Inside out: Bone marrow adipose tissue as a source of circulating adiponectin

Abstract

The adipocyte-derived hormone adiponectin mediates beneficial cardiometabolic effects, and hypoadiponectinemia is a biomarker for increased metabolic and cardiovascular risk. Indeed, circulating adiponectin decreases in obesity and insulin-resistance, likely because of impaired production from white adipose tissue (WAT). Conversely, lean states such as caloric restriction (CR) are characterized by hyperadiponectinemia, even without increased adiponectin production from WAT. The reasons underlying this paradox have remained elusive, but our recent research suggests that CR-associated hyperadiponectinemia derives from an unexpected source: bone marrow adipose tissue (MAT). Herein, we elaborate on this surprising discovery, including further discussion of potential mechanisms influencing adiponectin production from MAT; additional evidence both for and against our conclusions; and observations suggesting that the relationship between MAT and adiponectin might extend beyond CR. While many questions remain, the burgeoning study of MAT promises to reveal further key insights into MAT biology, both as a source of adiponectin and beyond.

Keywords: adiponectin; anorexia nervosa; bone marrow adipose tissue; caloric restriction; lipodystrophy; obesity; white adipose tissue.

Figure 1.

MAT is under-studied, despite being a major adipose depot. Numbers of publications featuring WAT, BAT or MAT were determined by searching the PubMed database in October 2015 with the following terms: WAT, “adipose tissue” OR “adipocyte” NOT “brown adipose tissue” NOT “brown adipocyte;” BAT, brown adipose tissue OR brown adipocyte; MAT, “marrow adipose tissue” OR “marrow adipocyte” OR “yellow marrow” OR “yellow bone marrow.” Values for WAT, BAT or MAT as percentage of total adipose mass in lean, healthy humans, are based on previous publications.

Figure 2.

Characteristics of MAT in comparison to WAT. Expression or secretion of each factor, relative to WAT, is indicated as follows: greater than WAT, red circle with upward arrow; lower than WAT, green circle with downward arrow; similar to WAT, amber circle with ‘∼’; unknown, gray circle with ‘?’. Where differences refer to mRNA expression, official transcript names are used as follows: Adipoq, adiponectin; Pparg, PPARγ; Fabp4, FABP4; Cebpa, C/EBPα; Lep, leptin; Plin1, Perilipin A. All other differences refer to expression or secretion of proteins. Micrographs for caudal and tibial MAT are H&E-stained sections from mice, rabbits or humans, as indicated. The micrograph of isolated adipocytes is a phase-contrast image of adipocytes from femoral MAT, post-isolation, and is presented for schematic purposes only. Characteristics of caudal and tibial MAT are based on our previously published observations. Characteristics of isolated BM adipocytes are based studies of MAT obtained from tibiae/femurs of mice or the iliac crest of humans. These observations demonstrate that MAT expresses and secretes adiponectin, but many questions remain to be addressed. Abbreviations and other details are given in the main text.

Figure 3.

Adiponectin expression in human femoral MAT. Subcutaneous WAT and MAT were isolated from the femoral heads of patients undergoing hip-replacement (Patients 1–3) or from the femoral diaphysis of an amputation patient (Patient 4). (A) Representative micrographs of H&E-stained tissue sections. Scale bar = 200 µm. (B) Total protein was isolated from scWAT and MAT of each patient and expression of the indicated proteins was assessed by immunoblotting; similar results were observed for tissue samples obtained from two other hip-replacement patients (data not shown). Expression of α-tubulin was analyzed as a loading control, although expression was sometimes variable between each tissue type. For patients 1–3, MAT and scWAT lysates were run on non-adjacent lanes of the same gel, and therefore intervening lanes have been removed for ease of comparison. Both the Institutional Review Boards of the University of Michigan and of the Veterans Affairs Hospital of Ann Arbor, MI, approved the study involving hip-replacement patients (IRB number: HUM00053722). The University of Michigan Medical School Institutional Review Boards approved the study involving lower-limb amputation patients (IRB number: HUM00060733). Methods for histology and immunoblotting are as described previously.

Figure 4.

Ocn-Wnt10b mice resist MAT expansion during CR. To determine if Ocn-Wnt10b mice resist MAT formation, we stained tibiae with osmium tetroxide and analyzed MAT volume in situ by micro-CT scanning, as described previously. Representative images of osmium-stained tibiae from wild-type mice, fed a control or CR diet, are shown on the left of the figure; osmium-stained MAT appears as darker regions within the bones. MAT volume was then quantified for each of the indicated tibial regions and normalized to total marrow volume to give percentage of MAT volume, as shown in the graph on the right. Data in the graph are presented as mean +/− standard deviation of the following numbers of mice: WT control, n = 6; Wnt10b control, n = 4; WT CR, n = 5; and Wnt10b CR, n = 6. For each diet group, significant differences between WT and Wnt10b mice are indicated by *** (P < 0.001). Within each genotype, significant differences between control and CR diets are indicated by # (P < 0.05) or ### (P <0.001).

Figure 5.

Potential relationships between MAT and circulating adiponectin in health and disease. Circulating adiponectin is represented as low-molecular-weight (LMW) trimers, middle-molecular-weight (MMW) hexamers, and high-molecular-weight (HMW) dodecamers, although HMW forms may consist of even larger multimers. MAT content varies in ‘normal’ physiological and developmental contexts, with further decreases or increases occurring in adverse or pathological conditions, as indicated. In some cases decreases or increases in MAT are paralleled by similar changes in circulating adiponectin (e.g. decreases in CGL1, CGL2, CGL4; increases in CR, AN), while in other conditions MAT and circulating adiponectin change in opposite directions (e.g., heart failure; obesity). Future studies must address the relative contributions of rMAT and cMAT (represented here by micrographs of rMAT and cMAT from rabbits), as well as how these diverse physiological and clinical conditions impact not just MAT content, but also MAT function.