cat# 126633; RRID: AB_2800622
Anti-mouse CD8a – APC/Cyanine7 (clone 53-6.7)	BioLegend	Cat# 100714; RRID: AB_312753
Anti-mouse CD8a – Brilliant Violet 605 (clone 53–6.7)	BioLegend	Cat# 100744; RRID: AB_2562609
Anti-mouse/human CD44 – Brilliant Violet 421 (clone IM7)	BioLegend	Cat# 103040; RRID: AB_2616903
Anti-mouse CD62L – PE-Cyanine7 (clone MEL-14)	Invitrogen	Cat# 25-0621-82; RRID: AB_469633
Anti-mouse CD127 – PE/Dazzle 594 (clone A7R34)	BioLegend	Cat# 135032; RRID: AB_2564217
Anti-mouse KLRG1 – PerCP-eFluor 710 (clone 2F1)	Invitrogen	Cat# 46-5893-82; RRID: AB_10670282
Anti-mouse CD69 – Brilliant Violet 711 (clone H1.2F3)	BioLegend	Cat# 104537; RRID: AB_2566120
Anti-mouse CD69 – PE (clone H1.2F3)	Invitrogen	Cat# 12-0691-82; RRID: AB_465732
Anti-mouse CD103 – Brilliant Violet 785 (clone 2E7)	BioLegend	Cat# 121439; RRID: AB_2800588
Anti-mouse CD103 – APC (clone 2E7)	Invitrogen	Cat# 17-1031-82; RRID: AB_1106992
Anti-human/mouse Granzyme B – Pacific Blue (clone GB11)	BioLegend	Cat# 515408; RRID: AB_2562196
Anti-human/mouse Granzyme B – Alexa Fluor 700 (clone QA16A02)	BioLegend	Cat# 372222; RRID: AB_2728389
Anti-mouse IFN gamma – PerCP-Cyanine5.5 (clone XMG1.2)	Invitrogen	Cat# 45-7311-82; RRID: AB_1107020
Anti-mouse TNF – PE-Cyanine7 (clone MP6-XT22)	Invitrogen	Cat# 25-7321-82; RRID: AB_11042728
Anti-mouse Ki-67 – Brilliant Violet 605 (clone 16A8)	BioLegend	Cat# 652413; RRID: AB_2562664
Anti-mouse Ki-67 – Alexa Fluor 647 (clone B56)	BD	Cat# 558615; RRID: AB_647130
Anti-mouse Eomes – PE-Cyanine7 (clone Dan11mag)	Invitrogen	Cat# 25-4875-82; RRID: AB_2573454
Anti-mouse Ahr – eFluor 660 (clone 4MEJJ)	Invitrogen	Cat# 50-5925-82; RRID: AB_2574255
Anti-mouse Ahr – Alexa Fluor 488 (clone 4MEJJ)	Invitrogen	Cat# 53-5925-82; RRID: AB_2574425
Anti-mouse Blimp1 – APC (clone 5E7)	BioLegend	Cat# 150007; RRID: AB_2728186
Anti-Annexin V – APC	eBioscience	Cat# 17-8007-74
Anti-Annexin V – PE/Dazzle 594	BioLegend	Cat# 640955
Anti-human CD3 – FITC (clone SK7)	TONBO biosciences	Cat# 35-0036-T025; RRID: AB_2621661
Anti-human CD8 – APC-Cyanine7 (clone SK1)	TONBO biosciences	Cat# 25-0087-T025; RRID: AB_2848136
Anti-human CD45RA – PerCP-Cyanine5.5 (clone HI100)	TONBO biosciences	Cat# 65-0458-T100; RRID: AB_2621896
Anti-human CD45RO – Brilliant Violet 605 (clone UCHL1)	BioLegend	Cat# 304237; RRID: AB_2562143
Anti-human CD197 – Brilliant Violet 711 (clone G043H7)	BioLegend	Cat# 353227; RRID: AB_11219587
Anti-human CD127 – PE/Dazzle 594 (clone A019D5)	BioLegend	Cat# 351335; RRID: AB_2563636
Anti-human CD69 – Brilliant Violet 785 (clone FN50)	BioLegend	Cat# 310931; RRID: AB_2561370
Anti-human CD103 – APC (clone Ber-ACT8)	BioLegend	Cat# 350215; RRID: AB_2563906
Anti-human AhR – PE (clone T49-550)	BD	Cat# 565711; RRID: AB_2739336
Anti-human T-bet – Brilliant Violet 421 (clone 4B10)	BioLegend	Cat# 644815; RRID: AB_10896427
Anti-human CD2 – eFluor 450 (clone RPA-2.10)	Invitrogen	Cat# 48-0029-42; RRID: AB_2574006
anti-mouse Ahr (polyclonal)	Enzo Life Science	Cat# BML-SA210–0100; RRID: AB_10540536
Bacterial and virus strains
L.m.-OVA InlAM	Sheridan et al., 20143	N/A
Biological samples
Human PBMC isolated from buffy coats	Life South Community Blood Centers	https://www.lifesouth.org/
Human colon tissue biopsies	UFHealth Inflammatory Bowel and Celiac Disease Program	https://ufhealth.org/comprehensive-inflammatory-bowel-diseases-program/overview
Chemicals, peptides, and recombinant proteins
Recombinant Murine IL-2	PeproTech	Cat# 212-12
Recombinant Murine IL-33	PeproTech	Cat# 210-33
Recombinant Human TGF-β	PeproTech	Cat# 100-21
Recombinant Human IL-2	PeproTech	Cat# 200-02
Recombinant Human IL-33	PeproTech	Cat# 200-33
6-formylindolo(3.2b) carbazole (FICZ)	Sigma	Cat# SML 1489-5MG
CH223191	Sigma	C8124-5MG
Critical commercial assays
TRIzol Reagent	Invitrogen	Cat# 15596018
iQ SYBR Green Supermix	Biorad	Cat# 1708887
Nextera DNA Library Preparation Kit	Illumina	Cat# FC-121-1030
NEBNext High-Fidelity 2X PCR Master Mix	NEB	Cat# M0541
SMART-Seq HT Kit	Takara	Cat# 634456
iDeal ChIP-Seq	Diagenode	Cat# C01010055
GoScript Reverse Transcriptase	Promega	Cat# A5003
Zombie Aqua Fixable Viability Kit	BioLegend	Cat# 423102
Live and Dead Violet Viability Kit	Invitrogen	Cat# L34955
Foxp3/Transcription Factor Staining Buffer Set	eBioscience	Cat# 00-5523-00
Dynabeads Mouse T-Activator CD3/CD28	Thermo Fisher	Cat# 11452D
Dynabeads Human T-Activator CD3/CD28	Thermo Fisher	Cat# 11131D
Deposited data
RNA-Seq, ChIP-Seq, scRNA-Seq	This paper	GEO: GSE220944
Experimental models: Cell lines
B16F10	ATCC	CRL-6475
MC38	Kerafast	ENH204-FP
Experimental models: Organisms/strains
Mouse: CD45.1/1	The Jackson Laboratory	Cat# 002014
Mouse: Ahr−/−	Fernandez-Salguero et al.	N/A
Mouse: AhrdCAIR	Ye et al., 201736	N/A
Mouse: EI8Cre	The Jackson Laboratory	Cat# 008766
Mouse: Ahrf/f	The Jackson Laboratory	Cat# 035734
Mouse: OTI TCR transgenic	The Jackson Laboratory	Cat# 003831
Oligonucleotides
See Table S1 for list of quantitative RT-PCR primers	This paper	N/A
Recombinant DNA
Plasmid: MIG-EV	Xiong et al., 202051	N/A
Plasmid: MIG-Ahr	Xiong et al., 202051	N/A
Plasmid: MIG-Y9A	Xiong et al., 202051	N/A
Plasmid: MIG-ΔbHLH	Xiong et al., 202051	N/A
Plasmid: hCD2-EV	Laboratory of Dr. Weishan Huang	N/A
Plasmid: hCD2-Blimp1	Laboratory of Dr. Weishan Huang	N/A
Software and algorithms
FlowJo version 10.4.2	FlowJo	https://www.flowjo.com
Prism 8	GraphPad Software	https://www.graphpad.com/scientific-software/prism/

Highlights.

• Ahr acts as a promoter of resident memory CD8⁺ T cell differentiation and function

• Ahr suppresses the circulating but promotes the resident memory core gene program

• Ahr enhances polyfunctional CD8⁺ T cells, which drive anti-tumor immunity

• In human, AHR promotes in vitro T_RM differentiation and granzyme B production