Article

Selenoprotein P-mediated reductive stress impairs

cold-induced thermogenesis in brown fat
Graphical abstract Authors
Swe Mar Oo, Hein Ko Oo,
Hiroaki Takayama, ...,
Normal Diabetes
Yuko Okamatsu-Ogura, Masayuki Saito,
Se
Se Se
Se SeP Se
Se
Toshinari Takamura
COLD EXPOSURE Se
Se Se
Se

Se
Se Se
Se
Se SeP Se
Se Se
Se Se

Se
Se Se
Se
Correspondence
Se SeP Se Se Se
Sympathetic nerve Se
Se Se
Se Se Se ttakamura@med.kanazawa-u.ac.jp
Se SeP Se
Se Se
Se Se
Se
Se Se
Se SeP Se
Se In brief
Se Se
Se Se
Se
Se Se
Se
Noradrenaline Se SeP Se ROS-mediated sulfenylation is essential
Autocrine/ Se Se
LRP1 Se Se
Paracrine Se
Se Se
Se for the activation of UCP1 in brown
Se Se Se SeP Se Se Se
Se Se
LRP1 Se Se
Se SeP Se
Se
Se Se
Se
Se SeP Se adipose tissue (BAT). Selenoprotein P
Se Se Se Se Se
Se Se Se Se

Se
Se (SeP) is the diabetes-associated
Se
Se Se hepatokine. Oo et al. reveal that local and
circulating SeP impairs UCP1
GPX4
mtROS GPX4
sulfenylation and thermogenesis via
SOH
UCP1
SOH reductive stress in BAT and may be a
UCP1 therapeutic target against obesity and
SOH SOH
UCP1 UCP1 UCP1 UCP1 diabetes.
HEAT
HEAT Sulfenylation Sulfenylation PRODUCTION
PRODUCTION

Highlights
d High-serum SeP is associated with low BAT activity in
humans

d Selenop KO mice show higher UCP1 activation and

thermogenesis in BAT

d SeP impairs UCP1 activation by eliminating the mitochondrial

ROS via GPX4

d Elevated serum SeP in diabetes impairs BAT thermogenesis

in mice

Oo et al., 2022, Cell Reports 38, 110566

March 29, 2022 ª 2022 The Authors.
https://doi.org/10.1016/j.celrep.2022.110566 ll
 ll
OPEN ACCESS

Article
Selenoprotein P-mediated reductive stress impairs
cold-induced thermogenesis in brown fat
Swe Mar Oo,1,9 Hein Ko Oo,1,9 Hiroaki Takayama,1,2,9 Kiyo-aki Ishii,1,3 Yumie Takeshita,1 Hisanori Goto,1 Yujiro Nakano,1
Susumu Kohno,4 Chiaki Takahashi,4 Hiroyuki Nakamura,5 Yoshiro Saito,6 Mami Matsushita,7 Yuko Okamatsu-Ogura,8
Masayuki Saito,8 and Toshinari Takamura1,10,*
1Department of Endocrinology and Metabolism, Kanazawa University Graduate School of Medical Sciences, 13-1 Takara-machi, Kanazawa,

Ishikawa 920-8640, Japan

2Life Sciences Division, Engineering and Technology Department, Kanazawa University, Kanazawa, Ishikawa, Japan
3Department of System Biology, Kanazawa University Graduate School of Medical Sciences, Kanazawa, Ishikawa, Japan
4Division of Oncology and Molecular Biology, Cancer Research Institute, Kanazawa University, Kanazawa, Ishikawa, Japan
5Department of Environmental and Preventive Medicine, Graduate School of Advanced Preventive Medical Sciences, Kanazawa University,

Kanazawa, Ishikawa, Japan

6Laboratory of Molecular Biology and Metabolism, Graduate School of Pharmaceutical Sciences, Tohoku University, Sendai, Japan
7Department of Nutrition, Tenshi College, Sapporo, Japan
8Laboratory of Biochemistry, Faculty of Veterinary Medicine, Hokkaido University, Sapporo, Japan
9These authors contributed equally
10Lead contact

*Correspondence: ttakamura@med.kanazawa-u.ac.jp
https://doi.org/10.1016/j.celrep.2022.110566

SUMMARY

Reactive oxygen species (ROS) activate uncoupler protein 1 (UCP1) in brown adipose tissue (BAT) under
physiological cold exposure and noradrenaline (NA) stimulation to increase thermogenesis. However, the
endogenous regulator of ROS in activated BAT and its role in pathological conditions remain unclear. We
show that serum levels of selenoprotein P (SeP; encoded by SELENOP) negatively correlate with BAT activity
in humans. Physiological cold exposure downregulates Selenop in BAT. Selenop knockout mice show higher
rectal temperatures and UCP1 sulfenylation during cold exposure. SeP treatment to brown adipocytes elim-
inated the NA-induced mitochondrial ROS by upregulating glutathione peroxidase 4 and impaired cellular
thermogenesis. A high-fat/high-sucrose diet elevates serum SeP levels and diminishes the elevated NA-
induced thermogenesis in BAT-Selenop KO mice. Therefore, SeP is the intrinsic factor inducing reductive
stress that impairs thermogenesis in BAT and may be a potential therapeutic target for obesity and diabetes.

INTRODUCTION
Converting excess energy to heat in the brown adipose tissue
(BAT) will undoubtedly be an attractive solution to obesity and
diabetes (Chouchani et al., 2019; Saito, 2013). After a long-
standing prevailing belief of BAT’s existence only in neonates
and small rodents, unexpected evidence of active BAT in adult
humans (Cypess et al., 2009; van Marken Lichtenbelt et al.,
2009) has demanded more extensive studies of BAT meta-
bolism. BAT, a specialized form of adipose tissue, primarily
functions to promote the dissipation of chemical energy into
heat by the process called non-shivering thermogenesis
(Chouchani et al., 2019). This dissipation occurs because of
BAT’s richness in mitochondria and uncoupling protein 1
(UCP1). BAT thermogenesis is usually escalated by cold-
induced sympathetic activation, the neuronal release of
noradrenaline (NA), and its subsequent b3-adrenergic signaling.
Accumulating evidence in both animals and humans indicates
that BAT’s thermogenic endowment plays a vital role in whole-
body energy expenditure, metabolic homeostasis, and substrate
utilization (Lee et al., 2010; Stanford et al., 2013; Yoneshiro et al.,
2011). When fully activated, BAT can be a crucial metabolic
sink for the clearance of glucose, triglycerides, and other
specific metabolites from the circulation. This leads to
lowered serum glucose and lipid levels and enhanced energy
expenditure (Bartelt et al., 2011; Olsen et al., 2017). Therefore,
identifying the upstream regulators of UCP1 in BAT becomes
critical for developing BAT-targeted therapies for obesity and
diabetes.
Reactive oxygen species (ROS), such as superoxide anions
(O2
), hydrogen peroxide (H2O2), and hydroxyl radicals (OH,),
are the by-products of cellular respiration under aerobic
conditions. Electron leakage during substrate oxidation through
electron transport chains located in the inner mitochondrial
membrane produces substantial ROS within the mitochondria.
Immediately after synthesis, mitochondrial superoxide is
dismutated to less reactive H2O2 by superoxide dismutase 2
(SOD2) in the mitochondrial matrix (Miller, 2012). The H2O2 and
its derivative lipid hydroperoxide are further converted into
non-reactive metabolic products by selenium-dependent

Cell Reports 38, 110566, March 29, 2022 ª 2022 The Authors. 1
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
 ll
OPEN ACCESS
antioxidant enzymes such as glutathione peroxidases (GPX),
namely GPX1 and GPX4 (Marı́ et al., 2009).
Although ROS are often described as deleterious to various
macromolecules, they also act as critical signaling molecules
to provoke particular beneficial cellular adaptation in response
to various stress conditions (Takamura, 2020). Notably, ROS
increase mitochondrial proton conductance by activating
mitochondrial UCPs, which may be a mechanism for decreasing
the ROS concentrations inside the mitochondria (Echtay et al.,
2002). Chouchani et al. (2016) reported that acute BAT activation
substantially accompanies mitochondrial ROS formation
(hydrogen peroxide, lipid hydroperoxides, and superoxide).
This formation, in turn, alters the redox status of the cysteine
253 thiol residue of UCP1 protein, resulting in increased NA
sensitivity and accelerated thermogenesis. Furthermore,
antioxidant sentrin 2 overexpression in adipose tissues inhibits
both basal and cold-induced UCP1 expression, compromising
thermogenesis in mice (Ro et al., 2014). Conversely, adipose
tissue deletion of antioxidant enzyme SOD2 protects the
animals from diet-induced obesity by activating mitochondrial
thermogenesis (Han et al., 2016). These factors indicate that
ROS may activate thermogenesis either by direct redox modifi-
cation of UCP1 or by altering the thermogenic gene expression.
Selenoprotein P, denoted as SeP (in humans, encoded by
SELENOP, previously known as SEPP1), is a hepatokine. Its
primary function is antioxidation or redox regulation, either by
its direct enzymatic action or by supplying selenium to various
cells to synthesize other selenoproteins (Burk and Hill, 2009;
Saito and Takahashi, 2002). Although the liver is the main
contributor to circulating SeP (Schweizer et al., 2005), many
other tissues locally synthesize SeP to some extent (Hoffmann
et al., 2007). Previous studies reported that SeP causes the
pathology of type 2 diabetes, such as insulin resistance in the
liver and skeletal muscle (Misu et al., 2010), hypoadiponectine-
mia (Misu et al., 2012), impaired angiogenesis (Ishikura et al.,
2014), insulin secretory failure in the pancreatic b cells (Mita
et al., 2017), exercise resistance in the skeletal muscle (Misu
et al., 2017), and exacerbation of myocardial infarction
(Chadani et al., 2018). Physiological ROS are required for cellular
signal transduction, whereas excessive ROS-mediated oxida-
tive stress causes diseases. SeP mitigates the physiological
ROS burst required for intracellular signal transduction and
thereby causes resistance to various signals, the condition of
which may be referred to as the ‘‘reductive stress’’ (Takamura,
2020).
Unexpectedly, our previous study revealed that Selenop-
deficient mice have higher basal energy expenditure under
a high-fat, high-sucrose diet (Misu et al., 2010), indicating the
accelerated thermogenesis by Selenop deficiency. We
hypothesized that the antioxidant action of SeP might impair
BAT thermogenesis by interfering with cold-induced mitochon-
drial ROS formation and subsequent signaling. This study
investigated the role of SeP in cold- and NA-induced mitochon-
drial ROS formation, UCP1 activation, and thermogenesis in
BAT. We present evidence of previously underrecognized
reductive stress in energy homeostasis. SeP may be an
intrinsic factor inducing reductive stress, a potential therapeutic
target for obesity and diabetes.
2 Cell Reports 38, 110566, March 29, 2022
Article
RESULTS
Serum levels of SeP predict impaired BAT activity in
humans
We previously observed that Selenop-deficient mice have higher
basal energy expenditure under a high-fat, high-sucrose diet
(Misu et al., 2010), which led us to hypothesize that SeP might
impair BAT activity. To confirm this hypothesis in humans, we
assayed serum levels of SeP and investigated their association
with BAT activity. We analyzed the sera of 87 healthy male
subjects with high BAT activity (maximum standardized uptake
value [SUVmax] > 3, n = 43) and low BAT activity (SUVmax % 3,
n = 44) at thermoneutral (27 C) and following cold exposure
(19 C) (Table S1). Subjects with SUVmax over 3, the subjects’
median, were considered the high BAT activity group. BAT
activity inversely correlated with serum SeP levels (Figure 1A),
age (Figure 1B), weight (Figure 1C), serum alanine transaminase
(ALT) levels (Figure 1D), and serum total cholesterol levels (Fig-
ure 1E) in the high BAT activity group. Serum SeP levels posi-
tively correlated with age (Figure 1F) and serum total cholesterol
levels (Figure 1G) in low and high BAT activity groups. In the
multivariate regression analysis, the subject’s weight contrib-
uted most to higher BAT activity followed by age and serum
SeP levels (Table 1) in the high BAT activity group. Because
SeP levels positively correlated with age (Figure 1F) (Oo et al.,
2018), the human data alone do not clarify the direct role of
SeP in thermogenesis independent of aging. Therefore, we
investigated the function of SeP in adaptive thermogenesis by
establishing systemic and BAT-specific Selenop-deficient mice.
Deficiency in Selenop enhances UCP1 oxidation and
thermogenesis in mice
We examined the systemic Selenop-deficient mice under acute
cold exposure at 4 C for 2 h after overnight fasting. Following
acute cold exposure, the BAT and rectal temperatures of
the systemic Selenop-deficient mice were higher, with lower
post-exposure blood glucose levels, than those of wild-type
(WT) littermates (Figures 2A–2C). There were no differences in
sympathetic actions such as urinary NA, serum free fatty acid
(FFA), and shivering between the genotypes (Figures S1A–
S1C). However, the oxygen consumption rate (OCR) after
10 min of NA stimulation was significantly higher in the systemic
Selenop-deficient mice than in WT littermates (Figure 2D).
Therefore, systemic Selenop-deficient mice have a higher
sensitivity to NA in their BAT.
Next, we addressed whether reduced antioxidant capacity is
related to the phenotype of systemic Selenop-deficient mice.
We treated mice with antioxidant N-acetylcysteine (NAC) at
a dose of 1.25 mmol/kg body weight for 6 h before cold
exposure. The NAC pretreatment completely canceled the
cold- and glucose-tolerant phenotypes of systemic Selenop-
deficient mice (Figures 2E–2G). Indeed, BAT of systemic
Selenop-deficient mice tended to be higher in the oxidative
stress marker thiobarbituric acid reactive substances (TBARS)
than that of WT littermates after acute cold exposure (Fig-
ure 2H). Because ROS are responsible for sulfenylation and
activation of the UCP1 protein (Chouchani et al., 2016), we
evaluated the UCP1 sulfenylation status in BAT of mice. BAT
 ll
Article OPEN ACCESS

A B C D E
30 r=-0.310, p=0.043 30 r=-0.388, p=0.010 30 r=-0.435, p=0.004 30 r=-0.406, p=0.007 30 r=-0.323, p=0.034

25 25 25 25 25

20 20 20 20 20

SUVmax
SUVmax
SUVmax
SUVmax
SUVmax

15 15 15 15 15

10 10 10 10 10

5 5 5 5 5

0 0 0 0 0
2.5 3.0 3.5 4.0 4.5 10 20 30 40 50 40 60 80 100 120 0 20 40 60 80 100 120 100 150 200 250 300
Serum SeP (mg/L) Age Weight (kg) ALT Total cholesterol

F G
4.5 4.5
High BAT subject
Low BAT subject
Serum SeP (mg/L)
Serum SeP (mg/L)

4.0 4.0

3.5 3.5

3.0 3.0

r=0.301, p=0.050 r=0.385, p=0.011

2.5 r=0.344, p=0.022 2.5 r=0.557, p<0.001
10 20 30 40 50 100 150 200 250 300
Age Total cholesterol

Figure 1. Serum levels of SeP predict impaired BAT activity in humans

(A–E) Scatterplots showing the correlation of brown adipose tissue (BAT) activity and serum SeP concentrations (A), age (B), weight (C), serum ALT levels (D), and
serum total cholesterol levels (E) in high BAT or low BAT activity groups of human subjects who were divided by SUVmax = 3, which is the median of the subjects
(n = 43 subjects with high BAT, n = 44 subjects with low BAT activity).
(F, G) Scatterplots showing the correlation of serum SeP concentrations with age (F) and serum total cholesterol levels (G) in high BAT or low BAT activity groups
of human subjects who were divided by SUVmax = 3, which is the median of the subjects (n = 43 subjects with high BAT, n = 44 subjects with low BAT activity). The
statistical significances were calculated by Pearson correlation (A–G).

of systemic Selenop-deficient mice showed high accumulation
of sulfenylated UCP1 under cold exposure, which was
abolished with NAC pretreatment (Figures 2I, 2J, and S2).
There was no significant difference in total UCP1 protein level
(Figure 2K) or in Ucp1 mRNA level in BAT between genotypes
(Figure 2L). These findings suggest that SeP suppresses
adaptive thermogenesis by eliminating the ROS required for
sulfenylation and activation of UCP1 without altering Ucp1
expression in BAT.
SeP suppresses noradrenaline-induced ROS
generation, thermogenesis, and glucose uptake in
brown adipocytes
To investigate the detailed mechanisms underlying impaired ther-
mogenesis by SeP in BAT, we treated primary brown adipocytes
with purified human SeP (hSeP) protein. First, we evaluated hSeP
action on redox status during NA stimulation. The whole-cell ROS
Table 1. Multivariate regression analysis of SUVmax in BAT-
positive group as a dependent variable
Variable SE
Age 0.097
Weight (kg) 0.077
FL-SeP (mg/L) 1.722
n = 43. FL-SeP, full-length selenoprotein P.
a
p < 0.05.
indicator dihydroethidium (DHE) revealed that NA treatment
elevated ROS generation in the primary brown adipocytes, which
was canceled by hSeP pretreatment (Figure 3A). hSeP adminis-
tration elevated the protein level of GPX4, which is known for a
mitochondrial form, but not GPX1 and selenoprotein W, in the pri-
mary brown adipocytes (Figures 3B and S3). The mitochondrial
complex II inhibitor (TTFA) and complex III inhibitor (antimycin A)
suppressed the NA-induced ROS generation (Figure 3C), which
confirmed the mitochondrial oxidative phosphorylation (OX-
PHOS) origin of NA-induced ROS generation. Thus, we focused
on mitochondria-derived ROS as the target of SeP. Similar to
Figure 3A, NA-stimulated mitochondrial ROS generation was
abolished by hSeP pretreatment (Figure 3D). We then assessed
SeP action on heat generation in primary brown adipocytes using
a cellular thermoprobe, a cationic fluorescence polymeric ther-
mometer for measuring intracellular temperature (Uchiyama
et al., 2017). NA administration stimulated primary brown adipo-
cyte thermogenesis up to 3%, whereas hSeP pretreatment
completely canceled it (Figure 3E). The hSeP pretreatment also
canceled the NA-induced glucose uptake into primary brown ad-
ipocytes (Figure 3F). The hSeP-mediated suppression of glucose
Variance inflation
p value factor uptake was associated with reduced phosphorylation of AMPK
and mTOR (Figures 3G, 3H, and 3I).
0.052 1.116
0.002a 1.018
SeP is taken up by the LRP1 receptor and impairs
0.069 1.108
thermogenesis via GPX4
We previously identified low-density lipoprotein (LDL) receptor-
related protein 1 (LRP1) as the SeP receptor in skeletal muscle

Cell Reports 38, 110566, March 29, 2022 3

 ll
OPEN ACCESS Article

A B C
39 250
WT *
38 **
Selenop-/-

Blood glucose (mg/dl)

Rectal temperature (°C)
WT 200
37

Selenop-/- 36 * ** 150
***
35
34 100

33
50
32
0 0
0 30 60 90 120 Pre Post Pre Post
Time of cold exposure (min)
WT Selenop-/-

D E 39 F 250
2500 WT+NAC
** 38
** ** Selenop-/-+NAC
Rectal temperature (°C)

Blood glucose (mg/dl)

2000 * 200
37
OCR (ml/kg/hr)

36
1500 150
35
1000 WT 34 100
Selenop-/-
33
500 50
32
Baseline ↑ NA injection
0 0 0
-28 -18 -10 0 8 18 26 36 44 0 30 60 90 120 Pre Post Pre Post
Time after NA injection (min) Time of cold exposure (min)
WT+NAC Selenop-/-+NAC

G H I
WT+NAC p=0.08 Room
2500 4°C
Selenop-/-+NAC 40 Temperature
TBARS (μM/ mg protein)

2000 p -/-

p -/-

AC -/-
OCR (ml/kg/hr)

+N nop
30
no

no
le

le

le
T

1500 T
Se

Se

Se
W

20
1000 Sulfenylated UCP1

10
500
Baseline ↑ NA injection Native UCP1
0 0
-28 -18 -10 0 8 18 26 36 44 WT Selenop-/-
Time after NA injection (min)

J K L n.s.
Relative intensity of sulfenylated UCP1

6
* * *
Relative intensity of total UCP1

5 1.5
Relative Ucp1 mRNA levels

2.0 **
normalized to UCP1

normalized to Actb

4
1.5
1
3
1.0
2
0.5
0.5
1

0 0 0
p -/-

p -/-

AC -/-

p -/-

-/-
-/-

-/-

AC -/-

T
T

+N nop

p
op

op

+N op

W
W

no

no

no

no
n

le

le

le

le

le
le

le

le

Se

Se

Se

Se

Se
Se

Se

Se

Room Room Room 4°C

4°C 4°C
temperature temperature temperature

Figure 2. The deficiency of Selenop enhances UCP1 sulfenylation and thermogenesis

(A) Representative image of interscapular temperature by the thermal camera in WT and systemic Selenop-deficient mice after acute cold exposure.
(B) Rectal temperature during acute cold exposure in WT and Selenop-deficient mice (n = 11 mice for WT; n = 8 mice for Selenop
/
).
(C) Blood glucose levels before and after cold exposure in WT and systemic Selenop-deficient mice (n = 11 mice for WT; n = 8 mice for Selenop
/
).
(D) Oxygen consumption rate (OCR) after noradrenaline stimulation in WT and systemic Selenop-deficient mice (n = 6 mice per group).
(E) Rectal temperature during cold exposure in WT and systemic Selenop-deficient mice pretreated with N-acetylcysteine (NAC) for 6 h (n = 8 mice per group).
(F) Blood glucose levels before and after cold exposure in WT and systemic Selenop-deficient mice pretreated with NAC (n = 8 mice per group).
(G) OCR after noradrenaline stimulation in WT and systemic Selenop-deficient mice pretreated with NAC (n = 4 mice for WT; n = 8 mice for Selenop
/
).
(H) TBARS levels in brown fat of WT and systemic Selenop-deficient mice after cold exposure (n = 5 mice for WT; n = 6 mice for Selenop
/
).

(legend continued on next page)

4 Cell Reports 38, 110566, March 29, 2022


Article
(Misu et al., 2017). Also, LRP8 is the SeP receptor in the brain
(Burk et al., 2007) and testis (Olson et al., 2007) and LRP2 is
the receptor in the kidney (Olson et al., 2008). Therefore, we
hypothesized that one of the LDL receptor family members
acts as the SeP receptor in BAT. Among LDL receptor family
members, Lrp1 was expressed predominantly in primary brown
adipocytes (Figure 4A). The knockdown of Lrp1 in primary brown
adipocytes almost entirely reduced hSeP uptake into the cells
(Figures 4B and S4A). These findings indicate that the LRP1 re-
ceptor is required for SeP uptake in BAT.
To identify the downstream effector of hSeP, we analyzed
protein expression in the cytosol and mitochondrial fractions of
brown adipocytes after hSeP treatment. The hSeP treatment
significantly increased GPX4 in the cytosol and mitochondrial
fractions (Figure 4C). To confirm the mediator of hSeP on ROS
production, we performed Gpx1 and Gpx4 knockdown in
primary brown adipocytes (Figures S4B–S4E). The knockdown
of Gpx4, but not Gpx1, rescued the hSeP-induced impairment of
the NA-induced ROS production in the primary brown
adipocytes (Figure 4D). In addition, we treated the primary brown
adipocytes with a selective GPX4 inhibitor, 1S, 3R RSL3 (RSL3).
The cell viability assay suggests slight but significant cell death
with 1 nM RSL3 treatment after 1 h (Figure S4F). However, the
thermoprobe assay targets individual cells adjusted with control
fluorescence signal inside the cells, instead of the whole culture
well. Therefore, the slight decline in cell numbers does not bias
the cellular thermogenesis data. The RSL3 treatment at 1 nM did
not affect the basal cellular temperature (Figures S4G and S4H),
but enhanced the NA-induced thermogenesis of primary brown
adipocytes (Figures 4E and 4F). These findings indicate that SeP
is taken up into the brown adipocytes via its receptor Lrp1 and
upregulates GPX4, which eliminates the NA-induced generation
of ROS, thereby inducing resistance to the NA-induced
thermogenesis and glucose uptake in primary brown adipocytes.
Deficiency of Selenop in brown fat, but not in the liver,
induces cold resistance in mice
Although the liver is a major SeP-producing organ, we found that
BAT also expresses the Selenop gene to some extent (Figure 5A).
To identify the primary source of SeP responsible for impaired
BAT thermogenesis, we generated liver-specific Selenop-defi-
cient (L-Selenop KO) mice using Alb-Cre and BAT-specific Sele-
nop-deficient (BAT-Selenop KO) mice using Ucp1-Cre (Figures
S5A and S5B). There were no significant differences in the
body weight, fasting plasma glucose, or glucose tolerance in
conditional knockout mice compared with the floxed mice under
normal chow feeding (Figures S5C–S5H). Unexpectedly, there
was no difference in the rectal temperature during acute cold
(I and J) Representative western blot images (I) and densitometric analyses (J) of UCP1 sulfenylation in the BAT of WT and systemic Selenop-deficient mice after
cold exposure, with or without NAC pretreatment (n = 3 mice per group).
(K) Densitometric analyses of total UCP1 in the BAT of WT and systemic Selenop-deficient mice after cold exposure with or without NAC pretreatment (n = 3 mice
per group).
(L) Ucp1 gene expression in the BAT of WT and systemic Selenop-deficient mice after cold exposure (n = 8 mice per group). In boxplots (H, J, K, L), center lines
show the medians, and box limits indicate the 25th and 75th percentiles; whiskers extend 1.53 the interquartile range from the 25th and 75th percentiles; data
points are plotted as dots. In (B), (D), (E), and (G), data are represented as the mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001; n.s., not significant; by two-tailed
unpaired Student’s t test (B, D, E, G, H), by two-tailed paired Student’s t test (C, F), or by one-way analysis of variance (ANOVA) with Tukey’s post hoc test (J, K, L).
Full images are shown in Figure S2.
ll
OPEN ACCESS
exposure between the L-Selenop KO mice and the floxed mice
(Figure S5I). On the other hand, the rectal temperature tended
to be higher in the BAT-Selenop KO mice than in the floxed
mice during acute cold exposure (Figure 5B).
We measured the BAT and rectal temperatures directly during
a 0.2 mg/kg subcutaneous NA injection to mimic the cold-
induced thermogenesis. After NA injection, the BAT-Selenop
KO mice showed higher BAT and rectal temperatures than the
floxed mice fed normal chow (Figures 5C and 5D). Notably,
a high-fat/high-sucrose (HFHS) diet fed for 16 weeks
canceled the enhancement of NA-induced thermogenesis in
the BAT-Selenop KO mice (Figures 5E and 5F). Unexpectedly,
under this condition, BAT-Selenop KO mice exhibited elevated
serum levels of SeP with no significant difference between the
genotypes. These results suggest that organs other than BAT
contribute to SeP overproduction under HFHS feeding. Based
on these findings, we conclude that locally produced SeP is
mainly involved in the fine-tuning of BAT thermogenesis under
normal chow conditions. Excess SeP, induced by HFHS chow,
impairs BAT thermogenesis.
We then investigated the molecular signature in BAT and
liver tissues after cold exposure at 4 C for 2 days. After cold
exposure, the Selenop expression was significantly decreased in
BAT but not in the liver (Figure 5H). In addition, NA stimulation
reduced the expression of both Selenop and Gpx4 in primary
brown adipocytes (Figure 5I). Treatment with an adenylate cyclase
activator, forskolin (10 mM for 24 h), that mimics the NA action
showed similar results in primary brown adipocytes (Figure 5J),
but not in primary hepatocytes (Figure S5J). On the other hand,
the Selenop expression was increased in the liver but not in the
BAT muscle and subcutaneous WAT under HFHS diet feeding (Fig-
ure S5K). These results indicate that the regulation of Selenop
gene expression is variable among organs. Cold exposure and
NA stimulation downregulate Selenop specifically in BAT, which
might be an adaptation to cold stress to enhance thermogenesis.
Locally circulating SeP suppresses noradrenaline-
induced ROS generation and thermogenesis in brown
adipocytes
We found that selenoprotein family member genes, such as
Selenop, Gpx1, Gpx4, and Selenow, were expressed in
primary brown adipocytes (Figure 6A). To further confirm the
role of locally expressed Selenop in BAT, we examined
Selenop-deficient primary brown adipocytes that were
differentiated from the stromal vascular fraction cells of systemic
Selenop-deficient mice (Figure S6A).
Selenop deficiency potentiated insulin-induced Akt phosphory-
lation in the primary brown adipocytes (Figure S6B). The
Cell Reports 38, 110566, March 29, 2022 5
 ll
OPEN ACCESS Article

A B C
p=0.067 3
hSeP

Oxidized/reduced DHE
0.30 Vehicle 10 μg/ml

(relative changes)
** n.s.
hSeP
Oxidized/reduced DHE

0.25 2
(BD1) *
0.20 *
GPx4
0.15 1

GPx1
0.10
0
0.05 SeW 400 nM NA - + - + - + - +
0 β-Actin

V)
II)

pl n A
l

III
pl FA

pl cin
tro

ex

ex
- -

ex
om TT
+ +

om y
on

om yci
400 nM NA

(C om
C

(C tim

lig
- - + +

An
10 μg/ml hSeP

(C

O
D E F
* * PBS pretreat hSeP pretreat 4 n.s.
1.2 1.04 ***
Fluorescent ratio (fold change)

Glucose uptake from baseline

*
1.03 *** *
1.0 *
Mitochondrial ROS

(relative changes)
1.02
(relative change)

0.8 1.01
0.6 1.00 2
0.99
0.4
0.98 1
0.2 0.97 Vehicle
0.96 ↑ NA ↑
0 stimuli stimuli 0
0.95
1000 nM NA - + + -6 -2 2 6 10 14 18 -6 -2 2 6 10 14 18 1000 nM NA - + - +
10 μg/ml hSeP - - + Baseline (min) Baseline (min) 10 μg/ml hSeP - - + +

G H I
2.5 2.0
PBS hSeP

Relative intensity of p-mTOR

Relative intensity of p-AMPK

**

normalized to t-mTOR
PBS NA PBS NA 2.0
normalized to t-AMPK

p-mTOR 1.5 0.0959

(Ser 2481) n.s.
1.5
n.s.
t-mTOR 1.0
1.0
p-AMPK
(Thr 174) 0.5
0.5
t-AMPK
0 0
NA - + - + NA - + - +
PBS hSeP PBS hSeP

Figure 3. SeP suppresses noradrenaline-induced ROS generation, thermogenesis, and glucose uptake in brown adipocytes
(A and B) Effects of hSeP on superoxide-dependent oxidation of DHE (A; n = 6) and selenoprotein production in primary brown adipocytes (B).
(C–E) Effects of OXPHOS inhibitors on NA-induced ROS generation (n = 6). Effects of hSeP on mitochondrial ROS generation (D, n = 11) and intracellular
temperature (E, n = 23–27 cells each) in NA-stimulated primary brown adipocytes.
(F) Effects of hSeP on glucose uptake in noradrenaline-stimulated primary brown adipocytes (n = 7 wells each).
(G) Western blots of phosphorylated mTOR and AMPK in NA-stimulated primary brown adipocytes pretreated with hSeP.
(H and I) Densitometric analyses of western blots of phosphor-AMPK (H) and phosphor-mTOR (I) in NA-stimulated primary brown adipocytes pretreated with
hSeP. In boxplots (A, C, D, F, H, I), center lines show the medians, and box limits indicate the 25th and 75th percentiles; whiskers extend 1.53 the interquartile
range from the 25th and 75th percentiles; data points are plotted as dots. *p < 0.05, **p < 0.01, ***p < 0.001; n.s., not significant by one-way ANOVA with Tukey’s
post hoc test (A, C, D, F, H, I). Full images are shown in Figure S3.

Selenop-deficient primary brown adipocytes showed higher NA- DISCUSSION

induced ROS production (Figure 6B) and thermogenesis (Figures
6C and 6D) than the control brown adipocytes. Extracellular flux The current study found that SeP derived from BAT, rather than
analysis indicated enhanced oxygen consumption in Selenop- the liver, principally impairs BAT thermogenesis and glucose
knockdown primary brown adipocytes (Figure S7). Conversely, tolerance during physiological cold exposure or NA stimulation
pretreatment with sodium selenite abolished the enhanced NA- in mice. Remarkably, the regulation of Selenop expression varied
induced ROS generation (Figure 6E) and thermogenesis (Figures between BAT and liver in mice. Cold exposure and NA signaling
6F and 6G) in the Selenop-deficient primary brown adipocytes. downregulate Selenop expression in BAT but not in the liver. On
Finally, we investigated the significance of SeP in selenium- the other hand, the HFHS diet challenge to mice upregulates
induced toxicity in BAT. The Selenop-deficient primary brown Selenop expression in the liver but not in BAT. These findings
adipocytes showed intolerance to higher concentrations of sele- suggest that Selenop expression is locally regulated and
nium exposure (Figure 6H), suggesting that SeP protects BAT plays a specific role in each organ. BAT may downregulate
from selenium toxicity by buffering excess selenium exposure. endogenous Selenop expression and elevate NA sensitization

6 Cell Reports 38, 110566, March 29, 2022

 ll
Article OPEN ACCESS

A B C
Veh hSeP
NC siRNA Lrp1 siRNA

Cyto
Cyto
Mito

Mito
2.0
PBS hSeP PBS hSeP
normalized to Actb

1.5 hSeP
mRNA levels

(BD1)
LRP1
1.0
GPx4
0.5 hSeP
(BD1) Cyt c
0
Lrp1 Ldlr Lrp2 Lrp8 β-Actin β-Actin

D E F
1.5 1.10
n.s. n.s.

Fluorescent ratio (fold change)

*** * 2.0 ***
1.08
**
Mitochondrial ROS
(relative change)

1.06 1.5
1.0

AUC0-18min
1.04 *
1.0
1.02
0.5
1.00 0.5
↑ Vehicle
0.98 stimuli NA 0
0 0.96 NA+RSL3

le

NA

3
- + - + - + - +

SL
hic
1000 nM NA 0.94

+R
- -

Ve
10 μg/ml hSeP + + + + + + -6 -2 2 6 10 14 18

NA
Baseline (min)
NC Gpx1 Gpx4
siRNA siRNA siRNA

Figure 4. SeP is taken up by the LRP1 receptor and impairs thermogenesis via GPX4
(A) Gene expression levels of LDLR family members in primary brown adipocytes (n = 3 samples each).
(B) Representative western blots of hSeP uptake in Lrp1-knockdown primary brown adipocytes.
(C) Western blots of cytosol and mitochondrial fraction proteins after hSeP treatment in primary brown adipocytes.
(D) Effects of knockdown of Gpx1 or Gpx4 on hSeP-induced inhibition of mitochondrial ROS production in NA-stimulated primary brown adipocytes (n = 8 wells in
each group).
(E and F) Effects of 1S, 3R RSL3 pretreatment on the cellular temperature of NA-stimulated primary brown adipocytes (n = 50 cells in each group). Fold change of
fluorescence ratio (E) and area under the curve (AUC) of 0–18 min (F). In boxplots (A and D), center lines show the medians, and box limits indicate the 25th and
75th percentiles; whiskers extend 1.53 the interquartile range from the 25th and 75th percentiles; data points are plotted as dots. In bar graph (F), data are
represented as the mean ± SEM of biologically independent samples. *p < 0.05, **p < 0.01, ***p < 0.001; n.s., not significant by one-way ANOVA with Tukey’s post
hoc test (A, D, F).

in response to cold exposure. HFHS diet feeding canceled the
enhanced NA-induced BAT thermogenesis in BAT-Selenop KO
mice. This result suggests the contribution of SeP overproduc-
tion to lower BAT activity in diet-induced obesity. Indeed, the
higher the serum levels of SeP, the lower is the BAT activity in
humans. Therefore, it might be possible that reduction of SeP
increment in metabolic disorders is a therapeutic target in
obesity and type 2 diabetes.
Due to their high reactivity, ROS are considered harmful, as they
oxidize and damage various biomolecules. Indeed, lipid-induced
excessive ROS production mediates insulin resistance
(Matsuzawa-Nagata et al., 2008; Nakamura et al., 2009) and
non-alcoholic steatohepatitis (Matsuzawa et al., 2007; Ota et al.,
2007). On the other hand, ROS transiently generated from NADPH
oxidase enhance signal transduction via receptor tyrosine kinases
by inactivating phosphatases such as PTP1B and PTEN
(Chiarugi and Cirri, 2003). We have previously found that SeP is
overproduced in type 2 diabetic liver and causes various signal
disturbances by eliminating intracellular physiological ROS
required for signal transductions, the condition referred to as the
‘‘reductive stress’’ (Takamura, 2020). Interestingly, a series of re-
ports indicate that appropriate ROS stimulation activates UCP1
and enhances heat production in BAT (Chouchani et al., 2016,
2017; Mills et al., 2018). In the current study, we further proved
that SeP regulates BAT thermogenesis in mice by eliminating the
mitochondrial ROS produced during NA signal transduction,
thereby impairing UCP1 sulfenylation. Based on these findings,
we conclude that SeP is the intrinsic factor inducing reductive
stress that causes resistance to intracellular signal transduction.
SeP exerts an antioxidative property directly through its
selenocysteine residue and indirectly by supplying selenium to
intracellular antioxidant GPX (Saito et al., 1999; Takebe et al.,
2002). The present study identified GPX4, but not GPX1 or
selenoprotein W, as the SeP-targeted downstream effector
that inhibits the NA-induced mitochondrial ROS production
and heat production in primary brown adipocytes. SeP
upregulates the mitochondria-localizing GPX4 that eliminates
the NA-induced generation of ROS from mitochondria.
Selenium is an essential trace element, but is highly toxic at
higher doses (MacFarquhar, 2010). In the body, selenium mainly
exists in incorporated form as selenoproteins (Yang et al., 2017).
We hypothesized that storing selenium in the form of SeP in the
interstitial fluid is safer than storing free selenium intracellularly.
In the present study, the Selenop-deficient primary brown

Cell Reports 38, 110566, March 29, 2022 7

 ll
OPEN ACCESS Article

A B C D
1.2 39 Selenopfl/fl
normalized to Gapdh (fold change)

Selenop fl/fl 3000

38 BAT-Selenop -/- ***
Selenop mRNA expression

1.0 39.0 BAT-Selenop -/-

Rectal temperature (°C)

37

AUC0-25min
38.5 2000
0.8

Temperature (°C)
36
p=0.074 38.0
0.6 35
1000
34 37.5
0.4
33 37.0
0
0.2
32 36.5

-/-
fl
fl/
↑ NA injection op op
0 0 36.0 l en l en
Se Se
0 30 60 90 120 -5 0 5 10 15 20 25 T-
BAT Liver Muscle sWAT
Time of cold exposure (min) Time (min) BA

E F G Serum SeP
3000 3

Absorbance (OD)
Selenop fl/fl n.s.
40 BAT-Selenop -/- 2
AUC0-25min

2000
39
Temperature (°C)

1
38 1000
0
37

-/-
-/-

l
l

fl/f
fl/f

p
p

op
op

no
no

36
len
len
-/-
fl

ele
ele

op
fl/

↑ NA injection op
Se
en
Se

en
T -S
T-S

l l
35
Se Se
BA
T-
BA

-5 0 5 10 15 20 25
BA
Time (min)
Before After
HFHS HFHS

H I J
normalized to Actb (fold change)

2.0 2.0 1.5 1.5

n.s.
Relative mRNA expression

Relative mRNA expression

Selenop mRNA expression

*** ***
normalized to Actb

normalized to Actb

1.5 *** 1.5 *** ***

1.0 1.0

1.0 1.0
0.5 0.5
0.5 0.5

0.0 0.0
0 0
30°C 4°C 30°C 4°C NA - + - + Forskolin - + - +
BAT Liver Selenop Gpx4 Selenop Gpx4

Figure 5. Deficiency of Selenop in the brown fat, but not in the liver, induces cold resistance in mice
(A) Selenop mRNA expression in various tissues of C57BL/6J mice (n = 4–5 mice per group).
(B) The rectal temperature during cold exposure in the BAT-specific Selenop-deficient mice (n = 7 mice per group).
(C and D) BAT temperature after subcutaneous NA injection in the BAT-specific Selenop-deficient mice (C; n = 8 mice per group) and AUC of 0–25 min (D).
(E and F) BAT temperature after subcutaneous NA injection in the BAT-specific Selenop-deficient mice after HFHS feeding (E; n = 8 mice per group) and AUC of 0–
25 min (F).
(G) Circulating SeP levels before and after HFHS feeding.
(H) Selenop expression in BAT and liver tissue of WT mice after 2 days of chronic cold exposure at 4 C (n = 12 mice in 4 C group and 15 mice in 30 C group).
(I and J) Selenop and Gpx4 expression in primary brown adipocytes after NA treatment (I) and after forskolin treatment (J) (n = 3 per group). The 800 nM NA
medium for 6 h or 10 mM forskolin medium for 24 h was supplied to fully differentiated primary brown adipocytes. In boxplots (A, G, H, I, J), center lines show the
medians, and box limits indicate the 25th and 75th percentiles; whiskers extend 1.53 the interquartile range from the 25th and 75th percentiles; data points are
plotted as dots. In (B), (D), and (F), data are represented as the mean ± SD. In (C) and (E), data are represented as the mean ± SEM. ***p < 0.001; n.s., not
significant by two-tailed unpaired Student’s t test (B, D, F, H, I, J).

adipocytes were intolerant to higher concentrations of selenium locally produced in BAT, pools selenium in the extracellular
exposure. These findings suggest that locally expressed SeP interstitial fluid, is taken up via LRP1-mediated endocytosis, is
protects BAT from selenium toxicity by buffering excess transferred to the lysosome, supplies selenium into the cells for
selenium. We identified Lrp1 as an uptake receptor for SeP the on-demand synthesis of GPX4, and thereby fine-tunes
in brown adipocytes. We interpret this to mean that SeP is ROS-mediated excessive signal transduction.

8 Cell Reports 38, 110566, March 29, 2022

 ll
Article OPEN ACCESS

A B C D
1.08
*

fluorescent ratio (fold change)

2.5
Relative expression of mRNA

***
1.06 3
1.0 **

Mitochondrial ROS
2.0 *
nomalized to Actb

(relative change)
1.04
1.5 2

AUC0-40min
1.02
0.5 1.0
1.00 WT 1
0.5 Selenop-/-
0.98 ↑
0 0 stimuli 0
0
- + - + T

-/-
NA p
-4 0 4 8 12 16 20 24 28 32 36 40 W no
p

w
no

px

px

le
no

Se
G

G
le

WT Selenop-/-
le

Time after NA stimulation (min)

Se

Se

E F 1.08 G
fluorescent ratio (fold change)

2.0 n.s.
n.s. 1.06 2.5 n.s.
Mitochondrial ROS

n.s.
(relative change)

1.5
1.04 2.5

AUC0-40min
1.0 1.5
1.02
1.0
1.00
0.5
↑ 0.5
WT+SS
0.98 stimuli
Selenop-/-+SS
0 0
NA - + - + 0
SS S
-4 0 4 8 12 16 20 24 28 32 36 40 T+ +S

-/-
WT+SS Selenop-/-+SS Time after NA stimulation (min) W
n op
le
Se

H I
Cell viability (relative change)

1.5
n.s. Diabetes
Se
Se Se
Se SeP Se
Se
Se Se
Cold exposure
WT
Se
Se Se
Se Se Se
Se SeP Se
Se Se
Se Se Se Se
Se SeP Se Se Se

n.s. Selenop-/-
Se Se
Se Se

Se Se
1.0 Se
Se SeP Se
Se

Sympathetic nerve
Se Se
Se Se Se Se
Se Se
Se Se
Se SeP Se Se Se
Se SeP Se
Se Se

** ** Autocrine/
Se Se Se
Se Se
Se
Noradrenaline
0.5 Paracrine LRP1

Se Se Se
Se Se
Se SeP Se
Se Se Se Se Se
Se GPX4 mtROS
Se Se
0 Se
Se SeP Se
Se
Se
Se
SOH
Sodium selenite 0 μM 20 μM 40 μM 80 μM Se
Se Se
Se

Se
Se Se
Se UCP1
Se SeP Se
Se
Se Se
Se
Se Se
Sulfenylation
Se Se P P
Se SeP Se
Se Se
AMPK mTOR
Se Se

Heat production Glucose uptake

Obesity/Hyperglycemia

Figure 6. Locally circulating Selenop suppresses noradrenaline-induced ROS generation and thermogenesis in brown adipocytes
(A) Gene expression of selenoprotein family in primary brown adipocytes (n = 3).
(B–D) Effects of NA stimulation on WT and Selenop-deficient primary brown adipocytes assessed by MitoSOX (B; n = 8 wells per group), cellular temperature (C;
n = 50 cells per group), and AUC of 0–40 min (D).
(E–G) Effects of NA stimulation on WT and Selenop-deficient primary brown adipocytes with sodium selenite pretreatment as assessed by MitoSOX (E; n = 8 wells
per group), cellular temperature (F; n = 56 cells per group), and AUC of 0–40 min (G). The 5 mM sodium selenite medium and control medium were supplied to fully
differentiated primary brown adipocytes for 24 h. After that, the cells were used for the experiments.
(H) Cell viability of primary brown adipocytes with various concentrations of sodium selenite treatment (n = 8 wells per group).
(I) Schematic depicting that SeP blunts ROS-regulated thermogenesis in brown fat. In boxplots (A, B, E, H), center lines show the medians, and box limits indicate
the 25th and 75th percentiles; whiskers extend 1.53 the interquartile range from the 25th and 75th percentiles; data points are plotted as dots. In (D) and (G), data
are represented as the mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001; n.s., not significant; by one-way analysis of variance (ANOVA) with Tukey’s post hoc test
(B, E) or by two-tailed unpaired Student’s t test (D, G, H).

Very recently, Jedrychowski et al. reported that the selenited (Bleys et al., 2007). Long-term selenium supplementation of
UCP1 with its cysteine 253 replaced with selenium becomes 200 mg/day increased the cumulative incidence of type 2 dia-
redox sensitive (Jedrychowski et al., 2020). Dietary selenium betes significantly (Stranges et al., 2007). These inconsistent
supplementation in mice elevates energy expenditure findings may be associated with the epidemiological finding
through thermogenic adipose tissue and protects against that the optimal blood range of selenium is narrow and
obesity (Jedrychowski et al., 2020). On the other hand, previous nonlinear in humans (Jedrychowski et al., 2020). Maintaining
papers adduced that high serum selenium levels were selenium levels in the exact sweet spot for health may be chal-
positively associated with the prevalence of diabetes lenging. At least in our present study, BAT-Selenop-deficient

Cell Reports 38, 110566, March 29, 2022 9

 ll
OPEN ACCESS
mice showed enhanced NA-induced thermogenesis under
standard selenium intake, which may not induce UCP1 selena-
tion. Therefore, the current study highlights that SeP causes
reductive stress that impairs UCP1 activation and BAT
thermogenesis under physiological conditions.
In conclusion, SeP-mediated reductive stress impairs cold-
induced sulfenylation of UCP1, adaptive thermogenesis, and
glucose uptake, a condition called catecholamine resistance in
BAT (Figure 6I). SeP, as a hepatokine and BATkine, may be a
therapeutic target for countering obesity and diabetes. Optimal
regulation of excessive SeP may help convert excess energy
to heat by manipulating BAT thermogenesis. Therefore, SeP
may be the intrinsic factor inducing reductive stress that causes
resistance to intracellular signal transduction.
Limitations of the study
In this study, a BAT-specific Selenop-deficient mouse model
was used to investigate the tissue-specific role of SeP in
adaptive thermogenesis. Although we prove the involvement of
local SeP in BAT by directly measuring the BAT temperature in
BAT-specific Selenop knockout mice, additional experiments
such as BAT activity measurements by FDG-glucose uptake
and NA-induced OCRs in flox and knockout mice under normal
chow and HFHS conditions will further strengthen our results.
STAR+METHODS
Detailed methods are provided in the online version of this paper
and include the following:
d KEY RESOURCES TABLE
d RESOURCE AVAILABILITY
B Lead contact
B Materials availability
B Data and code availability
d EXPERIMENTAL MODEL AND SUBJECT DETAILS
B Human study
B Animals
d METHOD DETAILS
B Cold exposure and rectal temperature measurement
B Image of interscapular temperature by a thermal cam-
era in mice
B Oxygen consumption rate in mice
B Thiobarbituric acid reactive substances (TBARS)
B BAT and rectal temperature measurements
B RNA isolation, cDNA synthesis, and real-time PCR
analysis
B ELISA for mouse SeP
B Western blot studies
B Assessment of sulfenylated UCP1 by gel shift
B Purification of SeP
B Primary brown adipocytes
B Measurement of DHE oxidation
B Measurement of mitochondrial ROS by MitoSOX assay
B Measurement of cellular temperature
B Glucose uptake in brown adipocytes
B Extracellular flux analysis
d QUANTIFICATION AND STATISTICAL ANALYSIS
10 Cell Reports 38, 110566, March 29, 2022
Article
SUPPLEMENTAL INFORMATION
Supplemental information can be found online at https://doi.org/10.1016/j.
celrep.2022.110566.
ACKNOWLEDGMENTS
We thank M. Kawamura for technical assistance, Dr. H. Misu for helpful
discussion, and Prof. Shingo Kajimura, Prof. Edward T. Chouchani, and
Prof. Bruce M. Spiegelman for methodological supports. We are indebted to
K.E. Hill and R.F. Burk of Vanderbilt University School of Medicine for the
systemic Selenop-KO mice and Selenopfl/fl mice and Hiroshi Inoue of
Kanazawa University for Alb-Cre mice. This work was supported by the
following grants: JSPS KAKENHI 18K16224 (H.T.) and 17H04199 (T.T.);
Takeda Science Foundation (T.T.); MSD Life Science Foundation, Public Inter-
est Incorporated Foundation (H.T.); Kato Memorial Bioscience Foundation
(H.T.); Japan International Cooperation Agency Project for Enhancement
of Medical Education (J1410304, S.M.O.); MEXT Scholarship Program
(Research Student) for years 2019–2023 (H.K.O.); and JST Adaptable and
Seamless Technology Transfer Program (A-STEP) grants AS2311400F and
15im0302407 (Y.S.).
AUTHOR CONTRIBUTIONS
H.T. and T.T. conceived the study and designed experiments. H.T., K.-A.I.,
Y.T., H.G., Y.N., S.K., H.K.O., S.M.O., Y.S., and M.M. conducted experiments.
H.T., K.-A.I., H.K.O., S.M.O., C.T., H.N., and T.T. analyzed the data and pro-
vided the discussion. H.T., H.K.O., S.M.O., and T.T. wrote the manuscript
with input from the other authors.
DECLARATION OF INTERESTS
The authors declare no competing interests.
Received: May 19, 2021
Revised: February 3, 2022
Accepted: March 3, 2022
Published: March 29, 2022
REFERENCES
Aune, U.L., Ruiz, L., and Kajimura, S. (2013). Isolation and differentiation of
stromal vascular cells to beige/brite cells. J. Vis. Exp. (73), 501961.
Bartelt, A., Bruns, O.T., Reimer, R., Hohenberg, H., Ittrich, H., Peldschus, K.,
Kaul, M.G., Tromsdorf, U.I., Weller, H., Waurisch, C., et al. (2011). Brown
adipose tissue activity controls triglyceride clearance. Nat. Med. 17, 200–206.
Bleys, J., Navas-Acien, A., and Guallar, E. (2007). Serum selenium and dia-
betes in U.S. Adults. Diabetes Care 30, 829–834.
Burk, R.F., and Hill, K.E. (2009). Selenoprotein P-Expression, functions, and
roles in mammals. Biochim. Biophys. Acta Gen. Subj. 1790, 1441–1447.
Burk, R.F., Hill, K.E., Olson, G.E., Weeber, E.J., Motley, A.K., Winfrey, V.P., and
Austin, L.M. (2007). Deletion of apolipoprotein E receptor-2 in mice lowers
brain selenium and causes severe neurological dysfunction and death when
a low-selenium diet is fed. J. Neurosci. 27, 6207–6211.
Chadani, H., Usui, S., Inoue, O., Kusayama, T., Takashima, S.I., Kato, T.,
Murai, H., Furusho, H., Nomura, A., Misu, H., et al. (2018). Endogenous
selenoprotein P, a liver-derived secretory protein, mediates myocardial
ischemia/reperfusion injury in mice. Int. J. Mol. Sci. 19, 878.
Chiarugi, P., and Cirri, P. (2003). Redox regulation of protein tyrosine
phosphatases during receptor tyrosine kinase signal transduction. Trends
Biochem. Sci. 28, 509–514.
Chouchani, E.T., Kazak, L., Jedrychowski, M.P., Lu, G.Z., Erickson, B.K.,
Szpyt, J., Pierce, K.A., Laznik-Bogoslavski, D., Vetrivelan, R., Clish, C.B.,
et al. (2016). Mitochondrial ROS regulate thermogenic energy expenditure
and sulfenylation of UCP1. Nature 532, 112–116.

Article
Chouchani, E.T., Kazak, L., and Spiegelman, B.M. (2017). Mitochondrial
reactive oxygen species and adipose tissue thermogenesis: bridging
physiology and mechanisms. J. Biol. Chem. 292, 16810–16816.
Chouchani, E.T., Kazak, L., and Spiegelman, B.M. (2019). New advances in
adaptive thermogenesis: UCP1 and beyond. Cell Metab. 29, 27–37.
Cypess, A.M., Lehman, S., Williams, G., Tal, I., Rodman, D., Goldfine, A.B.,
Kuo, F.C., Palmer, E.L., Tseng, Y.-H., Doria, A., et al. (2009). Identification
and importance of brown adipose tissue in adult humans. N. Engl. J. Med.
360, 1509–1517.
Echtay, K.S., Roussel, D., St-Plerre, J., Jekabsons, M.B., Cadenas, S., Stuart,
J.A., Harper, J.A., Roebuck, S.J., Morrison, A., Pickering, S., et al. (2002).
Superoxide activates mitochondrial uncoupling proteins. Nature 415, 96–99.
Han, Y.H., Buffolo, M., Pires, K.M., Pei, S., Scherer, P.E., and Boudina, S.
(2016). Adipocyte-specific deletion of manganese superoxide dismutase
protects from diet-induced obesity through increased mitochondrial
uncoupling and biogenesis. Diabetes 65, 2639–2651.
Hill, K.E., Zhou, J., McMahan, W.J., Motley, A.K., Atkins, J.F., Gesteland, R.F.,
and Burk, R.F. (2003). Deletion of selenoprotein P alters distribution of
selenium in the mouse. J. Biol. Chem. 278, 13640–13646.
Hill, K.E., Wu, S., Motley, A.K., Stevenson, T.D., Winfrey, V.P., Capecchi, M.R.,
Atkins, J.F., and Burk, R.F. (2012). Production of selenoprotein P (Sepp1)
by hepatocytes is central to selenium homeostasis. J. Biol. Chem. 287,
40414–40424.
Hoffmann, P.R., Höge, S.C., Li, P.A., Hoffmann, F.W., Hashimoto, A.C., and
Berry, M.J. (2007). The selenoproteome exhibits widely varying, tissue-spe-
cific dependence on selenoprotein P for selenium supply. Nucleic Acids
Res. 35, 3963–3973.
Inokuma, K.I., Ogura-Okamatsu, Y., Toda, C., Kimura, K., Yamashita, H., and
Saito, M. (2005). Uncoupling protein 1 is necessary for norepinephrine-
induced glucose utilization in brown adipose tissue. Diabetes 54, 1385–1391.
Ishikura, K., Misu, H., Kumazaki, M., Takayama, H., Matsuzawa-Nagata, N.,
Tajima, N., Chikamoto, K., Lan, F., Ando, H., Ota, T., et al. (2014). Selenopro-
tein P as a diabetes-associated hepatokine that impairs angiogenesis by
inducing VEGF resistance in vascular endothelial cells. Diabetologia 57,
1968–1976.
Isidor, M.S., Winther, S., Basse, A.L., Petersen, M.C.H., Cannon, B.,
Nedergaard, J., and Hansen, J.B. (2016). An siRNA-based method for efficient
silencing of gene expression in mature brown adipocytes. Adipocyte 5,
175–185.
Jedrychowski, M.P., Lu, G.Z., Szpyt, J., Mariotti, M., Garrity, R., Paulo, J.A.,
Schweppe, D.K., Laznik-Bogoslavski, D., Kazak, L., Murphy, M.P., et al.
(2020). Facultative protein selenation regulates redox sensitivity, adipose
tissue thermogenesis, and obesity. Proc. Natl. Acad. Sci. U S A. 117,
10789–10796.
Lee, P., Greenfield, J.R., Ho, K.K.Y., and Fulham, M.J. (2010). A critical
appraisal of the prevalence and metabolic significance of brown adipose tis-
sue in adult humans. Am. J. Physiol. Endocrinol. Metab. 299, E601–E606.
MacFarquhar, J.K. (2010). Acute selenium toxicity associated with a dietary
supplement. Arch. Intern. Med. 170, 256.
Marı́, M., Morales, A., Colell, A., Garcı́a-Ruiz, C., and Fernández-Checa, J.C.
(2009). Mitochondrial glutathione, a key survival antioxidant. Antioxid. Redox
Signal. 11, 2685–2700.
van Marken Lichtenbelt, W.D., Vanhommerig, J.W., Smulders, N.M.,
Drossaerts, J.M.A.F.L., Kemerink, G.J., Bouvy, N.D., Schrauwen, P., and
Teule, G.J.J. (2009). Cold-activated Brown adipose tissue in healthy men.
N. Engl. J. Med. 360, 1500–1508.
Matsuzawa-Nagata, N., Takamura, T., Ando, H., Nakamura, S., Kurita, S.,
Misu, H., Ota, T., Yokoyama, M., Honda, M., Miyamoto, K., et al. (2008).
Increased oxidative stress precedes the onset of high-fat diet–induced insulin
resistance and obesity. Metabolism 57, 1071–1077.
Matsuzawa, N., Takamura, T., Kurita, S., Misu, H., Ota, T., Ando, H.,
Yokoyama, M., Honda, M., Zen, Y., Nakanuma, Y., et al. (2007). Lipid-induced
ll
OPEN ACCESS
oxidative stress causes steatohepatitis in mice fed an atherogenic diet.
Hepatology 46, 1392–1403.
Miller, A.F. (2012). Superoxide dismutases: ancient enzymes and new insights.
FEBS Lett. 586, 585–595.
Mills, E.L., Pierce, K.A., Jedrychowski, M.P., Garrity, R., Winther, S., Vidoni, S.,
Yoneshiro, T., Spinelli, J.B., Lu, G.Z., Kazak, L., et al. (2018). Accumulation of
succinate controls activation of adipose tissue thermogenesis. Nature 560,
102–106.
Misu, H., Takamura, T., Takayama, H., Hayashi, H., Matsuzawa-Nagata, N.,
Kurita, S., Ishikura, K., Ando, H., Takeshita, Y., Ota, T., et al. (2010). A liver-
derived secretory protein, selenoprotein P, causes insulin resistance. Cell
Metab. 12, 483–495.
Misu, H., Ishikura, K., Kurita, S., Takeshita, Y., Ota, T., Saito, Y., Takahashi, K.,
Kaneko, S., and Takamura, T. (2012). Inverse correlation between serum levels
of selenoprotein p and adiponectin in patients with type 2 diabetes. PLoS One
7, e34952.
Misu, H., Takayama, H., Saito, Y., Mita, Y., Kikuchi, A., Ishii, K.A., Chikamoto,
K., Kanamori, T., Tajima, N., Lan, F., et al. (2017). Deficiency of the hepatokine
selenoprotein P increases responsiveness to exercise in mice through
upregulation of reactive oxygen species and AMP-activated protein kinase
in muscle. Nat. Med. 23, 508–516.
Mita, Y., Nakayama, K., Inari, S., Nishito, Y., Yoshioka, Y., Sakai, N., Sotani,
K., Nagamura, T., Kuzuhara, Y., Inagaki, K., et al. (2017). Selenoprotein P-
neutralizing antibodies improve insulin secretion and glucose sensitivity in
type 2 diabetes mouse models. Nat. Commun. 8, 1–17.
Nakamura, S., Takamura, T., Matsuzawa-Nagata, N., Takayama, H., Misu, H.,
Noda, H., Nabemoto, S., Kurita, S., Ota, T., Ando, H., et al. (2009). Palmitate
induces insulin resistance in H4IIEC3 hepatocytes through reactive oxygen
species produced by mitochondria. J. Biol. Chem. 284, 14809–14818.
Olsen, J.M., Csikasz, R.I., Dehvari, N., Lu, L., Sandström, A., Öberg, A.I.,
Nedergaard, J., Stone-Elander, S., and Bengtsson, T. (2017). b3-Adrenergi-
cally induced glucose uptake in brown adipose tissue is independent of
UCP1 presence or activity: mediation through the mTOR pathway. Mol. Metab.
6, 611–619.
Olson, G.E., Winfrey, V.P., NagDas, S.K., Hill, K.E., and Burk, R.F. (2007).
Apolipoprotein E receptor-2 (ApoER2) mediates selenium uptake from
selenoprotein P by the mouse testis. J. Biol. Chem. 282, 12290–12297.
Olson, G.E., Winfrey, V.P., Hill, K.E., and Burk, R.F. (2008). Megalin mediates
selenoprotein p uptake by kidney proximal tubule epithelial cells. J. Biol.
Chem. 283, 6854–6860.
Oo, S.M., Misu, H., Saito, Y., Tanaka, M., Kato, S., Kita, Y., Takayama, H.,
Takeshita, Y., Kanamori, T., Nagano, T., et al. (2018). Serum selenoprotein
P, but not selenium, predicts future hyperglycemia in a general Japanese
population. Sci. Rep. 8, 1–10.
Ota, T., Takamura, T., Kurita, S., Matsuzawa, N., Kita, Y., Uno, M., Akahori, H.,
Misu, H., Sakurai, M., Zen, Y., et al. (2007). Insulin resistance accelerates a
dietary rat model of nonalcoholic steatohepatitis. Gastroenterology 132,
282–293.
Ro, S.H., Nam, M., Jang, I., Park, H.W., Park, H., Semple, I.A., Kim, M., Kim,
J.S., Park, H., Einat, P., et al. (2014). Sestrin2 inhibits uncoupling protein 1
expression through suppressing reactive oxygen species. Proc. Natl. Acad.
Sci. U S A. 111, 7849–7854.
Saito, M. (2013). Brown adipose tissue as a therapeutic target for human
obesity. Obes. Res. Clin. Pract. 7, e432-438.
Saito, Y., and Takahashi, K. (2002). Characterization of selenoprotein P as a
selenium supply protein. Eur. J. Biochem. 269, 5746–5751.
Saito, M., Okamatsu-Ogura, Y., Matsushita, M., Watanabe, K., Yoneshiro, T.,
Nio-Kobayashi, J., Iwanaga, T., Miyagawa, M., Kameya, T., Nakada, K., et al.
(2009). High incidence of metabolically active brown adipose tissue in healthy
adult humans: Effects of cold exposure and adiposity. Diabetes 58, 1526–
1531.
Saito, Y., Hayashi, T., Tanaka, A., Watanabe, Y., Suzuki, M., Saito, E., and
Takahashi, K. (1999). Selenoprotein P in human plasma as an extracellular
Cell Reports 38, 110566, March 29, 2022 11
 ll
OPEN ACCESS
phospholipid hydroperoxide glutathione peroxidase: isolation and enzymatic
characterization of human selenoprotein P. J. Biol. Chem. 274, 2866–2871.
Saito, Y., Watanabe, Y., Saito, E., Honjoh, T., and Takahashi, K. (2001).
Production and application of monoclonal antibodies to human selenoprotein
P. J. Heal. Sci. 47, 346–352.
Schweizer, U., Streckfuß, F., Pelt, P., Carlson, B.A., Hatfield, D.L., Köhrle, J.,
and Schomburg, L. (2005). Hepatically derived selenoprotein P is a key factor
for kidney but not for brain selenium supply. Biochem. J. 386, 221–226.
Spitzer, M., Wildenhain, J., Rappsilber, J., and Tyers, M. (2014). BoxPlotR: a
web tool for generation of box plots. Nat. Methods 11, 121–122.
Stanford, K.I., Middelbeek, R.J.W., Townsend, K.L., An, D., Nygaard, E.B.,
Hitchcox, K.M., Markan, K.R., Nakano, K., Hirshman, M.F., Tseng, Y.H.,
et al. (2013). Brown adipose tissue regulates glucose homeostasis and insulin
sensitivity. J. Clin. Invest. 123, 215–223.
Stranges, S., Marshall, J.R., Natarajan, R., Donahue, R.P., Trevisan, M.,
Combs, G.F., Cappuccio, F.P., Ceriello, A., and Reid, M.E. (2007). Effects of
long-term selenium supplementation on the incidence of type 2 diabetes: a
randomized trial. Ann. Intern. Med. 147, 217–223.
Tajima-Shirasaki, N., Ishii, K.A., Takayama, H., Shirasaki, T., Iwama, H.,
Chikamoto, K., Saito, Y., Iwasaki, Y., Teraguchi, A., Lan, F., et al. (2017).
Eicosapentaenoic acid down-regulates expression of the selenoprotein P
gene by inhibiting SREBP-1c protein independently of the AMP-activated
12 Cell Reports 38, 110566, March 29, 2022
Article
protein kinase pathway in H4IIEC3 hepatocytes. J. Biol. Chem. 292, 10791–
10800.
Takamura, T. (2020). Hepatokine selenoprotein P-mediated reductive stress
causes resistance to intracellular signal transduction. Antioxid. Redox Signal.
33, 517–524.
Takebe, G., Yarimizu, J., Saito, Y., Hayashi, T., Nakamura, H., Yodoi, J.,
Nagasawa, S., and Takahashi, K. (2002). A comparative study on the
hydroperoxide and thiol specificity of the glutathione peroxidase family and
selenoprotein P. J. Biol. Chem. 277, 41254–41258.
Tanaka, Mutsumi, Saito, Yoshiro, Misu, Hirofumi, Kato, Seiji, Kita, Yuki, Take-
shita, Yumie, Kanamori, Takehiro, Nagano, Toru, Nakagen, Masatoshi, Urabe,
Takeshi, et al. (2016). Development of a Sol Particle Homogeneous Immuno-
assay for Measuring Full-Length Selenoprotein P in Human Serum. J. Clin.
Lab. Anal. 30 (2), 114–122.
Uchiyama, S., Gota, C., Tsuji, T., and Inada, N. (2017). Intracellular
temperature measurements with fluorescent polymeric thermometers.
Chem. Commun. 53, 10976–10992.
Yang, R., Liu, Y., and Zhou, Z. (2017). Selenium and selenoproteins, from
structure, function to food resource and nutrition. Food Sci. Technol. Res.
23, 363–373.
Yoneshiro, T., Aita, S., Matsushita, M., Kameya, T., Nakada, K., Kawai, Y., and
Saito, M. (2011). Brown adipose tissue, whole-body energy expenditure, and
thermogenesis in healthy adult men. Obesity 19, 13–16.
 ll
Article OPEN ACCESS

STAR+METHODS

KEY RESOURCES TABLE

REAGENT or RESOURCE SOURCE IDENTIFIER

Antibodies
Rabbit anti-UCP1 polyclonal Abcam Cat# ab10983; RRID:AB_381214
rabbit anti-GPX4 monoclonal Abcam Cat# ab125066; RRID:AB_10973901
rabbit anti-GPX1 polyclonal Abcam Cat# ab22604; RRID:AB_2112120
rabbit anti- LRP1 monoclonal Abcam Cat# ab92544; RRID:AB_2234877
Rabbit anti-total-mTOR monoclonal Cell Signaling Cat# 2983; RRID:AB_2105622
Rabbit anti-phospho-mTOR (Ser2481) polyclonal Cell Signaling Cat# 2974; RRID:AB_2262884
Rabbit anti-phospho-AMPK (Thr172) monoclonal Cell Signaling Cat# 2535; RRID:AB_331250
Rabbit anti-total-AMPK polyclonal Cell Signaling Cat# 2603; RRID:AB_490795
Beta-actin antibodies Cell Signaling Cat# 4967; RRID:AB_330288
Rabbit anti-selenoprotein W polyclonal Rockland Cat# 600-401-A29; RRID:AB_2285666
Rat anti-human selenoprotein P Saito et al. (2001) NA
monoclonal antibody (BD-1)
Chemicals, peptides, and recombinant proteins
Collagenase D Roche Cat# 11088866001
Dispase Roche Cat# 04942078001
Indomethacin Sigma-Aldrich Cat# I7378
IBMX Sigma-Aldrich Cat# I5879
Dexamethasone Sigma-Aldrich Cat# D1756
Rosiglitazone Sigma-Aldrich Cat# R2408
T3 Sigma-Aldrich Cat# T2877
Insulin Sigma-Aldrich Cat# I9278
Methoxypolyethylene glycol maleimide (PEG-Mal) Sigma-Aldrich Cat# 63187
L-norepinephrine hydrochloride Sigma-Aldrich Cat# 74480
N-ethyl maleimide Wako Cat# 054-02063
TECP Thermo Scientific Cat# 77720
N-acetyl cysteine Sigma-Aldrich Cat# A9165
1S 3R-RSL3 Sigma-Aldrich Cat# SML2234
Phloretin Wako Cat# 160-17781
(±)-Norepinephrine (+)-bitartrate salt Sigma-Aldrich Cat# A0937
Carbonyl cyanide 4-(trifluoromethoxy) Sigma-Aldrich Cat# C2920
phenylhydrazone (FCCP)
Antimycin A Sigma-Aldrich Cat# A8674
Rotenone Sigma-Aldrich Cat# R8875
Advanced DMEM/F12 Gibco Cat# 12634010
phenol red-free DMEM/F12 Gibco Cat# 21041025
glucose-free DMEM Gibco Cat# A1443001
Critical commercial assays
TBARS Assay Kit Cayman Chemical
Dihydroethidium (Hydroethidine) Thermo Fisher Cat# D23107
MitoSOX red mitochondrial superoxide indicator Invitrogen Cat# M36008
Cellular Thermoprobe for Fluorescence Ratio Funakoshi, Tokyo, Japan FDV-0005
Glucose Uptake Cell-Based assay Kit Cayman Chemical
TaqMan assay for Actb Thermo Fisher 4352341E
TaqMan assay for Gapdh Thermo Fisher 4352339E
(Continued on next page)

Cell Reports 38, 110566, March 29, 2022 e1

 ll
OPEN ACCESS Article

Continued
REAGENT or RESOURCE SOURCE IDENTIFIER
TaqMan assay for 18S rRNA Thermo Fisher 4319413E
TaqMan assay for Sepp1 Thermo Fisher Mm00486048_m1
TaqMan assay for Lrp1 Thermo Fisher Mm00464608_m1
TaqMan assay for Ldlr Thermo Fisher Mm01177349_m1
TaqMan assay for Lrp2 Thermo Fisher Mm01328171_m1
TaqMan assay for Lrp8 Thermo Fisher Mm00474030_g1
TaqMan assay for Gpx1 Thermo Fisher Mm00656767_m1
TaqMan assay for Gpx4 Thermo Fisher Mm00515041_m1
TaqMan assay for Ucp1 Thermo Fisher Mm01244861_m1
TaqMan assay for Sepw1 Thermo Fisher Mm01268252_m1
Selenop specific siRNA Invitrogen Cat# MSS208936
Gpx1 specific siRNA Invitrogen Cat# MSS274652
Gpx4 specific siRNA Invitrogen Cat# MSS286702
Lrp1 specific siRNA Invitrogen Cat# NM008512.2
Negative control siRNAs Invitrogen REF-12935112
Experimental models: Organisms/strains
C57BL6/J Sankyo Lab Service
Software and Algorithms
Image Lab software Bio-Rad Laboratories, Inc. RRID:SCR_014210
Statistical Package for Social Science IBM RRID:SCR_019096
(SPSS) Version 21 advanced model
GraphPad Prism GraphPad RRID:SCR_002798
Other
weighing environment logger A&D Company Limited, Cat# AD-1687
Tokyo, Japan
endorectal probe A&D Company Limited, Cat# AX-KO4746-100
Tokyo, Japan
InfRec G100EX thermal camera Nippon Avionics, Japan
indirect calorimeter chamber (Oxymax) Columbus Instruments
multi-point temperature logger system Tateyama Science High LT-200SA-TB
Technologies Co., Ltd.
Glomax-Multi+ Detection System Promega
StepOne Plus real-time PCR system Life Technologies Corporation,
Carlsbad, CA

RESOURCE AVAILABILITY

Lead contact
Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Toshinari
Takamura (ttakamura@med.kanazawa-u.ac.jp).

Materials availability
This study did not generate new unique reagents.

Data and code availability

The data reported in this paper will be shared by the lead contact upon request.
This paper does not report original code.
Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

e2 Cell Reports 38, 110566, March 29, 2022

 ll
Article OPEN ACCESS

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Human study
Eighty-seven Japanese nondiabetic healthy male participants went to LSI Sapporo Clinic, Sapporo, Japan, for a complete physical
examination. As described previously, the subjects underwent the 18FDG-PET/CT scan procedure (Saito et al., 2009; Yoneshiro
et al., 2011). After fasting for 6–12 h, the subjects entered an air-conditioned room at 19 C with light clothing (usually a T-shirt
with underwear) and put their legs on an ice block intermittently (usually for 4 min after every 5 min). One hour after the mild cold
condition, they were given an intravenous injection of 259 megabecquerels (MBq) fluorodeoxyglucose (FDG) and maintained under
the same cold conditions. One hour after the FDG injection, whole-body PET/CT scans were performed using a PET/CT system
(Aquiduo; Toshiba Medical Systems, Tochigi, Japan) in a room at 24 C. With the CT parameters of 120 kV and real-exposure control,
unenhanced low-dose spiral axial 2-mm collimated images were obtained. The images were used for PET attenuation correction as
well as anatomic localization. Subsequently, full-ring PET was performed in six incremental table positions, each 15 cm thick. The
total time for these scans was 30 min.
PET and CT images were coregistered and analyzed using a VOX-BASE workstation (J-MAC System, Sapporo, Japan). Two
experienced, blinded observers assessed the FDG uptake, particularly on both sides of the neck and paravertebral regions, by
visually judging the presence of radioactivity greater than that of the background. BAT activity in the neck region was quantified
by calculating the maximal standardized uptake value (SUVmax), defined as the radioactivity per milliliter within the region of interest
divided by the injected dose in megabecquerels per gram of body weight. Serum samples were obtained before 18FDG-PET/CT scan
and used to analyze serum SeP levels as describe previously (Tanaka et al., 2016).
By using the median value of the SUVmax as the cut-off point, the subjects were divided into Low BAT group and High BAT group
with the mean age of 31.73 ± 9.55 and 27.16 ± 7.46 years respectively. The physical and serological parameters were analyzed in
comparison between these two groups.
All participants provided written informed consent for participation in this study. The Ethics Committees of Kanazawa University
approved all experimental protocols (approval no. 2017-158, UMIN000029276).

Animals
Selenop-deficient mice were produced by homologous recombination using genomic DNA cloned from an Sv-129 P1 library, as
described previously (Hill et al., 2003). BAT or Liver Selenop-deficient mice were maintained on a mixed B6/129 background.
Mice carrying floxed alleles of Selenop (C57Bl/6 and 129/sv background) were originally obtained from R. F. Burk (Vanderbilt
University School) (Hill et al., 2012). Mice transgenically expressing Cre recombinase under the albumin promoter were generous gifts
from Hiroshi Inoue (Kanazawa University). Mice transgenically expressing Cre recombinase under the Ucp1 promoter were
purchased from The Jackson Laboratory (J:206508). We backcrossed them with C57Bl/6 mice more than five times. Liver
Selenop-KO mice were generated by crossing Selenop floxed mice with Alb-Cre transgenic mice. BAT-Selenop-KO mice were
generated by crossing Selenop floxed mice with Ucp1-Cre transgenic mice. Because Alb-Cre transgenic mice or Ucp1-Cre
transgenic mice and Selenopfl/fl mice were phenotypically indistinguishable, Selenopfl/fl mice were used as negative controls. All
mice were genotyped by PCR using the following primers: 50 - TCCTAGATTGGCAGAGGATAGAATGAA -30 and 50 - TCAGAAA-
CACCTTCCAACTGTAATGC -30 for the floxed Selenop gene and 50 -CGCCGCATAACCAGTGAAAC-30 and 50 -ATGTCCAATTTACT-
GACCG-30 for the Cre transgene.
All the animal studies were carried out following the Guidelines on the Care and Use of Laboratory Animals issued by Kanazawa
University. The ethical committee of Kanazawa University approved the study protocol (Approval NO. 153678). C57BL/6J mice were
obtained from Sankyo Lab Service (Tokyo, Japan). The mice ate a standard rodent food diet CRF-1 containing 0.45 mg/kg of
selenium (Oriental Yeast, Tokyo, Japan) or 60% fat with increased sucrose rodent food D03062301 contains 0.16 mg/kg of selenium
(Research Diet, New Brunswick, USA). All mice used in the current study had a C57BL/6J genetic background. Because female mice
had inconsistent phenotypes, only male mice were used in this study. The mice at the age of 7–21 weeks were used unless otherwise
indicated.

METHOD DETAILS

Cold exposure and rectal temperature measurement

Six hours before cold exposure, all mice have fasted and injected intraperitoneally with PBS or NAC (1.25 mmol/kg body weight). A
pair of WT and Selenop-deficient mice were put into a cage in the cold chamber at 4 C for 2 h. Rectal temperatures were measured
every 30 min during cold exposure by using a rectal thermometer (weighing environment logger (AD-1687) & an endorectal probe for
mice (AX-KO4746-100); A&D Company Limited, Tokyo, Japan). Shivering activities were assessed by video recording and manual
counting at indicated time points during cold exposure. Blood glucose levels were measured before and after cold exposure by
using a handheld blood glucose meter Glutest mint (Sanwa Kagaku Kenkyusho Co. Ltd., Nagoya, Japan). Serum non-esterified fatty
acid (NEFA) levels were determined using the NEFA C Enzymatic assay kit (FUJIFILM Wako Diagnostics, USA), according to the
manufacturer’s instructions. Immediately after the completion of cold exposure, the mice were sacrificed to obtain tissue
samples. For urinary norepinephrine measurement, 24-h urine samples were collected while the mice were housed at the assigned

Cell Reports 38, 110566, March 29, 2022 e3

 ll
OPEN ACCESS Article

temperature. Hydrochloric acid was added into the urine sample to a final concentration of 0.1 M for the integrity of norepinephrine
until analysis. Urinary norepinephrine levels were measured by Norepinephrine ELISA kit (Abnova Corporation, Taipei, Taiwan), ac-
cording to the manufacturer’s instructions. Urinary norepinephrine levels were normalized by creatinine content measured by the
Creatinine colorimetric assay (Exocell, Philadelphia, PA).

Image of interscapular temperature by a thermal camera in mice

The image of mice interscapular temperature was captured using the InfRec G100EX thermal camera (Nippon Avionics, Japan) after
2 h of cold exposure at 4 C. All the mice were fasted for 6 h before putting into the cold chamber.

Oxygen consumption rate in mice

Six hours before the OCR measurement, all mice were fasted and injected intraperitoneally with PBS or NAC (1.25 mmol/kg body
weight). After that, mice were anesthetized and put in an indirect calorimeter chamber individually (Oxymax; Columbus Instruments,
Columbus, OH). After recording the baseline OCR for 30 minutes, the NA (1 mg/kg body weight) was injected subcutaneously and
continued recording for 44 min. VO2 was measured every 8 or 10 min throughout the experiment. All procedures were performed at
room temperature (24 C–26 C).

Thiobarbituric acid reactive substances (TBARS)

According to the manufacturer’s instructions, the levels of TBARS in BAT of mice were measured using TBARS Assay Kit (Cayman
Chemical).

BAT and rectal temperature measurements

The mice’s BAT and rectal temperatures under the NA stimulation were measured according to a previously described procedure
(Inokuma et al., 2005). All mice were fasted in the thermoneutral chamber (30 C) for 12 h before the experiment. All mice were anes-
thetized and kept on the heating plate. The sensor probes were inserted under the interscapular brown adipose tissues and in the
rectum of the mice. The BAT and rectal temperatures were recorded by the multi-point temperature logger system (LT-200SA-
TB, Tateyama Science High Technologies Co., Ltd.). At 5-min after starting temperature recording, the NA (0.2 mg/kg body weight)
was injected subcutaneously, and the temperature was recorded for 30 min.

RNA isolation, cDNA synthesis, and real-time PCR analysis

Total RNA isolation, cDNA synthesis, and real-time PCR analysis were performed as previously described (Tajima-Shirasaki et al.,
2017). Quantitative RT-PCR was performed using TaqMan probes (Actb 4352341E; Gapdh 4352339E; 18S rRNA 4319413E;
Sepp1 Mm00486048_m1; Lrp1 Mm00464608_m1; LDLR Mm01177349_m1; Lrp2 Mm01328171_m1; Lrp8 Mm00474030_g1;
Gpx1 Mm00656767_m1; Gpx4 Mm00515041_m1; Ucp1 Mm01244861_m1; Sepw1 Mm01268252_m1) and the StepOne Plus
real-time PCR system (Life Technologies Corporation, Carlsbad, CA).

ELISA for mouse SeP

Ninety-six well microtiter plates were coated for 18 h at 4 C with 100 mL of a coating antibody against mouse SeP (made in-house) in
0.05-M sodium bicarbonate buffer, pH 9.6. The wells were washed four times with 200-ml PBS containing 0.05% Tween 20 (PBS/T)
and incubated at 37 C with 150 ml of blocking solution (1-mg/mL BSA in PBS) for 1 h. After washing the wells four times, 50 mL of
plasma sample (diluted 50 times with PBS/T, containing 1-mg/mL BSA) was added to each well and incubated at 37 C for 1 h. After
washing the wells four times, 50 ml of 9S4 antibody for mouse SeP (Developmental Studies Hybridoma Bank deposited by Burke,
R.D., RRID: AB_2617215) labeled with the Peroxidase Labeling kit-NH2 (DONJINDO Molecular Technologies, Kumamoto, Japan)
was added and incubated at 37 C for 1 h. Finally, the plates were washed eight times. Fifty microliters of TMB were added to
each well, and the enzyme-substrate reaction was allowed to proceed for 30 min at room temperature. The color development
was stopped by adding 50 ml of 0.25 mol/L sulphuric acid, and the absorbance at 450 nm was measured immediately. Because there
are no commercially available mouse SeP protein standards, we compared the net value of absorbance.

Western blot studies

We performed western blotting, as previously reported (Misu et al., 2017). Densitometric analysis of blotted membranes was
performed using Image Lab software (Bio-Rad Laboratories, Inc.).

Assessment of sulfenylated UCP1 by gel shift

Sulfenylated UCP1 status in brown adipose tissues was measured as described previously (Chouchani et al., 2016). In brief, mice
BATs were isolated rapidly after indicated in vivo intervention and then homogenized in 100-nM NEM, 1-mM EGTA, 50-nM Tris-HCl,
pH 7.4. After incubation at 37 C for 5 min, the homogenate was further incubated with SDS (2% final) at 37 C for 10 min. A
thermomixer performed those incubation steps at 1,300 rpm. After that, excess NEM in the samples was removed by acetone
precipitation. The pellets were then resuspended with 1-mM EGTA, 2% SDS, 10-mM TCEP, 50-mM Tris-HCl, pH 7.4 containing
50 mM of polyethylene glycol polymer conjugated to maleimide (PEG-Mal). Resuspended samples were incubated at 37 C for

e4 Cell Reports 38, 110566, March 29, 2022

 ll
Article OPEN ACCESS

30 min, followed by second acetone precipitation to remove excess PEG-Mal. After resuspension of the pallet in RIPA lysis
buffer, insolubilized debris was removed by centrifuging at 15,000 rpm for 30 min at 4 C. The supernatants were subjected to
immunodetection by using the rabbit anti-UCP1 antibody. The conjugation of maleimide to sulfenylated UCP1 will result in apparent
10 kDa shifts on a gel per cysteine oxidation event.

Purification of SeP
SeP was purified from human plasma using conventional chromatographic methods, as previously described (Saito et al., 1999).
Homogeneity of purified human SeP was confirmed by analyzing both amino acid composition and sequence (Saito et al., 1999).

Primary brown adipocytes

Preparation and culture of primary brown adipocyte were performed as previously reported (Aune et al., 2013). For the knockdown of
the target gene, completely differentiated primary brown adipocytes were transfected with a 50 nM of indicated siRNA duplex
oligonucleotides with Lipofectamine RNAiMAX (Invitrogen), using the reverse transfection method as described previously (Isidor
et al., 2016). Selenop specific siRNA (Cat. no. #MSS208936), Gpx1 specific siRNA (Cat. no. #MSS274652), Gpx4 specific siRNA
(Cat. no. #MSS286702), Lrp1 specific siRNA (Cat. no. #NM008512.2), and negative control siRNAs were purchased from Invitrogen.
siRNA transfection was performed for 48 h, which was followed by the indicated intervention.

Measurement of DHE oxidation

ROS production in primary brown adipocytes was measured by the DHE oxidation method. In brief, primary brown adipocytes were
plated and differentiated in clear bottom black wall 96-well plates. After completion of differentiation, the culture medium was
aspirated, and the cells were washed with imaging buffer (156-mM NaCl, 3-mM KCl, 1.25-mM KH2PO4, 2-mM MgCl2, 10-mM
HEPES, pH 7.4) twice. It was then replaced with a fresh imaging buffer containing 1 mM sodium pyruvate, 5-mM DHE (cat. no.
#D23107, Invitrogen), and with or without 400 nM NA to detect NA-induced DHE oxidation. Fluorescence intensity was measured
at ex355/em460 for reduced DHE and ex544/em590 for oxidized DHE using a fluorescence plate reader (Glomax-Multi + Detection
System, Promega). Each sample’s oxidative stress levels were described as oxidized/reduced DHE at 30 min after NA stimulation
and then normalized by their respective baseline values.

Measurement of mitochondrial ROS by MitoSOX assay

Measurement of mitochondrial-specific superoxide levels was performed using MitoSOX red mitochondrial superoxide indicator
(cat. no. #M36008, Invitrogen) per the manufacturer’s instruction. Briefly, completely differentiated primary brown adipocytes
were pretreated with or without 10-mg/mL SeP for 24 h. After that, the cells were washed with PBS 3 times, and the culture medium
was changed with phenol red-free DMEM/F12 supplemented with 5-mM MitoSOX with or without 1000-nM NA. Fluorescence
intensity was measured at ex510/em580 at baseline and 30 min after NA stimulation. The mitochondrial ROS levels in each sample
were described after normalization with their respective baseline values.

Measurement of cellular temperature

NA-induced changes in intracellular temperature were measured in primary brown adipocytes by using Cellular Thermoprobe for
Fluorescence Ratio (FDV-0005; Funakoshi, Tokyo, Japan). Primary brown adipocytes were cultured on glass-bottom dishes until fully
confluent. The culture medium was then removed, and the cells were washed with a 5% glucose solution. 0.05% w/v of Cellular
Thermoprobe in 5% glucose solution was then added to the cells, and the cultured dish was incubated at 25 C for 10 min. After
washing with PBS three times, 100 mL of phenol red-free DMEM/F12 was added. For observation, excitation was carried out at
488 nm, and emission was monitored at 525 and 610 nm in a microscope cage incubation chamber at 37 C. Fluorescent ratio
Em610/Em525 was used to evaluate the temperature change. After recording baseline fluorescence intensity, the cells were
stimulated with 400-nM NA (final concentration) or vehicle prepared in 100 mL of phenol red-free medium, followed by measurement
of fluorescence intensity in a time-lapse manner. Fold changes in fluorescence intensity at each time point were calculated by
normalization with baseline fluorescence intensity.

Glucose uptake in brown adipocytes

Glucose uptake into primary brown adipocytes was assessed by using Glucose Uptake Cell-Based assay kit (Cayman Chemical)
according to the manufacturer’s instruction. To describe in brief, after full differentiation in 96-well plates, primary brown
adipocytes were glucose and serum fasted in glucose-free DMEM (Gibco, Thermo Scientific) supplemented with 0.5% BSA for
6 h. After that, cells were treated with glucose-free medium (with 0.5% BSA) containing 100-mg/mL 2-NBDG, 200 mM phloretin
with or without 1000-nM NA for 5 min. At the end of the treatment, the plate was washed with ice-cold PBS 3 times to stop further
glucose uptake into the cells. Then, 100 mL of Cell-Based Assay buffer was added to each well, and fluorescent intensity was
measured at ex485/em535. Reading of the well treated with glucose-free medium without 2-NBDG treatment was regarded as
blank.

Cell Reports 38, 110566, March 29, 2022 e5

 ll
OPEN ACCESS Article

Extracellular flux analysis

5.5 3 104 cells were plated onto XF24 cell culture plate (#100777-004, Agilent) with complete medium. For mitochondrial stress
test, the medium was changed to unbuffered DMEM (D5030, Sigma-Aldrich) containing 25-mM glucose, 1 3 glutamax, and
1-mM pyruvate, and cells were incubated at 37 C for 1 h without CO2 control. NE (1 mM, A0937, Sigma-Aldrich) was injected to
stimulate cells. Maximal OCR was induced by exposing cells to mitochondrial uncoupler, FCCP (0.5 mM, C2920, Sigma-Aldrich).
Antimycin A (1 mM, A8674, Sigma-Aldrich) and rotenone (1 mM, R8875, Sigma-Aldrich) was added to disrupt all mitochondria-depen-
dent respiration. XF24 (Agilent) was used to measure OCR and ECAR under a 2 min period protocol, followed by 4 min mixing.

QUANTIFICATION AND STATISTICAL ANALYSIS

All data were analyzed using the Statistical Package for Social Science (SPSS) Version 21 advanced model and GraphPad Prism 8
software. Experimental data were visualized as box plots with individual points by a web tool, BoxPlotR (Spitzer et al., 2014).
Statistical methods were not used to determine the sample size. The sample size was based on trial experiments or experiments
performed previously. We randomized neither animals nor cell samples and performed all the experiments without the blinding of
the investigator. All groups in the current experiments showed normal variance. Statistical differences between the two groups
were assessed using unpaired two-tailed student t-tests, except paired t-tests were used for blood glucose changes before and after
cold exposure (Figures 2C and 2F). Data involving more than two groups were assessed by analysis of variance (ANOVA) with Tukey’s
post hoc test. The AUC calculation and unpaired two-tailed student t-tests were executed by GraphPad Prism 8 software (Figures 4F,
5D, 5H, 6D, and 6G). Using the box-plot display tool in SPSS, we defined cases with values more than three times the interquartile
range as outliers that were routinely excluded from all the analyses.

e6 Cell Reports 38, 110566, March 29, 2022