Bone marrow lympho-myeloid malfunction in obesity requires precursor cell-autonomous TLR4

Abstract

Obesity, a prevalent condition in adults and children, impairs bone marrow (BM) function. However, the underlying mechanisms are unclear. Here, we show that obese mice exhibit poor emergency immune responses in a toll-like receptor 4 (TLR4)-dependent manner. Canonical myeloid genes (Csf1r, Spi1, Runx1) are enhanced, and lymphoid genes (Flt3, Tcf3, Ebf1) are reduced. Using adoptive transfer and mixed BM chimera approaches we demonstrate that myeloid>lymphoid bias arises after 6 weeks of high-fat diet and depends on precursor cell-autonomous TLR4. Further, lean mice exposed to the TLR4 ligand lipopolysaccharide (LPS) at doses similar to that detectable in obese serum recapitulates BM lympho-myeloid alterations. Together, these results establish a mechanistic contribution of BM cell-intrinsic TLR4 to obesity-driven BM malfunction and demonstrate the importance of LPS. Our findings raises important questions about the impact of maternal obesity and endotoxemia to fetal hematopoiesis, as fetal immune precursors are also sensitive to TLR4 signals.

Fig. 1

Obesity compromises hematopoietic potential. a Number and percentage of multipotent BM-LSK (lineage⁻Sca-1⁺c-kit⁺) and HSC (CD150⁺CD48⁻ LSK) in mice fed 40 kcal% high-fat diet (HFD) or nutrient-matched low fat control diet (NCD) for 16–18 weeks. In spleen, HSCs were identified using the CD150⁺flk2⁻ LSK definition as rare CD150⁺CD48⁻ LSKs were below the limit of detection by flow cytometry. Data represent the mean ± SEM of animals pooled from at least three independent experiments of NCD/HFD cohorts assayed in parallel for a total of n = 7 mice/group. b For analysis of BrdU incorporation, mice were injected twice daily with BrdU 48 h prior to killing. Following surface staining, cells were subsequently permeabilized for intracellular BrdU staining. Data are pooled from three independent experiments in which NCD/HFD mice were examined side-by-side with the total number of mice inset in the graph. c Mice fed the indicated high-fat diet for 16 weeks were exposed to the chemotherapy drug 5-fluorouracil (5-FU) and peripheral blood leukocyte rebound examined. Each diet (40 kcal%, diamond; and 60 kcal%,  square) was examined separately and the data pooled from a total of two independent experiments of 3–4 mice/group). d BM from B6 mice (CD45.2) fed either 40 kcal% high-fat diet or control diet for 16–18 weeks was separately mixed with competitor BM (CD45.1) and co-transferred to CD45.1/2 recipients. Following 16 weeks engraftment, a point at which hematopoiesis derives exclusively from donor HSCs, competitively engrafted chimeras were subject to two sequential rounds of 5-FU (day 0, 28) and peripheral blood leukocyte rebound of the CD45.2 donor partner examined. Data are pooled from two independent experiments (total n = 3–4 recipients/group). e Following 5-FU challenge as in c, BM subset BrdU incorporation was examined at the indicated time points. Data are pooled from 2–3 independent experiments of NCD/HFD mice assayed in parallel, and each symbol is an individual animal. Data are analyzed by Student’s t-test. Error bars represent s.e.m. **p < 0.05; ns, not significant

Fig. 2

BM LKS^neg and CLP functional defects emerge early in obesity. a B6 mice were fed the indicated high-fat or nutrient-matched control diet for 6 or 20 weeks, after which defined BM precursor subsets were sorted for analysis of differentiation potential under stroma-free, serum-free conditions. lineage⁻c-kit⁺Sca-1⁻ (LKS^neg), common lymphoid progenitor (CLP) and LSK subsets were assays in liquid cultures under myeloid- or lymphoid-supportive conditions for 8 or 12 days, respectively, after which cells were stained with antibodies to CD19, B220, Gr-1, or CD11b. Yield/cells input is shown. Bar graphs depict average ± SEM of data from two independent experiments of NCD/HFD mice assayed in parallel, with the total number of mice inset. b Sorted HSC (CD150⁺flk2⁻ LSK), MPP (flk2⁺ LSK), each at 100 cells per well, or 2 × 10⁴ BM cells (WBM) from NCD or HFD mice were cultured in Methocult and colonies enumerated on day 10. For scatter plots, data are pooled from three independent experiments of paired NCD/HFD mice. Each symbol is an individual animal. Bar graphs depict average ± SEM (duplicate experiments, total n = 4/group). a–b Data are analyzed by Student’s t-test with Bonferroni correction for multiple comparisons. Error bars represent s.e.m. **p < 0.05; ns, not significant. c NanoString expression analysis of canonical genes associated with lymphoid or myeloid potential in BM LKS^neg and CLP progenitors sorted from mice fed high-fat or control diet for 20 weeks. Each lane is a different mouse; one lane was lost due to an instrument technical problem

Fig. 3

Myeloid>lymphoid BM bias requires TLR4. a Mice fed the 60 kcal% (square symbols) high-fat diet or matched control diet for 16 weeks were exposed to the chemotherapy drug 5-FU and peripheral blood leukocyte rebound tracked at the indicated time points. Data are pooled from two independent experiments of NCD/HFD mice analyzed side-by-side. b WT or Tlr4^−/− mice were fed 40 kcal% (diamond symbols) or 60 kcal% (square symbols) high-fat or matched control diet for 16–18 weeks after which sorted LKS^neg and CLP outgrowth was assessed in liquid assays as in Fig. 2. Data are pooled from four independent experiments of NCD/HFD mice examined in parallel. c WT (CD45.1) or Tlr4^−/− (CD45.2) BM was mixed and co-transferred to CD45.1/2 recipients, after which mice were rested for 1 week and then fed 60 kcal% (square symbols) high-fat or control diet for 16–20 weeks. (B220 + CD3)/Gr-1 is shown as lymphoid/myeloid ratio. Data are pooled from two independent experiments performed with different BM donors and 2–3 recipients/group. Each symbol is an individual animal. Data are analyzed by Student’s t-test. Error bars represent s.e.m. **p < 0.05; ns, not significant

Fig. 4

Chronic low-dose LPS administered at levels found in obese mice induces myeloid bias. a Serum LPS levels in B6 mice fed the indicated high-fat or control diet for 16–18 weeks. Data are pooled from 3 independent experiments in which NCD/HFD mice were analyzed side-by-side (total n = 8–9 mice/group). b, c WT or Tlr4^−/− mice were exposed to chronic low-dose LPS or PBS vehicle for 6 weeks after which serum LPS levels were assessed and the outgrowth potential of sorted LKS^neg and CLP subsets examined. Data are pooled from two to three independent experiments in which PBS and LPS mice were examined in parallel. Bar graphs depict average ± SEM, with number of mice inset. For scatter plots, each symbol is an individual animal. Data are analyzed by Student’s t-test. Error bars represent s.e.m. **p < 0.05

Fig. 5

Requirement for BM cell-autonomous TLR4 in HFD-induced myeloid bias. a Left, cytokine levels in BM were assessed via multiplex ELISA following 6 weeks of 60 kcal% high-fat or control diet. NCD/HFD cohorts were examined in parallel and each symbol is a different animal. The thresholds of detection are: IL-6, 0.4 pg/ml; IL-1b, 5.4 pg/ml; TNF, 0.5 pg/ml; G-CSF, 0.5 pg/ml; MIP1b, 11.9 pg/ml. Right, Sca-1 mean fluorescence intensity (MFI) on BM-LSK. Data are analyzed by Student’s t-test. Error bars represent s.e.m. b WT (CD45.1) or Tlr4^−/− (CD45.2) BM was mixed and co-transferred to CD45.1/2 recipients, after which mice were rested for one week and then fed 60 kcal% high-fat or control diet for 6 weeks. At sacrifice, LKS^neg and CLP subsets from each donor were sorted for analysis of outgrowth potential as in Fig. 2. Data are pooled from two independent experiments performed with different BM donors and 2–3 recipients/experiment. Bar graphs depict average ± SEM, with number of mice inset. Data are analyzed by Student’s t-test with Bonferroni correction for multiple comparisons. **p < 0.05; ns, not significant