Aging Induces an Nlrp3 Inflammasome-Dependent Expansion of Adipose B Cells That Impairs Metabolic Homeostasis

Abstract

During aging, visceral adiposity is often associated with alterations in adipose tissue (AT) leukocytes, inflammation, and metabolic dysfunction. However, the contribution of AT B cells in immunometabolism during aging is unexplored. Here, we show that aging is associated with an expansion of a unique population of resident non-senescent aged adipose B cells (AABs) found in fat-associated lymphoid clusters (FALCs). AABs are transcriptionally distinct from splenic age-associated B cells (ABCs) and show greater expansion in female mice. Functionally, whole-body B cell depletion restores proper lipolysis and core body temperature maintenance during cold stress. Mechanistically, the age-induced FALC formation, AAB, and splenic ABC expansion is dependent on the Nlrp3 inflammasome. Furthermore, AABs express IL-1R, and inhibition of IL-1 signaling reduces their proliferation and increases lipolysis in aging. These data reveal that inhibiting Nlrp3-dependent B cell accumulation can be targeted to reverse metabolic impairment in aging AT.

Keywords: B cell depletion; IL-1 signaling; Nlrp3 inflammasome; adipose tissue B cells; age-associated B cells; aging; fat-associated lymphoid cluster; growth hormone receptor; inflammaging; lipolysis.

Conflict of interest statement

Declaration of Interests

The authors declare no conflict of financial interest.

Copyright © 2019 Elsevier Inc. All rights reserved.

Figures

Figure 1.

Aged white VAT shows a unique increase in memory-like B lymphocytes in FALCs (A) Contour plots gated through CD45+ iv antibody− live cells in stroma vascular fraction (SVF) of 3- and 24-month VAT. Values represent the mean percentage of B220+ cells. (B) Quantification of CD3+ and B220+ lymphocytes as a percentage of live CD45+ residents. (C) Quantification of resident and circulating cells as a percentage of total B220+ cells from 3- or 24-month mice. (A-C) n=3/group/cohort and pooled from two cohorts; WT female mice. Quantification of B220+ lymphocytes as (D) a percentage of live CD45+ iv− cells or (E) as cells per gram of AT in male or female mice at 5-, 10-, 15- or 20-months. (D-E) n= 5/group; WT male and female mice. (F) Quantification of resident (iv antibody−) B220+ cells as a percentage of live CD45+ iv− cells in brown (BAT), subcutaneous (SAT) and visceral AT (VAT). n=8 and 7/group; WT female mice. (G) Quantification of lymphocyte infiltration based on hematoxylin and eosin (H&E) staining in the VAT and SAT of 3- or 20-month-old mice. n=8/group; WT female. (H) Representative Masson’s trichrome staining showing a FALC in 22-month VAT. WT female. Representative H&E staining showing a (I) FALC in 22-month VAT or (J) inguinal lymph node in SAT of 22-month mouse. WT female. (K) Quantification of the percentage of Ki67+ cells, as a proportion of cell subsets, in spleen, VAT or mesenteric AT (MAT) cells of 24-month-old mice. n=3 to 9 samples/tissue; WT female mice. Representative whole mount confocal imaging in (L) 22-month-old VAT of Hec6ST mice to visualize HEVs (green) in B220+ FALCs (red). N=3; Female mice or (M) in 22-month-old VAT of ProxTom mice to visualize lymphatic vessels (tomato) in B220+ FALCs (green). N=3 female mice. Each symbol represents one mouse unless otherwise indicated. All error bars represent mean±SEM. *<0.05, **<0.01, ***<0.005. Statistical significance was determined by an ANOVA with posthoc test to adjust for multiple corrections. See also Figure S1.

Figure 2.

Age-expanded adipose B cells are a unique subset of tissue resident ABCs (A) Quantification of the percentage of CD11b+ and CD11b− out of the total B220+ residents in VAT from 3- or 24-month-old male and female mice. n=3/group; WT mice. (B) Linear support vector analysis showing gene signature enrichment of AABs that overlap with plasma (light blue), memory (dark blue) or naïve (green) B cells (Newman et al., 2015). (C) Quantification of B cell subsets in spleen, MAT and VAT from 24-month female WT mice. n=4. (D) Contour plots (gated through CD45+ B220+ live cells) of CD21 and CD23 expression on B cells from 3-month-old mice after no stimulation (UNTX) or stimulation with AT media from 3- or 24-month-old mice. n=3–4; Repeated multiple times; WT female cells and media. (E) Quantification of Tbet+ cells as a percentage of live CD45+ B220+ cells in spleen and VAT. N=2 or 5/group; WT female mice. (F) PCA revealing separation of AT B cells and splenic ABCs. (G) Volcano plot showing log2 fold change and −log10 (padj) comparing AT B cells and splenic ABCs. Red dots are genes that are significantly over-represented, blue dots are under-represented and grey symbols signify no change. Each symbol represents one mouse unless otherwise indicated. Sequencing analysis is described in the methods. Otherwise, all error bars represent mean±SEM. *<0.05, **<0.01, ***<0.005. Statistical significance was determined by an ANOVA with posthoc test to adjust for multiple corrections. See methods for description on RNA seq. See also Figure S2.

Figure 3.

Depletion of B cells restores AT metabolic capacity (A) Fold change (% of B cell from CD20 mAb tissue/ average % of B cell from ISO tissue). (B) Frequency of B220+ cells gated through live CD45+ cells in VAT of 20-month-old WT mice given a single intra-AT injection of ISO or CD20 mAb. (C) Blood glucose following an insulin tolerance test (ITT) or (D) glycerol per gram of VAT in an ex vivo lipolysis assay from mice given treatment as described. (A-D) n=7/group WT female mice; 20m. (E) Glycerol per gram of tissue in VAT from mice that were fasted for 24 hours. n=9 and 8/group WT female mice; 20-months. (F) Histogram plots showing B220+ cells in tissues from 4- or 21-month-old mice given intraperitoneal injection of ISO or CD20 mAb. Values represent the mean percentage for that condition. n=6/4/5 at 4-/21-/21-month WT female mice. (G) Whole mount imaging showing representative FALCs in mesenteric AT from 22-month-old mice. (Top) FALCs from AT of mice treated with ISO control, (bottom) FALCs from mice treated with CD20 mAb. (H) Quantification of ABCs, FO B cells and MZ B cells in spleen of 5- or 21-month-old mice given ISO control or 21-month-old mice treated with CD20 mAb. n=3/4/4 at 5m/21m/21m WT female mice (I) Quantification of the percentage of CD4+ T cells out of the total CD45+ live cells in spleen or VAT. (right top) Quantification of CD4+ FoxP3+ CD25+ cells out of the total CD4+ T cells. (right bottom) Quantification PD1+ cells out of total CD4+ cells in the spleen and VAT tissue of conditions described. n=6/4/5 at 4m/21m/21m WT female mice. (J) Western blot showing phosphorylated HSL, ATGL, total HSL and ACTIN levels in VAT of 3-month, 22-month-old give isotype control or CD20 mAb as described. Representative of two individual experiments. n=2/3/4 at 3-/22-/22-month WT female mice. (K) Glycerol release (mM Glycerol per gram of tissue) from the VAT of 3-month-old WT mice that receive donor B cells via intraperitoneal injection. n=5/8/5 pooled from two separate experiments; WT female mice. (L) Rectal temperature at day 0 and day 3 in WT mice treated as described and exposed to four degrees. n=6/7/9 at 3–8-/23-/23-month-old WT female. Each symbol represents one mouse unless otherwise indicated. All error bars represent mean±SEM. *<0.05, **<0.01, ***<0.005. Statistical significance was determined by an ANOVA with posthoc test to adjust for multiple corrections. See also Figure S3.

Figure 4.

Age-induced adipose B cell expansion and FALC formation in adipose is regulated by the Nlrp3 inflammasome (A) hole mount confocal imaging of FALC in VAT of 22-month-old mT/mG;LysMcre with B220 (yellow) and DAPI (blue) antibody staining. mTomato expressed on all cells and mGFP on myeloid cells. (B) Contour plots showing B220+ CD19+ AT B cells (gated through CD45+ live cells) from the VAT of 2- or 18–22-month-old WT and Nlrp3−/− mice. Percentage and cellularity (cells/g tissue) below. (C) Numbers of FALCs per AT in 24-month old WT or Nlrp3-deficient AT. n=4/5 female mice. (D) Whole mount confocal imaging to visualize Prox1+ (green) lymphatic endothelial cells in FALCs (B220+; red). (E) Quantification of Ki67+ cells (gated as a percentage of B220+ CD19+ cells) from the VAT of 19-month-old WT and Nlrp3−/− mice. n=3/group female mice. (F) Quantification of CD11b+ or CD11b− B cells (gated as a percentage of B220+ CD19+ cells). (B &F) n=(3) WT, (6) WT, or (6) Nlrp3−/− at 2-,19–22-, or 19–22-month-old female mice (G) Quantification of CD21−CD23− B cells (gated as a percentage of B220+ CD19+ cells) from the VAT of 19-month-old WT and Nlrp3−/− mice. N=3/3 female mice. (H) Quantification of CD11c MFI on B cells (gated as a percentage of B220+ MHCII+ cells) from the VAT of 24-month-old WT and Nlrp3−/− mice. N=4/5 female mice. (I) Differentially expressed genes in AABs from 19–22-month-old WT and Nlrp3−/− mice. Grey dots represent no significant change in gene expression, red dots represent significant changes in gene expression (FDR<0.1). (J) Volcano plot showing log2 fold change and −log10 (padj) comparing WT AABs and Nlrp3−/− AABs. Red dots are genes that are significantly over-represented, blue dots are under-represented and grey symbols signify no change. Each symbol represents one mouse unless otherwise indicated. All error bars represent mean±SEM. *<0.05, **<0.01, ***<0.005. Statistical significance was determined by a 2-sided, unpaired T-test or ANOVA with posthoc test to adjust for multiple corrections. RNA sequencing analysis is described in supplemental data. See also Figure S4.

Figure 5.

B cell expansion is associated with reduced healthspan and lifespan (A) Pseudocolor density plot gated through CD45+ B220+ MHCII+ splenocytes to show B cell subsets in 3- or 24-month-old WT and Nlrp3-deficient mice. (B) Quantification of splenic B cell subset frequency. MZ: marginal zone B cells; FO: follicular B cells. n= (10) 3-month WT, (6) 3-month Nlrp3−/−, (8) 24-month WT, or (5) 24-month Nlrp3−/− female mice. (C) Principal component analysis plot showing WT versus Nlrp3−/− splenic ABCs. (D) Volcano plot showing log2 fold change and mean of normalized counts comparing WT versus Nlrp3−/− splenic ABCs. Contour plots showing (E) CD21 and CD23 expression on B220+ CD19+ cells (gated through CD45+ live cells) from the spleen or (F) B220+ CD19+ B cells (gated through CD45+ live cells) from the VAT of 19-month-old WT and GHRKO female mice. Quantification of the percentage and cellularity (cells/g tissue) below each contour plot. (E-F) n=3 or 4/group female mice. Each symbol represents one mouse unless otherwise indicated. All error bars represent mean±SEM. *<0.05, **<0.01, ***<0.005. Statistical significance was determined by a 2-sided, unpaired T-test or ANOVA with posthoc test to adjust for multiple corrections. RNA sequencing analysis is described in supplemental data. See also Figure S5.

Figure 6.

IL-1 signaling drives adipose B cell proliferation, inflammation and impaired lipolysis (A) Mean expression of chemokine receptors in sorted B cell sequencing analysis. (B) Quantification of Ki67+ cells, as a percentage of B220+ cells in the VAT of aged mice. (C) Correlation analysis in VAT, comparing % of Ki67+ and % of B cells. (D) Ki67+ cells, as a percentage of splenic ABCs of aged mice. (E) Correlation analysis in spleen ABCs, comparing % of Ki67+ and % of B cells. (F) Densitometric relative values of pro-caspase/actin and p20 (active caspase)/actin from VAT of mice treated as described. (G) Glycerol release (mM glycerol per gram of AT) in ex vivo assay from mice treated as described. Each symbol represents one mouse unless otherwise indicated. Each symbol represents a separate animal. n=8 or 9/group at 20-month-old female mice. All error bars represent mean±SEM. *<0.05, **<0.01, ***<0.005. Statistical significance was determined by a 2-sided, unpaired T-test. Pearson’s test was used to determine R2 and P value for correlation analysis. See also Figure S6.