Fasting reshapes tissue-specific niches to improve NK cell-mediated anti-tumor immunity

Abstract

Fasting is associated with improved outcomes in cancer. Here, we investigated the impact of fasting on natural killer (NK) cell anti-tumor immunity. Cyclic fasting improved immunity against solid and metastatic tumors in an NK cell-dependent manner. During fasting, NK cells underwent redistribution from peripheral tissues to the bone marrow (BM). In humans, fasting also reduced circulating NK cell numbers. NK cells in the spleen of fasted mice were metabolically rewired by elevated concentrations of fatty acids and glucocorticoids, augmenting fatty acid metabolism via increased expression of the enzyme CPT1A, and Cpt1a deletion impaired NK cell survival and function in this setting. In parallel, redistribution of NK cells to the BM during fasting required the trafficking mediators S1PR5 and CXCR4. These cells were primed by an increased pool of interleukin (IL)-12-expressing BM myeloid cells, which improved IFN-γ production. Our findings identify a link between dietary restriction and optimized innate immune responses, with the potential to enhance immunotherapy strategies.

Keywords: NK cells; anti-tumor responses; fasting; immunometabolism; innate immunity; metabolism.

Figure 1.. Cyclic fasting promotes NK cell-mediated anti-tumor immunity

(A) Mice were injected with tumor cells and 2 days later subjected to a 48hr water-only fast (red arrow), allowed to refeed and then fasted for 24hr twice per week (black arrows). Control mice were fed Ad Lib throughout. (B) Weight of surgically resected MC38 tumors at endpoint from mice subjected to Ad Lib or CFD as in (A). (C) Number of metastases enumerated (left) and lung weight (right) at endpoint from mice subjected to Ad Lib or CFD as in (A). (D-E) Mice were injected with MC38 cells s.c. (D) or B16F10 cells i.v. (E) as in (A) and continually treated with saline or anti-NK1.1 on days −3, 0, 3, 6, 10, 14 and 18. Tumor volume was measured over time (D) or lung weights recorded on day 21 after tumor injection (E). (F) Mean fluorescent intensity of H-2Kb ligand from B16F10 tumor cells from tumor-bearing mice subjected to Ad Lib or that had been fasted for 24hr after 3 weeks on CFD. (G-H) Total number of NK1.1⁺CD49b⁺ NK cells per gram of MC38 tumor tissue (G) or B16F10 tumor and lung tissue (H) at day 21 endpoint from mice subjected to Ad Lib or CFD as in (A). (I-J) Frequency of IFN-γ⁺ or CD107a⁺ NK cells from PMA and ionomycin stimulated splenocytes from MC38 tumor-bearing mice (I) or the spleen of B16F10 tumor-bearing mice (J) at day 21 endpoint subjected to Ad Lib or CFD as in (A). Data are pooled from 2 or more experiments (B-C) or representative of at least 3 independent experiments (D-J). Each symbol represents individual mice (B-C, E-J) or the mean of individual mice (D). ***P<0.001, **P<0.01, *P<0.05 with unpaired t-test (B, C, F-J) or two-way ANOVA (D, E). Error bars, mean ± s.e.m. (B-I).

Figure 2.. Nutrient restriction induces metabolic changes to maintain survival of NK cells

(A-B) Total numbers of NK cells (left) and CD8⁺ T cells (right) were enumerated per μl of blood pre (Ad Lib) and post 48hr fasting (A) or from the spleen of mice subjected to Ad Lib or 48hr fasting (B). (C) Total dead cells were enumerated from the spleen of mice subjected to Ad Lib or 48hr fasting. Graph shows enumerated NK cells and CD8⁺ T cells expressed as the fold-change in fasted mice over the average number found in Ad Lib mice. (D) Concentration of glucose and free fatty acids were measured from serum of the same mice at indicated timepoints after fasting. (E-F) Oxygen consumption rate (OCR) was measured in splenic NK cells (E) and CD8⁺ T cells (F) from mice subjected to Ad Lib or 48hr fasting in response to indicated metabolic inhibitors (left graphs). Data from performed replicate experiments was used to quantify basal respiration, maximum respiration and spare respiratory capacity in NK cells (E) and CD8⁺ T cells (F). (G-H) Mean fluorescent intensity (MFI) of Bodipy FLC-16 in NK cells (G) and CD8⁺ T cells (H) with representative histograms from the spleen of mice subjected to Ad Lib or 48hr fasting. (I-J) MFI of Bodipy 493/503 (I) or LC3-GFP (J) in splenic NK cells from mice subjected to Ad Lib or 48hr fasting. (K) Maximum respiration, ATP-linked respiration and spare respiratory capacity was calculated on splenic NK cells from mice subjected to Ad Lib or 48hr fasting in the presence of etomoxir. Calculations were obtained from oxygen consumption rate (OCR) measured in Figure S3K. Data are representative of at least 2 independent experiments (A-K). Each symbol represents individual mice (A-C, E-K) or the mean of individual mice (D-F). ****P<0.0001, ***P<0.001, **P<0.01, *P<0.05 with paired t-test (A), unpaired t-test (B-C, G-J) or two-way ANOVA (E-F, K). Error bars, mean ± s.e.m. (A-K).

Figure 3.. Cyclic fasting metabolically reprograms NK cells

A-B) Mean fluorescent intensity (MFI) of Bodipy FLC-16 (A) and Bodipy 493/503 (B) in splenic NK cells from naïve Ad Lib mice or mice that had been fasted for 24hr after 3 weeks on CFD. (C) UMAP of splenic NK cells from tumor-bearing mice that were subjected to Ad Lib or had been fasted for 24hr after 3 weeks on CFD. (D) Barcode plot generated from indicated Enrichr pathway obtained from gene set enrichment analysis of upregulated differentially expressed genes in splenic NK cells from tumor-bearing mice that were fasted for 24hr after 3 weeks on CFD (E) Violin plots showing indicated relative gene expression using MAGIC imputed data from NK cells indicated in (C) (F) Volcano plot depicting differentially expressed genes in CFD NK cells indicated in (C). Blue dots represent genes highly expressed in CFD NK cells. (G) Violin plots showing indicated relative gene expression using MAGIC imputed data from NK cells indicated in (C) (H) Concentration of glucocorticoids were measured from the serum (left) and spleen (right) of mice at indicated timepoints after fasting. (I) MFI of CPT1A in fresh isolated splenic NK cells from mice that had been subjected to vehicle or 100 nM dexamethasone for 24hr. Representative histogram shown on right. Data are representative of at least 2 independent experiments (A-B), 1 experiment with at least 6 biological replicates (C-G), or 1 experiment with at least 3 biological replicates (H-I). Each symbol represents individual mice (A-B, H, I left panel), individual human donors (I right panel) or individual cells (C, F). ****P<0.0001, ***P<0.001, **P<0.01, *P<0.05 with unpaired t-test (A-B), ordinary one-way ANOVA (H) or paired t-test (I). Error bars, mean ± s.e.m. (A-B, H-I).

Figure 4.. CFD-induced FA oxidation improves the NK cell anti-tumor response

(A-B) Mean fluorescent intensity of CPT1A (A) and Bodipy FLC-16 (B) in tumor NK cells from B16F10-bearing Ad Lib mice or mice that had been fasted for 24hr after 3 weeks on CFD. (C) Diagram highlights internal and external sources of FAs, and CPT1A as a critical enzyme that shuttles long chain fatty acids from the cytoplasm into the mitochondria for fatty acid oxidation. Created with BioRender.com. (D-E) Number of metastases enumerated at endpoint (D) and total number of CD49b⁺ NK cells per gram of B16F10 tumor and lung tissue (E) at day 21 endpoint from NK-Cpt1a^+/+ and NK-Cpt1a^−/− mice injected with B16F10 cells i.v., subjected to CFD and refed for 24hr as in 1A. (F-G) MFI of BCL2 (F) and BIM (G) in tumor NK cells from NK-Cpt1a^+/+ and NK-Cpt1a^−/− mice injected with B16F10 cells i.v., subjected to CFD and refed for 24hr as in 1A. (H) BIM:BCL2 ratio was calculated by dividing the genotype-specific MFI of BIM in (G) by the matched MFI of BCL2 in (F) (I) MFI of Bodipy FLC-16 in tumor NK cells from NK-Cpt1a^+/+ and NK-Cpt1a^−/− mice injected with B16F10 cells i.v., subjected to CFD and refed for 24hr as in 1A. (J) Frequency of IFN-γ⁺ (left) or CD107a⁺ (right) NK cells from PMA and Ionomycin restimulated spleen of NK-Cpt1a^+/+ and NK-Cpt1a^−/− mice injected with B16F10 cells i.v., subjected to CFD and refed for 24hr as in 1A. Data are representative of at least 2 independent experiments (A-B, D-J), Each symbol represents individual mice (A-B, D-J). **P<0.01, *P<0.05 with unpaired t-test (A-B, D-J). Error bars, mean ± s.e.m. (A-B, D-J). FA: fatty acid; LD: lipid droplet; TG: triglyceride.

Figure 5.. Fasting promotes NK cell redistribution to the BM

(A) Naïve mice were subjected to 48hr fasting or fed Ad Lib and NK cells enumerated from the indicated organs. (B) Naïve mice were subjected to 24hr or 48hr fasting, 48hr fasting with 24hr refeeding or fed Ad Lib and NK cells enumerated from the indicated organs. (C) Mature splenic CD45.1⁺ NK cells were sorted and transferred into CD45.2⁺ host mice. NK cells were enumerated from the indicated organs after fasting for 48hr or Ad Lib conditions. Representative bone marrow flow plots with percentage of donor cells as a fraction of total NK cells (left). (D) iNkp46^tdTomato mice were gavaged with tamoxifen to permanently express TdTomato (Tom) in mature NK cells. 24hr later, Tom⁺ NK cells were enumerated from the indicated organs after fasting for 48hr or Ad Lib conditions. Representative bone marrow flow plots with percentage of Tom⁺ NK cells as a fraction of total NK cells (left). (E-F) Mean fluorescent intensity of Bodipy FLC-16 (E) and CPT1A (F) in bone marrow NK cells from mice subjected to Ad Lib or 48hr fasting. (G) Concentration of glucocorticoids were measured from indicated organs from Ad Lib mice or mice fasted for 48hr. Graphs show enumerated NK cells expressed as the fold-change in fasted mice over the average number found in Ad Lib mice (A-D). Data are representative of at least 2 independent experiments (A-F) or 1 experiment with 3 biological replicates (G). Each symbol represents individual mice (A, C-G) or the mean of individual mice (B). ****P<0.0001, ***P<0.001, **P<0.01, *P<0.05 with ordinary one-way ANOVA (A-D), unpaired t-test (E-F) or two-way ANOVA (G). Error bars, mean ± s.e.m. (A-G).

Figure 6.. Redistribution of NK cells to the BM during CFD promotes tumor control

(A) Mean fluorescent intensity (MFI) of S1PR5-GFP (right) with representative histogram (left) in bone marrow NK cells from mice subjected to Ad Lib, 24hr or 48hr fasting. (B-C) Total NK cell numbers (B) and NK cell maturation subsets (C) were enumerated from mixed S1pr5^+/+ and S1pr5^−/− bone marrow chimera mice subjected to 48hr fasting or Ad Lib and fold-change calculated within each genotype. (D) Naïve mice were subjected to 24hr or 48hr fasting or fed Ad Lib then injected with a fluorescent CD45 antibody i.v. to label sinusoid (CD45⁺) and parenchymal (CD45⁻) NK cells in the bone marrow. (E) Total NK cell numbers were enumerated from naïve mice treated with vehicle or AMD3100 subjected to 48hr fasting and compared to mice fed Ad Lib treated with vehicle. (F-G) S1pr5^−/− mice (F) or mice treated with AMD3100 (G) were injected with MC38 cells s.c. and subjected to CFD or Ad Lib as in Figure 1A. Tumor volume was measured over time. (H-I) Total number of NK cells per gram of MC38 tumor tissue in S1pr5^−/− mice (H) or mice treated with AMD3100 (I) at day 21 endpoint from mice subjected to Ad Lib or CFD as in Figure 1A. (J-K) Frequency of IFN-γ⁺ NK cells from PMA and Ionomycin stimulated tumor infiltrating lymphocytes and splenocytes of MC38 tumor-bearing S1pr5^−/− mice (J) or mice treated with AMD3100 (K) at day 21 endpoint from mice subjected to Ad Lib or CFD as in Figure 1A. Graphs show enumerated NK cells expressed as the fold-change in fasted mice over the average number found in Ad Lib mice (B-C, E). Data are representative of at least 2 independent experiments (A-K). Each symbol represents individual mice (A-E, H-K) or the mean value of individual mice (F-G). **P<0.01, *P<0.05 with ordinary one-way ANOVA (A, D), two-way ANOVA (C, J-K) or unpaired t-test (B, E-I). Error bars, mean ± s.e.m. (A-K).

Figure 7.. The fasted BM niche promotes NK cell priming via IL-12 production

(A) UMAP projections generated by scRNA-seq of bone marrow (BM) NK cells from tumor-bearing mice that were subjected to Ad Lib or had been fasted for 24hr after 3 weeks on CFD. (B) Violin plots showing indicated relative gene expression using MAGIC imputed data from NK cells indicated in (A) (C) Protein levels of IL-12 in homogenized BM, spleen and serum from healthy mice subjected to Ad Lib or 48hr fasting presented as fold-change in fasted protein levels over the average protein levels found in Ad Lib mice (D) CD45⁺IL-12-GFP⁺ cells from the BM and spleen of mice that had been fasted for 24hr after 3 weeks on CFD presented as fold change from mice subjected to Ad Lib conditions. (E-F)) Frequency of IFN-γ⁺ NK cells stimulated with IL-12 and IL-18 from BM (E) and spleen (F) of mice subjected to Ad Lib or had been fasted for 24hr (E) or fasted and refed for 24hr (F) after 3 weeks on CFD Data are representative of at least 1 independent experiment with 6 biological replicates (A-B) or 3 independent experiments (C-F). Each symbol represents individual mice (C-F) or individual cells (A). ***P<0.001, **P<0.01, *P<0.05 with ordinary one-way ANOVA (C) or unpaired t-test (D-F). Error bars, mean ± s.e.m. (C-F).