Diet-induced obesity does not impact the generation and maintenance of primary memory CD8 T cells

Abstract

The extent to which obesity compromises the differentiation and maintenance of protective memory CD8 T cell responses and renders obese individuals susceptible to infection remains unknown. Here, we show that diet-induced obesity did not impact the maintenance of pre-existing memory CD8 T cells including acquisition of a long-term memory phenotype (i.e. CD27^hi, CD62L^hi, KLRG1^low) and function (i.e. cytokine production, secondary expansion, and memory CD8 T cell-mediated protection). In addition, obesity did not influence the differentiation and maintenance of newly evoked memory CD8 T cell responses in inbred and outbred hosts generated in response to different types of systemic (LCMV, L. monocytogenes) and/or localized (influenza virus) infections. Interestingly, the rate of naïve-to-memory CD8 T cell differentiation after a peptide-coated dendritic cell (DC) immunization was similar in lean and obese hosts suggesting that obesity associated inflammation, unlike pathogen- or adjuvant-induced inflammation, did not influence the development of endogenous memory CD8 T cell responses. Therefore, our studies reveal that the obese environment does not influence the development or maintenance of memory CD8 T cell responses that are either primed before or after obesity is established, a surprising notion with important implications for future studies aiming to elucidate the role obesity plays in host susceptibility to infections.

Introduction

The obesity epidemic is a public health concern currently affecting more than 30% of the adult population in the United States (1, 2) and obese individuals have a higher chance of developing coronary heart disease, several types of cancer, and type 2 diabetes (3–9). The rate of childhood and adolescent obesity is also increasing at a startling rate (currently affecting 18% of children and 21% of adolescents in the United States), suggesting that as these youth age, they are more likely to become obese as adults and develop associated health problems (10). Alarmingly, overweight and obesity are considered leading risk factors for global deaths, indicating the worldwide impact of this condition (11).

Recent studies have shown that obesity increases susceptibility to infection. For example, during the H1N1 influenza pandemic, obesity was found to be an independent risk factor associated with increased morbidity and mortality, as obese individuals were more likely to become admitted into the ICU and succumb to death (12). Obese hosts are also predisposed to pneumonia because of reduced lung volumes and altered ventilation patterns (13). Improper wound healing renders obese patients susceptible to nosocomial infections at surgical sites (14, 15). In addition to being a factor that increases susceptibility to infection, obesity is also associated with decreased vaccine efficacy. The first documented association between obesity and poor antibody responses was found after hepatitis B plasma vaccine administration in hospital personnel (16). Compared to lean youth, obese children and adolescents were found to have reduced anti-tetanus antibody titers (17). Furthermore, after inactivated trivalent influenza vaccination, humans with a higher body mass index had a greater decline in influenza-specific antibody titers a year after immunization (18). These data suggest that obese individuals are not only more susceptible to infection, but also may not be receiving the full benefits of current vaccine protocols.

Recently, vaccination strategies aiming to elicit a robust memory CD8 T cell response have gained recognition as an avenue to control infectious diseases that are resistant to antibody-based vaccines (19–21). Following an acute infection or vaccination, memory CD8 T cells are generated from a low number of naïve Ag-specific precursors that have undergone vigorous clonal expansion and contraction phases(19, 20, 22, 23). In addition to persisting at higher numbers in a host, memory CD8 T cells possess the unique ability to rapidly respond to re-infection and kill pathogen-infected cells by initiating effector functions such as cytokine production and the release of perforin and granzyme (24). Importantly, the degree of CD8 T cell-mediated protection correlates with the number of memory CD8 T cells present in vivo and their functional characteristics (25–27).

Recent studies showed that the inflammation present early after Ag encounter influences the homeostasis of primary Ag-specific CD8 T cell responses, including the magnitude of proliferative expansion as well as the rat eat which Ag-specific CD8 T cells acquire memory CD8 T cell characteristics (19, 28–33). Interestingly, obesity as a multi-factorial disease is associated with low-grade systemic inflammation that can potentially influence naïve-to-memory CD8 T cell differentiation. It has been shown that visceral adipocytes and macrophages infiltrating growing adipose tissue of obese individuals secrete increased levels of cytokines and chemokines, such as TNFα and CCL2, triggering this systemic inflammation (34–36). The hypertrophic changes that allow adipose tissue expansion result in adipocyte cell death and the release of ATP and uric acid, which serve as danger signals to alert the immune system of metabolic changes taking place in the host(6, 37). The local adipose tissue environment of obese hosts not only initiates continuous infiltration of macrophages, but also promotes their acquisition of a pro-inflammatory M1 polarized status because of already increased levels of TNFα within expanding adipose tissue (38–40). Importantly, the extent to which obesity-associated changes in the environment influence pre-existing or de novo generated memory CD8 T cell responses is currently unknown.

Here, using a diet-induced obesity model we show that obesity does not influence the generation, long-term maintenance, and function of memory CD8 T cell responses induced either by systemic or localized viral and bacterial infections, or by immunization with peptide-coated dendritic cells.

Materials and Methods

Mice and Diet-Induced Obesity

Inbred male C57BL/6 (B6 Thy1.2⁺/1.2⁺), female BALB/c, and female outbred Swiss Webster (SW) mice, were purchased from the National Cancer Institute (NCI) and housed under specific pathogen-free conditions. At 6–12 weeks of age, mice were assigned at random to either receive ad libitum standard chow diet (Harlan Teklad number 7913, 6.2% kcal from fat) or high-fat diet (HFD, Research Diets number 12492, 60% kcal from fat) for indicated times before the start of an experiment, unless otherwise noted, and then throughout the duration of an experiment. Mice receiving HFD are considered obese if body weight is greater than 3 SD above the mean weight of the lean group (41). Thy1.1⁺/1.2⁻ P14 and Thy1.1⁺/1.2 OT-I⁺ (specific for lymphocytic choriomening it is virus (LCMV)-derived GP33 and chicken ovalbumin Ova257 epitopes, respectively) T-cell-receptor transgenic (TCR-Tg) mice were bred at the University of Iowa and described previously (29). All animal procedures followed approved Institutional Animal Care and Use Committee (ACURF) protocols.

Pathogens

The Armstrong strain of LCMV (LCMV, 2×10⁵ PFU/mouse, i.p.) and the virulent Listeria monocytogenes (LM) strains expressing Ova257 (Vir LM-Ova, 5×10⁴ CFU/mouse, i.v.), GP33 (Vir LM-GP33, 2×10⁵ CFU/mouse, i.v.) and attenuated actA-deficient L. monocytogenes strain expressing Ova257 (Att LM-OVA, 5×10⁶ CFU/mouse, i.v.) were grown, injected, and quantified as described (25, 26, 42, 43). Mouse-adapted Influenza A viruses A/PR/8/34 and X-31 (IAV, PR8 and X-31, 2.75×10² tissue culture infectious units (TCIU) of virus/mouse, i.n.) were prepared from stocks, as previously described(44). Mice were first anesthetized and given a primary infection with IAV strain PR8 and then challenged 30 days later with IAV strain X-31. Infected mice were housed at the University of Iowa under the appropriate biosafety level.

Adoptive transfer and isolation of lymphocytes from tissue

Thy1.1⁺/1.2⁻ P14 and Thy1.1⁺/1.2⁺ OT-I CD8 T cells were obtained from peripheral blood samples of young naïve TCR-Tg P14 and OT-I mice, respectively. Contaminating memory phenotype (CD44^hi CD11a^hi) P14 and OT-I cells were always <5%. Naïve P14 CD8 T cells were injected (5×10³ cells/mouse, i.v.) into naïve B6 (Thy1.2⁺/1.2⁺) recipients. Naïve OT-I CD8 T cells were injected (1×10³ cells/mouse, i.v.) into LCMV immune lean and obese B6 (Thy1.2⁺/1.2⁺) recipients.

Before removal of secondary lymphoid organs and tertiary tissues, samples of blood were obtained by retro-orbital puncture. Anesthetized mice were then perfused through the left ventricle with cold PBS and tissues including spleen, lymph node (LN), and lung were collected. Single-cell suspensions from spleen and LN were washed before Ab staining. Lungs were first cut into small pieces and treated with collagenase D (125U/mL) and DNAse I for 30 min at 37°C before further processing and Ab staining.

Antibodies and peptides

Flow cytometry data was acquired on a FACSCanto flow cytometer (Beckton-Dickinson Biosciences) and analyzed with FlowJo software (Tree Star). The following is a list of used mAbs with the indicated specificity and appropriate combinations in fluorochromes from eBioscience: CD8 (clone 53–6.7), Thy1.1 (HIS51), Thy1.2 (53–2.1), CD11a (M17/4), CD27 (LG.759), CD62L (MEL-14), KLRG1 (2F1), CD127 (A7R34), and appropriate isotype controls. Intracellular staining for mAb IFNγ (XMG1.2, Biolegend), TNFα (MP6-XT22, Biolegend), IL-2 (JES6–5H4, eBioscience), Granzyme B (Invitrogen, FGB12), and appropriate isotype control was performed after surface fixation and permeabilization of the cell membrane using cytofix/cytoperm solution. To assess degranulation, we added FITC-conjugated CD107a (1D48, eBioscience) or isotype control and monensin A to cells during the stimulation period as previously described (45). Synthetic GP33–41 and Ova257–264 peptides were used as previously described (30, 46).

Quantification of CD8 T cell responses and intracellular cytokine staining

P14 and OT-I cell responses in the peripheral blood and tissues were monitored by FACS analysis for Thy1.1⁺/1.2⁻ and Thy1.1⁺/1.2⁺ positive CD8 T cells. Endogenous GP33-and Ova257-specific CD8 T cells were detected by allophycocyanin-conjugated tetramer complexes as previously described (30). Endogenous IAV NP366 and PA224-specific CD8 T cells were detected by allophycocyanin-conjugated tetramer complexes obtained from the National Institutes of Health and Tetramer Core Facility (Atlanta, GA)(47). Additionally, pathogen-specific CD8 T cells were identified in infected outbred SW mice using a’surrogate activation marker approach’ as previously described (48). Briefly, differences in expression of CD11a and CD8α were used to distinguish Ag-experienced (CD11a^hi CD8α^low) from naïve (CD11a^low CD8α^hi) CD8 T cells. Cells were incubated with mAb at 4°C for 30 min, washed with FACS buffer (PBS containing 1% FCS and 0.1% NaN₃), and then fixed with cytofix/cytoperm solution. The percentage of CD8 T cells producing cytokines after stimulation with GP33 or Ova257 peptides was determined using intracellular cytokine staining for IFNγ, TNFα, IL-2, and Granzyme B, after 5h incubation in brefeldin A with or without indicated peptides. For intracellular cytokine staining of lymphocytes isolated from peripheral blood, thymoma tumor cell-line derived EL-4 cells (H-2^b MHC) were used as antigen-presenting cells.

Bacterial titers

Bacterial titers were determined as previously described (30). Briefly, tissue was harvested from infected mice at the indicated time after challenge and placed into 4 mL of 0.2% IGEPAL buffer (Sigma-Aldrich) and sterile deionized water. A homogenizer was used and resulting suspension was incubated at room temperature for 1h. Serial dilutions were plated in streptomycin agar plates to quantify bacterial load. The dashed line indicates the limit of detection, LOD.

Dendritic cell immunizations

Splenic dendritic cells (DCs) were isolated after s.c. injection of B6 or BALB/c mice with 5×10⁶ B16 cells expressing Flt3L as previously described(32, 49). When tumors were palpable (5 × 5 mm), mice were injected with 2 μg LPS (Sigma-Aldrich) i.v. to mature the DCs. Spleens were harvested 16h later and digested with DNAse and collagenase for 20 min at 37°C/5% CO₂ with shaking (120 RPM). Spleen pieces were smashed through a nylon cell strainer (70 μm) to generate a single-cell suspension, RBCs were lysed, and splenocytes were suspended in 2 parts 10% FCS RPMI 1640 to 1 part B16-Flt3L-conditioned media + rGM-CSF (1000 U/mL) plus 2 μM Ova257- 264 or CS252–260, respectively, and incubated 2h at 37°C/5%CO₂ with shaking (100 RPM). Spleen cells were washed three times and CD11c+ cells were isolated using anti-CD11c microbeads (Miltyni Biotec). Routinely, >90% pure CD11c+ DCs were obtained. DCs were resuspended in saline and injected i.v. at 5×10⁵ to 1×10⁶ per mouse.

Cytokine and chemokines evaluation by BioPlex

Plasma samples were obtained from lean and obese mice. Plasma concentrations of the indicated cytokines and chemokines were determined via Multiplex analysis (Procarta Plex; Affymetrix by eBioscience) on a Bio-Rad BioPlex.

Statistical Analyses

Data were analyzed with Prism4 GraphPad software to determine statistical significance as indicated in figure legends. Statistical significance was assessed using the two-tailed, unpaired student’s T test, with a confidence interval >95% (*p ≤ 0.05, **p ≤ 0.005, ***p ≤ 0.001. and n.s. as no significance). Data presented as bar graphs are presented as mean±SEM, unless otherwise noted. Representative dot plots and histograms are presented as mean±SEM. Weight graphs are presented as scatter dot plots with mean±SD.

Results

Diet-induced obesity does not impact the maintenance of pre-existing memory CD8 T cells

The maintenance of memory CD8 T cells is a dynamic process that includes time-dependent changes in phenotype and function (20, 50–52) and in the absence of cognate Ag stimulation memory CD8 T cells can remain stable in numbers for the life of the laboratory mouse (53, 54). However, memory CD8 T cells undergo attrition in total numbers when mice are infected with unrelated (heterologous) pathogens or exposed to inflammation (55–57). The extent to which the obese environment (characterized by low levels of systemic inflammation) (6) influences the maintenance of pre-existing memory CD8 T cell responses is currently unknown. To address this, immune B6 Thy1.2⁺ mice that contained primary memory GP33-specific P14 T-cell-receptor transgenic (TCR-Tg) Thy1.1⁺/1.2⁻ CD8 T cells were started on a high-fat diet (60% kcal from fat), a month after primary LCMV infection (Fig. 1A-C). Importantly, waiting for 30 days to introduce the high-fat diet ensured two things – clearance of the virus (which usually occurs in the first week post infection (58)) and the establishment of memory CD8 T cell responses upon completion of the contraction phase (19)

Figure 1. Obesity does not impact the maintenance of pre-existing memory CD8 T cells.

A) Experimental design. Naïve B6 Thy1.2⁺/1.2⁺ mice received a transfer of naïve Thy1.1⁺/1.2⁻ P14 CD8 T cells (5×10³ cells/mouse, i.v.) and were infected with LCMV (2×10⁵ PFU/mouse, i.p.). 30 days after infection, mice either continued to receive standard chow diet or were started on high-fat diet. B) Representative dot plot showing the percentage of memory P14 CD8 T cells in the PBL day 30 after infection. C) Blood samples were pooled, and the expression of CD27, CD62L, or KLRG1 molecules was evaluated individually on memory P14 CD8 T cells day 30 after infection. Shaded graph represents isotype control staining and open graphs represent specific Ab staining on gated P14 cells. D) Weight of lean and obese mice 120 days after high-fat diet. Dots represent individual mice and the lines represent the mean and SD. E) The percentage of memory P14 CD8 T cells in the PBL of individual lean and obese mice 120 days after high-fat diet. F) The percentage of memory P14 CD8 T cells in lean and obese mice expressing CD27, or CD62L, or KLRG1 120 days after high-fat diet. G) Representative dot plots showing cytokine production by memory P14 CD8 T cells, isolated from the PBL of lean and obese mice 120 days after high-fat diet, after short ex vivo incubation in the presence of EL-4 thymoma cells (H-2^b MHC) as antigen-presenting cells and GP33 peptide. Numbers represent the percentage of memory P14 CD8 T cells positive for IFNγ and TNFα or IL-2. Representative histograms showing the percentage of IFNγ+ memory P14 CD8 T cells in the PBL of lean and obese mice expressing H) Granzyme B or I) CD107a following ex vivo incubation with GP33. Shaded graph represents isotype control staining and open graphs represent staining on gated memory P14 CD8 T cells. Data are of 6–9 mice per group and representative of three independent and similar experiments. Weight graph is presented as scatter dot plot with mean±SD. Data presented as bar graphs are presented as mean±SEM. Representative dot plot is presented as mean±SEM. Representative histogram is presented as mean±SEM.

Within 120 days of receiving high-fat diet, mice had gained a significant amount of weight and become obese (as defined by body weight greater than 3 SD above the mean weight of age-matched LCMV-immune lean B6 mice that were given standard chow diet (6.2% kcal from fat) for the duration of the experiment (41, 59) (Fig. 1D)). Interestingly, similar frequencies of memory P14 CD8 T cells were detected in the peripheral blood (PBL) of lean and obese mice suggesting that diet-induced obesity did not influence the ability of the host to maintain its memory CD8 T cell numbers (Fig. 1E). In addition, time-dependent changes in the phenotype of memory P14 cells (increase in CD27 and CD62L and decrease in KLRG-1 expression (20, 28)) were observed from day 30 (start of the high-fat diet) to day 150 after LCMV infection. Importantly, those changes occurred at similar rates in memory CD8 T cells detected in both lean and obese groups (Fig. 1C,F). A similar percentage of P14 cells from lean and obese mice produced effector cytokines such as IFNγ, TNFα and IL-2 after ex vivo GP33 peptide stimulation (Fig. 1G). Finally, the cytolytic potential of memory CD8 T cells defined by the expression of Granzyme B and CD107a, a marker of degranulation (45), was similar in both groups of mice (Fig. 1H,I). Thus, these data suggest that diet-induced obesity does not influence the number, phenotype, or function of memory CD8 T cells generated before the onset of obesity.

Obesity does not impact the generation and maintenance of newly evoked memory CD8 T cell responses

In order to investigate the extent to which the obese environment impacts the generation and differentiation of newly evoked memory CD8 T cells, groups of LCMV-immune lean and diet-induced obese B6 mice (initially described in Figure 1) were adoptively transferred with naïve Ova257-specific OT-I TCR-Tg CD8 T cells (29). Low numbers of naïve Thy1.1⁺/1.2⁺ OT-I cells (1×10³ cells/mouse), obtained from lean TCR-Tg mice, were seeded into recipient mice one day before infection with attenuated L. monocytogenes strain expressing Ova257 epitope (Att Lm-Ova; Fig. 2A)(29). This experimental set-up also allowed for continuous monitoring of the LCMV-specific memory P14 cell responses (Thy1.1⁺/1.2⁻) that were generated before the onset of obesity (Fig. 2A,B). Examination of OT-I CD8 T cells in the blood revealed that the obese environment did not influence the magnitude of proliferative expansion and accumulation of Ova257-specific effector CD8 T cells (Fig. 2C). L. monocytogenes-specific memory CD8 T cells in lean and diet-induced obese hosts persisted at similar numbers (Fig. 2D) and underwent time-dependent changes in their phenotype at similar rates (Fig. 2E and Suppl. Fig. 1A). Also, continuous high-fat diet did not influence the long-term maintenance and function of LCMV-specific P14 CD8 T cells. The number (determined in PBL and secondary lymphoid organs- Fig. 2F, Suppl. Fig. 1C), phenotype (Fig. 2G and Suppl. Fig. 1B), and function (peptide stimulated IFNγ, TNFα and IL-2 production- Fig. 3A) of memory P14 CD8 T cells detected in lean and diet-induced obese hosts were indistinguishable. Finally, to test whether memory CD8 T cell in obese hosts have an impaired ability to control secondary infection, both groups of mice were challenged with attenuated L. monocytogenes strain expressing GP33 epitope (Att LM-GP33- Fig. 3B). L. monocytogenes was used for secondary infection since anti-Listerial immunity is predominately mediated by CD8 T cells (60). Since the number of memory CD8 T cells present at the time of re-infection correlates with the level of CD8 T cell-mediated protection (21, 25–27), total numbers of endogenous and TCR-Tg GP33-specific CD8 T cells were determined by D^bGP33 tetramers to document that LCMV-immune lean and diet-induced obese mice contained similar percentages of GP33-specific CD8 T cells (Fig. 3C,D). Importantly, both groups of immune mice were able to clear L. monocytogenes infection at day 2 post challenge (Fig. 3E), suggesting that CD8 T cell-mediated immunity was not impaired by obesity.

Figure 2. Obesity does not impact the generation and maintenance of newly evoked memory CD8 T cell responses.

A) Experimental design. LCMV immune lean and obese mice containing memory Thy1.1⁺/1.2⁻ P14 CD8 T cells received a transfer of naïve Thy1.1⁺/1.2⁺ OT-I CD8 T cells (1×10³ cells/mouse, i.v.) one day before Att LM-Ova infection (5×10⁶ CFU/mouse, i.v.). B) Representative dot plot showing the percentage of memory Thy1.1⁺/1.2⁻ P14 and effector Thy1.1⁺/1.2⁺ OT-I CD8 T cells in the PBL of lean and obese mice 7 days after LM challenge. The frequency of OT-I cells in the PBL of individual mice at C) day 7 and at D) day 60 after bacterial infection. E) The percentage of memory OT-I CD8 T cells in lean and obese mice expressing CD27, or CD62L, or KLRG1 at day 60 p.i. F) The percentage of memory P14 CD8 T cells in the PBL of individual lean and obese mice 216 days after LCMV infection. G) The percentage of memory P14 CD8 T cells in lean and obese mice expressing CD27, or CD62L, or KLRG1 at day 216 after LCMV infection. Data are of 6 mice per group and representative of three independent and similar experiments. Data presented as bar graphs are presented as mean±SEM.

Figure 3. The protective function of memory CD8 T cells is not impacted by obesity.

A) Representative dot plots showing cytokine production by memory P14 CD8 T cells, isolated from the spleen of individual lean and obese mice at day 236 after LCMV infection, after short ex vivo incubation in the presence of GP33 peptide. Numbers represent the percentage of memory P14 CD8 T cells that were positive for IFNγ and TNFα or IL-2. B) Experimental design. At day 236 after infection, LCMV immune lean and obese mice containing memory P14 CD8 T cells were challenged with Vir LM-GP33 (2×10⁵ CFU/mouse, i.v.). C) Representative dot plots showing the frequency of tetramer+ (D^bGP33) GP33-specific memory CD8 T cells in the PBL. D) The percentage of GP33-specific CD8 T cells in the PBL of individual lean and obese mice on the day of challenge. E) Bacterial titers in the spleen (CFU/spleen) were determined 2 days after Vir LM-GP33 infection. Data are presented as mean of 3–5 mice per group. Dashed line represents the limit of detection (LOD). Data are of 3–5 mice per group and representative of two independent and similar experiments. Representative dot plot is presented as mean±SEM. Data presented as bar graph is presented as mean±SEM.

Taken together, these data suggest that the obese environment does not influence the development or maintenance of memory CD8 T cell responses that are either primed before or after obesity is established.

Endogenous CD8 T cell responses to systemic and localized viral and bacterial infections are not influenced by diet-induced obesity

In the studies above, we utilized adoptive transfers of naïve TCR-Tg CD8 T cell precursors obtained from lean Tg hosts to determine the extent to which obesity influences the development and maintenance of memory CD8 T cell responses. To determine whether obesity alters the ability of naïve endogenous CD8 T cell precursors to respond to infection and develop into memory CD8 T cells, endogenous pathogen-specific CD8 T cell responses were analyzed after systemic and localized viral or bacterial infections. After receiving high-fat diet for 100 days, diet-induced obese B6 mice and age-matched lean control mice were infected with LCMV (Fig. 4A,B). Similar percentages of endogenous D^bGP33-specific CD8 T cells were detected in the blood of lean and obese hosts at a memory time point (day 110 post infection) (Fig. 4C,D). In addition, the expression of surface markers associated with memory CD8 T cell development (i.e. CD27, CD62L, and KLRG1) was indistinguishable between GP33-specific memory CD8 T cells from both groups of mice (Fig. 4E). Finally, to determine if endogenous memory CD8 T cells primed in a lean or obese environment differ in their ability to undergo Ag-stimulated secondary expansion, LCMV-immune mice were challenged with virulent L. monocytogenes expressing GP33 epitope (Vir LM-GP33) (Fig. 4F). The magnitude of secondary expansion of GP33-specific CD8 T cells, detected in the blood at day 5 post secondary challenge, was not significantly different between lean and obese mice (Fig. 4G,H), suggesting that the capacity of memory CD8 T cells to respond to secondary challenge was not impaired by diet-induced obesity. Similar data were obtained after systemic bacterial infection as well (Supp. Fig. 2). Obesity did not impact the number, phenotype, or function of bacteria-specific memory CD8 T cells generated after infection with virulent L. monocytogenes expressing Ova epitope (Vir LM-Ova) (Supp. Fig. 2A-F). There fore, obesity does not change the ability of naïve CD8 T cell precursors to respond to systemic infections and develop into functional memory CD8 T cells.

Figure 4. Obesity does not influence the generation and function of endogenous virus-specific memory CD8 T cells.

A) Experimental design. Naïve B6 lean and obese mice were infected with LCMV (2×10⁵ PFU/mouse, i.p.). B) Weight of lean and obese mice 100 days after high-fat diet. Dots represent individual mice and the lines represent the mean and SD. C) Representative dot plots showing the percentage of tetramer+ (D^bGP33) endogenous GP33-specific memory CD8 T cells in the PBL at day 110 post infection. D) The frequency of endogenous GP33-specific memory CD8 T cells in the PBL at day 110 p.i. E) The percentage of endogenous GP33-specific memory CD8 T cells expressing CD27, or CD62L, or KLRG1 molecules. F) Experimental design. At day 112 after infection, LCMV immune lean and obese mice were challenged with Vir LM-GP33 (1×10⁵ CFU/mouse, i.v.). G) Representative dot plots showing the percentage of endogenous GP33-specific secondary effector CD8 T cells in the PBL at day 5 after challenge. H) The percentage of GP33-specific secondary effector CD8 T cells in the PBL of individual lean and obese mice at day 5 after challenge. Data are of 3 mice per group and representative of three similar and independent experiments. Weight graph is presented as scatter dot plot with mean±SD. Data presented as bar graphs are presented as mean±SEM.

Previous studies suggested, although never directly tested, that diet-induced obesity influences the development of memory CD8 T cells following localized influenza virus (IAV) infection(61–63). Following IAV PR8 infection (Supp. Fig 3A,B), the magnitude of proliferative expansion of endogenous D^bNP366-specific (Supp. Fig. 3C, D) and D^bPA224-specific (data not shown) CD8 T cells was similar in the PBL of individual lean and obese mice. In addition, evaluating the total numbers of memory NP366 and PA224-specific CD8 T cells in various tissues, including the lung and spleen, day 30 after infection (Supp. Fig. 3E,F) revealed that the generation of influenza specific memory CD8 T cells was not impaired in obese hosts. Furthermore, when IAV PR8-immune mice were challenged with the IAV strain X-31 (Supp. Fig. 3G), the magnitude of secondary expansion of NP366-specific CD8 T cells, detected in the blood at day 7 after challenge, was not different between lean and obese mice (Supp. Fig. 3H, I). The accumulation of secondary effector NP366-specific CD8 T cells in the spleen and lungs of lean and obese mice was also similar at day 10 after challenge (Supp. Fig. 3J). Thus, these data suggest that obesity does not influence the generation of memory CD8 T cells after localized influenza A virus infection and their ability to respond to a secondary challenge with cognate Ag delivered in the context of infection.

Continuous high-fat diet does not impair the generation of memory CD8 T cell responses in outbred mice

Recently, we identified a ‘surrogate activation marker approach’ to identify CD8 T cell responses without a priori knowledge of specific epitopes or MHC restriction elements that enabled longitudinal analysis of pathogen-specific CD8 T cell responses in outbred populations(48). We observed substantial variability in the kinetics and magnitude of CD8 T cell responses to various types of infections in genetically diverse outbred mice(48, 64). In order to extend our analysis and verify results obtained in inbred obese B6 mice, we performed similar experiments in which endogenous memory CD8 T cell responses to systemic LCMV infection were analyzed in cohorts of outbred Swiss Webster (SW) mice (Fig. 5). Groups of SW mice were either placed on a high-fat diet or remained on a standard diet for 120 days before LCMV infection (Fig. 5A). Within this time frame, a majority of SW mice receiving the special diet had gained weight, although a substantial variability in weight in both groups of mice was observed (Fig. 5B). Using this ‘surrogate activation marker approach’, endogenous pathogen-specific CD8 T cells were detected within the CD11a^hi population of CD8 T cells (Fig. 5C and (48)). As previously observed, the magnitude of the memory CD8 T cell response (day 60 post infection) in LCMV immune outbred mice receiving a standard diet was not uniform and varied widely, however, a similar pattern was observed in individual SW mice that were kept on a high-fat diet (Fig. 5D). Interestingly, we found no correlation between the magnitude of the CD8 T cell response and the weight of mice either receiving the standard or high-fat diet (data not shown). This suggests that whether an outbred mouse gains excessive weight receiving a standard or high-fat diet, the magnitude of the CD8 T cell response after infection is not impacted. Additionally, a similar composition of Ag-experienced memory CD8 T cells expressing CD62L and CD27 were found in both groups of mice (Fig. 5E). Thus, these data show that despite the numerical variability in LCMV-induced memory CD8 T cells in outbred hosts, continuous high-fat diet and excessive weight gain does not influence their generation and differentiation after systemic viral infection.

Figure 5. Continuous high-fat diet does not impair the generation of memory CD8 T cell responses in outbred mice.

A) Experimental design. Naïve outbred SW mice either receiving high-fat diet (HFD) or standard diet (SD) were infected with LCMV (2×10⁵ PFU/mouse, i.p.). B) Weight of mice 120 days after high-fat diet. Dots represent individual mice and the lines represent the mean and SD. C) Representative dot plot showing Ag-experienced (CD11a^hi CD8α^low) and Ag-in experienced (CD11a^low CD8α^hi) CD8 T cells 60 days after infection. Representative histograms showing expression of CD62L and KLRG1 molecules on CD11a^hi and CD11a^low CD8 T cells at day 60 after infection. D) The frequency of CD11a^hi CD8α^low memory CD8 T cells in the PBL at day 60 after infection. Dots represent individual mice and the line represents the mean. E) The percentage of LCMV-specific memory CD8 T cells in the PBL of individual mice expressing CD62L or KLRG1. Data are of 10–14 mice per group. Weight graph is presented as scatter dot plot with mean±SD.

Obesity associated inflammation does not prevent accelerated memory differentiation of Ag-specific CD8 T cells after immunization with peptide-coated dendritic cells

Recent studies, including our own, demonstrate that pro-inflammatory cytokines act directly on Ag-specific CD8 T cells to affect crucial aspects of memory CD8 T cell generation(19). ‘Signal 3’ inflammatory cytokines, such as IL-12 and type I IFNs, can influence the proliferative expansion, program of contraction, and rate of memory differentiation of responding CD8 T cells(22, 24, 28, 29, 31, 33, 65). Infection of mice with intracellular pathogens such as LCMV and L. monocytogenes stimulate robust CD8 T cells that initially exhibit an effector phenotype (i.e. GranzymeB^hi, CD127^low, KLRG-1^hi) and acquire memory phenotype and function relatively slowly after the infection/inflammation is cleared(19, 20). In contrast, immunization with peptide-coated mature dendritic cells (DC) in the absence of additional adjuvants evokes CD8 T cells that display memory characteristics (i.e. GranzymeB^low, CD127^hi, KLRG-1^low expression and ability to undergo robust proliferative expansion in response to secondary Ag-challenge) within days of initial priming(30, 32, 66). Thus, the inflammatory environment at the time of naïve CD8 T cell priming determines the rate of naïve-to-memory CD8 T cell differentiation.

Numerous studies suggest that the extent of metabolic dysfunction observed in obese hosts directly correlates with increased levels of pro-inflammatory cytokines in the periphery(35, 37, 41, 67, 68). In order to confirm that prolonged high-fat diet leads to increases in systemic inflammation, plasma from naïve (non-immunized) lean and diet-induced obese B6 mice was tested for concentrations of the indicated cytokines/chemokines via multiplex array (Fig. 6A). The results showed that concentrations of most of the analyzed analytes were increased in the plasma of obese hosts (Fig. 6A). However, in lean and diet-induced obese hosts the classical ‘signal 3’ cytokine IL-12 was below the limit of detection, while it was significantly increased in the plasma of the control lean mice after L. monocytogenes infection (Fig. 6A and data not shown). Thus, the absence of measurable differences in plasma concentrations of ‘signal 3’ cytokines in the lean and obese hosts suggest that priming of naïve CD8 T cells and their differentiation into the memory CD8 T cell pool would occur at similar rates despite the overall differences in the systemic inflammation.

Figure 6. Obesity associated inflammation does not prevent accelerated memory differentiation of Ag-specific CD8 T cells after immunization with peptide-coated dendritic cells.

A) Plasma, collected from naïve lean and obese B6 mice 130 days after high-fat diet, was tested via Multiplex analysis to determine concentrations (pg/mL) of the indicated cytokines and chemokines. Data are presented as mean of 11–12 mice per group. Dashed line represents the limit of detection (LOD). Statistical significance was determined when concentrations of analytes were above the LOD in both groups. B) Experimental design. Naïve lean and obese B6 mice were immunized with DCs coated with Ova peptide (5×10⁵ DCs/mouse, i.v.). An additional control group of lean mice received DC-Ova with coinjection of CpG (100 μg, i.p.). Ag-specific CD8 T cells were analyzed in the PBL at day 7 post immunization. Mice were boosted with Att LM-Ova (1×10⁷ CFU/mouse, i.v.) and the percentage of endogenous tetramer+ (K^bOva) endogenous Ova-specific CD8 T cells were analyzed 5 days later. C) Weight of mice 130 days after high-fat diet. Dots represent individual mice and the lines represent the mean and SD. D) The frequency of tetramer+ (K^bOva) endogenous Ova-specific CD8 T cells in the PBL of individual mice at day 7 after immunization. E) The percentage of Ova-specific CD8 T cells expressing KLRG1 and Granzyme B at day 7 after immunization with DC-Ova in the presence or absence of CpG coinjection. F) Representative dot plots showing the percentage of Ova-specific CD8 T cells in the PBL at day 5 after LM infection. G) The percentage of Ova-specific CD8 T cells in the PBL of individual lean and obese mice 5 days after booster infection. Data in B–G are of 4–6 mice per group and representative of four independent and similar experiments. Weight graph is presented as scatter dot plot with mean±SD. Data presented as bar graphs are presented as mean±SEM. Representative histogram is presented as mean±SEM.

To explore that obesity-associated systemic inflammation does not influence naïve-to-memory CD8 T cell differentiation, we used a peptide-coated DC immunization strategy that induces substantially less inflammation than systemic LCMV or L. monocytogenes infections(30). Lean and obese groups of mice were immunized with lipopolysaccharide (LPS)-matured DCs coated with Ova257 peptide (DC-Ova). An additional group of lean mice was introduced into the experiment to serve as a positive control in which inflammation, including high levels of IL-12, was induced by injection of the TLR9 agonist CpG(30, 32) at the time of DC-Ova immunization(Fig. 6B,C). This experimental model allowed us to determine whether obesity-associated inflammation, like adjuvant-induced inflammation (CpG), influenced the rate at which Ag-specific CD8 T cells acquire memory CD8 T cell characteristics(28, 30, 32). Consistent with our previous findings using B6 mice, the magnitude of proliferative expansion of Ova257-specific CD8 T cells responding to DC-immunization in the presence of heightened inflammation (Lean + CpG group) was significantly increased compared to cells activated in the absence of inflammation (Lean group)(32)(Fig. 6D). Whereas Ova257-specific CD8 T cells primed in an obese environment underwent reduced levels of proliferative expansion (even when compared to Lean group alone), these data suggest that obesity-associated inflammatory signals did not act on Ag-specific CD8 T cells to increase their accumulation in vivo (Fig. 6D). Similarly, Ova257-specific CD8 T cells in obese mice displayed an early-memory phenotype (GranzymeB^low, KLRG-1^low) (Fig. 6E) and underwent vigorous secondary expansion in response to early (day 7) booster challenge with Att LM-Ova (Fig. 6F,G), indicating that differences in the inflammatory environments of lean and obese hosts receiving peptide-DC immunization alone did not influence the rate of memory CD8 T cell differentiation. Similar data was obtained using a model of diet-induced obesity in BALB/c mice following immunization with mature DCs coated with P. berghei-derived circumsporozoite protein CS252 (Supp. Fig. 4A,B)(49). Obesity-associated inflammation in BALB/c mice did not promote increased accumulation of Ag-specific CD8 T cells (Supp. Fig. 4C,D) and did not inhibit the accelerated memory CD8 T cell differentiation (Supp. Fig. 4E). Collectively, these data suggest that unlike pathogen- or adjuvant-induced inflammation, obesity-associated inflammatory signals do not increase the primary CD8 T cell expansion and delay naïve-to-memory CD8 T cell progression. Alternatively, obesity associated inflammation (as shown in both B6 and BALB/c obese mice) might influence primary CD8 T cell accumulation after peptide-DC immunization, an interesting notion that will be explored in future studies.

Discussion

Here, we show that: I) obesity does not alter the maintenance and protective capacity of pre-existing memory CD8 T cells; II) naïve Ag-specific CD8 T cells originating in obese inbred or outbred hosts are equally capable of responding to systemic or localized infections, suggesting that obesity does not impair the intrinsic ability of CD8 T cells to respond to cognate Ag and develop into memory CD8 T cells; and that III) obesity-mediated alterations in the cytokine milieu does not influence naïve-to-memory CD8 T cell differentiation after peptide-DC immunization.

Although substantial resources are devoted in elucidating obesity-associated factors that influence the overall health status of the host, surprisingly only a small number of studies have been published to date that explored the role of obesity in the development and maintenance of pathogen-specific CD8 T cell responses. Work by Beck’s group suggests that after localized influenza infection, memory CD8 T cell function may be impaired in diet-induced obese hosts. In contrast to our findings, Karlsson et al. found that the capacity of influenza-specific CD8 T cells to undergo secondary expansion after diet-induced obese mice were first primed with the H3N2 (X-31) strain of influenza and then challenged with the H1N1 (PR8) strain was reduced (63). Potential differences in experimental models used in both studies that can contribute to different results observed are dose and strain of IAV used as primary challenge, the absence of detailed characterization of X-31 induced memory CD8 T cells (quantity, phenotype and function), and the high-fat diet used throughout the experiments.

Our longitudinal analysis of pre-existing GP33-specific memory CD8 T cells in LCMV-immune mice suggests that time-dependent changes in phenotype and function, including differential expression of phenotypic markers and capacity to exert effector function, are not impacted by the onset of obesity. We also found that naïve TCR-Tg OT-I CD8 T cells that originated in a lean Tg host developed into functional memory CD8 T cells when primed in a lean or obese host, suggesting that the obese environment does not impact the ability of Ag-specific CD8 T cells to respond to cognate Ag.

Results from the Dixit group suggest that diet-induced obesity accelerates thymic aging and decreases the diversity of the T cell repertoire (69, 70). Since it has been previously shown that age-associated changes in naïve T cell precursors compromise infection-induced immune responses due to a specific loss in CD8 T cells(71), it is possible that diet-induced mice have an altered CD8 T cell repertoire, which may increase their susceptibility to infection. Recent studies from our lab also show that sepsis-stimulated systemic inflammatory responses induce rapid apoptosis of naïve CD8 T cells that can ultimately lead to stochastic changes in the composition of the naïve CD8 T cell repertoire and/or render post-septic hosts unable to effectively mount primary CD8 T cell responses to some but not all pathogen-derived antigens (72). To test if obese hosts are more susceptible to infection due to the changes in naïve CD8 T cell precursors present in the obese environment, we analyzed endogenous primary CD8 T cell responses after systemic and localized viral or bacterial infections. Interestingly, regardless of the type or route of infection, pathogen-specific primary CD8 T cell responses were similar in lean and obese hosts. Specifically, the magnitude of proliferative expansion and effector CD8 T cell generation, acquisition of memory phenotype and function, and capacity to undergo secondary expansion upon additional Ag-encounter were similar when endogenous CD8 T cell responses were compared in lean and obese mice. In addition, analysis of endogenous pathogen-specific CD8 T cell responses in cohorts of outbred SW mice also revealed that patterns of memory differentiation were undisturbed in outbred mice continuously receiving a high-fat diet. In summary, these data suggest that diet-induced obesity did not change the overall capacity of the host’s naïve CD8 T cells to respond to new infections.

As previously reported, we also found that obese hosts have increased systemic levels of various cytokines and/or chemokines (34–36). Interestingly, diet-induced obesity triggered low-grade systemic inflammation did not include detectable alterations in ‘signal 3’ cytokines. It has been shown by us and others that infection-induced ‘signal 3’ cytokines such as IL-12 provide signals that shape all phases of primary Ag-specific CD8 T cell responses including their expansion and effector CD8 T cells accumulation, degree of contraction, and memory CD8 T cell differentiation (19, 28–33). However, using a model of peptide-DC vaccination (used here as model of antigen delivery that does not evoke an excessive inflammatory response that is detected after infection (30)) to evoke endogenous CD8 T cells we showed that DC-primed CD8 T cells acquired memory-like characteristics at accelerated rates in lean and diet-induced obese hosts, despite differences in systemic inflammation. These preliminary data suggest that although levels of inflammatory signals are increased in obese hosts, these inflammatory signals are unlike pathogen- or adjuvant-induced inflammation and do not influence naïve-to-memory CD8 T cell differentiation. Studies in the future will be designed to further explore the contribution, if any, of various obesity-induced inflammatory signals in T cell homeostasis.

In conclusion, we utilized a variety of experimental models to study the effects of obesity on Ag-specific CD8 T cell responses, and found that obese hosts do not have an impaired ability to generate and sustain protective memory CD8 T cell responses after systemic or localized infections. Increasing our understanding of the obesity-induced effects on CD8 T cell homeostasis has important clinical implications, which may help design future immune-based approaches needed to efficiently fight against infections.