Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2017 Aug 16.
Published in final edited form as: Immunity. 2016 Aug 2;45(2):428–441. doi: 10.1016/j.immuni.2016.06.016

Adipose-resident group 1 innate lymphoid cells promote obesity-associated insulin resistance

Timothy E O’Sullivan 1, Moritz Rapp 1, Xiying Fan 1, Orr-El Weizman 1, Priya Bhardwaj 2, Nicholas M Adams 1, Thierry Walzer 3, Andrew J Dannenberg 2, Joseph C Sun 1,4
PMCID: PMC5004886  NIHMSID: NIHMS808507  PMID: 27496734

Abstract

Innate lymphoid cells (ILCs) function to protect epithelial barriers against pathogens, and maintain tissue homeostasis in both barrier and non-barrier tissues. Here, utilizing Eomes reporter mice, we identify a subset of adipose group 1 ILC (ILC1) and demonstrate a role for these cells in metabolic disease. Adipose ILC1 were dependent on the transcription factors Nfil3 and T-bet, but phenotypically and functionally distinct from adipose mature natural killer (NK) and immature NK cells. Analysis of parabiotic mice revealed that adipose ILC1 maintained long-term tissue residency. Diet-induced obesity drove early production of interleukin (IL)-12 in adipose tissue depots, and led to the selective proliferation and accumulation of adipose-resident ILC1 in a manner dependent on the IL-12 receptor and STAT4. ILC1-derived interferon-γ was necessary and sufficient to drive proinflammatory macrophage polarization to promote obesity-associated insulin resistance. Thus, adipose-resident ILC1 contribute to obesity-related pathology in response to dysregulated local proinflammatory cytokine production.

Graphical Abstract

graphic file with name nihms808507u1.jpg

Introduction

Innate lymphoid cells (ILCs) can be found in lymphoid and non-lymphoid tissues, and are enriched at epithelial barrier surfaces such as the intestine, lung, and skin (Artis and Spits, 2015). ILCs are rapid producers of both proinflammatory and regulatory cytokines in response to local production of cytokines, pathogen infection, or commensal microbiota perturbation (Artis and Spits, 2015; Serafini et al., 2015). Because certain ILCs have been shown to be tissue-resident at barrier surfaces (Gasteiger et al., 2015), their ability to quickly respond to tissue stress and inflammation underpins their critical role in regulating tissue homeostasis and repair during infection or injury (Artis and Spits, 2015; Serafini et al., 2015). However, persistent inflammatory signals can also lead to unrestrained activation of certain ILC populations at barrier surfaces, exacerbating colitis, dermatitis, and contributing to tumorigenesis (Buonocore et al., 2010; Chan et al., 2014; Fuchs et al., 2013; Kirchberger et al., 2013; Salimi et al., 2013). These studies have highlighted the delicate balance between ILC-mediated immune protection and pathology at barrier sites. Several studies have also identified certain ILC populations in non-barrier tissues such as the adipose tissue (Brestoff and Artis, 2015). While these adipose-associated ILC populations serve to limit inflammation and contribute to metabolic homeostasis (Brestoff et al., 2015), whether and how tissue-resident ILCs contribute to disease progression in non-barrier tissues is unknown.

Recent evidence has suggested that mature ILCs can be further classified into group 1, 2, and 3 ILCs based on different expression of transcription factors, cell surface markers, and effector cytokines (Artis and Spits, 2015). Group 1 ILCs can be distinguished from other ILCs based on their constitutive expression of the transcription factor T-bet, co-expression of activating receptors NKp46 and NK1.1, and rapid production of interferon (IFN)-γ following stimulation with interleukin (IL)-12 (Artis and Spits, 2015; Serafini et al., 2015). However, the phenotypic and functional heterogeneity of group 1 ILCs has only recently been appreciated. Mature natural killer (mNK) cells, the original member of group 1 ILCs (Kiessling et al., 1975), constitutively express the transcription factor Eomes and the integrin CD49b (also known as DX5 or α2β1 integrin), and survey peripheral tissues by constant recirculation through the vasculature (Daussy et al., 2014; Gasteiger et al., 2015). Other group 1 ILCs comprise cells referred to in the literature as tissue-resident NK cells or ILC1 that lack Eomes and CD49b expression, but express CD49a (α1β1 integrin) (Daussy et al., 2014; Gordon et al., 2012; Klose et al., 2014; Peng et al., 2013; Sojka et al., 2014). These cells reside in the liver and small intestine (Gasteiger et al., 2015). An intraepithelial subset of CD103+ ILC1 also resides in the small intestine; development of these cells is dependent on Nfil3 and T-bet, but not IL-15 (Fuchs et al., 2013), further underscoring the complex heterogeneity of group 1 ILCs in different anatomical locations. Although mature NK cells play a critical role in response to viral infection and transformed cells (O’Sullivan et al., 2012; Sun and Lanier, 2011), much less is known about the importance of other tissue-resident group 1 ILCs in immune protection or pathology.

Obesity is a cause of chronic inflammation that is estimated to impact nearly 1 in 5 adults globally by 2025 (Collaboration; Howe et al., 2013), representing one of the largest public health challenges in history. Obesity is associated with several other medical conditions such as insulin resistance, which can develop into diabetes mellitus type 2 (Danaei et al., 2011). In the healthy lean state, type 2 immune cells consisting of a network of ILC2, eosinophils, natural killer T (NKT) cells, and alternatively activated M2 macrophages support adipose tissue homeostasis to prevent the accumulation of proinflammatory immune cells and maintain glucose homeostasis (Brestoff and Artis, 2015). However, obesity results in the hypertrophy of adipocytes, limiting their nutrient and oxygen availability and resulting in a sustained stress response in both adipocytes and stromal cells (Hotamisligil, 2006; Khan et al., 2009; Ye et al., 2007). This stress response results in the recruitment and activation of type 1 immune cells such as CD8+ T cells, Th1 cells, and NK cells, with subsequent accumulation of proinflammatory M1 macrophages in the adipose tissue (Cho et al., 2014; Lee et al., 2016; Morris et al., 2013; Nishimura et al., 2009; Wensveen et al., 2015)). Specifically, stress ligands expressed on adipocytes can lead to enhanced accumulation, proliferation, and activation of adipose NK cells early during diet-induced obesity (Wensveen et al., 2015); however, the lineage specification and tissue residency of these adipose NK cells remains unknown. As a consequence of unrestrained type 1 responses, the cytokines tumor necrosis factor (TNF)-α and IL-6 are produced by proinflammatory M1 macrophages, which can reduce insulin sensitivity over time (de Luca and Olefsky, 2008; Shoelson et al., 2006; Weisberg et al., 2003). Although circulating type 1 lymphocytes and adipose NK cells can contribute to M1 macrophage polarization and insulin resistance through production of IFN-γ during diet-induced obesity (Cho et al., 2014; Nishimura et al., 2009; Wensveen et al., 2015), whether adipose-resident ILCs respond to persistent inflammation and contribute to this process remains unknown.

In this study, we use Eomes reporter mice to show that the adipose tissue contains phenotypically and functionally distinct subsets of Nfil3-dependent group 1 ILCs, consisting of phenotypically stable mNK, immature NK (iNK), and ILC1 populations during homeostasis and diet-induced obesity. Analysis of short- and long-term parabiotic mice revealed that adipose mNK cells were not tissue-resident, while ILC1 maintained long-term tissue residency in the adipose tissue. Following high-fat diet feeding, early production of IL-12 in adipose tissue lead to rapid proliferation and accumulation of group 1 ILCs selectively in the adipose tissue in an IL-12R and STAT4-dependent manner. Lastly, group 1 ILC production of IFN-γ was found to be both sufficient and necessary to drive pro-inflammatory macrophage polarization in the adipose tissue and contribute to insulin resistance during diet-induced obesity.

Results

Tissue-specific heterogeneity of group 1 ILCs

Lineage-negative NK1.1+Nkp46+ cells are found in abundance in the adipose tissue (Liou et al., 2014; O’Rourke et al., 2014; Wensveen et al., 2015); however, whether these cells represent a population similar to splenic NK cells, or a heterogeneous mixture of phenotypically and functionally distinct subsets, remained unknown. When analyzing NK cells (T-bet+ NK1.1+ NKp46+) in various tissues, we found that the spleen and lung consisted predominantly of DX5+Eomes+ mNK, whereas liver and thymus contained both DX5+Eomes+ mNK and DX5−Eomes− cells (Fig. 1A), consistent with previous reports (Daussy et al., 2014; Gordon et al., 2012; Peng et al., 2013; Sojka et al., 2014; Vosshenrich et al., 2006). The small intestine and adipose tissue contained a large percentage of DX5−Eomes+ iNK in addition to DX5+Eomes+ mNK and DX5−Eomes− populations of group 1 ILC (Fig. 1A, B). The adipose DX5−Eomes+ and DX5−Eomes− cells did not require the RAG recombinase and were not thymically derived, because these cells were present in similar numbers in both Rag2−/− and athymic mice, respectively (Fig. 1C and S1A, B). Furthermore, these subsets did not express the ILC2- or ILC3-defining transcription factors GATA3 or Rorγt, respectively, nor did they show a history of Rorγt expression (Fig. S1C, D). In contrast to DX5+Eomes+ mNK and DX5−Eomes+ iNK, DX5−Eomes− cells lacked expression of NK maturation/activation markers CD11b and KLRG1, and did not express activating receptors Ly49H or Ly49D (Fig. 1D, E). Although all three adipose group 1 ILC subsets expressed similar levels of T-bet, DX5−Eomes− cells expressed higher levels of CD69, CD49a, Ly6C, and CD90 than DX5+Eomes+ mNK (Fig. 1E and S1E), similar to tissue-resident Eomes− NK cells in the liver and skin (Sojka et al., 2014), and CD103− ILC1 in the small intestine (Klose et al., 2014). In contrast to mucosal-associated ILC1, adipose DX5−Eomes− cells did not express IL-7Rα or CD103 (Fig. 1E and S1F), but produced more IFN-γ than splenic or adipose DX5+Eomes+ mNK following stimulation with IL-12 and IL-18 (Fig. 1F), suggesting that these cells share functional but not phenotypic properties with ILC1 found in the small intestine (Fuchs et al., 2013; Klose et al., 2014).

Figure 1. Phenotypic and functional heterogeneity of adipose group 1 ILCs.

Figure 1

Open in a new tab

(A) Representative plots show Eomes and DX5 expression on gated lin−NKp46+NK1.1+T-bet+ (group 1 ILC) cells in indicated peripheral organs of WT mice. (lin = TCRβ+CD3ε+CD19+TCRγδ+Ly6G+F4/80+ cells) (B) Percentage of group 1 ILC subsets in bone marrow and peripheral organs, including visceral adipose tissue (VAT) and subcutaneous adipose tissue (SAT), are shown. (C) Representative plots show Eomes and DX5 expression on gated group 1 ILCs in the adipose tissue of WT and Rag2−/− mice. (D, E) Representative plots and histograms show indicated cell surface markers on subsets of group 1 ILCs in adipose tissue. (F) Percentage of IFN-γ+ resting splenic NK cells or indicated adipose group 1 ILC populations after stimulation with IL-12 and IL-18. Data are representative of three independent experiments, with n=5 per group. Samples were compared using an unpaired, two-tailed Student’s t test, and data presented as the mean ± s.e.m. (*p < 0.05). See also Figure S1.

Adipose group 1 ILCs collectively require Nfil3 and common γ chain cytokines, but differentially require T-bet for their development

ILCs require common γ chain (γc)-associated cytokines and expression of the transcription factor Nfil3 for their development (Artis and Spits, 2015). Certain tissue-resident NK1.1+ cell populations in peripheral organs can develop in the absence of Nfil3 expression, and may represent a separate developmental lineage from ILCs (Sojka et al., 2014). We found that total group 1 ILCs were significantly reduced in the adipose tissue of Rag2−/−Il2rg−/− and Nfil3−/− hosts compared to WT controls (Fig. 2A, B). Generation of mixed WT (CD45.1) and Nfil3−/− (CD45.2) bone marrow chimeric mice demonstrated a cell-intrinsic role for Nfil3 in the development of liver and adipose group 1 ILCs (Fig. 2C, D). Group 1 ILCs are defined by expression of and dependence on T-bet for their maturation and development (Deng et al., 2015; Fuchs et al., 2013; Gordon et al., 2012; Klose et al., 2014; Townsend et al., 2004). In T-bet-deficient mice, we observed that whereas DX5+Eomes+ and DX5−Eomes− populations were reduced in both percentage of group 1 ILC (Fig. 2E) and absolute density (Fig. 2F) in the adipose tissue, only a slight decrease in the absolute density of adipose DX5−Eomes+ iNK and an enrichment of these cells in peripheral organs (data not shown), indicating that this latter subset does not require T-bet for its development.

Figure 2. Adipose group 1 ILCs have differential developmental requirements for Nfil3 and T-bet.

Figure 2

Open in a new tab

(A) Representative plots and (B) absolute density of lin−NK1.1+ cells in the adipose tissue of WT, Nfil3−/−, and Rag2−/−Il2rg−/− mice. (C) Representative plots of indicated lymphocyte populations in the adipose tissue and (D) chimerism shown as percentage of indicated lymphocyte populations from the spleen, liver, and adipose tissue of WT (CD45.1) and Nfil3−/− (CD45.2) mixed bone marrow chimeric mice 8 weeks following reconstitution. (E) Representative plots and (F) absolute density of indicated group 1 ILC subsets in Rag2−/−Tbx21+/− or Rag2−/−Tbx21−/− mice. Data are representative of two independent experiments, with n=4 mice per group. Samples were compared using an unpaired, two-tailed Student’s t test, and data presented as the mean ± s.e.m. (*p < 0.05, ***p < 0.001, ****p < 0.0001).

Adipose tissue ILC1 represent a stable population distinct from mNK and iNK cells

To determine whether the observed heterogeneity of adipose group 1 ILCs was due to the presence of stable subsets or differences in the maturation/activation state of tissue-specific iNK or mNK cells, we adoptively transferred equal numbers of adipose DX5+Eomes+, DX5−Eomes+, and DX5−Eomes− group 1 ILCs into Rag2−/−Il2rg−/− mice (Fig. 3A). Two weeks following transfer, we observed that DX5−Eomes+ iNK and DX5−Eomes− cells accumulated in the adipose tissue to a greater extent than DX5+Eomes+ mNK (Fig. 3B), suggesting either enhanced recruitment or retention, or both. DX5+Eomes+ mNK cells sorted from either donor adipose tissue or bone marrow maintained expression of Eomes, DX5, and CD11b in recipient adipose tissue (Fig. 3C) and other peripheral organs after transfer (Fig. S2A, B), representing a stable subset. Transferred DX5−Eomes+ iNK cells increased expression of DX5 and CD11b while retaining Eomes expression, and gave rise to DX5+Eomes+ mNK cells in the adipose (Fig. 3C) and other peripheral tissues (S2A, B) after recovery, suggesting that this DX5−Eomes+ population indeed represented iNK cells. In contrast to iNK cells, adoptively transferred DX5−Eomes− cells did not express appreciable levels of DX5, CD11b, or Eomes in the adipose tissue or other peripheral organs analyzed (Fig. 3C and S2A, B), indicating that these cells, which we will hereafter refer to as ILC1, are a stable and distinct subset of group 1 ILCs during homeostasis.

Figure 3. Group 1 ILCs consist of immature and phenotypically stable mature subsets in the adipose tissue.

Figure 3

Open in a new tab

(A) Schematic of the experiment. Briefly, total adipose tissue was harvested from Eomes-GFP reporter mice and either 2 × 104 DX5+Eomes+, DX5−Eomes+, or DX5−Eomes− group 1 ILCs were sorted (>95% purity) and adoptively transferred i.v. into Rag2−/−Il2rg−/− recipients. (B) Representative plots show the percentage of each transferred group 1 ILC subset (NKp46+NK1.1+) recovered from the adipose tissue of recipient mice after 14 days. (C) In a separate experiment, 2 × 105 DX5+Eomes+, 2 × 104 DX5−Eomes+, or 2 × 104 DX5−Eomes− group 1 ILCs were sorted from adipose tissue or bone marrow of Eomes GFP mice and adoptively transferred i.v. into Rag2−/−Il2rg−/− recipients. Representative plots show the expression of Eomes and DX5 of indicated adoptively transferred NKp46+NK1.1+ cells in the adipose tissue of recipient mice 14 days after transfer. Data are representative of three independent experiments, with n=3 mice per group. See also Figure S2.

Adipose ILC1, but not mNK cells, are tissue-resident

Previous studies have demonstrated that innate lymphocytes are tissue-resident in certain peripheral organs, whereas DX5+Eomes+ mNK cells generally are not (Gasteiger et al., 2015; Peng et al., 2013; Sojka et al., 2014). To determine whether adipose group 1 ILC populations are tissue-resident, we generated parabiotic mice (Fig. 4A) and observed complete chimerism (~50:50 CD45.2+/CD45.1+ ratio) of major lymphocyte populations in the peripheral blood and spleens of parabionts within 30 days (Fig. 4B and data not shown). Although mNK populations had achieved near complete host/donor chimerism after 1 or 4 months, fewer than 10% of adipose tissue ILC1 or ILC2 cells were donor-derived, with similar results observed in all ILC subsets in the small intestine (Fig. 4C, D). These results demonstrate that adipose ILC1 and ILC2 cells are long-term tissue resident cells, whereas adipose mNK cells recirculate in a manner similar to mNK cells found in the spleen, liver and small intestine (Fig. 4B, D and S2A, B). Interestingly, iNK and NKT cells were resident in the adipose tissue and small intestine after 1 month, but gradually replaced by donor-derived cells after 4 months (Fig. 4D), suggesting that these populations may not be as long-term tissue resident or long-lived as ILCs in peripheral tissues. Because tissue-resident memory T cells have been shown to require TGF-β for their retention in the small intestine and skin (Mackay et al., 2013; Zhang and Bevan, 2013), we investigated whether TGF-β was necessary for the tissue retention of adipose iNK and ILC1. We observed equal chimerism between WT and TGF-β-deficient group 1 ILCs in the adipose tissue of mixed bone marrow chimeric mice (Fig. S3A, B), suggesting that TGF-β is dispensable for adipose group 1 ILC retention.

Figure 4. Adipose ILC1 are long-term tissue resident.

Figure 4

Open in a new tab

(A) Schematic of the parabiosis experiment. Briefly, CD45.1+ and CD45.2+ mice were surgically connected for 30 or 130 days and (B) chimerism of T and mNK cells (shown as percentage of donor-derived populations) was analyzed in the spleens of parabionts. (C) Representative plots and (D) quantitation of indicated host and donor derived lymphocytes in the adipose tissue or small intestine. Data are representative of two independent experiments, with n=3–4 (parabiotic pairs) per time point. See also Figure S3.

Adipose-resident ILC1 proliferate and differentially accumulate in fat depots during diet-induced obesity

Previous studies have shown that circulating CD8+ T and NK cells accumulate in the visceral adipose tissue (VAT) of mice fed a high fat diet (HFD) compared to normal diet controls (Nishimura et al., 2009; Wensveen et al., 2015). However, whether adipose-resident cells accumulate by selectively proliferating within the fat depot during HFD feeding is unknown. Because other adipose-resident cells, such as NKT cells, ILC2, and Tregs are reduced in the VAT following HFD feeding (Brestoff and Artis, 2015), we investigated whether the absolute density (number of cells per mg) of adipose-resident iNK and ILC1 populations modulate in different adipose depots during diet-induced obesity. During HFD feeding, we observed a slight and transient increase (~1.5 fold) in mNK, iNK, and ILC1 density in the VAT (Fig. 5A), which was not observed in the spleen or liver (Fig. S4A, B). In contrast to the VAT, we observed a robust and sustained accumulation (~4–6 fold) of mNK, iNK, and ILC1 populations in the subcutaneous adipose tissue (SAT) of HFD-fed mice, suggesting that the SAT is the main site of group 1 ILC accumulation during diet-induced obesity. The observed accumulation of group 1 ILCs in the adipose tissue was not restricted to male mice, because female ovariectomized mice fed a HFD also displayed preferential accumulation of group 1 ILCs in the SAT, with the greatest accumulation occurring in ILC1, compared to iNK or mNK cells (Fig. 5B).

Figure 5. Robust proliferation and persistent accumulation of group 1 ILCs in the subcutaneous adipose tissue during high fat diet feeding.

Figure 5

Open in a new tab

WT mice were fed either a control low fat diet (LFD, 10%) or a high fat diet (HFD, 60%) and peripheral tissues were harvested and analyzed for group 1 ILCs at each time point following administration of diet. (A) The absolute density (total cell number per mg of tissue) of each indicated group 1 ILC in the VAT or SAT of HFD fed mice was calculated and absolute HFD cell density was normalized to LFD controls and presented as relative cell density. (B) WT female mice were ovariectomized and 1 week later fed either a LFD or HFD for 10 weeks. Graph shows the relative cell density of indicated group 1 ILCs in the adipose tissue at week 10 after diet administration. (C) Representative plots show Ki67 intracellular staining of lin−NKp46+NK1.1+ cells in the spleen or SAT of LFD or HFD mice on week 4 following diet administration. (D) Absolute density of Ki67+ group 1 ILCs in the SAT of LFD or HFD mice 4 weeks after diet administration. (E) Absolute density of BrdU+ group 1 ILCs in the SAT of LFD or HFD mice 4 weeks after diet administration. Data are representative of two independent experiments, with n=3–4 per cohort. Samples were compared using an unpaired, two-tailed Student’s t test, and data presented as the mean ± s.e.m. (*p < 0.05, **p < 0.01, ****p < 0.0001). See also Figure S4.

To investigate whether the accumulation of group 1 ILCs during diet-induced obesity was due to proliferation, we first analyzed Ki67 expression. At week 4 after HFD administration, we observed increased Ki67 staining in total NKp46+NK1.1+ cells compared low fat diet (LFD) controls (Fig. 5C), with an increase in Ki67+ mNK, iNK, and ILC1 in the SAT and VAT (Fig. 5D and S4C). To further determine that group 1 ILCs proliferate locally within the adipose tissue, and not at other tissue sites, BrdU incorporation was analyzed in mice fed a HFD or LFD. Although the number of BrdU+ mNK or ILC1 from the spleen or liver of HFD mice did not increase compared to LFD controls, the overall density of BrdU+ group 1 ILCs increased in both the VAT and the SAT (Fig. 5E and S4D, E), suggesting that group 1 ILCs selectively proliferate and accumulate within the adipose tissue during diet-induced obesity.

Production of IFN-γ by adipose-resident ILC1 drives proinflammatory macrophage polarization and obesity-associated insulin resistance

Previous studies have demonstrated that IFN-γ deficiency can decrease insulin resistance in mice fed a HFD (O’Rourke et al., 2012; Rocha et al., 2008; Wong et al., 2011), and that circulating CD8+ T and NK cells can increase expression of IFN-γ during diet-induced obesity (Nishimura et al., 2009; Wensveen et al., 2015). Whether adipose-resident innate lymphocytes produce IFN-γ, and contribute to disease progression, remained unclear. Utilizing IFN-γ-YFP reporter mice (referred to as “GREAT” mice) fed a HFD or LFD, we observed a significant increase in YFP fluorescence in mNK, iNK, and ILC1 from the SAT and VAT of HFD compared to LFD-fed controls, with highest levels found in adipose ILC1 (Fig. 6A, B), suggesting this subset is the most poised to produce IFN-γ among the group 1 ILCs. However, no increase in YFP fluorescence was observed in mNK or ILC1 in the spleen or liver of HFD-fed GREAT mice (Fig. 6A, B and S5A). Consistent with these findings, mNK, iNK, and ILC1 populations from the SAT of HFD-fed mice, but not mNK or ILC1 cells from other peripheral organs, produced more IFN-γ during stimulation with IL-12 and IL-18 (Fig. 6C and S5B–D), suggesting that IFN-γ can be produced early during obesity, and is sustained in group 1 ILCs in the adipose tissues, but not in other peripheral organs. Furthermore, we did not observe differences in splenic NK cell cytotoxicity (Fig. S5E) after 12 weeks of HFD feeding, confirming that effector function of circulating mNK cells is unaffected during diet-induced obesity. Interestingly, adipose ILC1 produced the greatest percentage of IFN-γ following stimulation (relative to mNK and iNK) (Fig. 6C) and were found to be the most abundant among IFN-γhi cells in the adipose tissue following HFD compared to other known IFN-γ producing cells (Fig. S5F), highlighting that adipose-resident ILC1 can likely produce the most IFN-γ in situ on a per cell basis early during diet-induced obesity.

Figure 6. IFN-γ production by group 1 ILCs polarizes proinflammatory macrophages in the adipose tissue and contributes to obesity-associated insulin resistance.

Figure 6

Open in a new tab

IFN-γ YFP reporter (GREAT) or WT mice were fed either a HFD or control LFD for 3 weeks and peripheral organs were harvested. (A) Representative histogram and (B) quantitation of YFP fluorescence in indicated group 1 ILCs from LFD or HFD cohorts. (C) Quantitation of intracellular IFN-γ staining of splenic NK cells and indicated adipose group 1 ILC subsets from LFD or HFD mice after stimulation with IL-12 and IL-18 (D, E) Briefly, WT or Ifng−/− mice were fed a HFD for 3 weeks, and 2 × 104 indicated group 1 ILCs were sort purified from adipose tissue of each group and adoptively transferred into separate Rag2−/−Il2rg−/− mice that had previously received HFD for two weeks. Recipient mice were fed a HFD for an additional 2 weeks (D) Schematic of experiment. (E) The absolute density (cells/mg) of M1 macrophages (lin−F4/80+CD11b+CD11c+) was analyzed in the SAT of recipient mice compared to control LFD or untreated controls. (F–J) 2 × 104 iNK or ILC1 were sort purified from the adipose tissue of WT mice fed a HFD for 4 weeks and adoptively transferred into Rag2−/−Il2rg−/− recipients starting on week 2 of HFD. iNK or ILC1 were adoptively transferred into recipients every two weeks for 6 weeks. Mice were analyzed on week 12 of HFD.(F) Schematic of experiment. (G) Absolute density of M1 macrophages in the SAT, (H) fasting plasma insulin concentration, (I) glucose tolerance test, and (J) insulin tolerance test of each indicated cohort are shown. Data are representative of two independent experiments, with n=3–5 per cohort. Samples were compared using an unpaired, two-tailed Student’s t test, and data presented as the mean ± s.e.m. (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001). See also Figures S5 and S6.

To study the contribution of adipose–resident group 1 ILCs and IFN-γ to proinflammatory macrophage polarization and insulin resistance, without the potential confounding effects of adaptive immunity, we utilized Rag2−/− mice, which lack T, B, and NKT cells. Analysis of NK1.1-depleted or IFN-γ-depleted Rag2−/− mice fed a HFD for 10 weeks revealed a reduction in the density of proinflammatory M1 macrophages in the SAT and VAT compared to untreated Rag2−/− mice (Fig. S5G–I), and an increased sensitivity to glucose as measured by a glucose tolerance test (Fig. S5J). These findings indicate that NK1.1+ cells and non-T/NKT cell derived IFN-γ contribute to M1 macrophage polarization and contribute to insulin resistance following HFD feeding, consistent with previous findings (Wensveen et al., 2015; O’Rourke et al., 2012), These results were not due to inherent differences in weight gain between cohorts or influenced by density of ILC2 because weight gain and VAT ILC2 density were similar between all groups (Fig. S5K and data not shown).

To test whether adipose-resident iNK or ILC1 were sufficient to drive early M1 macrophage polarization during HFD feeding, we adoptively transferred purified adipose iNK or ILC1 into Rag2−/−Il2rg−/− hosts (Fig. 6D). Two weeks after cell transfer and HFD feeding, we observed a significant increase in the density of M1 macrophages in mice receiving group 1 ILCs compared to untreated controls, with ILC1 contributing greater to M1 macrophage polarization than iNK on a per-cell basis (Fig. 6E). To confirm that group 1 ILC-derived IFN-γ (and not another factor) was responsible for M1 macrophage polarization early during HFD feeding, we adoptively transferred purified iNK and ILC1 populations from Ifng−/− mice fed a HFD for 3 weeks into Rag2−/−Il2rg−/− hosts fed a HFD for 2 weeks (Fig. 6D). Following Ifng−/− group 1 ILC transfer and HFD feeding, we did not observe an increase in M1 macrophage polarization compared to untreated controls (Fig. 6E), indicating that production of IFN-γ by group 1 ILCs enables M1 macrophage polarization early during HFD feeding. These results were not due to unequal recruitment of adoptively transferred cells or differential weight gain for each cohort, because iNK (which differentiate into mNK), ILC1, and Ifng−/− NK1.1+ cells were present in the adipose tissue at equal amounts (Fig. S6A), and each cohort of mice gained similar amounts of weight following HFD administration (data not shown). Furthermore, our observations were not confounded by potential conversion of adoptively transferred subsets by obesity-associated inflammation, because adoptively transferred ILC1 in recipient adipose tissue maintained phenotypic stability during diet-induced obesity, whereas iNK cells were still able to differentiate into mNK cells in recipient adipose tissue (Fig S6B–D).

To determine whether adipose-resident group 1 ILC-dependent M1 macrophage polarization could contribute to diet-induced obesity associated insulin resistance, we adoptively transferred purified adipose iNK or ILC1 from HFD-fed WT mice into Rag2−/−Il2rg−/− hosts maintained on a HFD for 12 weeks (Fig. 6F). Consistent with our previous results, we observed a significant increase in the density of M1 macrophages 10 weeks after initial cell transfer in the SAT of mice receiving iNK or ILC1 compared to untreated controls, with ILC1 contributing to greater M1 macrophage polarization compared to iNK on a per-cell basis (Fig. 6G). The observed difference in adipose M1 macrophage polarization can likely be explained by the enhanced IFN-γ production of transferred adipose ILC1, along with their superior long-term retention or longevity in the adipose tissue compared to iNK cells (Fig. S6E). Additional analysis of macrophage populations revealed a HFD-dependent decrease in the density of M2 macrophages in the SAT that was independent of transferred iNK or ILC1 (Fig. S6F). Furthermore, metabolic analysis of recipient mice revealed significant increases in fasting plasma insulin (Fig. 6H), glucose intolerance (Fig. 6I), and insulin resistance (Fig. 6J) in iNK and ILC1 adoptively transferred groups compared to untreated controls. These results were not due to differential weight gain for each cohort, because each group of mice gained similar amounts of weight following HFD administration (Fig. S6G). Together, these results suggest that adipose-resident group 1 ILCs are sufficient to contribute to obesity-associated insulin resistance through polarization of adipose M1 macrophages, with greatest metabolic dysfunction attributable to the adipose-resident ILC1 population.

IL-12 and STAT4 signaling drives proliferation and accumulation of adipose-resident ILC1, contributing to proinflammatory macrophage polarization

To determine the signals that were driving group 1 ILC production of IFN-γ, we harvested VAT and SAT tissue from WT mice fed either LFD or HFD and analyzed poly-A enriched mRNA from tissue homogenates. We found that IL-12p35 and IL-12p40 transcripts were increased in both the SAT and VAT after 3 weeks of HFD feeding compared to LFD controls, with higher levels observed in the SAT compared to the VAT (Fig. 7A). To investigate whether IL-12 signaling drove the proliferation and accumulation of adipose group 1 ILCs following HFD feeding, we utilized mice deficient in the IL-12 receptor (IL-12R) or STAT4, which signals downstream of the IL-12R. Whereas WT mice displayed increased density and percentage of Ki67+ NKp46+NK1.1+ cells in the SAT of HFD-fed mice compared to LFD-fed controls at week 4 after diet administration (Fig. 7B, C), IL-12R or STAT4-deficient mice fed a HFD did not accumulate Ki67+ mNK, iNK, or ILC1 in the SAT (Fig. 7B, C), indicating that IL-12 signaling is required for the proliferation of all group 1 ILC subsets during diet-induced obesity. Additionally, HFD-fed IL-12R and STAT4-deficient HFD mice did not accumulate M1 macrophages in the SAT to the same degree as WT HFD controls (Fig. 7D). These results were not due to differential weight gain for each cohort, because all three groups of mice gained similar amounts of weight following HFD administration (Fig. S7A, B). Furthermore, IL-12R and STAT4-deficient adipose mNK, iNK, and ILC1 displayed normal development in mixed bone marrow chimeric mice (Fig. S7C, D), confirming that these results were not due to developmental defects in these cells. These findings suggest that IL-12 and STAT4 signaling in adipose mNK, iNK, and ILC1 drives IFN-γ production that is required for M1 macrophage polarization during diet-induced obesity.

Figure 7. Production of IL-12 in adipose tissues following HFD drives group 1 ILC proliferation, accumulation, and proinflammatory macrophage polarization.

Figure 7

Open in a new tab

WT mice were either fed a control LFD or HFD for 1, 2, 3, or 4 weeks and adipose tissue was harvested. (A) Graphs show IL-12p35 and IL-12p40 mRNA levels in SAT and VAT for each cohort. (B–D) WT mice were fed a LFD; or WT, Il12rb2−/−, and Stat4−/− mice were fed a HFD for 4 weeks. (B) Absolute density of group 1 ILCs in the SAT (C) percentage of Ki67+ cells and (D) absolute density of M1 macrophages in the SAT of each cohort are shown. Data are representative of two independent experiments, with n=3–5 per cohort. Samples were compared using an unpaired, two-tailed Student’s t test, and data presented as the mean ± s.e.m. (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001). See also Figure S7.

Discussion

Recently it has been appreciated that NK cells represent only one member of a family of innate lymphoid cells, consisting of group 1, group 2, and group 3 ILCs. Although NK cells have been considered a member of the group 1 innate lymphoid cells (Artis and Spits, 2015; Serafini et al., 2015), currently there are no definitive criteria to discriminate NK cells from other NKp46+ group 1 ILCs. Expression of Eomes, CD49a, CD49b, CXCR6, and TRAIL are expressed by different subsets of group 1 ILCs within different tissues (e.g. liver, small intestine, salivary gland, spleen, etc.) at steady state (Cortez et al., 2014; Daussy et al., 2014; Gordon et al., 2012; Klose et al., 2014; Sojka et al., 2014); however, it is unknown whether these are stable traits or can be modulated by the particular tissue microenvironment or activation state of the cells. Given the recently reported reversible differentiation of human RORγt+ ILC3 into T-bet+ IFN-γ-producing ILC1 in the small and large intestine during homeostasis (Bernink et al., 2015), further validation of stable heterogeneity within tissue-specific group 1 ILCs is warranted. In this study, we demonstrated that the expression of DX5 and Eomes differentiates functionally distinct adipose mNK, iNK, and ILC1 during homeostasis. Adipose iNK cells increased expression of DX5 and CD49a following homeostatic proliferation, giving rise to mNK cells in peripheral tissues, cautioning the usage of CD49a as a lineage-defining marker. In contrast, adoptively transferred adipose mNK and ILC1 maintained stable phenotypes during homeostasis and obesity-associated inflammation in the adipose tissue and liver, and did not express RORγt, indicating that mature adipose group 1 ILCs do not display the phenotypic and functional plasticity observed in human mucosal ILC1. Furthermore, adipose ILC1 did not display a history of Rorγt expression, nor did we observe the presence of ILC3 in the adipose depots during homeostasis or during diet-induced obesity, suggesting that the accumulation of adipose ILC1 during obesity is likely not due to a conversion from ILC3. We propose that adipose ILC1 are not iNK cells, nor an activation state of a mNK or ILC3 cells influenced by tissue-specific microenvironments or inflammation, because these cells maintain tissue residency during MCMV-induced viremia (data not shown), and do not display evidence of phenotypic conversion during homeostasis or obesity-associated inflammation. Therefore adipose ILC1 likely represent a stable lineage distinct from circulating mNK cells during homeostasis and disease. Future lineage-tracing experiments will be necessary to determine whether CD49a+DX5− cell found in other peripheral organs, which have been referred to as tissue-resident NK cells (Sojka et al., 2014), are not in fact tissue-resident iNK cells.

Nfil3 is now recognized as a master transcription factor regulating the development of all known ILC subsets (Artis and Spits, 2015). Although recent studies have identified tissue-resident CD49a+DX5−Eomes− NK-like cells that develop independently of Nfil3 in the liver, uterus, skin, and tumor microenvironment (Dadi et al., 2016; Sojka et al., 2014), our results indicate that both liver and adipose ILC1 (which do not express CD103) require Nfil3 in a cell-intrinsic manner for their development. Thus, we believe that these tissue-resident NK cell populations should not be considered independently derived populations distinct from the Nfil3-dependent precursor that gives rise to all other ILC lineages. Because mature NK cells have been shown to appear in the absence of Nfil3 during viral infection (Firth et al., 2013), it remains unknown whether pathogen, tumor, or microbiota-derived inflammatory cues present in peripheral organs may be promoting the appearance of Nfil3-deficient CD49a+DX5−Eomes− cells, contributing to these conflicting observations.

Previous studies have shown that systemic depletion of NK1.1+ or NKp46+ cells reduces diet-induced obesity-associated insulin resistance by limiting the polarization of proinflammatory macrophages (Lee et al., 2016; O’Rourke et al., 2014; Wensveen et al., 2015). Although the genetic or antibody depletion methods utilized in these studies also reduced adipose NKT cells, which can suppress insulin resistance during diet-induced obesity (Hams et al., 2013; Lynch et al., 2012), adoptive transfer of DX5+ splenic NK cells into the VAT of Ifng−/− mice is sufficient to exacerbate insulin resistance following HFD administration (Wensveen et al., 2015). Similarly, adoptive transfer of splenic DX5+ NK cells into Nfil3−/− mice is sufficient to enhance insulin resistance during diet-induced obesity (Lee et al., 2016), demonstrating that DX5+ NK cells can contribute to insulin resistance in the presence of an intact adaptive immune system. However, recent studies show that DX5+ NK cells are not tissue resident in peripheral organs (Gasteiger et al., 2015; Peng et al., 2013; Sojka et al., 2014). Therefore, the contribution of adipose-resident NK cells to obesity-associated insulin resistance remains unclear from these studies. In contrast, our results demonstrated that “adipose-resident NK” cells are not a homogenous population, but rather a heterogeneous mixture of iNK, mNK, and ILC1 cells. Because our results show that DX5+ adipose mNK cells are not tissue resident, our study revealed the contribution of individual tissue resident group 1 ILCs and IFN-γ in obesity-associated insulin resistance in the absence of complicated opposing influences of T and NKT cells. We discovered that adipose-resident ILC1 produced the highest amounts of IFN-γ early during HFD feeding compared to other known IFN-γ producing lymphocytes in the adipose tissue. Furthermore, we found that adipose ILC1 polarized adipose proinflammatory macrophages and contributed to metabolic dysfunction to a greater extent on a per cell basis (relative to iNK or mNK), suggesting that adipose ILC1 responses to HFD-induced inflammation are the most potent amongst group 1 ILCs in driving disease progression. Although the genetic tools to definitively differentiate between the contribution of each subset of group 1 ILCs to obesity-associated insulin resistance do not currently exist, our results suggest that each member of the group 1 ILCs can differentially polarize adipose proinflammatory macrophages through HFD-induced IFN-γ production, and therefore represent a functional spectrum of cells able to contribute to insulin resistance during diet-induced obesity.

In human studies, VAT inflammation in obese patients is thought to be the primary cause of obesity-induced insulin resistance (Van Gaal et al., 2006; Xu et al., 2003). However, recent studies have also found that inflammation and the presence crown like-structures (CLS) in the SAT of obese patients correlates with metabolic dysfunction (Iyengar et al., 2015; Le et al., 2011; Wentworth et al., 2010), suggesting that inflammation in both adipose depots may contribute to obesity-associated metabolic dysfunction. In support of this hypothesis, surgical removal of VAT in lean mice does not completely abrogate diet-induced obesity-associated insulin resistance (Wensveen et al., 2015). Our results indicate that group 1 ILCs produce IFN-γ in both adipose depots during diet-induced obesity and contribute to the polarization of M1 macrophages in the SAT, and to a lesser extent in the VAT, during diet-induced obesity. Therefore, it is likely that IFN-γ dependent inflammation in both adipose depots contributes to obesity-associated metabolic dysfunction.

Although our study has implicated IL-12 production in contributing to obesity associated inflammation and insulin resistance, the upstream signals and cellular sources of IL-12 remain poorly defined. A previous study demonstrated that MGL−CD11c+ macrophages from the VAT of HFD mice produce higher amounts of IL-12p40 transcripts compared to MGL1+CD11c− macrophages, and actively produce IL-12 protein in vivo (Strissel et al., 2010). However, the relative contribution of adipose dendritic cells or other myeloid lineage cells to IL-12 production during diet-induced obesity will need to be determined in future studies. Because obesity is associated with elevated amounts of circulating LPS through metabolic endotoxemia (Cani et al., 2007), it may be possible that circulating LPS (via TLR4 activation) or stress factors produced by adipocytes contribute to IL-12 production in the adipose tissue. Future work will be needed to determine the signals that potentiate IL-12 production in the adipose depots during diet-induced obesity.

In summary, our study identifies a unique population of phenotypically stable and functionally distinct adipose-resident ILC1 that represent a separate population from tissue-resident iNK and circulating mNK cells. We demonstrate that the proinflammatory cytokine IL-12 can contribute to adipose ILC1 activation and exacerbation of obesity-associated insulin resistance, demonstrating that tissue-resident ILCs present in non-barrier organs can also contribute to pathology. It will be of interest to identify human adipose ILC1 populations, and determine whether these cells also accumulate and correlate with metabolic dysfunction in obese individuals. Future work will elucidate the stability of other innate lymphoid cell subsets at non-barrier sites, and determine whether dysregulated responses of these cells can contribute to disease.

Experimental Procedures

Mice

Mice were bred at Memorial Sloan Kettering Cancer Center in accordance with the guidelines of the Institutional Animal Care and Use Committee (IACUC). The following strains were used in this study: C57BL/6 (CD45.2), B6.SJL (CD45.1), Rag2−/−Il2rg−/−, EomesGFP transgenic (Daussy et al., 2014), Nfil3−/−, Nude mice (nu/nu), Rag2−/−, Rag2−/−Tbx21+/−, Rag2−/−Tbx21−/−, Il12rb2−/−, Stat4−/−, NKp46iCre × TGFβRIIfl/fl, IFN-γ YFP (GREAT), Rorccre × Rosa26lslTdtomato, and Ifng−/−. Experiments were conducted using age- and gender-matched mice in accordance with approved institutional protocols.

Parabiosis

Parabiosis surgery was performed, as previously described (Gasteiger et al., 2015), and parabiotic mice generated in accordance with approved institutional protocols. A more detailed protocol can be found in Supplemental Experimental Procedures.

Adoptive transfer of adipose group 1 ILCs

In adoptive transfer experiments, mNK (lin−NK1.1+NKp46+DX5+Eomes+CD49a−CD69−), iNK (lin−NK1.1+NKp46+DX5−Eomes+CD49a−CD69+), and ILC1 (lin−NK1.1+NKp46+DX5−Eomes−CD49a+CD69+) were purified from adipose tissue or bone marrow from EomesGFP mice by flow cytometric cell sorting to high purity and transferred into male Rag2−/−Il2rg−/− mice on LFD or HFD for 2 weeks. After an additional 2 weeks after transfer and on respective diets, recipient peripheral organs were harvested and adoptively transferred cells were analyzed by flow cytometry. In other experiments, WT or Ifng−/− male mice were fed a high fat diet for 3 or 4 weeks and adipose tissue was harvested. iNK, and ILC1 were purified from adipose tissue and transferred into male Rag2−/−Il2rg−/− mice fed a HFD for 2 weeks. 2 or 4 weeks after final transfer and additional HFD feeding, recipient peripheral organs were harvested and analyzed by flow cytometry.

Calculation of absolute and relative density of adipose lymphocytes

In order to normalize the variation in HFD-induced weight gain among male mice, we individually housed each mouse and administered HFD (60% of caloric intake) or low fat diet (LFD, 10% of caloric intake) ad libitum. We then harvested either VAT or SAT and analyzed the composition of group 1 ILCs by flow cytometry. Because individual mice displayed variation in the amount of weight gain and fat pad mass, we calculated the density of each adipose group 1 ILC (absolute number/fat pad weight) and normalized the density of HFD mice to the average density of LFD controls for each adipose depot in certain experiments.

Flow cytometry and qRT-PCR on isolated lymphocytes

Various lymphocyte populations were isolated from multiple tissues, and isolation procedures described in Supplemental Experimental Procedures. The flow cytometry antibodies and qRT-PCR primers used in analysis of purified lymphocytes or whole tissues can be found in Supplemental Experimental Procedures.

Ex vivo stimulation of group 1 ILCs

Approximately 1 × 105 splenic or adipose group 1 ILCs were stimulated for 5 hours in RPMI containing 5% fetal bovine serum with recombinant mouse IL-12 (20 ng/ml; R&D Systems) plus IL-18 (10 ng/ml; R&D Systems). Cells were cultured in media alone as a negative control.

Glucose and insulin tolerance tests

For glucose tolerance tests, male mice were fasted for 16 hours and then injected with 1g/kg D-glucose (Sigma) i.p. For insulin tolerance tests, mice were fasted for 4 hours and injected i.p. with 0.75 U/kg of human insulin (Eli Lilly) dissolved in saline solution. Blood glucose was sampled from the mouse tail vein every 15–30 minutes following injection and measured using TrueTest glucose strips inserted into a TrueTest glucometer.

Statistical analyses

For graphs, data are shown as mean ± s.e.m. and, unless otherwise indicated, statistical differences were evaluated using a two-tailed unpaired Student’s t-test, assuming equal sample variance. P < 0.05 was considered significant. Graphs were produced and statistical analyses were performed using GraphPad Prism.

Supplementary Material

supplement

Acknowledgments

We thank members of the Sun lab for comments, discussions, technical support and experimental assistance. We thank Lewis Lanier and Alexander Rudensky for helpful discussions. T.E.O. was supported by the American Cancer Society. M.R. was supported by a fellowship from the DAAD (Germany). X.F. and N.M.A. were supported by the NIH Medical Scientist Training Program grant T32GM07739 to the Weill Cornell/Rockefeller/Sloan-Kettering Tri-Institutional MD-PhD Program. A.D. was supported by grants from the National Institutes of Health (1R01CA154481-01A1), the Breast Cancer Research Foundation, and the Botwinick-Wolfensohn Foundation. J.C.S. was supported by the Ludwig Center for Cancer Immunotherapy, the Searle Scholars Program, the Cancer Research Institute, the Starr Cancer Consortium, and grants from the NIH (AI100874 and P30CA008748).

Footnotes

The authors declare no financial conflicts of interest.

Author Contributions

T.E.O., A.J.D., and J.C.S. designed the study; T.W. provided Eomes GFP mice; T.E.O., O.E.W, and M.R. performed cell sorting; X.F. and T.E.O generated parabiotic mice; P.B. performed RNA extraction and generated cDNA; P.B., O.E.W., N.M.A, and T.E.O performed tissue harvest and processing. T.E.O. performed all additional experiments; T.E.O. and J.C.S. wrote the manuscript.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Artis D, Spits H. The biology of innate lymphoid cells. Nature. 2015;517:293–301. doi: 10.1038/nature14189. [DOI] [PubMed] [Google Scholar]
  2. Bernink JH, Krabbendam L, Germar K, de Jong E, Gronke K, Kofoed-Nielsen M, Munneke JM, Hazenberg MD, Villaudy J, Buskens CJ, et al. Interleukin-12 and -23 Control Plasticity of CD127(+) Group 1 and Group 3 Innate Lymphoid Cells in the Intestinal Lamina Propria. Immunity. 2015;43:146–160. doi: 10.1016/j.immuni.2015.06.019. [DOI] [PubMed] [Google Scholar]
  3. Brestoff JR, Artis D. Immune regulation of metabolic homeostasis in health and disease. Cell. 2015;161:146–160. doi: 10.1016/j.cell.2015.02.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Brestoff JR, Kim BS, Saenz SA, Stine RR, Monticelli LA, Sonnenberg GF, Thome JJ, Farber DL, Lutfy K, Seale P, Artis D. Group 2 innate lymphoid cells promote beiging of white adipose tissue and limit obesity. Nature. 2015;519:242–246. doi: 10.1038/nature14115. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Buonocore S, Ahern PP, Uhlig HH, Ivanov II, Littman DR, Maloy KJ, Powrie F. Innate lymphoid cells drive interleukin-23-dependent innate intestinal pathology. Nature. 2010;464:1371–1375. doi: 10.1038/nature08949. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Cani PD, Amar J, Iglesias MA, Poggi M, Knauf C, Bastelica D, Neyrinck AM, Fava F, Tuohy KM, Chabo C, et al. Metabolic endotoxemia initiates obesity and insulin resistance. Diabetes. 2007;56:1761–1772. doi: 10.2337/db06-1491. [DOI] [PubMed] [Google Scholar]
  7. Chan IH, Jain R, Tessmer MS, Gorman D, Mangadu R, Sathe M, Vives F, Moon C, Penaflor E, Turner S, et al. Interleukin-23 is sufficient to induce rapid de novo gut tumorigenesis, independent of carcinogens, through activation of innate lymphoid cells. Mucosal immunology. 2014;7:842–856. doi: 10.1038/mi.2013.101. [DOI] [PubMed] [Google Scholar]
  8. Cho KW, Morris DL, DelProposto JL, Geletka L, Zamarron B, Martinez-Santibanez G, Meyer KA, Singer K, O’Rourke RW, Lumeng CN. An MHC II-dependent activation loop between adipose tissue macrophages and CD4+ T cells controls obesity-induced inflammation. Cell reports. 2014;9:605–617. doi: 10.1016/j.celrep.2014.09.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Collaboration N.C.D.R.F. Trends in adult body-mass index in 200 countries from 1975 to 2014: a pooled analysis of 1698 population-based measurement studies with 19·2 million participants. The Lancet. 387:1377–1396. doi: 10.1016/S0140-6736(16)30054-X. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Cortez VS, Fuchs A, Cella M, Gilfillan S, Colonna M. Cutting edge: Salivary gland NK cells develop independently of Nfil3 in steady-state. Journal of immunology. 2014;192:4487–4491. doi: 10.4049/jimmunol.1303469. [DOI] [PubMed] [Google Scholar]
  11. Dadi S, Chhangawala S, Whitlock BM, Franklin RA, Luo CT, Oh SA, Toure A, Pritykin Y, Huse M, Leslie CS, Li MO. Cancer Immunosurveillance by Tissue-Resident Innate Lymphoid Cells and Innate-like T Cells. Cell. 2016;164:365–377. doi: 10.1016/j.cell.2016.01.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Danaei G, Finucane MM, Lu Y, Singh GM, Cowan MJ, Paciorek CJ, Lin JK, Farzadfar F, Khang YH, Stevens GA, et al. National, regional, and global trends in fasting plasma glucose and diabetes prevalence since 1980: systematic analysis of health examination surveys and epidemiological studies with 370 country-years and 2.7 million participants. Lancet. 2011;378:31–40. doi: 10.1016/S0140-6736(11)60679-X. [DOI] [PubMed] [Google Scholar]
  13. Daussy C, Faure F, Mayol K, Viel S, Gasteiger G, Charrier E, Bienvenu J, Henry T, Debien E, Hasan UA, et al. T-bet and Eomes instruct the development of two distinct natural killer cell lineages in the liver and in the bone marrow. The Journal of experimental medicine. 2014;211:563–577. doi: 10.1084/jem.20131560. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. de Luca C, Olefsky JM. Inflammation and insulin resistance. FEBS letters. 2008;582:97–105. doi: 10.1016/j.febslet.2007.11.057. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Deng Y, Kerdiles Y, Chu J, Yuan S, Wang Y, Chen X, Mao H, Zhang L, Zhang J, Hughes T, et al. Transcription factor Foxo1 is a negative regulator of natural killer cell maturation and function. Immunity. 2015;42:457–470. doi: 10.1016/j.immuni.2015.02.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Firth MA, Madera S, Beaulieu AM, Gasteiger G, Castillo EF, Schluns KS, Kubo M, Rothman PB, Vivier E, Sun JC. Nfil3-independent lineage maintenance and antiviral response of natural killer cells. The Journal of experimental medicine. 2013;210:2981–2990. doi: 10.1084/jem.20130417. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Fuchs A, Vermi W, Lee JS, Lonardi S, Gilfillan S, Newberry RD, Cella M, Colonna M. Intraepithelial type 1 innate lymphoid cells are a unique subset of IL-12- and IL-15-responsive IFN-gamma-producing cells. Immunity. 2013;38:769–781. doi: 10.1016/j.immuni.2013.02.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Gasteiger G, Fan X, Dikiy S, Lee SY, Rudensky AY. Tissue residency of innate lymphoid cells in lymphoid and non-lymphoid organs. Science. 2015 doi: 10.1126/science.aac9593. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Gordon SM, Chaix J, Rupp LJ, Wu J, Madera S, Sun JC, Lindsten T, Reiner SL. The transcription factors T-bet and Eomes control key checkpoints of natural killer cell maturation. Immunity. 2012;36:55–67. doi: 10.1016/j.immuni.2011.11.016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Hams E, Locksley RM, McKenzie AN, Fallon PG. Cutting edge: IL-25 elicits innate lymphoid type 2 and type II NKT cells that regulate obesity in mice. Journal of immunology. 2013;191:5349–5353. doi: 10.4049/jimmunol.1301176. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Hotamisligil GS. Inflammation and metabolic disorders. Nature. 2006;444:860–867. doi: 10.1038/nature05485. [DOI] [PubMed] [Google Scholar]
  22. Howe LR, Subbaramaiah K, Hudis CA, Dannenberg AJ. Molecular pathways: adipose inflammation as a mediator of obesity-associated cancer. Clinical cancer research: an official journal of the American Association for Cancer Research. 2013;19:6074–6083. doi: 10.1158/1078-0432.CCR-12-2603. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Iyengar NM, Zhou XK, Gucalp A, Morris PG, Howe LR, Giri DD, Morrow M, Wang H, Pollak M, Jones LW, et al. Systemic Correlates of White Adipose Tissue Inflammation in Early-Stage Breast Cancer. Clinical cancer research: an official journal of the American Association for Cancer Research. 2015 doi: 10.1158/1078-0432.CCR-15-2239. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Khan T, Muise ES, Iyengar P, Wang ZV, Chandalia M, Abate N, Zhang BB, Bonaldo P, Chua S, Scherer PE. Metabolic dysregulation and adipose tissue fibrosis: role of collagen VI. Molecular and cellular biology. 2009;29:1575–1591. doi: 10.1128/MCB.01300-08. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Kiessling R, Klein E, Pross H, Wigzell H. “Natural” killer cells in the mouse. II. Cytotoxic cells with specificity for mouse Moloney leukemia cells. Characteristics of the killer cell. European journal of immunology. 1975;5:117–121. doi: 10.1002/eji.1830050209. [DOI] [PubMed] [Google Scholar]
  26. Kirchberger S, Royston DJ, Boulard O, Thornton E, Franchini F, Szabady RL, Harrison O, Powrie F. Innate lymphoid cells sustain colon cancer through production of interleukin-22 in a mouse model. The Journal of experimental medicine. 2013;210:917–931. doi: 10.1084/jem.20122308. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Klose CS, Flach M, Mohle L, Rogell L, Hoyler T, Ebert K, Fabiunke C, Pfeifer D, Sexl V, Fonseca-Pereira D, et al. Differentiation of type 1 ILCs from a common progenitor to all helper-like innate lymphoid cell lineages. Cell. 2014;157:340–356. doi: 10.1016/j.cell.2014.03.030. [DOI] [PubMed] [Google Scholar]
  28. Le KA, Mahurkar S, Alderete TL, Hasson RE, Adam TC, Kim JS, Beale E, Xie C, Greenberg AS, Allayee H, Goran MI. Subcutaneous adipose tissue macrophage infiltration is associated with hepatic and visceral fat deposition, hyperinsulinemia, and stimulation of NF-kappaB stress pathway. Diabetes. 2011;60:2802–2809. doi: 10.2337/db10-1263. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Lee BC, Kim MS, Pae M, Yamamoto Y, Eberle D, Shimada T, Kamei N, Park HS, Sasorith S, Woo JR, et al. Adipose Natural Killer Cells Regulate Adipose Tissue Macrophages to Promote Insulin Resistance in Obesity. Cell metabolism. 2016;23:685–698. doi: 10.1016/j.cmet.2016.03.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Liou YH, Wang SW, Chang CL, Huang PL, Hou MS, Lai YG, Lee GA, Jiang ST, Tsai CY, Liao NS. Adipocyte IL-15 regulates local and systemic NK cell development. Journal of immunology. 2014;193:1747–1758. doi: 10.4049/jimmunol.1400868. [DOI] [PubMed] [Google Scholar]
  31. Lynch L, Nowak M, Varghese B, Clark J, Hogan AE, Toxavidis V, Balk SP, O’Shea D, O’Farrelly C, Exley MA. Adipose tissue invariant NKT cells protect against diet-induced obesity and metabolic disorder through regulatory cytokine production. Immunity. 2012;37:574–587. doi: 10.1016/j.immuni.2012.06.016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Mackay LK, Rahimpour A, Ma JZ, Collins N, Stock AT, Hafon ML, Vega-Ramos J, Lauzurica P, Mueller SN, Stefanovic T, et al. The developmental pathway for CD103(+)CD8+ tissue-resident memory T cells of skin. Nature immunology. 2013;14:1294–1301. doi: 10.1038/ni.2744. [DOI] [PubMed] [Google Scholar]
  33. Morris DL, Cho KW, Delproposto JL, Oatmen KE, Geletka LM, Martinez-Santibanez G, Singer K, Lumeng CN. Adipose tissue macrophages function as antigen-presenting cells and regulate adipose tissue CD4+ T cells in mice. Diabetes. 2013;62:2762–2772. doi: 10.2337/db12-1404. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Nishimura S, Manabe I, Nagasaki M, Eto K, Yamashita H, Ohsugi M, Otsu M, Hara K, Ueki K, Sugiura S, et al. CD8+ effector T cells contribute to macrophage recruitment and adipose tissue inflammation in obesity. Nature medicine. 2009;15:914–920. doi: 10.1038/nm.1964. [DOI] [PubMed] [Google Scholar]
  35. O’Rourke RW, Meyer KA, Neeley CK, Gaston GD, Sekhri P, Szumowski M, Zamarron B, Lumeng CN, Marks DL. Systemic NK cell ablation attenuates intra-abdominal adipose tissue macrophage infiltration in murine obesity. Obesity. 2014;22:2109–2114. doi: 10.1002/oby.20823. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. O’Rourke RW, White AE, Metcalf MD, Winters BR, Diggs BS, Zhu X, Marks DL. Systemic inflammation and insulin sensitivity in obese IFN-gamma knockout mice. Metabolism: clinical and experimental. 2012;61:1152–1161. doi: 10.1016/j.metabol.2012.01.018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. O’Sullivan T, Saddawi-Konefka R, Vermi W, Koebel CM, Arthur C, White JM, Uppaluri R, Andrews DM, Ngiow SF, Teng MW, et al. Cancer immunoediting by the innate immune system in the absence of adaptive immunity. The Journal of experimental medicine. 2012;209:1869–1882. doi: 10.1084/jem.20112738. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Peng H, Jiang X, Chen Y, Sojka DK, Wei H, Gao X, Sun R, Yokoyama WM, Tian Z. Liver-resident NK cells confer adaptive immunity in skin-contact inflammation. The Journal of clinical investigation. 2013;123:1444–1456. doi: 10.1172/JCI66381. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Rocha VZ, Folco EJ, Sukhova G, Shimizu K, Gotsman I, Vernon AH, Libby P. Interferon-gamma, a Th1 cytokine, regulates fat inflammation: a role for adaptive immunity in obesity. Circulation research. 2008;103:467–476. doi: 10.1161/CIRCRESAHA.108.177105. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Salimi M, Barlow JL, Saunders SP, Xue L, Gutowska-Owsiak D, Wang X, Huang LC, Johnson D, Scanlon ST, McKenzie AN, et al. A role for IL-25 and IL-33-driven type-2 innate lymphoid cells in atopic dermatitis. The Journal of experimental medicine. 2013;210:2939–2950. doi: 10.1084/jem.20130351. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Serafini N, Vosshenrich CA, Di Santo JP. Transcriptional regulation of innate lymphoid cell fate. Nature reviews Immunology. 2015;15:415–428. doi: 10.1038/nri3855. [DOI] [PubMed] [Google Scholar]
  42. Shoelson SE, Lee J, Goldfine AB. Inflammation and insulin resistance. The Journal of clinical investigation. 2006;116:1793–1801. doi: 10.1172/JCI29069. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Sojka DK, Plougastel-Douglas B, Yang L, Pak-Wittel MA, Artyomov MN, Ivanova Y, Zhong C, Chase JM, Rothman PB, Yu J, et al. Tissue-resident natural killer (NK) cells are cell lineages distinct from thymic and conventional splenic NK cells. eLife. 2014;3:e01659. doi: 10.7554/eLife.01659. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Strissel KJ, DeFuria J, Shaul ME, Bennett G, Greenberg AS, Obin MS. T-cell recruitment and Th1 polarization in adipose tissue during diet-induced obesity in C57BL/6 mice. Obesity. 2010;18:1918–1925. doi: 10.1038/oby.2010.1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Sun JC, Lanier LL. NK cell development, homeostasis and function: parallels with CD8(+) T cells. Nature reviews Immunology. 2011;11:645–657. doi: 10.1038/nri3044. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Townsend MJ, Weinmann AS, Matsuda JL, Salomon R, Farnham PJ, Biron CA, Gapin L, Glimcher LH. T-bet regulates the terminal maturation and homeostasis of NK and Valpha14i NKT cells. Immunity. 2004;20:477–494. doi: 10.1016/s1074-7613(04)00076-7. [DOI] [PubMed] [Google Scholar]
  47. Van Gaal LF, Mertens IL, De Block CE. Mechanisms linking obesity with cardiovascular disease. Nature. 2006;444:875–880. doi: 10.1038/nature05487. [DOI] [PubMed] [Google Scholar]
  48. Vosshenrich CA, Garcia-Ojeda ME, Samson-Villeger SI, Pasqualetto V, Enault L, Richard-Le Goff O, Corcuff E, Guy-Grand D, Rocha B, Cumano A, et al. A thymic pathway of mouse natural killer cell development characterized by expression of GATA-3 and CD127. Nature immunology. 2006;7:1217–1224. doi: 10.1038/ni1395. [DOI] [PubMed] [Google Scholar]
  49. Weisberg SP, McCann D, Desai M, Rosenbaum M, Leibel RL, Ferrante AW., Jr Obesity is associated with macrophage accumulation in adipose tissue. The Journal of clinical investigation. 2003;112:1796–1808. doi: 10.1172/JCI19246. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Wensveen FM, Jelencic V, Valentic S, Sestan M, Wensveen TT, Theurich S, Glasner A, Mendrila D, Stimac D, Wunderlich FT, et al. NK cells link obesity-induced adipose stress to inflammation and insulin resistance. Nature immunology. 2015;16:376–385. doi: 10.1038/ni.3120. [DOI] [PubMed] [Google Scholar]
  51. Wentworth JM, Naselli G, Brown WA, Doyle L, Phipson B, Smyth GK, Wabitsch M, O’Brien PE, Harrison LC. Pro-inflammatory CD11c+CD206+ adipose tissue macrophages are associated with insulin resistance in human obesity. Diabetes. 2010;59:1648–1656. doi: 10.2337/db09-0287. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Wong N, Fam BC, Cempako GR, Steinberg GR, Walder K, Kay TW, Proietto J, Andrikopoulos S. Deficiency in interferon-gamma results in reduced body weight and better glucose tolerance in mice. Endocrinology. 2011;152:3690–3699. doi: 10.1210/en.2011-0288. [DOI] [PubMed] [Google Scholar]
  53. Xu H, Barnes GT, Yang Q, Tan G, Yang D, Chou CJ, Sole J, Nichols A, Ross JS, Tartaglia LA, Chen H. Chronic inflammation in fat plays a crucial role in the development of obesity-related insulin resistance. The Journal of clinical investigation. 2003;112:1821–1830. doi: 10.1172/JCI19451. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Ye J, Gao Z, Yin J, He Q. Hypoxia is a potential risk factor for chronic inflammation and adiponectin reduction in adipose tissue of ob/ob and dietary obese mice. American journal of physiology Endocrinology and metabolism. 2007;293:E1118–1128. doi: 10.1152/ajpendo.00435.2007. [DOI] [PubMed] [Google Scholar]
  55. Zhang N, Bevan MJ. Transforming growth factor-beta signaling controls the formation and maintenance of gut-resident memory T cells by regulating migration and retention. Immunity. 2013;39:687–696. doi: 10.1016/j.immuni.2013.08.019. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

supplement
NIHMS808507-supplement.pdf (4MB, pdf)

RESOURCES