Obesity Modulates Intestinal Intraepithelial T Cell Persistence, CD103 and CCR9 Expression, and Outcome in Dextran Sulfate Sodium–Induced Colitis

Rates of obesity have tripled worldwide in the last three decades (1). In the United States, 39.8% of adults and 18.5% of youth were classified as obese in 2015–2016 (2). Overnutrition and reduced activity are known contributors to the growth in obesity (3). At the forefront of mediating the influx of nutrients into the body is the intestinal barrier, which is also responsible for guarding against pathogens and mediating symbiosis with the microbiome. The intestinal environment requires constant maintenance and rapid repair to preserve the integrity of the epithelium. Obese mice exhibit decreased mucosal barrier function, altered tight junctions, and increased permeability in both genetic and diet-induced obesity models such as leptin-deficient (Lep^ob) mice, leptin receptor–deficient (Lepr^db) mice, and mice fed a high-fat diet (HFD) (4–6). This disruption in intestinal barrier function leads to permeability and leakage of bacterial pathogens, which can contribute to chronic low levels of inflammation (4, 6). Obesity has been linked to systemic inflammation and increased intestinal pathologies such as inflammatory bowel disease and colorectal cancer (7, 8). In addition, obesity exacerbates colitis in dextran sulfate sodium (DSS)–treated mice (9), and inflammatory bowel disease in obese patients results in increased perianal complications and requires more frequent treatment with surgery (7, 10, 11).

A complex population of lymphocytes reside within the epithelia at the intestinal barrier where they regulate epithelial homeostasis, protection, and repair. Intraepithelial lymphocytes (IELs) include natural (thymus differentiated) and induced (peripherally differentiated) T cells that balance barrier maintenance and pathogen control. Natural IELs, such as γδ T cells, control transepithelial migration of pathogens in the intestine (12) and mediate epithelial repair after DSS-induced damage (13). Induced IELs include memory CD4⁺αβ and CD8⁺αβ T cell populations, which provide protective antimicrobial responses (14). Early in life, the intestinal epithelium is dominated by natural IELs, which acquire their intestinal homing properties in the thymus. However, induced IELs increase with age and eventually surpass natural IEL numbers as they become activated in the periphery and obtain signals to migrate to the intestine (15). Epithelial localization of both natural and induced IELs is controlled by CD103/E-cadherin, occludins, and CCR9 (16, 17). Together, these receptors contribute to IEL persistence and population heterogeneity in the intestinal epithelia.

Obesity alters the induced Th17 and regulatory T cells in the lamina propria suggesting that mucosal immunity is sensitive to fluctuations in adiposity, and this may exacerbate inflammation leading to type 2 diabetes (18, 19). Although important in barrier repair and protection, epithelial T lymphocytes have been less extensively studied. In this study, we examine how obesity and diet-induced weight loss modulate the intestinal IEL compartment. Mice fed an HFD exhibit a reduction in both αβ and γδ T lymphocytes in the intestinal epithelia. Induced CD4⁺ subsets, which are already somewhat rare in the epithelia, are especially vulnerable, showing impaired homeostatic proliferation. Neither apoptosis nor TNFR signaling regulate the reduction of IELs in obesity, but instead, the IELs downregulate CD103 and CCR9, which are required for retention in the epithelia. Diet-induced weight loss restores the number of IELs, upregulates CD103 and CCR9 expression, and protects mice from the severe effects of DSS-induced colitis.

Male C57BL/6J and C57BL/6N mice were purchased from The Jackson Laboratory and Taconic Biosciences, respectively, and housed at California State University San Marcos. TNFR1^−/− and TNFR2^−/− mice were purchased from The Jackson Laboratory and bred at California State University San Marcos. B6.BKS(D)-Lepr^db/J mice were purchased from The Jackson Laboratory and studied at 6 wk (early obesity) and 4–6 mo (late obesity). For diet-induced obesity, either 3-, 6-, or 10-wk-old mice were fed an HFD (60% kcal from fat) (Research Diets, Brunswick, NJ) or a normal chow diet (NCD; 17% kcal from fat) (7012; Harlan Envigo) for 7–46 wk. In additional experiments, control mice were fed a different NCD (29% kcal from fat) (7004; Harlan Envigo) with similar results leading to use of 7012 chow. The HFD administered was Research Diets D12492, which contains 60% kcal from fat (lard and soybean oil), 20% kcal from protein (casein and cystine), 20% kcal from carbohydrate (lodex 10 and sucrose). NCDs do not exactly match HFDs with regard to nutritional content. In some cases, 6-wk-old mice were fed an HFD and then switched to an NCD for 4–7 wk (dieted mice). All mice were periodically weighed, and blood glucose was monitored by a Contour blood glucose monitor (Bayer). Fat mass was calculated using wet and dry mass with an estimate for fat proportion: Y = 1.0844x − 0.2818, x = dry mass/wet mass (20). The animal procedures were carried out in accordance with the animal research protocols approved by the Institutional Animal Care and Use Committee of California State University San Marcos (no. 18-008).

Abs specific for CD3ε (145-2C11), CD4 (RM4-5), CD8α (53-6.7), CD8b.2 (53-5.8), γδ TCR (GL3), αβ TCR (H57-597), CD11b (M1/70), CD11c (HL3), CD69 (H1.2F3), CD103 (2E7), and CCR9 (CW-1.2) were purchased from BioLegend. Annexin V Staining Kit (BD Biosciences, San Diego, CA) was used per manufacturer’s instructions. For BrdU studies, following an initial i.p injection of BrdU, mice were administered BrdU in drinking water at 0.8 mg/ml for 7 d. BrdU Staining Kit (BD Biosciences) was used in accordance with manufacturer’s instructions. Cells were run on a LSR II (BD Biosciences) or Accuri C6 (BD Biosciences) and analyzed with FlowJo software (BD Biosciences). Gating strategies are shown in Supplemental Fig. 1.

The small intestine was harvested, cleaned, and cut open longitudinally. Peyer’s patches were removed, and the intestine was cut into small pieces. Sections of intestine were shaken twice in HBSS without calcium/magnesium/phenol red, with 1 mM dithioerythritol, 1 mM EDTA, and 10% FCS at 165 rpm for 30 min at 37°C. IELs were enriched using a 40/70% Percoll gradient. For intracellular cytokine staining, cells were stimulated for 5 h with 100 ng/ml PMA, fixed and permeabilized using BD Cytofix/Cytoperm (BD Biosciences), and stained with Abs prior to running samples on a Accuri C6 (BD Biosciences) and analysis with FlowJo software (BD Biosciences). Live cells were gated and CD3⁺ cells analyzed.

The spleen and thymus were harvested and cells disassociated and spun through a Lympholyte-M gradient (Cedarlane, Burlington, NC) prior to counting and staining for flow cytometry. Cells were stained with Abs specific for CD3ε, γδ TCR, αβ TCR, CD4, and CD8 prior to running samples on a Accuri C6 (BD Biosciences) and performing analysis using FlowJo software (BD Biosciences). Live cells were gated and CD3⁺ or CD3⁺γδ TCR⁺ cells analyzed.

The small intestine was subdivided into three sections approximating the proximal, middle, and distal portions. The intestine was cleaned, cut open longitudinally, and rolled using a wooden dowel. Rolled sections were frozen using OCT compound on dry ice and stored at −80°C. Eight-micrometer sections were cut, fixed with 4% methanol-free formaldehyde and blocked for 30 min with gelatin block solution (2.5% goat serum, 2.5% donkey serum, 1.0% BSA, 2.0% gelatin from cold water fish skin, 0.1% Triton X-100, and 0.3 g glycine in PBS). Sections were stained with 10 μg/ml Ab or 1 μM BODIPY FL dye (Thermo Fisher Scientific, San Diego, CA) for 1 h, washed in PBS, and mounted using Antifade Medium with DAPI (Thermo Fisher Scientific). Twenty five to thirty digital images were acquired for each tissue using a Zeiss Axiocam at original magnification ×20. Images were analyzed, and cells were counted using ImageJ software (National Institutes of Health). Villi length was measured using ImageJ software (National Institutes of Health). For experiments quantifying cells in the upper, middle, and lower villus, each villus total length was measured using ImageJ software and split into three equal sections. For lipid droplet quantification, lipid droplets >0.7 μm in diameter within an epithelial cell were recorded as positive for lipid droplets.

Mice were given 2.5% DSS in their drinking water ad libitum for 6 d. This was followed by administration of drinking water without DSS for 5 d, allowing for tissue repair. Weight, rectal prolapse, and bleeding were monitored daily. At the end of the experiment, the small intestine and colon were harvested for further histological analysis by immunofluorescence and H&E staining.

Statistical analysis was performed using Prism Graph Pad and Excel software. Repeated measures ANOVA with post hoc Tukey test was used to determine the significance in body mass and T cell number over time. Two-way ANOVA with a post hoc Tukey test or Student unpaired t test were used to analyze T cell numbers and cytokine production in weanling, adolescent, and adult mice administered an NCD or HFD. To assess the effect on a given sample size, we performed power analysis by using preliminary datasets. A p value <0.05 was the cutoff for statistical significance.

When measured for body weight, fat storage, and leptin levels, C57BL/6J mice fed an HFD develop obesity and the associated complications in three stages: an early stage (1 wk on HFD), a middle stage (7–9 wk on HFD), and a late stage (19 wk on HFD) (21). In this study, we started by examining the impact of the late stage of obesity on IEL number in the small intestine. T cells within the epithelia of the proximal, middle, and distal sections of the small intestine were quantified in mice receiving an HFD versus NCD at this later timepoint of obesity development. The number of CD3⁺ IELs is significantly reduced in all three sections of the small intestine, with the greatest difference observed in the distal portion (Fig. 1A, 1B). This reduction in IEL numbers is significant per millimeter of the villi and per villus and is observed in mice obtained from both The Jackson Laboratory and Taconic Biosciences. Although there is a slight increase in villus height in the proximal section, it does not reach significance (Fig. 1C). IELs were isolated from the small intestine of mice administered an HFD versus NCD for 18–22 wk, and flow cytometric analysis further confirmed that IELs are reduced in mice receiving an HFD (Fig. 1D).

Both natural and induced populations of T lymphocytes reside in the intestinal epithelia (15). These IEL subsets can be defined by their expression of the αβ or γδ TCR with or without CD4/8αα/8αβ (Table I). In the small intestine of lean mice, γδ and αβ IELs are equally represented (Fig. 2A). Both γδ and αβ IEL numbers are decreased in mice fed an HFD (Fig. 2A, bottom), but the relative proportion of γδ to αβ IELs remains similar in mice fed an HFD or NCD (Fig. 2A, top). When further subdivided into specific subsets, it becomes clear that certain subsets, such as CD4⁺CD8α⁺ DP cells, are negatively impacted more than others (Fig. 2C, Table I). To determine the mechanism for the lack of IEL persistence in mice fed an HFD, IELs were examined for increased apoptosis or reduced homeostatic proliferation (Fig. 2B). Although none of the IEL subsets tested exhibit an increase in apoptosis, the ability of the induced CD4⁺ or CD4⁺CD8⁺ αβ T cell populations to incorporate BrdU is reduced (Fig. 2B). In contrast, natural γδ T cells exhibit an increase in their ability to proliferate in mice fed an HFD (Fig. 2B). These data suggest that whereas the persistence of both αβ and γδ T cell populations within the intestinal epithelium is impacted by administration of a long-term HFD, induced CD4⁺ or CD4⁺CD8⁺ αβ T cell populations exhibit a reduced ability to proliferate and maintain their numbers.

To further define the mechanism by which IELs become reduced in number, receptors required for intestinal epithelial adhesion and chemotaxis were examined. CD103 is highly expressed on IELs and mediates lymphocyte adhesion to the epithelium via E-cadherin (16). In addition, the expression of CCL25/CCR9 regulates CD103 and homing of T cells to the intestinal epithelium (22, 23). In mice fed an HFD, CD103 and CCR9 become downregulated on both natural γδ and induced CD4⁺ αβ IELs (Fig. 2D). Together, these data suggest that the mechanisms that contribute to the reduction in IEL subsets in mice fed an HFD are complex and subset dependent, with induced CD4⁺ or CD4⁺CD8⁺ αβ T cell populations more impacted because of problems with both homeostatic proliferation and adhesion molecule expression.

To investigate the influence of the early impacts of diet on IEL persistence, C57BL/6J mice were fed an HFD for 1 wk. Mice did not exhibit any changes in CD3⁺ IEL number whether they were fed an NCD or HFD (Fig. 3A, 3B), although we have shown that eosinophil number in the intestine is reduced at this early timepoint of obesity (24). To compare the impact of obesity attained by HFD versus obesity attained by overeating an NCD, a genetic obesity model was examined. Lepr^db on the B6 background [B6.BKS(D)-Lepr^db/J] exhibit obesity and type 2 diabetes because of the lack of satiety. Lepr^db mice are fed the same low-fat normal control chow as controls but begin to show signs of obesity by 4 wk of age, similar to the HFD model. Lepr^db mice at the early stage of obesity (6 wk of age) have the same number of IELs as control mice (Fig. 3C). As obesity progresses to the late stage in this mouse model (4–6 mo), Lepr^db mice show a reduction in CD3⁺ IEL number as compared with their heterozygous controls (Fig. 3D). Lepr^db mice do not exhibit a significant increase in villi length as compared with control mice (Fig. 3E). Although leptin is a well-known modulator of hormones, immune responses, and fertility, and thus, the Lepr^db model has caveats in the study of T cell function, this experiment provides additional evidence to support the premise that obesity due to an HFD and obesity due to overeating show a similar reduction in IEL number.

Our previous study showed that inhibition of TNF-α was sufficient to restore function to skin γδ T cells in the epidermis of obese and diabetic mice, suggesting that TNF-α plays a key role in regulating epithelial T cell function in obesity (25). To elucidate the impact of TNFR signaling on the reduction in IEL number, TNFR1^−/−2^−/− mice were administered either an NCD or HFD, and CD3⁺ IELs were quantified in the villi of the intestine. TNFR is not required for normal IEL homeostasis, as the number of IELs is similar between wild-type and TNFR1^−/−2^−/− mice fed an NCD (Fig. 3F). Upon HFD administration, the number of CD3⁺ IELs is reduced similarly in TNFR1^−/−2^−/− mice and wild-type mice, suggesting that the reduction of IELs in the intestine is independent of TNFR signaling (Fig. 3F).

To identify whether changes in IEL persistence due to obesity occur in young mice that are still establishing their IELs and 3- (weanling), 6- (adolescent), and 10-wk-old (adult) mice were examined (Fig. 4A). For this set of experiments, we focused on the middle stage of obesity (HFD for 7 wk), as it appears to be a key timepoint in which IEL numbers are reduced, and mice are showing significant weight gain and fat storage characteristic of obesity (Fig. 4B, 4C, 4E). All cohorts of mice experienced significant weight gain and fat accumulation as compared with mice fed NCD (Fig. 4B, 4C). Weanling mice at 3 wk of age do not yet have their full complement of IELs; thus, after 7 wk on an NCD, they exhibit a significant increase in IEL number (Fig. 4D). However, weanling mice fed an HFD for 7 wk exhibit similar IEL numbers as their 3-wk-old counterparts, suggesting that the normal seeding and homeostatic accumulation of IELs is impaired (Fig. 4D). Similarly, adolescent mice at 6 wk of age are also unable to establish normal IEL numbers when fed an HFD for 7 wk (Fig. 4E). Furthermore, when 10-wk-old adult mice are placed on an HFD for 7 wk, they are unable to gain and sustain normal IEL numbers (Fig. 4F). To determine whether obesity was causing the IELs to home a specific part of the villus, we quantified the number of IEL in the upper, middle, and lower third of each villus. IELs were reduced in all parts of the villus in mice fed an HFD, with significant reductions in the middle and lower parts of the villus, whereas the upper part is reaching significance (Fig. 4H). These results suggest that the T cells do not accumulate in a specific region of the villi, but still freely migrate throughout and possibly exit the villus entirely (Fig. 4H).

These findings led us to consider whether mice would eventually lose their IELs in the intestinal compartment, leading to a complete absence of IELs with advanced age. To study this, IEL numbers were quantified in mice aged 52 wk old. Starting at 6 wk of age, mice were either fed an NCD or HFD for 46 wk. Mice fed an NCD for 46 wk (age = 52 wk) exhibit fewer IELs than mice fed an NCD for 7 wk (age = 13 wk), suggesting that aging leads to a reduction in IEL number even in the absence of an HFD (Fig. 4E, 4G). However, the number of IEL found in mice fed an HFD for 46 wk is significantly lower than mice fed an NCD for 46 wk (Fig. 4G). In fact, it is interesting to note that the number of IELs remaining in mice fed an HFD for 46 wk is similar to the number found in weanling mice at 3 wk of age (Fig. 4D). Ultimately, it is interesting to note that some IELs still remain in aged mice fed an HFD. These data have implications for understanding the impact of obesity in children and the elderly on the development and maintenance of the intestinal IEL compartment.

To determine whether nonepithelial sites such as the spleen and thymus also exhibit reduced numbers of T cells, we performed flow cytometry to examine splenocytes and thymocytes from mice fed an HFD for 16–20 wk as compared with NCD. There are similar numbers and percentages of CD3⁺ T cells in the spleen of obese and lean mice (Fig. 5A, 5C). Similarly, no changes in CD3⁺ T cell populations were observed in the thymus (Fig. 5B, 5D). Proportionally, there is no skewing toward or away from γδ T cells in either the spleen or thymus (Fig. 5E, 5F).

IELs were isolated from mice fed an HFD for 7 wk starting as a weanling, adolescent, or adult (Fig. 4A). IELs were stimulated ex vivo and examined for cytokine production by flow cytometry. αβ IELs exhibit more robust TNF-α and IFN-γ production as compared with γδ IELs (Fig. 6A–D). γδ IELs do not exhibit much TNF-α upon stimulation in these conditions; however, γδ IELs from obese mice exhibit increased TNF-α production (Fig. 6A). The increase in TNF-α production only occurs in γδ IELs isolated from adult mice fed an HFD, whereas γδ IELs from weanling mice fed an HFD did not show elevated TNF-α (Fig. 6A). αβ IELs were able to produce TNF-α regardless of diet. When the αβ T cells were further gated on either CD4⁺ or CD8α⁺ T cells, a significant increase in TNF-α was observed in adolescent and adult mice fed an HFD. No significant differences in IFN-γ production were observed by either γδ or αβ IELs isolated from mice fed an HFD (Fig. 6B, 6D). Together, these data suggest that the IELs remaining in obese mice are functional, but γδ IELs in obese adult mice are skewed toward TNF-α production.

Obesity compromises the integrity of the intestinal barrier, as demonstrated by the increased susceptibility of mice fed an HFD to DSS-induced colitis (Fig. 7B–F) (9, 26). DSS treatment leads to the disruption of the intestinal epithelium, resulting in crypt loss, ulceration, and infiltration by inflammatory cells (Fig. 7F). Following administration of DSS, mice fed an HFD had increased rectal bleeding and rectal prolapse and had to be sacrificed per Institutional Animal Care and Use Committee protocol because of weight loss and excessive blood in the stool (Fig. 7B–D). To determine whether diet-induced weight loss would reduce the severity of colitis, mice were fed an HFD for 14–20 wk and then placed on a diet with NCD for 4 wk (Fig. 7A). On average, the dieted mice lost ∼10 g in the 4 wk (Fig. 7E). These dieted mice weighed less than their HFD counterparts but more than the controls on NCD. During DSS-induced colitis, dieted mice experience less bleeding than the obese mice (Fig. 7B). In addition, dieted mice have a reduced susceptibility to anal prolapse, exhibit a 100% survival rate (Fig. 7C), and lose the least amount of weight in response to DSS (Fig. 7D). Dieted mice exhibit less disruption of the intestinal epithelium, crypt loss, ulceration, and infiltration by inflammatory cells (Fig. 7F). Together, these data show that diet-induced weight loss protects mice from severe intestinal damage by DSS-induced colitis in obesity.

Interestingly, the expression of CD103 on γδ and CD4⁺ IELs, which is decreased in HFD mice, is almost fully restored following weight loss (Fig. 7G). Both γδ and CD4⁺ T cells isolated from the intestine of dieted mice exhibit higher expression of CCR9 and CD103 than the IELs isolated from mice administered an HFD (Fig. 7G). The upregulation of these molecules may enable more IELs to be retained in the intestine, indicating that weight loss could lead to the restoration of IEL numbers. Indeed, mice dieted for 4 wk exhibit a partial restoration in IEL numbers in the intestine (Fig. 7H). To determine whether IEL numbers in the intestine of mice fed an HFD can fully recover with a middle stage of obesity (7 wk) and longer duration of diet-induced weight loss (7 wk), mice were administered an HFD for 7 wk, followed by either 7 wk on HFD or 7 wk on NCD (Fig. 7A). Diet-induced weight loss of 7 wk restores IEL numbers back to levels of NCD-fed mice, indicating that the IELs are able to recover back to normal numbers (Fig. 7I). Mice dieted for 7 wk have significantly fewer numbers of epithelial cells with lipid droplets (Fig. 7J, 7K), suggesting changes in lipid droplets locally near the IELs may contribute to their recovery. However, the dieted mice have not yet reached normoglycemia (Fig. 7L) or weight equivalent to NCD-fed mice (Fig. 7M), suggesting that a complete return to normal glucose levels or weight is not required for the IEL to fully return. Together, these data show that diet-induced weight loss in obesity has a restorative impact on the intestinal immune system and promotes recovery from DSS-induced damage to the epithelia.

In the intestine, natural and induced IELs monitor the single-celled epithelium for damage and regulate inflammatory responses to pathogens (13, 15, 27). Extensive cross-talk occurs between epithelial cells and the IELs that monitor them for damage or infection. The data presented in this study support a detrimental role for obesity on the persistence of both natural and induced IELs. Natural γδ IELs normally produce growth factors and cytokines including keratinocyte growth factor-1 (KGF-1) and TGF-β that regulate the epithelial environment (13, 15). Mice deficient in γδ T cells exhibit more extensive mucosal injury and delayed repair during DSS-induced colitis because of the absence of KGF-1 (13). In addition, induced CD4⁺ αβ IELs play key roles in regulating inflammation and maintaining barrier integrity (27). In the absence of αβ T cells, γδ T cells can cause spontaneous colitis in the intestine, suggesting that αβ T cells control aberrant γδ T cell function (28). Thus, a reduction in IEL number and T cell composition during obesity impacts the ability of these cells to participate in tissue repair and control inflammation.

Villus hyperplasia has been reported in humans with obesity (29). The combination of increased villi length and the loss of IELs in obesity indicates that the remaining IELs have to mediate their protective function over a larger epithelial surface area. This may interfere with the ability of IELs to maintain epithelial integrity and ultimately contribute to barrier dysfunction in obesity. Parabiosis studies have shown that IELs do not recirculate but, instead, are tissue resident (30, 31). Within the tissue, IELs actively monitor the epithelia by migrating cell to cell to assess for infection or damage (12). IEL homing and migration are regulated by CCR9 and CD103. IELs are reduced in mice lacking CD103, and CCR9 regulates CD103 expression, suggesting a key role for these receptors in the homing and persistence of IELs (16, 22, 23). In this study, we show that CD4⁺ αβ T cells are susceptible to both disruptions in homeostatic proliferation and reduced expression of intestinal epithelial homing receptors CD103 and CCR9. γδ T cells are resistant to disruptions in homeostatic proliferation but sensitive to the downregulation of intestinal epithelial homing receptor expression. Thus, obesity reduces the ability of IELs to persist in the intestinal epithelial environment, which has negative consequences for monitoring of the intestinal barrier.

Obesity increases permeability of the intestinal barrier, which has been implicated in the initiation and perpetuation of systemic inflammation (4, 5, 24). Systemic TNF-α levels are elevated in obese patients (32, 33). Additionally, there are elevated levels of TNF-α in the small intestine and colon of obese mice (34). In this study, we show that γδ IELs isolated from adult obese mice exhibit elevated TNF-α production, further exacerbating the chronic inflammation associated with obesity. In addition, the reduction in CD4⁺ T cells that normally regulate T cells at the epithelium may enhance the proinflammatory environment. Previous work has shown that lamina proprial CD4⁺ regulatory T and Th17 cells also become reduced in the small intestine during obesity, suggesting that the loss of functional regulatory cells has negative impacts that go further than the epithelial layer (18).

The age of obesity onset has important implications on the composition of natural and induced IELs within the intestine as aging occurs. Although natural IELs seed and populate the intestinal epithelium early in life, induced IELs become activated in the periphery and migrate to the intestine (15). We have shown, in this study, that weanlings have few IELs seeding the intestine. Thus, obesity in weanlings may have a more severe long-term impact than adult-onset obesity because of impaired αβ CD4⁺ T cell proliferation and T cell homing/retention. This has serious implications because worldwide childhood obesity has risen from under 1% of 5–19-y-olds in 1975 to 6% (girls) and 8% (boys) in 2016 (1). In the United States, obesity rates are even more dire, with 18.5% of United States youth obese in 2015–2016 (2).

Although the mechanism by which obesity induces immunomodulation of the intestine is not completely clear, it occurs within 7 wk in a manner that is independent of TNFR signaling. Elevated glucose and leptin levels may modulate the IELs as both become elevated in the middle stage of obesity, although the effect of leptin on T lymphocytes may be indirect, as previously shown (35). It is important to note that the type of dietary fat and other nutritional components impacts obesity and type 2 diabetes; thus, the use of lard and soybean oil and the lack of a perfect nutritional control in this study limits the ability to broadly ascertain the impact of nutrition in the IELs in the intestine. However, the impact of obesity on T cell persistence and colitis is reversible with diet-induced weight loss, suggesting that administering a low-fat diet or inducing caloric restriction may improve barrier immunity in obesity. Weight loss in obese patients is associated with reduced intestinal inflammation and a decreased risk of colorectal cancer in patients (36). Future studies are necessary to identify the impact of weight loss on recovering normal immune composition within the intestine to reduce the risk of severe outcomes in colitis.

We thank Peggy Han, Zerlina Clementsmith, Shelley Dutt, and Kristen Taylor for advice and technical help. We also thank Drs. Deborah Witherden and Kerri Mowen at The Scripps Research Institute for helpful suggestions regarding experiments and Benjamin Maderazo for help with figures.

The authors have no financial conflicts of interest.