Normalization of obesity-associated insulin resistance through immunotherapy

Abstract

Obesity and its associated metabolic syndromes represent a growing global challenge, yet mechanistic understanding of this pathology and current therapeutics are unsatisfactory. We discovered that CD4(+) T lymphocytes, resident in visceral adipose tissue (VAT), control insulin resistance in mice with diet-induced obesity (DIO). Analyses of human tissue suggest that a similar process may also occur in humans. DIO VAT-associated T cells show severely biased T cell receptor V(alpha) repertoires, suggesting antigen-specific expansion. CD4(+) T lymphocyte control of glucose homeostasis is compromised in DIO progression, when VAT accumulates pathogenic interferon-gamma (IFN-gamma)-secreting T helper type 1 (T(H)1) cells, overwhelming static numbers of T(H)2 (CD4(+)GATA-binding protein-3 (GATA-3)(+)) and regulatory forkhead box P3 (Foxp3)(+) T cells. CD4(+) (but not CD8(+)) T cell transfer into lymphocyte-free Rag1-null DIO mice reversed weight gain and insulin resistance, predominantly through T(H)2 cells. In obese WT and ob/ob (leptin-deficient) mice, brief treatment with CD3-specific antibody or its F(ab')(2) fragment, reduces the predominance of T(H)1 cells over Foxp3(+) cells, reversing insulin resistance for months, despite continuation of a high-fat diet. Our data suggest that the progression of obesity-associated metabolic abnormalities is under the pathophysiological control of CD4(+) T cells. The eventual failure of this control, with expanding adiposity and pathogenic VAT T cells, can successfully be reversed by immunotherapy.

Figure 1. Phenotype of T^fat (all mice 14–16 wk old, each contour plot represents 3–4 independent experiments)

a) CD4 and CD8 expression on CD3+ cells in spleen and SAT and VAT of B6 high fat diet (HFD) and normal chow diet (NCD) mice. b) Proportion of CD4+IFNγ+ (Th1) and CD4+IL-17+ (Th17) T-cells in spleen, SAT and VAT. c) Proportion of CD4+Foxp3+ (Treg) cells in spleen, SAT and VAT (upper panel: FACS plots; lower panel: pooled data; *p<0.03, t-test). d) Total number of CD3+CD4+, CD4+IFNγ+ (Th1), CD4+IL-17+ (Th17), CD4+Foxp3+ (Treg) T cells/g of SAT and VAT in normal chow diet and HFD mice (n=4–5/group, *p<0.05, t-test). e) T-bet (Th1):Foxp3 ratio in human VAT (upper left panel, r²=0.62, p<0.05). Lower left panel: VAT T-bet:Foxp3 ratio in patients with BMI>30 vs. <25 (*p<0.02, t-test). Upper right panel: Foxp3 (brown) and T-bet (blue, Th1 cells) in human VAT (BMI>30); lower right panel: BMI<25, bar 50 μm; T-bet:Foxp3 ratios from >200 stained cells/2 tissue levels. f) CD3+CD4+ T cells (%) with secondary re-rearranged. non-OVA TCRα detected by reduced OT2 TCRα:CD3 ratio in spleens, SAT and VAT of 6 or 16 wk old, regular diet or HFD B6 OT2 mice. g) TCRVα gene usage of CD4+ T cells isolated from spleens of 16 wk old obese wild type (WT) or OT2 mice, and from VAT of HFD and NCD WT and OT2 mice. CD4+ OT2 T cells were identified and sorted as in (f).

Figure 2. Impact of lymphocyte deficiency on weight gain, fat distribution, glucose tolerance and insulin resistance

a) Body weights of >6 wk old WT and RAG^null B6 mice on NCD (n=20/group) or HFD (n=20/group, *p<0.03, Mann-Whitney, WT vs. RAG^null). b) Weights of single epididymal VAT and inguinal SAT pads from 14–15 wk WT and RAG^null mice on NCD or HFD (left panel: n=6 mice/group, *p<0.03, t-test), right panel: VAT:SAT ratio, *p<0.001, t-test. c) Relative fat cell diameter (see methods, n=3 mice/group, *p<0.0001, Mann-Whitney). d) Food intake (top panel) and respiratory exchange ratio (RER, bottom panel) in HFD WT or RAG^null mice (n=4/group). e) Glucose tolerance of RAG^null or WT mice on NCD or HFD (n=10/group, *p<0.02, 2-way ANOVA). f) Fasting glucose (left panel) or insulin blood levels (right panel) in 14 wk old HFD and NCD WT or RAG^null mice (n=10 mice/group, *p<0.05, t-test). g) Insulin tolerance test (ITT) in 14 wk old WT RAG^null mice on NCD or HFD (n=6–10/group, *p<0.05, two-way ANOVA).

Figure 3. CD4+ T cell grafts reverse obesity-associated metabolic abnormalities in RAG^null mice

a) FACS plots analyzing purity of CD3+CD8+ (top) and CD3+CD4+ (bottom) T cells in spleen, SAT, and VAT, 2 wk following transfer into 12 wk old HFD RAG^null recipients. All subsequent experiments were performed 2–4 wk post-transfer and compared to non-transferred, age-matched, HFD RAG^null mice. b) Left panel: body weights (n=5/group) of recipient mice were monitored post-transfer (*p<0.05, t-test, data from one of four similar experiments). Right panel: weights of individual VAT and SAT pads (n=5/group). CD4+ T cell transfer reduces VAT mass (*p<0.05, t-test). c) Adipocyte diameter following transfer of CD4+ or CD8+ T cells into HFD RAG^null mice (n=3 mice/group, *p<0.01, Mann-Whitney). d) CD4+ T cell transfer improves glucose tolerance in HFD RAG^null recipients (n=10/group, *p<0.0001, 2-way ANOVA). e) Fasting glucose (left panel) and insulin (right panel) in CD4+ or CD8+ transferred HFD RAG^null mice (n=5–10 mice/group, *p<0.03, **P<0.01, t-test). f) HFD RAG^null mice reconstituted with CD4+ T-cells show improved insulin tolerance (n=10, *p<0.05, 2-way ANOVA). g) FACS plots of OT2 TCRVα re-rearrangements in spleen and VAT, 2 wk following transfer. The purity of CD4⁺OT2^hi cells transferred was 99.5%. h) CD4+OT2:TCRα^hi (CD4⁺OT2^hi) T cells fail to improve glucose and (i) insulin tolerance following transfer (n=5/group, *p<0.0001, 2-way ANOVA). j) Fasting glucose (left panel) and fasting insulin (middle panel) in CD4+ or CD4+OT2:TCRα^hi (OT2^hi) transferred HFD RAG^null mice (n=5 mice/group, *p<0.05, t-test). Right panel, CD4+OT2^hi:TCRα^hi T cells fail to improve weight post transfer (n=5 mice/group, *p<0.004, t-test).

Figure 4. CD4+Foxp3− T cells reverse metabolic abnormalities following transfer

a) FACS plots of CD4+Foxp3+ T cells (%) in spleen and VAT, 3 wk after transfer of purified CD4+ or purified CD4+Foxp3− (EGFP^neg, 99.1% pure) cells, one of two similar plots shown. b) Change of HFD-RAG^null body weight 3 wk post transfer of CD4+, CD4+Foxp3−, and CD4+IL-10^null T cells (n=5/group, *p<0.01, t-test). c) GTT (n=5/group, *p<0.0001, 2-way ANOVA). d) fasting glucose (left panel, n=5 mice/group, *p<0.01, t-test) and fasting insulin (right panel, n=5 mice/group, *p<0.01, t-test), 3 wk post transfer of CD4+, CD4+Foxp3−, or CD4+IL-10^null T cells into HFD-RAG^null recipients. e) IL-4 (left panel) and IL-13 (right panel) produced by anti-CD3 plus anti-CD28 stimulated VAT T cells of 16 wk old HFD WT, and HFD RAG^null recipients 3 wk post transfer of CD4+ T cells (n=3/group). f) FACS plots for CD4+ gated T cells from VAT of HFD RAG^null mice that received either CD4+STAT6^null (left panel) or CD4+WT (right panel) T cells 3 wk previously. g) Glucose tolerance 3 wk following transfer of CD4+ or CD4+STAT6^null T cells (n=5/group, *p<0.05, 2-way ANOVA). h) Fasting glucose (left panel), fasting insulin (middle panel), and weight change (right panel) in HFD RAG^null recipients 3 wk post transfer of purified CD4+ or CD4+STAT6^null T cells (n=5 mice/group, *p<0.01, t-test). i, Left panel: representative FACS plot CD4+GATA-3+ (%Th2) cells in spleen and VAT of 14–16 wk old regular diet and HFD B6 mice. j, Left panel: pooled data from (i, n=4/group, *p<0.03, t-test), right panel: total number of CD4+GATA-3+ (Th2) T cells/g of VAT in regular diet and HFD B6 mice (n=4/group, p>0.4, t-test).

Figure 5. Anti-CD3 and its F(ab′)₂ fragment improve obesity-induced insulin resistance

16 wk old obese (HFD) B6 mice received anti-CD3 antibody (αCD3, a–e), Isotype-matched control IgG or anti-CD3-(F(ab′)₂ (f–j), maintaining HFD for 6 or 9 wk. a. FACS plots (%) of CD4+Foxp3+ Treg cells in spleen and VAT 9 weeks post αCD3 (one of 3 similar experiments shown). b) Fasting glucose (top panel) and insulin (bottom panel) after αCD3 (n=8, *p<0.05, t-test). c) Glucose tolerance (n=8, *p<0.0004, 2-way ANOVA) and d) ITT (n=8, *p<0.0005, 2-way ANOVA) post αCD3. e) Left panel: body weights of HFD mice following αCD3 (n=8, *p<0.05, t-test). Right panel: adipocyte diameter post αCD3 (n=3 mice/group). f) Glucose tolerance (left panel) 6 wk post F(ab′)₂ (n=5, *p<0.01, 2-way ANOVA, one of 2 similar experiments). Fasting insulin (middle panel) and glucose (right panel) in HFD mice 6 wk following F(ab′)₂ (n=5, *p<0.04, t-test). g) Body weights of HFD mice following F(ab′)₂. h) CD4+Foxp3+ Treg cells in spleen and VAT of HFD B6 mice 6 wk post F(ab′)₂. (one of 3 similar experiments). i) IFNγ, IL-17, and IL-13, levels following stimulation of B6 VAT T cells with αCD3 plus anti-CD28, 8 wk post F(ab′)₂ (n=3/group). j) FACS plot of CD4+ gated, GATA-3 stained VAT T cells, 8 wk post F(ab′)₂).

Figure 6. F(ab′)₂ therapy alters VAT resident macrophage phenotype

a) MMR^hi (%, upper gate), MMR^lo (middle gate), and MMR⁻ (lower gate) macrophages from VAT of 16 wk old HFD or lean NCD B6 mice (representative data from 4 similar experiments). b) IL-10 (left panel), MCP-1 (middle panel) and TNFα (right panel) produced by LPS stimulated, F4/80+ macrophages sorted from VAT of HFD (top panel) or NCD (lower panel) mice into MMR⁻, MMR^lo, and MMR^hi cell populations (one of 4 similar experiments). We consistently failed to obtain sufficient cell numbers for analysis of MMR⁻ macrophages from VAT from lean mice. c) FACS plots from 2 independent experiments, measuring MMR expression in F4/80+ macrophages from HFD B6 VAT, 6 wk post F(ab′)₂. d) IL-10 (left panel), MCP-1 (middle panel) and TNFα (right panel) produced by LPS stimulated F4/80+ macrophages sorted from VAT 6 wk post F(ab′)₂ (n=3/group, *p<0.05, t-test).