NanoSPECT imaging reveals the uptake of ¹²³I-labelled oxidized low-density lipoprotein in the brown adipose tissue of mice via CD36

Abstract

Aims

The liver is the major organ shown to remove oxidized low-density lipoprotein (oxLDL) from the circulation. Given increased evidence that thermogenic adipose tissue has anti-effects, we used ¹²³I-labelled oxLDL as a tracer to reveal oxLDL accumulation in the brown adipose tissue (BAT) of mice. We also explored the mechanisms of oxLDL accumulation in BAT.

Methods and results

We used high-resolution nanoSPECT/CT to investigate the tissue distribution of ¹²³I-oxLDL and ¹²³I-LDL (control) following intravenous injection into conscious mice. ¹²³I-oxLDL distribution was discovered in BAT at an intensity equivalent to that in the liver, whereas ¹²³I-LDL was detected mostly in the liver. Consistent with the function of BAT related to sympathetic nerve activity, administering anaesthesia in mice almost completely eliminated the accumulation of ¹²³I-oxLDL in BAT, and this effect was reversed by administering β₃-agonist. Furthermore, exposing mice to cold stress at 4°C enhanced ¹²³I-oxLDL accumulation in BAT. Because in ¹²³I-oxLDL, the protein of oxLDL was labelled, we performed additional experiments with DiI-oxLDL in which the lipid phase of oxLDL was fluorescently labelled and observed similar results, suggesting that the whole oxLDL particle was taken up by BAT. To identify the receptor responsible for oxLDL uptake in BAT, we analysed the expression of known oxLDL receptors (e.g. SR-A, CD36, and LOX-1) in cultured brown adipocyte cell line and primary brown adipocytes and found that CD36 was the major receptor expressed. Treatment of cells with CD36 siRNA or CD36 neutralizing antibody significantly inhibited DiI-oxLDL uptake. Finally, CD36 deletion in mice abolished the accumulation of ¹²³I-oxLDL and DiI-oxLDL in BAT, indicating that CD36 is the major receptor for oxLDL in BAT.

Conclusion

We show novel evidence for the CD36-mediated accumulation of oxLDL in BAT, suggesting that BAT may exert its anti-atherogenic effects by removing atherogenic LDL from the circulation.

Graphical Abstract

Visualization of brown adipose tissue (BAT) with nanoSPECT after the intravenous injection of a ¹²³I-labelled oxidized low-density lipoprotein (oxLDL) probe (¹²³I-oxLDL). Intravenously injected ¹²³I-oxLDL accumulated in BAT at a high density similar to that in the liver, whereas the control ¹²³I-labelled LDL probe accumulated mostly in the liver. Stimulation of BAT by β-3 agonist and cold stress enhanced ¹²³I-oxLDL accumulation in BAT. In contrast, anaesthesia suppressed the accumulation of ¹²³I-oxLDL in BAT. Deleting CD36 in mice completely abolished the accumulation of ¹²³I-oxLDL in BAT, indicating that CD36 is the receptor that mediates oxLDL uptake in BAT. These data support a possible role of BAT in protecting against atherosclerosis.

Oxidized LDL, CD36, Brown adipose tissue, Imaging, NanoSPECT/CT

1. Introduction

Elevated low-density lipoprotein (LDL) levels are a major cause of atherogenesis and is known to gain proatherogenic properties by modification such as oxidation.^1,2 The proatherogenic effects of modified LDL are mediated by LDL receptors including scavenger receptor A (SR-A), CD36, and lectin-like oxidized LDL receptor-1 (LOX-1).^3–5 For example, foam cell formation is mediated by SR-A and CD36, and the induction of endothelial dysfunction is mediated by LOX-1.^6–10

Although the modification of LDL is believed to occur in the vascular wall, a detectable concentration of modified LDL is present in the circulation. Levels of oxidized LDL (oxLDL) measured by using anti-oxidized phosphatidylcholine antibody are higher in patients with vulnerable plaque than in patients with stable plaque.¹¹ In addition, circulating oxLDL concentration is significantly correlated with the onset of metabolic syndrome.¹² Regardless of hypercholesterolemia, oxLDL concentration in the blood is high during the acute phase of coronary syndrome and acute myocardial infarction.¹³ LDL can be modified in various ways. LDL is composed of phospholipids, cholesterol, triglycerides, and proteins with a carbohydrate chain; each of these components can be chemically modified via oxidation, hyperchlorination, carbamylation, glycation, or desialylation.^5,14–18 Therefore, oxidized and modified LDL is actually a mixture of heterogeneous lipoprotein particles that have undergone various modifications. Furthermore, dietary lipoproteins (e.g. chylomicrons or remnants) and the most negatively charged LDL subfraction (i.e. L5) are known to have properties similar to those of oxLDL.^19,20 Comprehensively detecting these various types of modified LDL has been shown to be useful for predicting cardiovascular risk.^15,20–25

Individuals with high levels of modified LDL are at increased risk of cardiovascular disease, even after adjusting for various confounding factors.²¹ Furthermore, the progression of atherosclerosis was suppressed in mice with decreased levels of circulating modified LDL that artificially overexpressed LOX-1 in the liver to increase the uptake of modified LDL.²⁶ This suggested that the concentration of circulating modified LDL affects atherogenesis.

Given the diverse types of LDL modification, more than one type of modified LDL is likely to be important for the progression of arteriosclerosis. However, oxLDL that is oxidized in vitro has historically provided crucial information for understanding the pathophysiology of atherosclerosis and continues to be useful. OxLDL labelled with radioisotope used for in vivo organ distribution studies showed that, similar to LDL, oxLDL was cleared from the circulation by the liver, but much more rapidly than LDL, with a half-life of several minutes.²⁷ In the present study, we analysed oxLDL distribution in tissues after the injection of ¹²³I-labelled oxLDL by using a high-resolution nanoSPECT/CT imaging system. Surprisingly, we found that oxLDL accumulates in BAT at a density equal to or higher than that of the liver. Thus, we further explored the mechanism of oxLDL accumulation in BAT.

2. Methods

2.1 Animals

All procedures conformed to the guidelines from the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Wild-type (WT) C57BL/6J mice and CD36 knockout (KO) mice (details in the supplementary material online) used in this study were treated in accordance with the Guidelines for the Care and Use of Laboratory Animals at Shinshu University, Hokkaido University, and the National Cerebral and Cardiovascular Center (NCVC). This study was approved by the Institutional Animal Care and Use Committee (approval number 020015), the Committee for Safe Handling of Living Modified Organisms (approval number 20-005) of Shinshu University, and the Animal Care and Use Committee (approval number 18-0015) of Hokkaido University, whose laboratory animal facilities comply with the Fundamental Guidelines for Proper Conduct of Animal Experiments and Related Activities in Academic Research Institutions under the jurisdiction of the Ministry of Education, Culture, Sports, Science, and Technology. This study was also approved by the Institutional Animal Care and Use Committee (approval numbers 12009, 13036) and by the Committee for Safe Handling of Living Modified Organisms (approval numbers 12–44, 13–46) of the NCVC, whose laboratory animal facilities comply with the Basic Policies for the Conduct of Animal Experimentation in the Ministry of Health, Labour, and Welfare according to an assessment by the Center for Accreditation of Laboratory Animal Care and Use, Japan Health Sciences Foundation. All experiments were performed in accordance with the approved guidelines. Mice were anaesthetized with inhalation of 3% isoflurane. Mice were euthanized by inhalation of isoflurane at a saturated vapour pressure of approximately 32% in a closed container.

2.2 Preparation of ¹²³I-oxLDL and ¹²³I-LDL

¹²³I-oxLDL and ¹²³I-LDL were prepared as described in the supplementary material online. Native albumin-free LDL was isolated by performing the sequential ultracentrifugation of plasma obtained from healthy volunteers. Written informed consent was obtained from all volunteers, and the study protocol was approved by the Institutional Committee on Human Research of Shinshu University (Permission number 3566, 5093). This study conformed to the ethical guidelines of the Declaration of Helsinki.

2.3 NanoSPECT/CT whole-body imaging

NanoSPECT/CT was used to analyse the tissue distribution of ¹²³I-labelled LDL and oxLDL in mice as described in the supplementary material online. SPECT images were displayed with the standardized uptake value (SUV) scale. The SUV in SPECT imaging is a value of tissue radioactivity per tissue weight divided by the total injected radioactivity per mouse body weight. In other words, the SUV is an index that represents the ratio of tissue radioactivity to systemic radioactivity, assuming that the injected ¹²³I-oxLDL is uniformly distributed throughout the body.

2.4 Evaluation of the dose uptake ratio of ¹²³I-oxLDL and ¹²³I-LDL in mouse tissues

The tissue accumulation of ¹²³I-oxLDL and ¹²³I-LDL in mice was determined by calculating the DUR as described in the supplementary material online. The DUR represents the uptake of radioactivity in tissue per tissue weight corrected by the total radioactivity dose per mouse body weight. The effects of anaesthesia and cold stimulus on ¹²³I-oxLDL uptake in tissues of conscious mice were also examined.

2.5 Tissue distribution of DiI-oxLDL

The preparation of fluorescently labelled DiI-oxLDL and DiI-LDL and the experimental procedures for examining their tissue distribution are described in the supplementary material online.

2.6 Immunofluorescence staining

Immunofluorescence was examined in specimens stained with anti-CD36 and anti-UCP-1 as described in the supplementary material online.

2.7 Cell lines

COS7 cells were maintained in culture, and HB2 brown predipocyte cells²⁸ were maintained and induced to differentiate as described in the supplementary material online.

2.8 Isolation and differentiation of primary brown preadipocytes from mice

Primary brown and white preadipocytes were isolated from male C57BL/6J mice, maintained in culture, and induced to differentiate as described in the supplementary material online.

2.9 Oil Red O staining

HB2 cells and primary brown preadipocytes were induced to differentiate in 24-well plates and used for Oil Red O staining as described in the supplementary material online.

2.10 mRNA expression analysis using quantitative real-time PCR (q-RTPCR)

mRNA expression in HB2 cells and primary brown adipocytes was analysed by performing q-RTPCR as described in supplementary material online.

2.11 Western blot analysis

CD36 and β-actin protein were detected in HB2 cells and primary brown adipocytes as described in the supplementary material online.

2.12 Preparation of mouse CD36 expression vector

The preparation of the mouse CD36 cDNA expression vector is described in the supplementary material online.

2.13 Detection of Dii-oxLDL uptake in CD36-expressing COS7 cells

DiI-oxLDL uptake and CD36 staining in CD36-expressing COS7 cells were analysed as described in the supplementary material online.

2.14 siRNA-mediated knockdown of CD36 in primary brown adipocytes

HB2 cells and primary brown adipocytes were treated with two CD36 siRNAs and analysed for CD36 staining and DiI-oxLDL uptake as described in the supplemental material online.

2.15 Treatment of primary brown adipocytes with anti-CD36 neutralizing antibody

CD36-expressing COS7 cells, HB2 cells, and primary brown adipocytes were treated with anti-CD36 IgA neutralizing antibody and analysed for DiI-oxLDL uptake as described in the supplementary material online.

2.16 Statistical analysis

All data were presented as the mean ± SEM. Statistical differences were analysed by using the unpaired t-test for paired data or one-way analysis of variance followed by Tukey’s post-test for multiple comparisons (GraphPad, San Diego, CA, USA, Prism 8). A P value < 0.05 was considered statistically significant.

3. Results

3.1 OxLDL accumulates not only in the liver but also in BAT

In unanaesthetized mice, nanoSPECT/CT was used to analyse the tissue distribution of ¹²³I-labelled LDL and oxLDL 10 min after injection via the tail vein. SUV imaging showed that ¹²³I-LDL accumulated in the liver, as expected, while revealing that ¹²³I-oxLDL accumulated in BAT and the liver at similar levels (Figure 1A). Because it was not previously known that oxLDL accumulates in BAT, we calculated the tissue accumulation of the trace radioactivity as the DUR after organ removal. The DUR comparison among BAT, white adipose tissue (WAT), and liver tissues showed that, similar to our nanoSPECT/CT data, radioactive ¹²³I-LDL was predominantly distributed in the liver, and BAT and WAT contained less than one-tenth of ¹²³I-LDL distributed in the liver (liver: 10.63 ± 0.17*, BAT: 0.91 ± 0.11, WAT: 0.35 ± 0.02, *P < 0.0001 vs. BAT or WAT) (Figure 1B). In contrast, radioactive ¹²³I-oxLDL was almost equal between BAT and the liver, with no statistically significant difference (BAT: 5.66 ± 0.49, WAT: 0.30 ± 0.03, liver: 5.99 ± 0.23).

Figure 1

The accumulation of oxLDL in the brown adipose tissue (BAT) of conscious mice. (A) Representative nanoSPECT/CT images showing the whole-body distribution of ¹²³I-LDL and ¹²³I-oxLDL in conscious C57BL/6J mice. (B) The accumulation of radioactivity is shown based on the dose uptake ratio (DUR) of ¹²³I-LDL (n = 7) and ¹²³I-oxLDL (n = 8) in BAT, white adipose tissue (WAT), and liver. (C) Representative fluorescence microscopy images showing the tissue accumulation of DiI-LDL (n = 4) and DiI-oxLDL (n = 5) in conscious C57BL/6J mice. (D) The fluorescence intensity of DiI-LDL (n = 4) and DiI-oxLDL (n = 5) is shown for BAT, WAT, and liver. Scale bar, 20 µm. *P < 0.05, ****P < 0.0001 (one-way analysis of variance). Bars represent the mean ± SEM.

The lipoprotein tracers used in these experiments were labelled with ¹²³I at the aromatic amino acid residues of the protein. Because lipoproteins are lipid-protein complexes, we performed similar experiments using lipoprotein tracers labelled with DiI at the lipid moiety. LDL and oxLDL fluorescently labelled at the lipid portion with DiI-C18 (DiI-LDL and DiI-oxLDL) were injected intravenously into mice tails. Ten minutes later, mice were euthanized, and tissues were removed and examined by using fluorescence microscopy. In mice injected with DiI-LDL, DiI fluorescence was observed in the liver but not in BAT or WAT (Figure 1C). In mice injected with DiI-oxLDL, however, DiI fluorescence was observed in both the liver and BAT, as in the case of ¹²³I-oxLDL. Fluorescence intensity quantification in tissue sections revealed that, in the case of DiI-LDL, DiI fluorescence detected in BAT and WAT was less than one-thirtieth of that in the liver (liver: 2.70 ± 0.24 × 10⁷*, BAT: 0.07 ± 0.02 × 10⁷, WAT: 0.04 ± 0.03 × 10⁷, *P < 0.0001 vs. BAT or WAT) (Figure 1D). In the case of DiI-oxLDL, DiI fluorescence was almost equal between the liver and BAT, with no statistically significant difference (liver: 2.41 ± 0.17 × 10⁷, BAT: 2.26 ± 0.16 × 10⁷, WAT: 0.47 ± 0.03 × 10⁷). This is the first time to our knowledge that oxLDL, but not LDL, has been shown to accumulate in BAT at the same density as in the liver.

3.2 OxLDL accumulation in BAT is enhanced by stimulation of BAT

To investigate what factors affect the accumulation of oxLDL in BAT, the DUR of removed organs was measured after mice were administered ¹²³I-oxLDL under experimental conditions. Because BAT function is activated by β₃-stimulation of the sympathetic nervous system, isoflurane anaesthesia completely abolished the accumulation of ¹²³I-oxLDL in BAT (awake: 9.05 ± 1.21; isoflurane: 0.58 ± 0.18, P < 0.0001) (Figure 2A, left). The inhibitory effect of the anaesthesia was significantly restored by administering β₃-agonist (isoflurane + β₃-agonist: 4.12 ± 1.04, P < 0.05 vs. awake) (Figure 2A, left). In contrast, the anaesthesia had no effect on oxLDL accumulation in WAT or liver (Figure 2A, middle and right). We next analysed the effect of cold stimuli on oxLDL accumulation in BAT. The accumulation of ¹²³I-oxLDL in BAT was more than two-fold higher when mice were kept at 4°C overnight than when kept at 25°C. (25°C: 7.16 ± 1.44, 4°C: 14.87 ± 0.99, P < 0.01) (Figure 2B left). In this case, no change in ¹²³I-oxLDL accumulation was observed in WAT; however, we observed a mild but significant decrease in ¹²³I-oxLDL accumulation in the liver of mice kept at 4°C compared with 25°C (25°C: 7.27 ± 0.96, 4°C: 4.83 ± 0.14, P < 0.05) (Figure 2B, middle and right). This result may be due to the increased accumulation of ¹²³I-oxLDL in BAT decreasing the accumulation of ¹²³I-oxLDL in the liver. These findings suggest that oxLDL uptake in BAT is important for the state of BAT activation.

Figure 2

Factors affecting oxLDL accumulation in brown adipose tissue (BAT). (A–C) OxLDL accumulation in BAT was affected by anaesthesia, β₃ stimulation, and cold stimulation and was suppressed by pretreatment with L-NAME (n-nitro-L-arginine methylester). (A) The effects of anaesthesia and β₃-agonist on ¹²³I-oxLDL accumulation in BAT (left panel), white adipose tissue (WAT) (middle panel), and liver (right panel) in C57BL/6J mice. Data are shown for the control unanesthetized group (n = 4), anaesthesia group (n = 5), and anaesthesia + β₃-agonist group (n = 4). *P < 0.05, **P < 0.01, ****P < 0.0001 (one-way analysis of variance); ns, not significant. (B) The effect of 4°C cold stimulation on ¹²³I-oxLDL accumulation in BAT (left panel), WAT (middle panel), and liver (right panel). *P < 0.05, **P < 0.01 (unpaired t-test); ns, not significant.

3.3 CD36 is expressed as a receptor for oxLDL in HB2 cells

To elucidate the mechanisms involved in the accumulation of oxLDL in BAT, we performed additional studies in the HB2 brown adipocyte cell line. Induction of HB2 cell differentiation resulted in adipose-like cells with several fat droplets that stained positive for Oil Red O (Figure 3A). In addition, HB2 cell differentiation was accompanied by an induction in the gene expression of adipocyte markers Fabp4 and Pparg, as well as the brown adipocyte markers Ucp1 and Cidea (Figure 3B). Upon HB2 cell differentiation into brown adipocytes, DiI-oxLDL uptake was observed, which was completely absent in the undifferentiated state (Figure 3C). During the transition from undifferentiated to differentiated HB2 cells, we examined changes in the expression of genes encoding oxLDL uptake receptors CD36 (CD36), SR-B1 (Scarb1), SR-A1 (Msr1), and LOX-1 (Olr1). Compared with other receptors, CD36 mRNA was upregulated by nearly 40 000-fold (Figure 3D). Western blot analysis showed that CD36 protein levels were also significantly upregulated with differentiation, showing 60-fold greater levels in cells on differentiation Day 6 than in undifferentiated cells when quantified as a ratio to β-actin (Day 0: 0.014 ± 0.004; Day 6: 0.854 ± 0.163, P < 0.001) (Figure 3E). In differentiated HB2 cells, immunostaining showed the colocalization of DiI-oxLDL uptake and CD36 on the plasma membrane, observed by using confocal laser microscopy (Figure 3F).

Figure 3

CD36 expression and oxLDL uptake in differentiated brown adipocyte HB2 cells. (A) Oil Red O staining of undifferentiated and differentiated HB2 cells. Scale bar, 100 µm. (B) q-RTPCR results showing changes in mRNA expression of Fabp4, Pparg, Ucp1, and Cidea in undifferentiated and differentiated HB2 cells (n = 4 each). ***P < 0.001, ****P < 0.0001 (unpaired t-test). (C) Fluorescence microscopy images showing DiI-oxLDL accumulation in undifferentiated and differentiated HB2 cells. Scale bar, 100 µm. (D) q-RTPCR results showing changes in mRNA expression of 4 oxLDL receptors in undifferentiated and differentiated HB2 cells (n = 4). *P < 0.05, **P < 0.01, ****P < 0.0001 (unpaired t-test). (E) Western blot showing changes in protein levels of CD36 and β-actin in HB2 cells after differentiation (n = 3). ***P < 0.001 compared with Day 0 (one-way analysis of variance); ns, not significant. (F) Fluorescence microscopy images showing DiI-oxLDL and CD36 in differentiated HB2 cells. Scale bar, 20 µm. Blue indicates nuclei (DAPI), red indicates DiI-oxLDL, and green indicates CD36. Experiments were repeated three times for A and E, twice for B and D, and five times for C and F under equivalent conditions. Bars represent mean ± SEM.

3.4 CD36 is required for the uptake of oxLDL in differentiated HB2 cells

To clarify whether CD36 is responsible for oxLDL uptake in differentiated HB2 cells, we examined the effects of siRNA-based CD36 knockdown and anti-CD36 neutralizing antibody on DiI-oxLDL uptake in differentiated HB2 cells. In differentiated HB2 cells, treatment with two siRNAs against CD36 reduced CD36 protein levels by half or less compared with control scrambled RNA (control: 3.55 ± 0.22; CD36 siRNA #1: 1.56 ± 0.53*; #2: 1.53 ± 0.11*; *P < 0.001 vs. control) (Figure 4A). CD36 knockdown by siRNAs also significantly suppressed DiI-oxLDL uptake in differentiated HB2 cells (control: 2.54 ± 0.12 × 10⁶; CD36 siRNA #1: 0.16 ± 0.07 × 10⁶*, #2: 0.15 ± 0.01 × 10⁶*; *P < 0.001 vs. control) (Figure 4B).

Figure 4

CD36-mediated uptake of oxLDL in differentiated brown adipocyte HB2 cells. (A) Western blot analysis showing the effect of two different siRNAs on CD36 knockdown in differentiated HB2 cells (n = 4 each). ***P < 0.001 (one-way analysis of variance). (B) Representative fluorescence microscopy images and quantification showing DiI-oxLDL uptake in differentiated HB2 cells with siRNA-based CD36 knockdown (n = 4 each). Scale bar, 20 µm. ***P < 0.001 (one-way analysis of variance). (C) Representative fluorescence microscopy images and quantification showing changes in DiI-oxLDL uptake induced by anti-CD36 neutralizing antibody in CD36-expressing COS7 cells (n = 3 each). **P < 0.01, ***P < 0.001 compared with control IgA antibody (unpaired t-test). (D) Representative fluorescence microscopy images and quantification showing changes in DiI-oxLDL uptake induced by anti-CD36 neutralizing antibody in differentiated HB2 cells (n = 3 each). Scale bar, 100 µm. Blue indicates nuclei (DAPI), red indicates DiI-oxLDL, and green indicates CD36. ***P < 0.001 compared with control IgA antibody (unpaired t-test). Experiments were repeated three times for A and C, four times for B, and twice for D under equivalent conditions. Bars represent the mean ± SEM.

Next, in CD36-expressing COS-7 cells, treatment with an anti-CD36 neutralizing antibody inhibited DiI-oxLDL uptake in a concentration-dependent manner (Figure 4C). In the same way, anti-CD36 neutralizing antibody inhibited DiI-oxLDL uptake in differentiated HB2 cells in a concentration-dependent manner (Figure 4D). These results indicate that CD36 is required for oxLDL uptake in HB2 cells.

3.5 CD36 is required for oxLDL uptake in primary cultured brown adipocytes

We next examined CD36 expression and oxLDL uptake in primary cultured brown adipocytes. After differentiation was induced in primary cultured brown adipocytes from mice, staining with Oil Red O showed the formation of fat droplets (Figure 5A). In addition, q-RTPCR analysis of differentiated cultured brown adipocytes revealed the markedly increased expression of adipocyte marker genes (Fabp4, Pparg) and brown adipocyte marker genes (Ucp1, Cidea) (Figure 5B). Differentiated primary brown adipocytes also showed the colocalization of DiI-oxLDL and CD36 on confocal laser microscopy (Figure 5C). The siRNA-mediated knockdown of CD36 with two types of siRNA significantly reduced CD36 protein levels to less than one-third of levels in differentiated brown adipocytes treated with scrambled siRNA(control: 3.31 ± 0.61; CD36 siRNA #1: 0.99 ± 0.38*, #2: 0.97 ± 0.12*; *P < 0.05 vs. control) (Figure 5D). In a similar manner, DiI-oxLDL uptake was also significantly reduced in cells treated with CD36 siRNA compared with cells treated with scrambled siRNA (control: 1.15 ± 0.17 × 10⁶; CD36 siRNA #1: 0.35 ± 0.02 × 10⁶*, #2: 0.28 ± 0.02 × 10⁶*; *P < 0.05 vs. control) (Figure 5E). Additionally, the treatment of differentiated brown adipocytes with anti-CD36 neutralizing antibody inhibited DiI-oxLDL uptake in a concentration-dependent manner (Figure 5F). Thus, our findings indicate that oxLDL uptake is mainly mediated by CD36 in primary brown adipocytes.

Figure 5

OxLDL uptake mediated by CD36 in primary brown adipocytes. (A) Oil Red O staining of undifferentiated and differentiated primary brown adipocytes. Scale bar, 100 µm. (B) q-RTPCR results showing changes in the mRNA expression of Fabp4, Pparg, Ucp1, and Cidea in undifferentiated and differentiated primary brown adipocytes (n = 3 each). *P < 0.05, **P < 0.01, ***P < 0 .001 (unpaired t-test). (C) Fluorescence microscopy analysis of differentiated primary brown adipocytes. Blue indicates nuclei (DAPI), red indicates DiI-oxLDL, and green indicates CD36. Scale bar, 20 µm. (D) Western blot analysis showing the effect of two different siRNAs on CD36 knockdown in differentiated primary brown adipocytes (n = 3 each). *P < 0.05 (one-way analysis of variance). (E) Quantification of fluorescence intensity showing the effects of the siRNA-based knockdown of CD36 on DiI-oxLDL uptake in differentiated primary brown adipocytes (n = 3 each). *P < 0.05 (one-way analysis of variance). (F) Quantification of fluorescence intensity showing changes in DiI-oxLDL uptake induced by anti-CD36 neutralizing antibody in differentiated primary brown adipocytes (n = 3 each). *P < 0.05, **P < 0.01 compared with control antibody (unpaired t-test). Each experiment was repeated twice under equivalent conditions. Bars represent the mean ± SEM.

3.6 OxLDL accumulation in BAT is abolished in CD36 KO mice

We examined whether CD36 is required for oxLDL uptake in BAT in vivo by using CD36 knockout mice. NanoSPECT/CT imaging after the intravenous injection of ¹²³I-oxLDL showed that ¹²³I-oxLDL accumulated in BAT of WT mice but not in that of CD36 KO mice. However, ¹²³I-oxLDL accumulated in the liver of CD36 KO mice (Figure 6A). When tissues were removed and the radioactivity level was analysed after the intravenous injection of ¹²³I-oxLDL, the BAT in CD36 KO mice showed less than one-tenth of the radioactivity of BAT in WT mice (WT: 9.95 ± 1.75; CD36 KO: 0.64 ± 0.12; P < 0.0001) (Figure 6B, left). In addition, the accumulation of ¹²³I-oxLDL in the WAT of WT mice was small, but the WAT of CD36 KO mice showed even less (WT: 0.33 ± 0.04; CD36 KO: 0.10 ± 0.01; P < 0.0001) (Figure 6B, middle). On the other hand, ¹²³I-oxLDL accumulation in the liver was significantly higher in CD36 KO mice than in WT mice (WT: 5.88 ± 0.33; CD36 KO: 13.20 ± 0.38; P < 0.0001) (Figure 6B, right). CD36 expression and DiI-oxLDL accumulation were observed in the BAT of WT mice but not in that of CD36 KO mice (Figure 6C upper). Furthermore, by quantifying fluorescence intensity, we found that DiI-oxLDL uptake observed in the BAT of WT mice was abolished in the BAT of CD36 KO mice (WT: 2.24 ± 0.16 × 10⁷; CD36 KO: 0.05 ± 0.02 × 107; P < 0.0001) (Figure 6C, right).

Figure 6

Suppression of oxLDL accumulation in the brown adipose tissue (BAT) of CD36 KO mice. (A) Representative nanoSPECT/CT images showing the whole-body distribution of ¹²³I-oxLDL in conscious WT and CD36 KO mice. (B) The accumulation of radioactivity based on the dose uptake ratio (DUR) of ¹²³I-oxLDL in BAT, white adipose tissue (WAT), and liver in WT (n = 10) and CD36 KO mice (n = 9). ****P < 0.0001 (unpaired t-test). (C) In conscious WT (n = 8) and CD36 KO mice (n = 6), the representative accumulation of DiI-LDL fluorescence in tissue is shown for BAT (upper panel), WAT (middle panel), and liver (lower panel). The fluorescence intensity of DiI-oxLDL is also shown (right panel). Blue indicates nuclei (DAPI), red indicates DiI-oxLDL, and green indicates CD36. Scale bar, 20 µm. *P < 0.05, ****P < 0.0001 (unpaired t-test). Bars represent the mean ± SEM.

In WAT, CD36 expression and a low level of DiI-oxLDL uptake were detected in WT mice but not in CD36 KO mice. The quantification of fluorescence intensity showed a significant decrease in DiI-oxLDL uptake (P < 0.0001) in the WAT of CD36 KO mice (0.04 ± 0.02 × 10⁷) compared with the WAT of WT mice (0.38 ± 0.05 × 10⁷).

No CD36 expression was observed in the liver of either WT or CD36 KO mice. However, DiI-oxLDL uptake was observed in the liver of both types of mice (Figure 6C, bottom). The quantification of fluorescence intensity showed the increased accumulation of DiI-oxLDL in the liver of CD36 KO mice compared with that of WT mice (WT: 2.14 ± 0.20 × 10⁷; CD36 KO mice: 3.38 ± 0.30 × 10⁷; P < 0.05). These in vivo experiments corroborated that CD36 is involved in the uptake of oxLDL into BAT.

4. Discussion

4.1 A novel mechanism of modified LDL metabolism in vivo

To date, research has improved our understanding of the pathologic role of modified LDL receptors in vascular cells and macrophages that leads to atherosclerotic diseases. Chemical modifications such as oxidation allow LDL to bind to modified LDL-specific receptors, which in turn leads to proatherogenic effects. Ultimately, these effects result in the progression of atherosclerosis and the development of myocardial and cerebral infarction.

However, the biodistribution of circulating modified LDL has remained poorly understood. In the present study, we used high-resolution nanoSPECT/CT with ¹²³I-labelled LDL and oxLDL to reveal the biodistribution of intravenously injected modified LDL. We show that oxLDL is distributed in BAT and liver at similar densities, whereas native, unmodified LDL is distributed mostly in the liver but not in BAT. Thus, for the first time to our knowledge, we describe the accumulation of oxLDL in BAT. Furthermore, we show that this accumulation is facilitated by β₃-stimuli and cold stress but eliminated by anaesthesia, suggesting a relationship between oxLDL and the physiologic function of BAT.

Studies from the 80 and 90 s previously suggested the importance of the liver in oxLDL uptake.^29–31 However, the accumulation of oxLDL in BAT may have been overlooked because of the elimination of oxLDL in BAT by anaesthesia. Alternatively, BAT may not have been analysed in these studies.

The in vivo kinetics of intravenously injected oxLDL have been studied in knockout mice with the deletion of oxLDL receptor genes. According to reports, however, the deletion of the genes encoding SR-A or LOX-1 did not significantly change the half-life or kinetics of oxLDL injected into the circulation.^32,33 In the case of CD36, the half-life of injected oxLDL was slightly increased in CD36 KO mice compared with WT mice,³⁴ and it was postulated that CD36 may be involved in the uptake of oxLDL in the liver. However, considering the results of the present study, the increased half-life of oxLDL in CD36 KO mice may have been the result of decreased oxLDL uptake in extra-hepatic organs such as BAT. Furthermore, the large uptake capacity and compensatory uptake of oxLDL by the liver may make the degree of change minimal. Notably, in the present study, the distribution of oxLDL in the liver was actually greater in CD36 KO mice than in WT mice.

4.2 Relationship between BAT function and atherosclerosis

Fat accumulates in WAT, which causes obesity and diabetes, whereas BAT metabolizes fatty acids to produce heat and maintain body temperature. It is known that the dysfunction of BAT and beige adipocytes may cause obesity and metabolic syndrome in humans. Recently, the possibility has been raised that improving BAT function may suppress the progression of atherosclerosis.^35–38

BAT and beige adipocytes have been understood in relation to fatty acid metabolism. However, the findings of our present study suggest that BAT function may directly affect the metabolism of oxLDL or modified LDL and thus atherogenesis. CD36 is a major fatty acid transporter, along with fatty acid transport proteins and fatty acid binding proteins . In mice, CD36 is essential, especially in BAT, for maintaining body temperature under cold conditions.³⁹ Given the dual function of CD36 as a fatty acid transporter and modified LDL receptor, it is possible that CD36 is important as an oxLDL receptor in tissues in which it is physiologically essential as a fatty acid transporter.

Receptors that bind oxidatively modified LDL include SR-A, CD36, and LOX-1. In a brown adipocyte cell line and primary cells, we observed a remarkable increase in CD36 expression along with differentiation into brown adipocytes. Furthermore, CD36 was required for the uptake of oxLDL into cultured brown adipocytes. These findings support our in vivo observations of oxLDL distribution in BAT.

In addition to the uptake of oxLDL into BAT via CD36, as shown in the present study, BAT secretes adiponectin, which is a so-called ‘good’ adipocytokine. Although adiponectin has a broad spectrum of actions, we previously found that adiponectin directly binds to oxLDL to inhibit its receptor binding.⁴⁰ In that study, adiponectin inhibited the binding of oxLDL to SR-A and LOX-1 but not to CD36. Thus, BAT may exert anti-atherogenic effects by clearing oxLDL from the circulation via CD36 while simultaneously blocking oxLDL’s effects via SR-A and LOX-1.

4.3 Significance of BAT imaging

[¹⁸F]FDG positron emission tomography/computed tomography (PET/CT) has been commonly used for the detection of BAT. However, given the potential clinical benefits of BAT on obesity, type II diabetes mellitus, and cardiovascular diseases, attempts have been made to improve the methodology of BAT imaging. For example, the combined method of fluorescence optical imaging and PET has been described, in which an improved fluorescent probe with a glucose backbone is used together with RediJect 2-DG and [¹⁸F]FDG.⁴¹

In addition, in light of the fact that BAT is a lipid-metabolizing organ, [¹⁸F]BODIPY-C16/triglyceride and [¹⁸F]BODIPY-triglyceride-containing chylomicron-like particles have been used as imaging agents for BAT.^42,43 The latter method resembles those in the present study in that the use of lipoprotein-like particle is involved. Similar to our study, the authors reported that CD36 was involved in the accumulation of the lipoprotein-like tracer in BAT. However, it was postulated that the lipoprotein-like particle was metabolized with lipoprotein lipase to produce a fatty acid tracer, and then [¹⁸F]fatty acids were taken up by BAT via CD36.

In the present study, we used [¹²³I]oxLDL SPECT/CT to visualize BAT at a much higher resolution than can be achieved with [¹⁸F]FDG PET/CT. The changes in the accumulation of the tracer induced by cold stress, β₃-agonist, and anaesthesia suggest that the tracer accumulation reflects physiologic BAT functions. Furthermore, we detected the accumulation of both [¹²³I]-labelled (protein) and DiI-C18-labelled (lipids) oxLDL in BAT. Thus, our findings suggest that both the lipid and protein components of oxLDL accumulated in BAT, implying that the whole lipoprotein particle of oxLDL is taken up by BAT. In the future, it may be useful to perform multimode imaging of SPECT and fluorescence imaging by simultaneously using the [¹²³I]- and DiI-C18–labelled oxLDL. Interestingly, it was recently shown that CD36 internalizes entire chylomicron remnant particles into BAT, which was probably stimulated by fatty acids released by lipoprotein lipase.^39,44

4.4 Clinical significance

In humans, BAT present during the neonatal period is believed to regress by adulthood. However, studies in which [¹⁸F]FDG PET/CT was used have shown that BAT-like tissue is present in adult humans.^45–47 Therefore, it is conceivable that oxLDL accumulates in human BAT-like tissue.

Findings of recent gene-expression profile studies have indicated that human thermogenic adipocytes are mostly beige adipocytes that are transdifferentiated from white adipocytes, while some brown adipocytes are found in deep tissues.^48,49 Regardless of whether the thermogenic adipocytes in human adults are brown or beige, defining BAT as the tissue that accumulates [¹⁸F]FDG led to the discovery that people who had a larger amount of BAT also had a significantly lower prevalence of cardiometabolic diseases (e.g. type 2 diabetes, dyslipidemia, coronary artery disease, cerebrovascular disease, congestive heart failure, and hypertension).⁵⁰ Although the authors speculated that lower blood glucose and triglyceride levels and higher high-density lipoprotein concentrations in these individuals may have contributed to the observed improvement, the present study raises the possibility that the removal of atherogenic LDL by BAT may contribute to the lower incidence of atherogenic diseases in this population.

4.5 Study limitations

Our study has a few limitations. First, we assessed the in vivo distribution of intravenously injected exogenous LDL that was oxidized beforehand in vitro in the presence of cupric ion. Therefore, our study does not address the kinetics or distribution of endogenously modified LDL. Second, although we believe oxLDL uptake by BAT may decrease circulating oxLDL and result in anti-atherogenic effects, the possibility remains that oxLDL may induce BAT dysfunction that leads to dyslipidemia and the promotion of atherogenesis. Third, although this study focused on CD36 in the brown adipocytes of BAT, other types of cells, such as macrophages or endothelial cells, may also be involved in the uptake of oxLDL in BAT. Recently, it was reported that CD36 is expressed on endothelial cells in activated BAT and plays a role in the particle uptake of chylomicron remnants.⁵¹ Thus, the possibility remains that orchestration occurs between CD36 of endothelial cells and CD36 of brown adipocytes. Finally, because our study was performed in mice, it remains unknown whether our results are applicable to human brown and beige adipocytes. Further study is warranted to elucidate the relationship among the function of human thermogenic adipocytes, the in vivo distribution of modified atherogenic LDL, and the progression of atherosclerosis.

Author contributions

H.H., H.I., and T.S. conceived the idea of the study. H.H. and A.K. developed the statistical analysis plan and conducted statistical analyses. K.H., H.K., A.N., and A.K. contributed to the operation of experiments and data collection, and the interpretation of the results. Y.Y., M.S., and A.K. prepared DiI-labelled oxLDL and mouse CD36 expression vectors, performed mRNA expression experiments, and maintained CD36 KO mice. H.H., A.K., and T.S. drafted the original manuscript. S.Y., H.I., and T.S. supervised the conduct of this study. Y.O. provided HB2 brown preadipocytes and isolated brown preadipocytes. D.M. and S.Y. provided CD36 KO mice. All authors reviewed the manuscript draft and revised it critically with respect to intellectual content. All authors approved the final version of the manuscript to be published.

Supplementary material

Supplementary material is available at Cardiovascular Research online.

Acknowledgements

We thank Nicole Stancel, PhD, ELS (D), of Scientific Publications at Texas Heart Institute, for providing editorial support, and Tomoyuki Nakajima, PhD, of the Department of Laboratory Medicine at Shinshu University Hospital, for supporting the preparation of sections for histological analyses.

Funding

This work was supported by the Japan Society for the Promotion of Science (grant no. 15K01309 to A.N.; no. 18H02578 to T.S.; no. 16K19192 to A.K.), and the Aiba Works Medical Research Grant (grant no. GOEE1070 to A.K.).

Data availability

The data that support the findings of this study are available from the corresponding author T.S., upon reasonable request.

Translational perspective

The function of thermogenic adipocytes, including brown and beige adipocytes, is believed to be crucial in preventing atherogenesis and resultant cardiovascular disease. We demonstrated for the first time to our knowledge the eminent uptake of atherogenic ox-LDL in the brown adipose tissue (BAT) of mice and provide evidence of the mechanism underlying the direct link between BAT insufficiency and atherosclerosis progression. Furthermore, our use of the novel technique nanoSPECT/CT to detect ox-LDL in BAT opens a new avenue for imaging and evaluating BAT that may be useful in evaluating cardiovascular disease risk in the clinical setting.

Author notes

These authors were equally contributed to the work.

Conflict of interest: H.H. received research grants from Tosoh Bioscience, Inc. T.S. is an inventor of a patents and received consulting fee from PreMedica Inc.