Sleep Fragmentation in Mice Induces Nicotinamide Adenine Dinucleotide Phosphate Oxidase 2-Dependent Mobilization, Proliferation, and Differentiation of Adipocyte Progenitors in Visceral White Adipose Tissue

Abstract

Background:

Chronic sleep fragmentation (SF) without sleep curtailment induces increased adiposity. However, it remains unclear whether mobilization, proliferation, and differentiation of adipocyte progenitors (APs) occurs in visceral white adipose tissue (VWAT), and whether nicotinamide adenine dinucleotide phosphate (NADPH) oxidase 2 (Nox2) activity plays a role.

Methods:

Changes in VWAT depot cell size and AP proliferation were assessed in wild-type and Nox2 null male mice exposed to SF and control sleep (SC). To assess mobilization, proliferation, and differentiation of bone marrow mesenchymal stem cells (BM-MSC), Sca-1+ bone marrow progenitors were isolated from GFP+ or RFP+ mice, and injected intravenously to adult male mice (C57BL/6) previously exposed to SF or SC.

Results:

In comparison with SC, SF was associated with increased weight accrual at 3 w and thereafter, larger subcutaneous and visceral fat depots, and overall adipocyte size at 8 w. Increased global AP numbers in VWAT along with enhanced AP BrDU labeling in vitro and in vivo emerged in SF. Systemic injections of GFP+ BM-MSC resulted in increased AP in VWAT, as well as in enhanced differentiation into adipocytes in SF-exposed mice. No differences occurred between SF and SC in Nox2 null mice for any of these measurements.

Conclusions:

Chronic sleep fragmentation (SF) induces obesity in mice and increased proliferation and differentiation of adipocyte progenitors (AP) in visceral white adipose tissue (VWAT) that are mediated by increased Nox2 activity. In addition, enhanced migration of bone marrow mesenchymal stem cells from the systemic circulation into VWAT, along with AP differentiation, proliferation, and adipocyte formation occur in SF-exposed wild-type but not in oxidase 2 (Nox2) null mice. Thus, Nox2 may provide a therapeutic target to prevent obesity in the context of sleep disorders.

Citation:

Khalyfa A, Wang Y, Zhang SX, Qiao Z, Abdelkarim A, Gozal D. Sleep fragmentation in mice induces nicotinamide adenine dinucleotide phosphate oxidase 2-dependent mobilization, proliferation, and differentiation of adipocyte progenitors in visceral white adipose tissue. SLEEP 2014;37(5):999-1009.

INTRODUCTION

Reduced sleep duration is a frequent and pervasive phenomenon in modern societies and leads to increased prevalence of excessive daytime sleepiness.¹ In recent years, it has also become apparent that insufficient sleep may modify the propensity for development of obesity in humans.^2–7 Another frequent condition that has received significantly less attention, fragmented sleep, is a cardinal and characteristic feature of a variety of highly prevalent human diseases (e.g., sleep apnea, depression, asthma). Although fragmented sleep is not necessarily accompanied by reduced sleep duration, it is associated with excessive daytime sleepiness and with neurocognitive and mood deficits.⁸ Studies in humans and mice have shown that sleep fragmentation (SF) leads to adverse metabolic consequences such as increased food consumption and insulin resistance.^9–12 However, because human experiments are obviously restricted in duration, we recently explored whether long-term SF ultimately leads to obesity in a mouse model. These experiments confirmed that mice exposed to chronic SF showed increases in food consumption, and developed exaggerated trajectories in body weight accrual and insulin resistance that appeared to be mediated, at least in part, by increased oxidative stress starting at 3–4 w, and culminating in frank obesity at 8–12 w.^13,14 Interestingly, transgenic mice that do not express the nicotinamide adenine dinucleotide phosphate (NADPH) oxidase 2 gene (NOX2) and are exposed to SF manifest similar increases in sleep propensity but do not develop insulin resistance.^13,15

Obesity is a state of chronic low-grade inflammation characterized by elevated concentrations of circulating inflammatory markers. Several reports suggest that the adipose tissue itself, particularly the visceral white adipose tissue (VWAT), might be a source of pro-inflammatory factors and a target of inflammatory processes.¹⁶ Increased accumulation of VWAT is considered a strong and independent predictor of adverse health outcomes associated with obesity, such as hypertension, insulin resistance, and atherosclerosis.¹⁷ It is currently assumed that the liver is directly exposed to increasing amounts of free fatty acids and pro-inflammatory factors released from VWAT into the portal vein of obese patients (i.e., the so- called portal theory), promoting the onset and progression of hepatic insulin resistance and liver steatosis.^18,19

A large body of evidence indicates that distinct adipocyte populations play a role in obesity, and originate from diverse sources including different germ cell layers, the formation of singular preadipocyte populations from mesenchymal progenitors, and the production of adipocytes from hematopoietic stem cells from the bone marrow.^20–25 The generation of new adipocytes from progenitor cells has been a topic of great interest, in terms of understanding normal adipose tissue development and turnover, and the expansion of adipose tissue that occurs with obesity. The stromal vascular fraction (SVF) in adipose tissues is composed of heterogeneous cell types, including macrophages, preadipocytes, fibroblasts, and non-differentiated mesenchymal stem cells.^20–24 Based on the important roles played by adipocyte progenitor (AP) cells in obesity,^21,23 we hypothesized that SF would lead to increased oxidative stress via activation of Nox2, and promote mobilization, proliferation, and differentiation of AP in visceral white adipose tissue (VWAT), ultimately culminating in frank obesity (Figure 1).

Figure 1.

Schematic diagram illustrating the working hypothesis of the current study. Sleep fragmentation will induce the generation of oxidative stress via activation of nicotinamide adenine dinucleotide phosphate (NADPH) oxidase in visceral white adipose tissue. Increased reactive oxygen species (ROS) in adipose tissue lead to mobilization of adipocyte progenitors residing in adipose tissue or from bone marrow and their proliferation and differentiation into adipocytes. Increased visceral white adipose tissue mass will then promote obesity and appearance of insulin resistance.

MATERIALS AND METHODS

Animals

Male C57BL/6J mice (termed wild type; WT), hemizygous gp91^phox-/Y mice on a C57BL/6J background (termed Nox2), TgN-actin-EGFP mice (termed GFP), and B6.Cg-Tg(ACTB-mRFP1)1F1Hadj/J mice (termed RFP), weighing 22-25 g, were purchased from Jackson Laboratories (Bar Harbor, ME), housed in a 12-h light/dark cycle (light on 07:00 to 19:00) at a constant temperature (24 ± 1^°C) and allowed access to food and water ad libitum. Of note, C57BL/6 mice and Nox2 null mice were exposed to sleep fragmentation, whereas C57BL/6, GFP, and ACTB-mRFP1 mice were used to isolate Sca1+ from bone marrow. The experimental protocols were approved by the Institutional Animal Use and Care Committee and are in close agreement with the National Institutes of Health Guide in the Care and Use of Animals. All efforts were made to minimize animal suffering and to reduce the number of animals used.

Sleep Fragmentation

The custom-designed SF approach used to induce SF in rodents has been previously reported in detail^8,26,27 and relies on automated intermittent tactile stimulation of freely behaving mice in a standard laboratory mouse cage, using a near-silent motorized mechanical sweeper. This method obviates the need for human contact and intervention, and does not involve introduction of foreign objects or touching of the animals during sleep. To induce SF, we chose a 2-min interval between each sweep, implemented during the light period (07:00 to 19:00) for 2–8 w based on the experimental paradigms. Four to five mice were housed in each cage to prevent isolation stress. SF exposures varied from 2–8 w as dictated by the experimental design.

Body Weight

Body weight was measured twice a week for a period of 8 w, always at the same time of the day (middle of the light cycle period). Body weight gain was determined by subtracting the body weight on first day of SF exposure from the body weight on subsequent days.

Magnetic Resonance Imaging Quantitation of Visceral and Subcutaneous Adipose Tissues

To enable improved quantification of fat deposition in mice, we conducted magnetic resonance imaging (MRI) studies to obtain high-resolution images of the abdominal cavity for unbiased quantification of visceral and subcutaneous fat compartments in mice exposed to SF paradigms. Images were acquired on a 9.4 Tesla (T)/20 USR Bruker BioSpec (Ettlingen, Germany) equipped with BGA12S actively shielded gradients and ParaVision 4.0 software using a 50-mm i.d., 90 mm-long quadrature resonator (m2m Imaging, Brisbane, Australia). Acquisition was synchronized with the respiratory cycle to minimize physiological artifacts (SA Instruments, Stony Brook, NY). Two sets of proton density high-resolution scans (echo time/repetition time [TE/TR] 4000/27 ms, field of view [FOV] 40 × 40 mm, covered the abdominal cavity. A 20° flip angle gave optimal contrast between background and tissue, Image sets underwent manual segmentation procedures using AMIRA software (http://www.amira.com/amira/quantification.html; version 5.4) as previously described.²⁸

Isolation of AP Cells

Visceral fat tissues were dissected and washed with phosphate buffered saline (PBS), and immediately incubated in Hank balanced solution (HBSS, Invitrogen, Carlsbad, CA) containing 1% collagenase type 1A for 1 h at 37°C. Collagenase activity was neutralized with an equal volume of Dulbecco Modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS). The digested solutions were filtered through a 70-μm and 40-μm screen mesh followed by centrifugation at 300g for 10 min. After enrichment of lineage negative cells (LNC) using the lineage cell depletion kit (Miltenyi Biotec Inc., Auburn, CA), the SVF from three mice/sample was incubated with monoclonal antibodies to CD29 APC, CD34 FITC, Sca1 Pacific Blue, and CD24 PE antibodies in 100 μL staining buffer for 15 min on ice. Isotype controls relevant for each antibody were used to establish background fluorescence. Dead cells were excluded by propidium iodide (PI), and cell numbers and viabilities were assessed with an automated cell counter (Cellometer, Nexcelom Bioscience, Lawrence, MA). Data were acquired on a FACS CantoII cytometer using the FACS Diva 5.5 software (BD Biosciences, San Jose, CA). The results were analyzed by FlowJo software (Tree Star, San Carlos, CA). We used this approach to sort AP cells from adult male mice exposed to 7 days of SF and sleep control (SC) mice, with AP cells being represented as Lin⁻:CD29⁺:CD34⁺:Sca-1⁺:CD24⁺.²⁹

AP Cell Culture

The sorted AP cells from the SVF or visceral white adipose tissues (VWAT) were cultured in DMEM and F12 (1:1, vol/ vol) with 10% FBS, 100 units/mL penicillin, 100 μg/mL streptomycin as described by Rodeheffer and colleagues.³⁰ After reaching 80% of confluence, the cells were transferred into differentiated media containing 100 μM 3-isobutyl-1-methylxanthine (IBMX), 60 μM indomethacin, and 1 μg/mL insulin, all from Sigma-Aldrich, as previously described.³⁰. The differentiating medium was replaced every other day, and mature adipocytes were formed after 1 week.

Oil Red O Staining

Progenitor cells were stained with Oil Red O to assess lipid accumulation, as described previously.³¹ Briefly, cells were washed with 1x PBS, fixed in 4% paraformaldehyde solution for 30 min at room temperature, rinsed twice with PBS, and followed by washing with 60% isopropanol for 2 min. Cells were then stained with Oil Red O solution for 30 min, followed by a gentle rinse with water. The accumulated lipid with Oil Red O complexes were eluted with 100% isopropanol and the extracted lipids were measured at 510 nm.

5-Bromo-2'-Deoxyuridine Labeling and Immunocytochemistry

In vivo labeling

Mice (n = 7-10/experimental group) were injected intraperitoneally (i.p.) with 200 μL of 5-bromo-2'-deoxyuridine (BrDU) solution (1 mg/mL) twice a day for 3 days. Mice were divided into three groups: (1) control mice that sleep normally and injected with 1× PBS solution (negative control); (2) control mice injected with BrdU; and (3) mice exposed to SF for 2 w injected with BrDU. The BrDU mixture was prepared fresh and changed daily. VWAT were dissected and prepared for immunohistochemistry by fixing in 4% paraformaldehyde for 24 h. Dissected visceral fat tissues were embedded in paraffin and sectioned at 5 μm, and sections were incubated with anti-BrDU antibody (RDI) at a concentration of 1:5,000 for 1 h at room temperature. Tissue sections were then incubated with antisheep biotinylated secondary antibody (Vector Laboratories, BA-6000), and then the Elite ABC kit was used (Vector Laboratories, PK-6100). The antigen-antibody was detected by 3,3'-diaminobenzidine tetrahydrochloride (DAB) substrate chromogen system (DAKO, K3466). The slides were either briefly immersed in hematoxylin-eosin for counterstaining or were not stained with hematoxylin-eosin, and were evaluated under a microscope by investigators who were blinded to the experimental setting. BrdU expression was imaged using NIS-Elements AR vs. 3.10 imaging software (http://nikon.com/products/instruments/lineup/bioscience/img_soft/index.htm). Sections were visualized with a Nikon Elipse Ti microscope (Nikon Inc, Melville, NY) using 10 × objectives and photographed with a Cool SNAP EZ camera (Photometrics, Tucson, AZ). In addition, adipocytes were visualized and their diameters were determined by using OpenLab 4 software (Improvision, Waltham, MA, http://www.improvision.com).

BrDU Incorporation In Vitro

AP cells were isolated from VWAT of mice exposed to SF and SC. The BrDU Cell Proliferation Assay Kit was used to detect BrDU incorporated into cellular DNA during cell proliferation (Cell Signaling Technologies, Danvers, MA). Cells were plated into 24-well plates in equal number (3 × 10⁵) per well and allowed to grow such as to reach 80% confluence as described previously. Cells were pulsed with 160 μM BrDU for 48 h at 37°C. After labeling medium was removed, cells were fixed with 37% formaldehyde for 15 min. Cells were blocked for 15 min at room temperature. Cells were incubated with BrDU mouse monoclonal antibody for 1 h at 37°C, and subsequently were incubated with secondary goat-antimouse antibody at room temperature for 30 min. Cells were visualized using a Nikon microscope.

Bone Marrow Isolation

Bone marrow cells (BMCs) were flushed from the femur bone marrow of either GFP or RFP-8-w-old mice. Sca-1+ bone marrow progenitors (hematopoietic stem cell [HSC]) were sorted using anti-Sca1 antibody conjugated to Sca-1-micro-magnetic beads (Miltenyi Biotec, Inc., Auburn, CA), and injected into adult male mice (C57BL/6, Jackson Laboratories) that were exposed to SF or SC conditions. Briefly, BMCs were obtained by flushing the femurs of GFP-expressing transgenic mice with cold washing buffer containing PBS, 0.5% bovine serum albumin (BSA), and 2 mM ethylene diaminetetraacetic (EDTA). The cells were filtered through 40-μm screen mesh (BD Biosciences, San Jose, CA), and subsequently centrifuged at 300g for 10 min. Red blood cells (RBCs) were removed using RBC lysis buffer (Miltenyi Biotec, Inc., Auburn, CA). The cells were counted using an automated cell counter (Cellometer), and the concentration of BMC was prepared to be 1 × 10⁷ cells per mL. The single cell suspension was then incubated with anti-Sca-1-FITC antibody, followed by incubation with mouse Sca-1 microbeads (Miltenyi Biotec, Inc., Auburn, CA) for 15 min at 4°C refrigerator. The cell preparation was then incubated with magnetic microbeads conjugated with antibody specific to Sca-1 and applied twice to an automated magnetic separation column according to the manufacturer's instructions (Miltenyi Biotec, Inc., Auburn, CA). Sca-1+ cell-enriched preparations were measured by automated cell count of viable cells as determined by trypan blue dye exclusion. To assess enrichment efficiency, aliquots of each cell preparation were incubated with either phycoerythrin (PE)-conjugated Sca-1-specific antibody (Biosciences) and analyzed for Sca-1 and/or GFP expression with a FACS Calibur System (BD Biosciences, San Jose, CA). The percentage of Sca-1+ cells was calculated by subtracting the value obtained with the PE-conjugated rat isotype control antibody from that obtained with the PE-conjugated Sca-1-specific antibody.

WT and Nox2 mice exposed to SF or SC for 2-4 w were injected with 4 × 10⁵ GFP + bone marrow Sca-1+ cells via the tail vein, after which exposures were continued for 1 and 3 w. VWATs were then collected for flow cytometry and immunostaining.

VWAT Sca-1+ Cell Enrichment

VWATs were incubated in Hank balanced solution (HBSS, Invitrogen, Carlsbad, CA) containing 1% collagenase/HBSS (Fisher Scientific, Waltham, MA) for 45-60 min at 37°C. The pellets containing the SVF fraction were further treated with RBC lysis for 10 min at room temperature. Cell pellets were suspended in 400 μL washing buffer and analyzed for Sca-1, RFP, or GFP expression by flow cytometry.

Immunohistochemistry

Fresh VATs from mice injected with Sca-1+ cells were fixed in 4% paraformaldehyde for 24 h. Formalin-fixed and paraffin-embedded tissues were cut into 5-μm slices. Sections then were incubated with the primary antibody against perilipin (Cell Signaling, Danvers, MA; 1:100) for 1 h at room temperature. After subsequent washes in PBS for 20 min, tissue sections were incubated with antisheep biotinylated secondary antibody (Vector Laboratories, BA-6000), and then Elite ABC Kit (Vector Laboratories, PK-6100). The slides were either immersed in hematoxylin for counterstaining or without hematoxylin and evaluated under a microscope, and were imaged using NIS-Elements AR vs. 3.10 imaging software. Sections were visualized with a Nikon Elipse Ti microscope using 10× objectives and photographed with a Cool SNAP EZ camera.

Statistical Analyses

Data are reported as mean ± standard deviation, and all analyses were conducted using SPSS software (version 17; Chicago, IL). Comparisons according to group assignment were determined with independent t-tests or analysis of variance followed by post hoc comparisons, with P values adjusted for unequal variances when appropriate (Levene test for equality of variances), or χ² analyses with Fisher exact test (dichotomous outcomes). A two-tailed P < 0.05 was considered statistically significant.

RESULTS

SF Induces Increases in Body Weight and Visceral Fat Mass in WT But Not in Nox2 Null Mice

WT mice exposed to SF exhibited substantial accelerations in body weight accrual, starting at week 3 of SF and thereafter. These changes did not occur in SF-exposed Nox2 null mice (Figure 2A). At the end of 8 w of SF, marked increases in both subcutaneous and visceral body fat were apparent in SF-exposed WT mice, but not in Nox2 null mice (Figure 2B). Indeed, for WT mice, VAT and subcutaneous adipose tissue (SAT) volumes were 2,376 ± 244 mm³ and 548 ± 126 mm³ respectively, compared with 487 ± 78 mm³ and 204 ± 41 mm³ in sleep control conditions (P < 0.0001). In contrast, VAT and SAT volumes in SF-exposed Nox2-deficient mice were 299 ± 63 mm³ and 68 ± 17 mm³, respectively, compared with 330 ± 72 mm³ and 113 ± 35 mm³ in sleep control Nox2-deficient mice (P > 0.05). For three animals in each experimental group, adipocyte diameters were measured from six random fields of hematoxylin and eosin (H&E)-stained paraffin sections from VWAT (n > 150 adipocytes per animal) and counted based on specified diameter range categories. These analyses showed enlarged size of adipocytes after 8 w in WT mice exposed to SF for 8 w (Figure 2C; P < 0.01) compared with no changes in adipocyte size patterns in Nox2 transgenic mice (Figure 2C; WT vs. Nox2, P < 0.01).

Figure 2.

(A) Time course of body weight gain in wild-type (WT) and Nox2 null mice (Nox2) exposed to either sleep fragmentation (SF) or control sleep condition (SC) for 8 w. Increased weight gain became apparent starting at 3 w after initiation of SF and thereafter. (B) Representative visceral (yellow) and subcutaneous (green) fat content in the abdominal section of WT and Nox2 mice subjected to either SF or SC for 8 w. Magnetic resonance images underwent three-dimensional reconstruction using AMIRA software. These images are representative of five to six mice/experimental group. (C) Mean distribution of adipocyte diameters as measured from three random fields of hematoxylin and eosin-stained paraffin sections from visceral white adipose tissue (n > 150 adipocytes per animal, per group) averaged from three mice/group after 6 w of each experimental condition. Shifts toward larger adipocyte diameters are clearly apparent in SF-exposed wild type mice, but no changes emerged in Nox2 null mice.

SF Increases the Number of AP Cells VWAT of WT Mice, But Not in Nox2 Null Mice

Based on the putatively critical role of AP cells in obesity, we first examined AP cell counts in the SVF of VWAT in mice subjected to SF. Unlike bone marrow mesenchymal stem cells (BM-MSC), AP cells express CD34,³² which enables their enumeration by flow cytometry as Lin⁻:CD29⁺:CD34⁺:Sca-1⁺:CD24⁺ cells.^29,33 For VWAT, which encompasses the major AP cell reservoir,³⁴ the AP frequency in SVF was 56.7 ± 9.4% in WT mice exposed to SF for 2 w (i.e., before differences in somatic weight gain emerge) and 38.7 ± 8.6% in SC-exposed WT mice, respectively (Figure 3; n = 6 separate experiments; P < 0.001). Furthermore, these findings persisted throughout the duration of exposures up to 8 w (data not shown), at which time the obese phenotype was fully apparent (Figure 2B). Similar experiments in Nox2 null mice revealed no differences in AP cell counts between SF and SC after 2-w exposures (Figure 3; 36.2 ± 7.8% in SC and 39.5 ± 8.2% in SF; n = 6 mice/condition; P > 0.05).

Figure 3.

Gating strategy for adipocyte progenitor (AP) cells enumerated as Lin⁻:CD29⁺:CD34⁺:Sca-1⁺:CD24⁺ cells, and representative experiment of the percentage of AP cells in visceral white adipose tissue (VWAT) of a wild-type (WT) mouse exposed to sleep fragmentation (SF) for 2 w and a time-matched sleep control (SC). The right panel depicts the summary of six different experiments in WT and Nox2 null mice exposed to either SF or SC for 2 w. Significant increases in AP cell counts in VWAT emerged only in WT mice exposed to SF.

Upon plating cells from SVF of VWAT, adherent AP cells appeared as large fibroblasts with well-defined nuclei and nucleoli, and were easily distinguishable from smaller myelomonocytic cells (Figure 4). Quantification of adherent AP from the SVF of VWAT upon enzymatic tissue digestion showed increased efficiency in their recovery from 2-w SF-exposed mice (Figure 4A and 4B; 53.9 ± 11.7% in SF versus 31.3 ± 8.3% in SC; n = 9/group; P < 0.001). Furthermore, under undifferentiating media conditions, the number of proliferating cells (i.e., BrDU-positive) derived from the SVF of VWAT of 2-w SF-exposed mice was significantly higher than in SC sleep conditions (Figure 4G and 4H; P < 0.001; n = 9/group). In addition, there were increased numbers of adipocyte-looking cells (Figure 4C and 4D) and Oil-Red O positively labeled cells after 3 days exposures of AP to differentiating medium (Figure 4E and 4F, P < 0.001; n = 9/group). When these experiments were performed using SVF from the VWAT of Nox2 null mice, there were no significant differences between WT SC-exposed mice and Nox2 null mice (P > 0.05; n = 5/group), and more importantly, all the differences reported between SF and SC conditions were absent in Nox2 null mice (n = 5; P > 0.05).

Figure 4.

Representative images of adipocyte progenitor (AP) cells harvested from the stromal vascular fraction (SVF) of visceral white adipose tissue in mice exposed to sleep control (SC, upper panels) or sleep fragmentation (SF, lower panels) in cell culture under undifferentiating media conditions (A, B), differentiating media (C, D), Oil Red O staining (E, F), and after pulsing with 5-bromo-2'-deoxyuridine (BrDU; nuclei are stained in blue and BrDU in green) (G, H). The bar graph shows the mean BrDU cell counts/field for AP cells cultured from wild-type (WT) and Nox2 mice exposed to either SF or SC (n = 6 experiments/group).

SF Increases the Number of Proliferating Cells in VWATs of WT Mice, But Not of Nox2 Null Mice

Weekly pulsed injections of BrDU to mice exposed to either 2 w SF or SC revealed increased numbers of BrDU positively labeled cells in VWAT following SF (Figure 5, P < 0.001). These cells were located in the vascular regions of VWAT surrounding adipocytes. However, after 4–6 w of SF, some of the adipocytes showed the presence of distinct BrDU labeling, suggesting that newly formed adipocytes may have originated from such proliferating cells (Figure 5).

Figure 5.

Representative images of BrDU immunoreactive cells in visceral white adipose tissue (VWAT) after phosphate buffered saline (PBS) or intraperitoneal BrDU treatment of wild-type (WT) mice exposed to either sleep fragmentation (SF) or sleep control (SC) for 2 w. The bar graph shows the mean number of BrDU+ cells/section in VWAT (n = 6 sections/mouse and n = 5 mice/group). Increased BrDU incorporation was apparent only in WT mice exposed to SF. Arrows in D represent examples of positive labeling. Horizontal bars in A-D indicate magnification scale and represent 100 micrometers.

SF Induces Increased Sca1+-Cell Migration to VWATs of WT Mice, But Not of Nox2 Null Mice

In addition to resident AP, the critical contribution of BM-MSC to VWAT expansion has been proposed.³⁵ To further examine the effect of SF on the homing of BM-MSC to VWAT, intravenous injections of Sca1+ bone marrow cells from GFP or RFP mice to WT mice exposed to SC conditions or to 1 w of SF were performed, and revealed marked increases in the number of GFP-positive or RFP-positive cells in the VWAT of mice exposed to 1 w of SF, particularly 3 w after injection (Figure 6). Intravenous injections of Sca1+ bone marrow cells to Nox2 null mice exposed to either SF or SC failed to reveal any differences in the counts of GFP or RFP labeled cells within the VWAT (Figure 6). Administration of BrDU to these mice showed increased numbers of GFP+ BrDU+ co-labeled cells in SF condition but only in WT mice, because no changes occurred in Nox2 mice (Figure 6). Furthermore, FACS analysis of GFP-+ cells for Sca-1⁺ immunoreactivity revealed a significantly larger number of such cells at 1 w after injection, but not 3 w following injection, suggesting that the disappearance of Sca-1+ immunoreactive cell phenotype may reflect differentiation of the BM-MSC, as indicated by the enhanced perilipin immunoreactivity³⁶ (Figure 7).

Figure 6.

(I) Representative images of visceral white adipose tissue (VWAT) from 1-w sleep fragmentation (SF)- or sleep control (SC)-exposed wild-type (WT) mice 1 w (C and D) and 3 w (E and F) after intravenous injection of 4 × 10⁵ GFP + bone marrow Sca1+ cells via the tail vein. Panels A and B represent unbiased quantitative assessment of mean fluorescence intensity (MFI) for sections after 1 w after injection of Sca-1+ cells and panels G and H for sections obtained 3 weeks after intravenous injection. (II) The right upper graph depicts the number of BrDU+GFP+ or BrDU+RFP+ cells in VWAT from WT and Nox2 null mice exposed to SF or SC for 1 w and injected with Sca-1+ cells, 1 w after injection. (III) The right lower panel shows MFI of VWAT sections from WT and Nox2 null mice exposed to SF or SC for 1 w and injected with Sca-1+ cells at 1 w (filled columns) and 3 w (hashed columns) after injection. BrDU, 5-bromo-2'-deoxyuridine. Horizontal bars in E and F indicate magnification scale and represent 100 micrometers.

Figure 7.

Representative perilipin immunostaining and hematoxylin and eosin counterstaining of visceral white adipose tissue sections from wild-type mice exposed to sleep fragmentation (SF) or sleep control (SC) for 1 w and injected with Sca-1+ cells, 1 w and 3 w after injection. (n = 3/condition.) Horizontal bars in A-D indicate magnification scale and represent 100 micrometers.

DISCUSSION

This study shows that long-term SF during the rest period, as occurs in many human diseases, elicits the appearance of increased somatic weight accrual and obesity in mice. As such, changes are accompanied by evidence of recruitment, proliferation, and differentiation of AP cells (probably both VWAT resident AP and BM-MSC-derived AP cells) during the early and subsequent stages of chronic sleep perturbation. Furthermore, these SF-induced alterations in overall adipose tissue accrual, and more specifically in VWAT, were abrogated in mice lacking NADPH oxidase activity, suggesting that oxidative stress pathways are implicated in SF-induced obesogenic changes.

The strength and scope of the evidence linking sleep duration and integrity and a variety of metabolic consequences including the onset and development of obesity and insulin resistance as well as metabolic syndrome has been steadily increasing in recent years in both epidemiological studies and experimental settings.^2,7,37,38 However, the mechanisms specifically linking altered sleep to the onset and sustainability of obesity have not been thoroughly investigated, and the current study provides initial cues to the underlying processes contributing to the sleep-obesity associations. We have recently shown that in addition to restricted sleep, SF-inducing paradigms will lead to emergence of increased food consumption and ultimately onset of obesity in the absence of sleep curtailment.¹⁴ We further showed that in the early phases of chronic SF (< 2–3 w), insulin resistance and potentially leptin resistance became apparent even before the divergence in body weight became apparent, and that such changes in homeostatic glycemic regulation were abrogated in Nox2 null mice.¹³

NADPH oxidase (Nox) is a multisubunit enzyme complex (composed of membrane-bound (p22phox and gp91phox) and cytoplasmic subunits (p40phox, p47phox, and p67phox) that localizes to both the plasma membrane and membranes of subcellular organelles, and catalyzes electron transfer from NADPH to molecular oxygen, producing superoxide.^39,40 The Nox system is considered a key contributor to generation of reactive oxygen species (ROS) in many cell types and tissues.⁴¹ The first Nox isoform to be identified was Nox2 (i.e., gp91phox) in neutrophils and macrophages.⁴² However, other studies have identified Nox2 at varying expression levels in numerous cell types, including endothelial, vascular smooth muscle, mesangial, and tubular epithelial cells.^43,44 In adipocytes, Nox has recently emerged as a critical regulator of insulin resistance, and its downstream-induced oxidative stress and vascular dysfunction, such that inhibition of Nox2 in vivo led to significant improvements in endothelial cell function in mice with insulin resistance.^45–47

We now show that in addition to prevention of insulin resistance induced by SF, the absence of Nox2 activity further prevents onset of obesity. Considering the well-known patterns of Nox2 expression in inflammatory cells, such as monocytes and macrophages, and the role played by these inflammatory cells in both obesity and insulin resistance,^23,48–51 it will be worthwhile to explore in future studies the specific functional and Nox2 activity changes in these important regulators of adipose tissue properties. A fundamental question is whether or not SF leads to an increase in Nox2 cellular activity. In the context of Nox2 null mice, we have reported that SF fails to increases NADPH oxidase activity in the brain as opposed to marked increases in WT mice.¹⁵ Furthermore, Nox2 null mice are protected from SF-induced cognitive dysfunction, but not from increases in sleep propensity, suggesting that this enzyme plays a role in mechanisms underlying end-organ morbidities associated with disrupted sleep architecture.¹⁵ In addition, work from our laboratory indicated that SF induces increased activity of Nox2 in VWAT, along with polarization and infiltration of M1-type macrophages in VWAT, and these findings are conspicuously absent in Nox2 null mice.¹³

The contributions of AP to obesity through the promotion of adipocyte numbers in several adipose tissue depots is now unquestionable.⁵² Although white adipocytes were originally thought to arise exclusively from progenitors residing in fat stroma,⁵³ such notions have been quite conclusively dispelled with the demonstration that AP cells exhibit different and distinct subpopulation markers attesting to their various potential sources, primarily mesenchymal sources,^35,54 as well as by evidence indicating that bone marrow progenitor cells can be mobilized to VWAT to differentiate into mature adipocytes.^20,55 Some dissenting observations have also been reported, whereby BM-MSCs fail to transdifferentiate into adipocytes.⁵⁶ In the current study, we do not contend nor have we specifically shown that BM-MSCs are the only and exclusive source of AP cells in the emergence of increased adipose tissue mass associated with long-term SF exposures. However, our results demonstrating the increased in vivo mobilization, replication (increased GFP+/ BrDU+ cells), and differentiation of GFP+Sca-1+ BM-MSC in VWAT of SF-exposed mice along with the increased presence of perilipin expression suggest that sleep perturbations do indeed functionally recruit BM-MSC to the VWAT depot. Of course, based on the increased BrDU+ cells in VWAT and the increased proliferative rates of AP harvested form VWAT, we surmise that both visceral fat resident AP cells and BM-MSCs underlie the increases in fat mass observed in SF-exposed WT mice.

Of the multiple mechanisms that have been implicated in the expansion of adipose tissue, we explored the potential contributions by increased Nox2-dependent oxidative stress, and found that absence of Nox2 abrogated the development of SF-induced obesity, and insulin resistance.¹³ These findings occurred despite previous findings whereby Nox2 null mice are not protected from developing excessive sleepiness when subjected to the SF paradigm.¹⁵ Thus, our study illustrates the effect of sleep disruption on end-organ structure and function in tissues beyond the central nervous system such as VWAT. It is possible that depending on the circumstances, increased ROS levels may promote AP cell proliferation and differentiation to adipocytes,^57–60 but may also inhibit this process, particularly when obesity and insulin resistance are already firmly established. Under these assumptions, early generation of excessive ROS during SF in nonobese mice would enhance the differentiation and proliferation of resident AP in VWAT, and further lead to increased mobilization of BM-derived MSC and their conversion to AP. Indeed, ROS derived from overexpression of Nox4 in human-derived adipose stem cells was shown to promote adipogenesis and accumulation of intracellular lipids, whereas treatment with N-acetyl-cysteine markedly attenuated such an effect.⁶¹ Thus, once such early oxidative stress-mediated processes have resulted in frank obesity, the increased oxidative burden generated by recruitment of inflammatory cells to VWAT, and increased ROS generation by adipocytes themselves would potentially lead to reduced adipogenic capacity, and contribute the putative missing link between “dysfunctional” fat expansion and insulin resistance.⁶²

Some methodological limitations should be mentioned. First, in this set of experiments, we did not perform sleep recordings, primarily because of our decision to avoid any disruption of the abdominal cavity by implantation of the telemetry device and other electrodes required for polysomnographic recordings that could have altered the biological processes under investigation.^63,64 Second, we did not explore the identity of the putative mediators released under SF conditions that could potentially lead to the observed changes in resident AP cell, as well as to enhance BM-MSC migration, proliferation, and differentiation into adipocytes when the latter were systemically injected.^65–68 Finally, we have not established whether cessation of SF will be accompanied by reversal of the processes described herein, or whether a window of reversibility is present whereby complete return to baseline conditions is possible upon discontinuation of SF, but only for a limited time, and once such processes are firmly established as indicated by emergence of accelerated body weight gain the reversibility is progressively hampered.

In summary, we have shown that chronic fragmented sleep as occurs in many human disorders is accompanied by the appearance of increased fat mass and obesity that are preceded by increased proliferation and differentiation of AP in VWAT, possibly by induction of local resident AP but potentially also by recruiting BM-MSC to VWAT. Furthermore, such processes appear to require the presence of increased SF-induced Nox2 activity. The pathways uncovered in this murine model of sleep disorders may also be pertinent to the obesogenic effects of chronic sleep curtailment, and indicate potential therapeutic targets aiming to circumvent one of the major adverse consequences of the sleepless modern society in which we live.

DISCLOSURE STATEMENT

This was not an industry supported study. The authors have indicated no financial conflicts of interest. Dr. Gozal is supported by National Institutes of Health grants HL65270 and HL86662. Dr. Khalyfa and Dr. Zhang were supported by Comer Kids Classic grants. Drs. Khalyfa and Wang were equal contributors to the work.