Diet-induced obesity promotes myelopoiesis in hematopoietic stem cells

Abstract

Obesity is associated with an activated macrophage phenotype in multiple tissues that contributes to tissue inflammation and metabolic disease. To evaluate the mechanisms by which obesity potentiates myeloid activation, we evaluated the hypothesis that obesity activates myeloid cell production from bone marrow progenitors to potentiate inflammatory responses in metabolic tissues. High fat diet-induced obesity generated both quantitative increases in myeloid progenitors as well as a potentiation of inflammation in macrophages derived from these progenitors. In vivo, hematopoietic stem cells from obese mice demonstrated the sustained capacity to preferentially generate inflammatory CD11c(+) adipose tissue macrophages after serial bone marrow transplantation. We identified that hematopoietic MyD88 was important for the accumulation of CD11c(+) adipose tissue macrophage accumulation by regulating the generation of myeloid progenitors from HSCs. These findings demonstrate that obesity and metabolic signals potentiate leukocyte production and that dietary priming of hematopoietic progenitors contributes to adipose tissue inflammation.

Keywords: Hematopoietic stem cells; Myelopoiesis; Obesity.

Supplementary Figure 1

(A) Gaiting strategy for ATMs from singlets to CD45⁺ to CD64⁺ total ATMs, CD64⁺CD11c⁺ ATMs (B) Gating strategy for HSCs using singlets, lineage negative cells and then LK and LSK populations for sub-gating of GMPs, Pre-GMs, MPP, and LT-HSC cells.

Supplementary Figure 2

Characteristics of ND and HFD fed mice and weight loss model (A) C57Bl/6J mice fed ND or HFD for 16 weeks. Flow quantitation of Ly6c^hi circulating monocytes expressed as a percentage of total CD115⁺ monocytes. (B) Quantitation of CD11c⁺ ATMs from epididymal adipose tissue expressed as a percent of total ATMs (n = 8 per group). (C) Body composition analysis on a representative group of 4 animals each diet. (D) Glucose tolerance testing demonstrating glucose intolerance in HFD fed mice. (E) Mice were fed a HFD for 12 weeks and then switched to a ND for 8 weeks (HFD off). (E) Body weight and (F) glucose tolerance is shown relative to age matched ND fed controls. (n = 8 ND and 12 HFD off). Two-sided Student's t-test with *p < 0.05.

Supplementary Figure 3

Tissue weight and RT-PCR analysis of serial BMT mice. (A) Tissue weights of mice undergoing serial BMT assessed after the 2nd transplant. (B) Gene expression from epididymal fat pads from HFD fed mice after the 2nd BMT.

Figure 1

Bone marrow derived cells from obese mice are polarized towards an activated state. BM cells from ND or HFD fed mice were differentiated in GM-CSF (n = 4 wells). After (A) vehicle or (B) LPS stimulation for 6 h, M1 gene expression analyzed by RT-PCR. (C and D) Representative flow cytometry assessment for MHCII CD11c in BMDC derived from ND and HFD fed mice (n = 3 per group). (E) Results of pathway analysis from microarray of genes overexpressed in BMDC derived from obese mice compared to lean. Table of most-significant pathways shown. Two-sided Student's t-test. *p < 0.05, **p < 0.01, ****p < 0.001.

Figure 2

Obese BM derived macrophages are primed for inflammatory trafficking. (A) BM from ND (CD45.1) and HFD (CD45.2) mice were mixed in a 1:1 ratio prior to injection into CD45.1 wild-type mice. (B–D) IP injection of thioglycollate (Thio) was then used to induce PM recruitment and donor origin of the PMs (F4/80⁺CD11b⁺) assessed by flow cytometry using CD45 specific antibodies. Data reported as ratio of HFD derived (CD45.2) to ND derived (CD45.1) cells. (E) Cytokine production of FACS purified PMs from the different donor sources. This is representative data from two different experiments with similar results. Two-sided Student's t-test. *p < 0.05, N = 4.

Figure 3

Long-term HSC and myeloid progenitors are increased with HFD. Flow cytometry quantitation of BM resident (A) lineage neg. (Lin⁻) cells, (B,C) LSK cells, (D) GMP and Pre-GM, and (E) LT-HSC and MPP populations in lean and DIO mice. (F) Granulocyte and macrophage colonies from BM of ND and HFD fed mice. (G) Analysis of splenic progenitors in lean and DIO mice. (H) Quantitation of BM progenitors in animals fed HFD for 12 weeks and then taken off diet and age matched ND fed controls. (I) Assessment of BM progenitors after short term HFD feeding. Two-sided Student's t-test for two group comparisons and one-way ANOVA with Tukey's multiple comparisons where three groups are present. *p < 0.05, **p < 0.01, ***p < 0.005. n = 4–8 per group.

Figure 4

BM from HFD mice produce more Lyc^hi monocytes and CD11c⁺ ATMs than ND mice after BMT. BM chimeras generated from ND or HFD donors. Groups are noted by the donor source followed by the dietary challenge (N = normal diet, H = high fat diet for 16 weeks). (A) Flow cytometry quantitation of Ly-6c^hi classical monocytes as a percentage of all CD115⁺ monocytes. N = 4. (B) Colony forming unit assays enumerating CFU-G/M/GM (N = 4 independent mice). (C) Flow cytometry quantitation of CD11c⁺ ATMs in gonadal white adipose tissue expressed as percentage of total ATMs and ratio of CD11c⁺/CD11c⁻ ATMs. (D) Quantitation of T cells in adipose tissue expressed as percentage of total CD3⁺ lymphocytes (ATT). (E) Fasting insulin and (F) glucose tolerance tests (N = 7 per group). (G–H) BM chimeras generated with equal numbers of FACS purified LSK cells from ND or 16-week HFD donor animals. (G) Flow analysis of blood monocytes 4 weeks after BMT. (H) Colony forming unit assays from LSK donor chimeras 16 weeks after BMT (N = 6 plates per group). When two groups are analyzed Student's t-test was used. When more then one group is present ANOVA was done first followed by Students t-test comparing groups with ND control. *p < 0.05.

Figure 5

BM from HFD mice has the sustained potential to generate inflammatory macrophages after serial transplantation. (A) Experimental design of serial BMT experiments. Groups designated by initial donor source (N vs H) followed by diet challenge after the 2nd BMT. (B) Body weight of mice after the 2nd BMT. HFD initiated 6 weeks after BM transfer. (C) Quantitation of BM progenitors after the 2nd transplant in mice fed a HFD. No differences noted in mice fed ND. (D) Quantitation of CD11c⁺ BM cells, total CD64⁺ ATMs and CD11c⁺ ATMs. (E) Immunofluorescence of epididymal fat pads stained for CD11c highlighting crown-like structures. Representative images shown from similar results from 4 mice per group. (F) Fasting insulin and (G) glucose tolerance testing in ND and HFD donor groups after 2nd BMT. N = 8 in ND groups, N = 10 in HFD groups.

Figure 6

Generation of myeloid progenitors and adipose tissue macrophages is dependent on MyD88. (A) A competitive BMT was performed with MyD88^−/− (CD45.2) and WT (CD45.1) marrow in a 1:1 ratio into CD45 1.2 heterozygous animals. After 16 weeks of HFD the ratio of donor contribution to HSC, GMP, monocytes, and ATMs was determined by flow cytometry (N = 8). Data expressed as a ratio of CD45.2 to CD45.1 (CD45.2 negative staining) cells in each pool. (B) A competitive BMT was performed with TLR4^−/− (CD45.2) and WT (CD45.1) marrow in a 1:1 ratio into CD45 1.2 heterozygous animals. After 16 weeks of HFD the ratio of donor contribution to HSC, GMP, monocytes, and ATMs was determined (N = 7). *p < 0.05, **p < 0.01, ***p < 0.005, ****p < 0.001.