Insights into Brown Adipose Tissue Physiology as Revealed by Imaging Studies

Abstract

There has been resurgence in interest in brown adipose tissue (BAT) following radiological and histological identification of metabolically active BAT in adult humans. Imaging enables BAT to be studied non-invasively and therefore imaging studies have contributed a significant amount to what is known about BAT function in humans. In this review the current knowledge (derived from imaging studies) about the prevalence, function, activity and regulation of BAT in humans (as well as relevant rodent studies), will be summarized.

Introduction

Recent publications have unequivocally demonstrated the presence of thermogenically active brown adipose tissue (BAT) in adult humans and have led to renewed interest in the study of this type of adipose tissue. When activated, brown adipocytes release energy in the form of heat by uncoupling the protons generated by substrate oxidation from adenosine triphosphate (ATP) production. BAT cells express a special protein called UCP1 (uncoupling protein1/thermogenin) which enables them to do this. Since activated BAT increases energy expenditure, it may play an important role in energy homeostasis and thus could be utilised in the treatment of obesity. Many techniques have been employed to study this unique tissue and imaging techniques in particular have enabled in vivo studies to be performed. This review will highlight the main imaging modalities that have been used to study BAT and summarise how each of these modalities has contributed to our knowledge of the characteristics and function of BAT in humans.

Positron emission tomography - computed tomography (PET-CT)

¹⁸F-FDG ([18F]-fluorodeoxyglucose)

PET-CT is the most widely used imaging modality currently used to study BAT. It consists of a functional scan in which metabolically or biochemically active tissues are detected (i.e. the PET scan) and an anatomic scan (i.e., CT scan) performed at the same time. Following acquisition and processing of the images from both scans, they can be viewed individually or superimposed on each other to produce a single fused (or co-registered) image. [18F]-fluorodeoxyglucose (¹⁸F-FDG) is a tracer that is used to detect highly metabolically active tissue(s). ¹⁸F-FDG enters the metabolically active cells via specific glucose transporters and is then phosphorylated by hexokinase to its 6-phosphate. The 6-phosphate cannot be metabolised any further and therefore it is effectively trapped within the cell. The radioactive fluorine component of the tracer decays, and the products of its decay are detected by the PET scanner. The metabolically active tissues that have taken up the tracer can then be identified.¹

PET-CT was initially used in clinical practice for identifying and staging malignant tumors. However, on these PET scans bilateral symmetric uptake was often noted in the neck and shoulder regions. Initially, this was thought to be due to active muscle, but CT scans of the same regions demonstrated that the tissues with this symmetrical uptake had the density of adipose tissue not muscle. These areas were called "USA-fat" (uptake of ¹⁸F-FDG localizing to the supraclavicular area)¹ and some authors felt that this represented BAT^2,3 especially as the prevalence of “USA-fat” was found to be 3 times higher in winter (when outdoor temperatures were low) than the rest of the year.⁴

In 2009, Virtanen et al. demonstrated that the cold-induced increased ¹⁸F-FDG uptake seen on PET scans was due to paracervical and supraclavicular adipose tissue.⁵ These tissues were biopsied and found to have the cellular morphology of BAT and expressed UCP1 protein and mRNA.⁵ This study proved that not only is BAT present in adult humans, it is metabolically active and can be stimulated by cold.⁵ Two retrospective studies, which examined different series of over 3600 consecutive PET-CT scans, found a prevalence of active BAT of ∼3% in men and 7.2–7.5% in women.^6,7 A more recent and much larger retrospective study found a prevalence of active BAT of 1.32% in 31,088 PET-CT scans performed for medical check-ups (n = 16,699) and cancer surveillance (n = 14,389).⁸ Smaller cohort studies have reported a higher prevalence of cold-activated BAT (confirmed histologically) of 34% (19/56) in healthy volunteers aged 23–65 y in winter,⁹ 48% (125/260) in healthy volunteers aged 20–72 y in winter,¹⁰ 96% (23/24) of healthy men aged 20–32 y with BMIs (body mass indexes) ranging from 21.3–38.8,¹¹ and 20% (3/15) in morbidly obese volunteers.¹²

In these studies (which were performed in temperate regions) the prevalence and/or activity of detectable BAT was higher in winter/lower outdoor temperatures, younger subjects, females, in people with lower BMIs, lower blood glucose levels, lower percentages of body fat and lower quantities of visceral fat. In a study performed in a subtropical area, an inverse association between BAT ¹⁸F-FDG uptake and outdoor temperature was also observed.¹³ Furthermore the mass of BAT detected on ¹⁸F-FDG PET-CT scans decreases with increasing outdoor temperature, age and BMI, as well as being lower in men and in people with diabetes.¹⁴ Although lower BMIs are associated with a higher prevalence of active BAT, none of the 14 people with anorexia nervosa (BMI 15.3±0.8 in those with ongoing reduced calorie intake and BMI 18.8±1.1 in those who had been re-fed) imaged with ¹⁸F-FDG PET-CT in winter had increased uptake in BAT, but all of the 7 constitutionally lean people (BMI 16.2±0.9) had increased uptake in BAT.¹⁵ Therefore a low BMI does not necessarily predict the presence of active BAT, and chronic starvation in anorexia may reduce BAT activity.

Some authors have investigated whether ethnicity influences BAT activity. A retrospective study of 386 scans did not reveal any differences in BAT uptake between Caucasians and Black Africans.¹³ Cold-stimulated BAT activity was found to be similar in Caucasians (n = 10) and South Asians (n = 10).¹⁶ In another study, although cold-induced BAT activity was similar between age- and BMI-matched male Caucasians (n = 11) and South Asians (n = 12), the volume of activated BAT was significantly lower in South Asians.¹⁷ Additionally, cold-induced/non-shivering thermogenesis increased in Caucasians by 20% but did not increase in South Asians exposed to the same conditions.¹⁷ Larger studies are required to determine whether ethnic differences in BAT activity exist, and whether there are any correlations between BAT activity/thermogenesis and metabolic phenotypes between different ethnicities.

In children, the reported prevalence of active BAT in one retrospective review of 385 PET-CT scans performed on oncology patients aged 5-21 y was similar for boys (43.3%) and girls (45.3%), with peak activity seen in the 13–15 age group.¹⁸ Similarly to adults, BAT activity was inversely correlated with BMI, but in contrast to adults there was no correlation between BAT activity and outdoor temperature.¹⁸ However, in another study of scans performed during winter on people aged less than 21 years, ¹⁸F-FDG uptake was reduced from 31% to 5% by maintaining the room temperature where the subjects were placed at 24°C from 30 mins pre-tracer injection to 1-hour post-injection.¹⁹

The absence of increased ¹⁸F-FDG uptake may not necessarily mean that BAT is absent. In a prospective study of 17 people with preoperative PET-CT staging scans who had head and neck surgery, 3/17 (17.6%) had BAT detected by ¹⁸F-FDG with histological confirmation of BAT in regions corresponding to increased ¹⁸F-FDG uptake. However, histological analysis of the supraclavicular fat in the other 14 subjects who did not have increased ¹⁸F-FDG uptake on PET-CT revealed a mixture of adipocytes indistinguishable from white/subcutaneous fat and islands of adipocytes which contained multilobulated lipid droplets and expressed UCP1 (i.e. brown fat cells).²⁰ It is feasible that in some subjects, islands of brown fat exist that are metabolically active but are below the spatial resolution threshold required for detection by the scanner.

Therefore the prevalence of active BAT as detected by ¹⁸F-FDG PET-CT ranges from 3 to 100% (Table 1) depending on the cohort studied, season, outdoor/indoor temperature and the nature of the study (retrospective review of scans vs. interventional study), and BAT may be present even though it is not detected by ¹⁸F-FDG PET-CT. Additionally, there may be a diurnal or meal-related variation in ¹⁸F-FDG uptake by BAT since in mice, ¹⁸F-FDG uptake by BAT was found to peak 9 hours into the light phase (i.e., when mice are less active) of a 12-hour light/12-hour dark day.³¹

Table 1.

¹⁸F-FDG PET-CT Studies in Humans using Cold Stimulation of Brown Adipose Tissue

Study	Number of subjects	Subject characteristics (mean ± s.d.)	Cold Exposure	Prevalence of active BAT post-cold stimulation	Comments
van Marken Lichtenbelt et al. 200911	24♂	10 lean: -Age: 24.3 ± 3.6 y -BMI:23.2 ± 1.2 14 overweight/obese: -Age: 23.5 ± 3.4 y -BMI:30.3 ± 4.2	Subjects kept in room temperature of 16°C for 2 hours prior to scan	96% (23/24)	Mean BAT activity was significantly lower in the group with BMI >25. Negative correlations between BAT activity and BMI/percentage body fat. Positive correlation between BAT activity and resting metabolic rate.
Orava et al. 201121	7♂, 20♀	Age: 40.2 ± 9.4 y BMI: 22.8 ± 2.2	Subjects kept in room temperature of 17°C ± 1°C for 2 hours prior to scan, then moved into PET-CT room (23°C). During scan one foot placed intermittently (5 mins in, 5 mins out) in cold water (8°C ± 1°C).	70% (19/27)	Cold stimulation increased BAT glucose uptake fold12- and doubled BAT perfusion. Insulin stimulation increased BAT glucose uptake fold5- but did not increase BAT perfusion. Positive correlation between energy expenditure (EE) and BAT perfusion. GLUT4 expression higher in BAT vs. WAT.
Vijgen et al. 201112	2♂, 13♀	Age: 39.2 ± 8.1 y BMI: 42.1 ± 3.8	Room temperature and cooling mattress temperature maintained 1oC above temperature that induced shivering for 2 hours prior to scan.	20% (3/15, 0♂ and 3♀)	Non-significant increase in post-cold exposure EE from 41.9 ± 3.3 J/s to 43.7 ± 4.8 J/s (P= 0.100).
Yoneshiro et al. 201122	13♂	BAT +ve: -Age: 22.7 ± 3.0 y -BMI: 20.0± 2.0 BAT –ve:
-Age: 22.9 ± 4.6 y -BMI: 21.4 ± 1.7	Subjects kept in room temperature of 19°C for 2 hours prior to scan. During scan one foot placed intermittently (for 4 mins every 5 mins) on ice-cooled footrest.	46% (6/13)	Post-cold exposure increase in EE was significant in BAT +ve group (410 ± 293 kcal/day, P<0.05) and non-significant in BAT –ve group (42 ± 114 kcal/day, p=0 .37). Positive correlation between cold-induced rise in EE and BAT activity.
Cypess et al. 201223	4♂, 6♀	Age (mean ± s.e.m.): 27.1 ± 1.7 BMI (mean ± s.e.m.): 23.7 ± 1.4	Subjects kept in room temperature of 20°C for 2 hours prior to scan while wearing a cooling vest with circulating water at 14°C.	100% (10/10)	No BAT activation seen 120min post-ephedrine administration at room temperature of 23oC. EE increased by 79 kcal/day post-cold (p=0 .033) and by 136 kcal/day post-ephedrine administration (p=0 .01).
Ouellet et al. 201224	6♂	Age (range): 23 – 42 y BMI (range): 23.7 - 31	Subjects wore a liquid conditioned tube suit with circulating water at 18°C for 180 mins.	Cold induced BAT oxidative metabolism was seen in all subjects.	EE increased from 2592 ± 115.2 kcal/day to 4593 ± 446.4 kcal/day (P<0.05) post-cold exposure.
Admiraal et al. 201316	20♂ 10 Cauca-sians 10 South Asians	Caucasians - Median Age (IQR) 22.4 y (21.2–25.1)
-Median BMI (IQR) 22.6 (21.2–23.1) South Asians - Median Age (IQR) 23.2 y (21.1–26.1) -Median BMI (IQR) 22.3 (21.0–23.5)	Subjects kept in room temperature of 17°C for 2 hours prior to scan.	80% (8/10) Caucasians 80% (8/10) South Asians	No significant difference in BAT activity or BAT volume between the 2 ethnic groups.
Chen et al. 201325	14♂, 10♀	Age: 28.1 ± 7.3 years
BMI (range): 20–27	Subjects kept in room temperature of 19°C for 12 hours prior to scan.	50% (7/14, 4♂ and 3♀)	Cold exposure increased EE from 1694.4 ± 296.16 kcal/day to 1782.48 ± 277.92 kcal/day (P<0.001)
Matsushita et al. 201310	184♂, 76♀
	Median Age (IQR) 26 y (22–39) Median BMI (IQR) 21.6 (20.1–23.5)	Subjects kept in room temperature of 19°C for 2 hours prior to scan and placed one foot intermittently (for 4 mins every 5 mins) on a cloth-wrapped iceblock.	48% (125/260, 100♂ and 25♀)	BAT +ve group was significantly younger, with lower BMI, body fat mass, abdominal fat, HbA1c, total cholesterol, and LDL cholesterol.
Muzik et al. 201326	10♂, 15♀	Age: 30 ± 7 y BMI: -High-BAT group 22.1 ± 3.1 -Low-BAT group 24.7±3.9	Subjects kept in room temperature of 15.5°C for total of 150 mins.	36% (9/25, 1♂ and 8♀) had >10g active BAT (high-BAT group)	Post-cold exposure, EE increased by 17.4 ± 15% in the high-BAT group vs. 0.4 ± 15.6% in low-BAT group (p=0 .04). Lean body mass was lower in the high-BAT group (46. Five± 7.1 vs. 54.1 ± 11.0, p=0 .04).
van der Lans et al. 201327	8♂, 9♀	Age: 23 ± 3.2 years
BMI: 21.6 ± 2.2
	30 mins of cooling to just above shivering using water-perfused suit prior to each scan, which continued during each 60-min scan. Increasing duration of daily cold exposure (15–16°C) for 10 consecutive days (Day 1: 2 hours, Day 2:4 hours, Days 3–10: 6 hours). Suit temperature: -pre-cold acclimation 25.4 ± 1.8°C -post-cold acclimation 25.8 ± 1.6°C	Pre-cold acclimation: 94% (16/17) Post-cold acclimation: 100% (17/17)	Cold acclimation increased BAT activity and volume, as well as increased non-shivering thermogenesis (NST) from 10.8 ± 7.5% to 17.8 ± 11.1% (P<0.01). EE was higher in males both pre- and post-cold acclimation, but there were no significant differences between genders in NST pre- and post-cold acclimation. Cold acclimation did not increase glucose uptake rate in BAT.
van Rooijen et al. 201328	4♂, 7♀	Age: 23.9 ± 2.2 y BMI: 22.9 ± 3.2	30 mins of cooling to just above shivering using water-perfused suit (suit temperature 24.1± 1.4°C) prior to each scan, which continued during each 60-min scan.	73% (8/11)
Yoneshiro et al. 201329	51♂	Age: 24.4 ± 0.5 y BMI: 22.0 ± 0.4	Acute: Subjects kept in room temperature of 19°C for 2 hours prior to each scan.
Acclimation in 12/22 subjects with initially undetectable or low cold-activated BAT: Subjects kept in room temperature of 17°C for 2 hours/day for 6 weeks.	Acute: 53% (27/51) Post-cold acclimation: 75% (6/8)	Cold-induced thermogenesis (CIT) was significantly higher in BAT +ve subjects compared with BAT –ve subjects (252.0 ± 41.1 kcal/day vs. 78.4 ± 23.8 kcal/day, P<0.01). BAT activity was independently associated with CIT. Cold acclimation resulted in higher CIT at week 6 than week 0 (289.0 ± 70.0 kcal/day vs. 108.4 ± 22.8 kcal/day, P<0.05), and a decrease in body fat mass (-5.2 ± 1.9%, P<0.05) without altering body weight and fat-free mass. Changes in CIT and body fat mass were significantly greater in the cold-acclimated group compared with controls.
Blondin et al. 201430	6♂	Age (mean ± s.e.m.): 23 ± 1 years
BMI (mean ± s.e.m.): 24.5 ± 1.2	Acute: Liquid-conditioned suit (water temperature 18oC) for a total of 3 hours.
Acclimation: Liquid-conditioned suit (water temperature 10°C) 2 hours/day 4–5 days/week for 4 weeks.		Cold acclimation increased the volume of active BAT by 45% (from 66 ± 30 mL to 95 ± 28 mL, P<0.05), increased cold-induced BAT oxidative metabolism from 0.725 ± 0.3 mL/s to 1.591± 0.326 mL/s (P<0.05), and decreased circulating glucose and cortisol levels at 25oC room temperature and after acute cold exposure.

The use of PET-CT has facilitated the identification and study of various stimulators and inhibitors of BAT activity, which are summarised in Figure 1. Cold exposure is the best known stimulator of BAT activity, and several ¹⁸F-FDG PET-CT studies in humans using acute and/or chronic cold exposure are summarized in Table 1. The issue of whether cold-activated BAT actually results in significantly increased energy expenditure in humans has been addressed by several studies. In a randomized single-blind crossover study,²⁵ 24 participants (14 male and 10 female, mean age 28 years) spent 12 hours in a whole-room indirect calorimeter maintained at 24°C or 19°C after which a PET-CT scan was performed. Thirty-six hours later the participants crossed over to the alternate study temperature followed by a second scan. During the study the participants received an individualized calorie-calculated caffeine-free diet consisting of 50% carbohydrate, 20% protein and 20% fat, wore hospital scrubs, slept with cotton sheets (and without blankets), and care was taken to ensure the participants did not shiver. This study demonstrated that cold-activated BAT results in a 5% increase in energy expenditure (with higher increases in energy expenditure in women and an inverse relationship between age and increased energy expenditure).²⁵ Other studies have demonstrated a mean increase in energy expenditure of 3.1 kcal/day after cold exposure (15.5°C for 1 hour) in adults with a mean age of 29.6 years,²⁶ and 250 kcal/day (after exposure to 19°C for 2 hours) in healthy young men with a mean age of 24.4 y²² Also in the latter study, chronic cold exposure (17°C for 2 hours a day for 6 weeks) in subjects with absent or low ¹⁸F-FDG uptake on baseline PET scanning resulted in increased ¹⁸F-FDG uptake and increased energy expenditure at week 6.²² In a different study, chronic cold exposure (10°C for 2 hours per day for 4 weeks) increased BAT volume by 45%, as well increased BAT oxidative metabolism and fractional glucose uptake.³⁰ These studies have demonstrated that BAT in adult humans can be activated (and recruited) resulting in significant increases in energy expenditure and metabolic activity, which can potentially be exploited to develop agents to treat obesity.

Figure 1.

: Inhibitors and stimulators of brown adipose tissue activation identified by PET-CT studies. Warm temperatures, β-adrenoceptor antagonists, reserpine, fentanyl, inhaled anesthetics and obesity inhibit (- - -) brown adipose tissue thermogenesis. Cold temperatures, catecholamines, the sympathetic nervous system, β-adrenoceptor agoinists, nicotine, insulin and thyroid hormone stimulate ( ) brown adipose tissue.

¹⁸F-FDG imaging has also provided useful information about the pharmacological regulation of BAT activation. An ex-vivo rodent study demonstrated that ¹⁸F-FDG uptake in BAT is increased 7.9-fold by nicotine, 3.7-fold by ephedrine, 12-fold by a combination of nicotine and ephedrine, and slightly but non-significantly by caffeine.³² The increased ¹⁸F-FDG uptake caused by nicotine administration is prevented by prior administration of propranolol or reserpine.³² Cold-induced ¹⁸F-FDG uptake is inhibited by both propranolol and reserpine in rodents, with diazepam having little or no effect.³³ This data suggests that the effects of nicotine and cold on BAT activity are mediated by the sympathetic nervous system. Similarly, in humans, ¹⁸F-FDG uptake by BAT can be reduced by intravenous fentanyl³⁴ and oral propranolol,^35,36 but it is not significantly affected by diazepam.³⁴

The crucial role that the sympathetic nervous system plays in BAT activation may be utilized therapeutically. Atomoxetine (a potent highly selective inhibitor of presynaptic norepinephrine transport that increases synaptic concentrations of norepinephrine) increased ¹⁸F-FDG uptake by BAT at ambient temperatures in fasted and non-fasted rats with corresponding decreases in blood glucose levels and increases in interscapular BAT temperature compared to controls.³⁷ Additionally, it is important to be aware that inhaled anesthetics (including halothane, isoflurane and chloroform) have been shown to directly inhibit norepinephrine-induced BAT activation in cultured cells and in in vivo experiments.^38–40

BAT is richly innervated by the sympathetic nervous system and norepinephrine increases glucose transport in brown adipocytes and increases the numbers of brown adipocytes.^41,42 Therefore conditions in which circulating catecholamine levels are chronically elevated (such as in patients with phaeochromocytomas) provide natural experiments for the study of BAT. Case reports have demonstrated greatly increased (non-cold stimulated) ¹⁸F-FDG uptake in cervical, axillary, mediastinal, abdominal (including omental and mesenteric), paravertebral, and perirenal fat depots of patients with phaeochromocytomas and concomitantly increased catecholamines/metanephrines,^43–49 and surgical resection of the phaeochromocytomas with restoration of normal circulating catecholamine levels^43,46 or propranolol administration⁴⁷ abolished the increased uptake in these regions. In patients with phaeochromocytomas whose total plasma metanephrines were not elevated, BAT uptake of ¹⁸F-FDG was not increased. Instead it was the same as the uptake seen in normal (non-phaeochromocytoma) controls.⁵⁰ Additionally in this study (which also included phaeochromocytoma patients with elevated total plasma metanephrine levels) BAT activity was inversely correlated with BMI and waist circumference, and positively correlated with total plasma metanephrine levels.⁵⁰

Histological analysis of intra-abdominal fat obtained from 3 patients with phaeochromocytomas with high circulating catecholamine levels revealed that this fat possessed all the unique features of thermogenically active brown adipose tissue.⁵¹ In a more recent study in which the omental fat from 12 patients with phaeochromocytoma was analyzed, UCP1-immunoreactivity (UCP1-ir) positive multilocular cells were present in samples from 6 patients and islands of BAT were identified within WAT from 4 of these 6 patients.⁵² In contrast, none of the samples from 20 non-phaeochromocytoma patients who underwent cholecystectomy were UCP1-ir positive.⁵² Therefore evidence from imaging case reports and histological studies indicate that chronic catecholamine excess increases BAT activity and may promote transdifferentiation of WAT to BAT.

Other tracers

¹⁵O-H₂O (radiolabelled water) has been used as a tracer in PET-CT scans to detect and quantify the perfusion of BAT. The oxygen consumption of a tissue is matched by the perfusion rate of the tissue, therefore perfusion rate can be used as an approximation of oxygen consumption.⁵³A study in which PET-CT scans performed in the same individuals using both ¹⁵O-H₂O and ¹⁸F-FDG as tracers demonstrated that in response to cold exposure both the perfusion and glucose uptake of BAT increased.²¹ However in response to insulin infusion (with concurrent glucose infusion to maintain euglycaemia), BAT glucose uptake increased without a concomitant increase in BAT perfusion.²¹ Glucose enters brown adipocytes via GLUT1 (an insulin-independent glucose transporter) and GLUT4 (an insulin-regulated glucose transporter). So the insulin-mediated increase in ¹⁸F-FDG uptake by BAT does not appear to be accompanied by increased oxidation in BAT, and may represent increased glucose transport due to higher insulin concentrations during an insulin infusion. In a different study using these 2 tracers, hyperthyroidism increased glucose uptake in BAT (and muscle, but not in WAT) with increased lipid oxidation and whole body energy expenditure, but BAT perfusion was not increased (unlike muscle which had increased perfusion).⁵⁴ These changes were reversed by restoration of euthyroidism.⁵⁴ Therefore increased ¹⁸F-FDG uptake by BAT in PET scans provides limited information about the metabolic activity of BAT.

Another PET-CT tracer that has been used to study BAT is 11C-meta-hydroxyephedrine (¹¹C-MHED). ¹¹C-MHED is a radioactive analog of norepinephrine, which emits positrons as it decays.¹¹C-MHED uptake by BAT was higher in mice kept at an ambient temperature of 21°C compared with mice kept at 26°C.⁵⁵ No ¹¹C-MHED uptake was seen in denervated BAT and increased ¹¹C-MHED uptake in sham-operated BAT was blocked by SR59230A (a selective β3-adrenoreceptor antagonist).⁵⁵ Furthermore, in mice treated with 1mg/kg/day CL316,243 (a potent β3-adrenoreceptor agonist) for 4 weeks, BAT ¹¹C-MHED uptake was 2.5 times higher than in vehicle-treated mice.⁵⁵ This provides further evidence that cold-induced BAT thermogenesis is mediated by the sympathetic nervous system.

Other experiments performed by the same group demonstrated reduced β3-adrenoreceptor mediated ¹¹C-MHED and ¹⁸F-FDG uptake by BAT in mice fed a high-fat diet compared to those fed a standard chow diet.⁵⁵ Additionally, a 3-fold increase in ¹¹C-MHED and ¹⁸F-FDG uptake was detected in the inguinal subcutaneous white adipose tissue (WAT) depot of chow-fed mice that were treated with 1mg/kg/day CL316,243 for 4 weeks compared with vehicle-treated controls. Post-mortem histological analysis of these inguinal adipose tissue depots revealed tissue similar to BAT (i.e. clusters of multi-locular adipocytes that expressed UCP-1 and peroxisome-proliferator activated receptor γ co-activator 1 α, which are hallmarks of BAT).⁵⁵ These experiments indicate that diet-induced obesity can adversely affect BAT function and chronic β3-adrenoreceptor activation induces transformation of WAT to BAT-like tissue.

Single-photon emission computed tomography (SPECT) – CT (CT)

In SPECT a gamma camera is used to obtain multiple 2 dimensional images from multiple angles, which are used by a computer to reconstruct a 3 dimensional image. Furthermore the emissions from the radioactive ligand reveal the capillary blood flow of the imaged region, and can also provide information about metabolism within the imaged tissue. Images from a CT scan can then be co-registered with the SPECT images to provide more detailed anatomical information.

There have been reported cases of activated BAT identified on SPECT-CT scans using technetium sestamibi (^99mTc-sestamibi) as this tracer is taken up in the same distribution as ^18FFDG uptake on a PET-CT scan.^56,57 Furthermore the uptake of ^99mTc-sestamibi is 4 times higher in BAT than in WAT, and in rats its uptake is inversely correlated to body weight.⁵⁸ A retrospective review of 205 ^99mTc-sestamibi SPECT-CT scans performed for evaluation of parathyroid adenomas, active BAT was detected in 6.3% of the scans (a prevalence similar to that obtained from retrospective reviews of ¹⁸F-FDG PET-CT scans).⁵⁹ In another study, active BAT was detected in 5.4% of 74 ^99mTc-sestamibi SPECT-CT scans that had been performed for parathyroid imaging, and histological examination of biopsies taken from the cervical and supraclavicular regions of increased tracer uptake revealed adipocytes that possessed the morphological and genetic characteristics of BAT.²³ In mice injected with a β3-adrenoreceptor agonist, perfusion of BAT (as revealed by ^99mTc-sestamibi scanning) increased by 61%, while glucose uptake (as demonstrated by ¹⁸F-FDG PET-CT scanning) increased by 440%.²³ Therefore sympathetic activation of BAT results in both increased perfusion and glucose uptake. However the increased glucose uptake far exceeds the increased perfusion, which raises the possibility that glucose may be used for other non-thermogenic processes in active brown adipocytes.

Another tracer used in SPECT-CT is^99mTc-tetrofosmin (technetium tetrofosmin), which is a tracer that is absorbed by functional mitochondria. In a retrospective study of 385 scans performed on 329 children (with cardiac conditions) with a mean age of 9.1 y (range 0–19), ^99mTc-tetrofosmin uptake was increased in the neck and shoulder regions (i.e., areas known to contain deposits of BAT) in 17% (65/385) of the scans.⁶⁰ There was no relationship between the underlying cardiac disorder and increased tracer uptake. Similar to ¹⁸F-FDG PET scans, the frequency of active BAT detection was significantly higher (P<0 .05) in winter (29%) than in spring (16%) or summer (9%).⁶⁰ However, the prevalence of active BAT as detected by this tracer (i.e. 17%) is much lower than that reported in children of similar ages using ¹⁸F-FDG PET scans (43–45%).¹⁸

^99mTc-sestamibi and ^99mTc-tetrofosmin uptake are determined by the perfusion of a tissue (i.e., the higher the perfusion rate the greater the uptake of these tracers), and the accumulation of these tracers is proportional to the density and metabolic activity of the mitochondria within the tissues. Therefore SPECT-CT scans utilizing these tracers may offer more accurate information about the presence of activated BAT, since activated BAT requires increased perfusion as well as increased mitochondrial activity.

Metaiodobenzylguanidine (MIBG) is derived from guanethidine and it acts like norepinephrine and accumulates in neuroendocrine tumors. In clinical practice, radiolabelled MIBG is used in SPECT scans for diagnosis and staging of these types of tumors. However, a rodent study (in which ¹²³I-MIBG was administered to rats that were subsequently killed and their tissues examined) demonstrated that MIBG also accumulates in BAT (at concentrations 20–30 times that of WAT) and this accumulation increases after stimulation with a β3-receptor agonist.⁶¹ Similarly, ¹²³I-MIBG uptake was noted on the left neck and shoulder region of a child with a neuroblastoma but no uptake was seen on the right where there had been surgical disruption of the sympathetic nerves in the neck and upper thorax (as evidenced by right sided Horner's syndrome).⁶² Furthermore, in a study of 10 lean young men (BMI 19–25 aged 18–32 years) who were exposed to cold (17°C) for 2 hours prior to ¹²³I-MIBG SPECT-CT and ¹⁸F-FDG PET-CT scanning, ¹²³I-MIBG uptake occurred in the same anatomic areas (i.e. cervical, supraclavicular and superior mediastinal regions) as ¹⁸F-FDG uptake.⁶³ This is consistent with evidence from other studies that BAT activation due to cold exposure is mediated by the sympathetic nervous system.

Magnetic resonance imaging (MRI)

MRI scanners are used to create a strong magnetic field around an area to be imaged. Oscillation of this magnetic field provides energy to hydrogen atoms. This results in the excitation of the hydrogen atoms, which emit a radiofrequency signal. This signal is detected by the MRI scanner and is processed to produce an image. The differing rate at which excited hydrogen atoms (protons) return to equilibrium, enables different body tissues to be distinguished from each other. Furthermore, fat protons resonate at a frequency that is 3.5 ppm higher than water protons.⁶⁴ The MRI signals of BAT are produced by both water protons and fat protons. In contrast, the MR signals from WAT are produced predominantly by fat protons, while that from muscle is produced almost entirely by water protons.⁶⁵ Furthermore, in rats BAT is composed of 20–50% fat, while WAT is made up of 70–90% fat, and the MRI estimate of the fat fraction of BAT post-mortem was 10% higher than in live rats.⁶⁶ In mice similar results have been obtained. In one mouse study the fat fraction of BAT was reported to be 40–80% and that of WAT was 90–93%,⁶⁷ while in another study (which used different MRI protocols) the mean fat fraction of BAT was 51.5% and the mean fat fraction of WAT was 92.9%.⁶⁴ In mice housed at different room temperatures, the mean fat fraction of BAT decreased with decreasing room temperature (79.4% at 30°C, 61.8% at 23°C and 50.9% at 16°C).⁶⁸ This is in keeping with the fact that cold exposure activates BAT, leading to increased lipolysis in brown adipocytes.

In humans, the water-to-fat ratios of BAT and WAT have also been found to be significantly different. In human neonates, the mean fat fraction of BAT was 30.2%, while that of WAT was 67.7% (n = 22).⁶⁹ Furthermore, in children, in MRIs performed at a room temperature of 25°C, the fat fractions of both supraclavicular BAT and subcutaneous WAT increase with increasing BMI, and the fat fraction of supraclavicular BAT increases with increasing age.⁷⁰ In a study of 11 healthy volunteers aged 18–30 years, in response to cold stimulation 8/11 subjects had active BAT on ¹⁸F-FDG scans, and in these subjects water-fat MRI showed that the mean fat fraction of BAT was 65.2% while the mean fat fraction of WAT was 81.5%.²⁸

Fat-fraction MRI (performed at a room temperature of 22°C) has been used to identify BAT in a person in whom no BAT activity was detected on an ¹⁸F-FDG PET-CT scan performed at 22°C.⁷¹ In another study carried out in a patient with a parathyroid tumor, 83% of retrospectively identified regions of activated BAT from an ¹⁸F-FDG PET-CT scan showed corresponding low signal (i.e., adipose tissue with a low fat fraction) on MRI, while prospectively 87% of low fat-fraction adipose tissue depots on a subsequent MRI scan demonstrated increased uptake on the second ¹⁸F-FDG PET-CT scan.⁷² Therefore these 2 imaging modalities appear to identify most, but not all of the BAT depots that are present in an adult human.

MRI has also been used to study the perfusion of BAT before and after stimulation. Following an injection of 5mcg/kg of adrenaline (and a gadolinium-based contrast agent) to anaesthetized male rats, the signal intensity of interscapular BAT was found to be significantly higher than in rats that had not received adrenaline, indicating that adrenaline increases BAT perfusion.⁷³ Furthermore, in this study, following adrenaline administration, the perfusion of dorsal muscles (i.e. muscles adjacent to BAT) was also increased in comparison to limb muscles.⁷³

From a functional point of view, blood oxygen level dependent (BOLD) contrast has been used in MRI for studying BAT. BOLD is based on the principle that activation of a highly vascularised specific tissue or organ results in increased blood flow to and oxygen consumption within that region with a corresponding change in the relative levels of oxy- and deoxyhaemoglobin and MRI signal intensity. BOLD has been proposed as a means of functionally assessing BAT because when BAT is activated, its oxygen consumption increases to a much greater extent than the increase in its perfusion,^21,23,74 resulting in higher levels of deoxyhaemoglobin. In anesthetized female mice injected with an intraperitoneal bolus of 2.5mg/kg of norepinephrine, the signal intensity in BAT changed significantly (by ∼20%) and this was accompanied by a 5.5°C increase in BAT temperature (measured directly with a probe), which occurred within 40–50 minutes post-injection and declined to pre-injection levels after 140 minutes.⁷⁵ The signal intensities of WAT and skeletal muscle did not change significantly following the norepinephrine injection.⁷⁵ In a study in humans, water-fat MRI identified non-stimulated BAT in 5 adults in the same supraclavicular areas as cold-activated BAT was detected by previous ¹⁸F-FDG PET-CT scans, and both types of scans yielded similar BAT volumes.⁷⁶ In this same study BOLD MRI detected a 10% change in signal intensity in BAT after it had been activated by cold.⁷⁶

An alternative method of functionally imaging BAT uses MRI with hyperpolarized [1–¹³C]-pyruvate as a contrast agent. ¹³C is a non-radioactive isotope of ¹²C and it produces a different signal from ¹²C on MRI scans. Oxidative phosphorylation increases in activated BAT, therefore within activated BAT there will be increased conversion of [1–¹³C]-pyruvate to ¹³C-bicarbonate and [1–¹³C]-lactate. A study using this technique demonstrated that following intravenous administration of [1–¹³C]-pyruvate and subsequent intraperitoneal injection of 2.5mg/kg of norepinephrine to anesthetized rats, compared to baseline, there was a 6.3-fold increase in BAT hyperpolarized bicarbonate signal and a 3.9-fold increase in hyperpolarized lactate signal.⁷⁷ Significant increases in hyperpolarised lactate signals in the heart (3.4-fold) and kidneys (2.0-fold) were also reported.⁷⁷ Since oxidative metabolism in BAT accounts for about 88% of its total oxygen consumption,⁷⁸ this technique could also be used to investigate fatty acid metabolism in activated BAT.

MRI may prove to be invaluable in the study of BAT because it has the potential to differentiate BAT from WAT, to identify non-activated BAT and activated BAT, as well as allow functional studies to be performed on BAT. Furthermore, since it is not dependent on the use of ionizing radiation it may be used for repeated imaging of the same subjects as well as imaging of groups of people who are often excluded (e.g. children, women of childbearing age) from studies involving imaging techniques utilizing ionizing radiation.

Infrared thermography/thermal imaging

Infrared radiation is emitted by objects at temperatures above absolute zero, and the amount of radiation an object emits increases as the temperature of that object rises. Thermal imaging cameras detect radiation in the infrared range of the electromagnetic spectrum and produce thermograms (i.e., images of variations in temperature). Since the temperature of BAT increases when it is activated,⁷⁵ and some of this heat will be transferred to surrounding tissues including overlying skin, thermal imaging has been proposed as means of studying BAT activation in vivo in response to different stimuli, especially as it does not require the use of ionizing radiation.

This non-invasive technique has been used to demonstrate BAT activation in rodent studies. Mice with an R384C mutation in the central thyroid hormone receptor (TRα1) are hypermetabolic due to central overactivation of BAT.⁷⁹ These mice had a significantly higher maximum interscapular skin temperature at thermoneutral room temperature (with increased whole body weight-adjusted oxygen consumption) when compared to wild type mice.⁸⁰ The body temperature of p62 knockout mice (as measured with a thermal camera), as well as energy expenditure, following cold exposure was lower than that of wild type mice.⁸¹ In this same study, p62 knockout mice gained more weight than wild type mice when both groups were fed normal chow diet and high fat diet.⁸¹ This provides further support for the theory that BAT activity plays a significant role in energy homeostasis.

In a study in which shaved mice where either kept in a room at 4°C for 24hours, stressed by causing full-thickness non-lethal burn injuries to 30% of their total body surface area or removal of a 1cm² section of skin to cause a full thickness wound, accumulation of ¹⁸F-FDG in BAT increased by 15-fold in the cold and cutaneous wound animals and by 5-fold in the cold stress animals (in comparison to control mice).⁸² The temperature of skin overlying interscapular BAT (in comparison to adjacent non-BAT tissue), measured with a thermal imaging camera, was on average 1.7°C higher in the cold stress group, 1.5°C higher in the skin wound group, and 1.5°C higher in the skin burn group. This was significantly different from the 0.9°C difference between skin temperature overlying BAT and adjacent skin temperature that was not overlying BAT seen in the control group.⁸² Also, there was a strong linear correlation between ¹⁸F-FDG accumulation in BAT and the increase in skin temperature detected by the thermal imaging camera (r² = 0.835, P<0 .0001).⁸²

Several human studies using thermal imaging have been reported. In one study, volunteers sat in a room (with the temperature maintained at 19–21°C) and were asked to place one hand in water (maintained at a temperature of 19–20°C) for 5 minutes and thermal images were recorded during the 5 minutes immediately preceding the cold stimulus as well as throughout cold stimulation. The change (from baseline) in the mean of the upper 10^th percentile of readings in the supraclavicular area was reported as +0.62±0.14°C in 3–8 y olds (n = 7), +0.25±0.08°C in 13–18 y olds (n = 12), and +0.20±0.08°C in 35–58 y olds (n = 7) (P<0 .05).⁸³ These areas of increased temperature in response to cold stimulation corresponded with areas of increased ¹⁸F-FDG uptake on PET-CT scans.⁸³ In a second study performed by the same group during winter months, the median value of the hottest 25% of pixels in the thermal image of the skin in the supraclavicular area was measured. An inverse relationship was found between the baseline value and BMI percentile in 55 healthy children aged 6–11 y In 24 of these 55 children, the median temperature of the hottest 25% of the region of interest increased by 0.28±0.06°C within 5 minutes of immersing one hand in water maintained at a temperature of 20.1°C.⁸⁴ There was no significant change in the temperature of the sternum, which served as a control area.⁸⁴

A case report of an 11 y old girl with untreated acquired primary hypothyroidism demonstrated a 1.7°C temperature difference (using infrared thermography) between the skin overlying the supraclavicular fat depot (36.0°C±0.16) and the skin overlying the sternum (34.3°C±0.19) at a room temperature of 22°C.⁸⁵ Interestingly, after 2 months of treatment with thyroid hormone replacement, which resultant clinical and biochemical euthyroidism, the skin temperature difference (at the same room temperature) fell to 0.7°C (supraclavicular skin temperature 37.1°C±0.13 vs. sternal skin temperature 36.4°C±0.13).⁸⁵ This raises the possibility that TSH (which was markedly elevated when her hypothyroidism was untreated with a subsequent reduction to normal levels post-treatment) may stimulate BAT thermogenesis, especially since brown adipocytes are known to have TSH receptors and TSH increases UCP1 mRNA in brown adipocytes.⁸⁶

Ultrasound

Contrast ultrasound is used to estimate blood flow noninvasively by visualizing and quantifying intravenous echogenic microbubbles. These microbubbles are destroyed by a high-energy ultrasound pulse and the rate at which they are replenished can be used to estimate the blood flow to a tissue. Following the intravenous administration of 1mcg/kg/min of norepinephrine to anesthetized mice (kept at a room temperature of 37°C), blood flow to BAT as measured with contrast ultrasound was found to increase 15 to 20-fold compared to baseline in wild-type mice, and this effect of norepinephrine was present but diminished in UCP-1 deficient mice.⁸⁷ In a different study, infusion of the same dose of norepinephrine resulted in a similar increase in BAT blood flow (15–fold20- increase) in wild type mice fed a low fat diet or high fat diet for 2 months, with a higher increase in blood flow in the low fat diet mice.⁸⁸ In db/db (obese diabetic leptin receptor-deficient) mice, baseline blood flow to BAT was significantly lower than wild type and norepinephrine increased BAT blood flow to a lesser extent (about fold10- increase).⁸⁸ Additionally, these authors were able to estimate BAT mass using contrast ultrasound in the wild type mice, and the estimates correlated well with BAT mass measured at necropsy (r² = 0.83, P<0 .001).⁸⁸

Contrast ultrasound is an established technique used in cardiology and its advantages include no use of ionizing radiation and its ability to quantify BAT blood flow and mass. However, one limitation is its inability to estimate BAT mass in poorly vascularised BAT (for instance in severely obese and diabetic db/db mice),⁸⁸ which may limit its utility in obese and/or diabetic humans.

Conclusions

PET-CT is currently the most widely used imaging modality employed in studies of BAT prevalence and activity. It has provided important information about the prevalence and regulation of BAT in humans and with the use of different tracers it has facilitated the investigation of differential effects of stimuli such as cold, insulin and thyroid hormone have on BAT perfusion and glucose uptake. However, its main limitations are: it is expensive, it involves the use of ionizing radiation, it can only be used to detect BAT that is metabolically active and it may not be able to detect small depots of BAT that are below the spatial resolution threshold of the scanner. SPECT-CT utilizing technetium-based tracers may provide more accurate estimates of active BAT since the uptake of these tracers is proportional to the perfusion of a tissue as well as the density and metabolic activity of the intra-tissue mitochondria, and active BAT requires both increased perfusion and increased mitochondrial activity. The use of ¹²³I-MIBG with SPECT-CT has provided further evidence that cold-induced BAT activation is mediated by the sympathetic nervous system. Its applications may be limited by the expense of this type of scan and the use of ionizing radiation.

MRI has been used in the anatomical and functional imaging of BAT. Using a variety of tracers it is possible to study the changes in blood flow to and oxygen consumption by BAT in response to different stimuli, and has the potential to be used to study fatty acid metabolism in activated BAT. Additionally, it has the potential to be used to detect non-activated BAT (since WAT and BAT have distinct fat fractions), and MRI has the added advantage of not requiring the use of ionizing radiation.

Infrared thermography and contrast ultrasound are other imaging modalities that do not require the use of ionizing radiation. There are much fewer published studies utilizing these techniques compared with other imaging modalities. Although its use in the study of BAT is in its infancy, thermal imaging has been used to successfully detect activated BAT (using the increase in the skin temperature overlying BAT as a marker of BAT activation) in rodents, children and adults in distributions similar to that seen in PET-CT studies. The use of contrast ultrasound is well established in several medical specialties, and it has been used to quantify BAT mass and BAT blood flow in response to sympathetic nervous system activation. However, its use might be limited by its inability to quantify BAT mass in poorly vascularised BAT such as in severely obese and diabetic rodents.

Imaging studies have made significant contributions to our understanding of BAT function. Although each imaging modality has its limitations, refinements of existing imaging techniques and the development of new techniques are likely to improve the non-invasive investigation of BAT. This is important because in vivo studies will be required if modulation of BAT activity is to be employed in the treatment of obesity.

Disclosure of Potential Conflicts of Interest

All authors have completed the Unified Competing Interest form at www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declare that (1) none of the authors have any relationships with companies that might have an interest in the submitted work in the previous 3 years; (2) none of the authors’ spouses, partners, or children have a financial relationship that may be relevant to the submitted work; and (3) none of the authors have a non-financial interest that may be relevant to the submitted work.

Funding

The Section of Investigative Medicine is funded by grants from the MRC, BBSRC, NIHR, an Integrative Mammalian Biology (IMB) Capacity Building Award, an FP7- HEALTH- 2009–241592 EuroCHIP grant and is supported by the NIHR Imperial Biomedical Research Centre Funding Scheme. CI is supported by an Imperial College Healthcare NHS Trust Charity Research Fellowship, VS is supported by an NIHR Clinical Lectureship, WSD is supported by an NIHR Career Development Fellowship.