Night eating and obesity in the EP3R-deficient mouseManuel Sanchez-Alavez, Izabella Klein, Sara E. Brownell, Iustin V. Tabarean, Christopher N. Davis, Bruno Conti,and Tamas Bartfai*

The Harold L. Dorris Neurological Research Institute and Molecular and Integrative Neurosciences Department, The Scripps Research Institute,10550 North Torrey Pines Road, La Jolla, CA 92037

Communicated by Julius Rebek, Jr., The Scripps Research Institute, La Jolla, CA, December 21, 2006 (received for review November 7, 2006)

Adult mice carrying a null mutation of the prostanoid receptorEP3R (EP3R�/� mice) exhibit increased frequency of feeding duringthe light cycle of the day and develop an obese phenotype undera normal fat diet fed ad libitum. EP3R�/� mice show increasedmotor activity, which is not sufficient to offset the increasedfeeding leading to increased body weight. Altered ‘‘nocturnal’’activity and feeding behavior is present from a very early age anddoes not seem to require age-dependent factors for the develop-ment of obesity. Obesity in EP3R�/� mice is characterized byelevated leptin and insulin levels and >20% higher body weightcompared with WT littermates. Abdominal and subcutaneous fatand increased liver weight account for the weight increase inEP3R�/� mice. These observations expand the roles of prostaglan-din E2 signaling in metabolic regulation beyond the reportedstimulation of leptin release from adipose tissue to involve actionsmediated by EP3R in the regulation of sleep architecture andfeeding behavior. The findings add to the growing literature onlinks between inflammatory signaling and obesity.

body weight � insulin � leptin � prostanoid receptor � signaling

Prostaglandin E2 (PGE2) is one of the most important inflam-matory mediators in the periphery and in the CNS (1–7). PGE2

is synthesized locally in the brain by neurons and glia expressingCOX1, COX2, and PGE synthase (8, 9). The activation of specificToll-like receptors (TLR) by IL-1 or LPS and the subsequentproduction of PGE2 mediates different CNS responses such as thefebrile response (3, 5, 10, 11), activation of the hypothalamic–pituitary–adrenal axis (12–15), and the central activation of brownfat metabolism/uncoupling (16–18). The central and peripheralactions of PGE2 are mediated by four G protein-coupled receptor-type prostanoid receptors (EP1R–EP4R) (for review, see refs. 19and 20). Transgenic mice carrying a null mutation for each of thefour prostanoid receptors have been generated, and differentphenotypes have been described for EP1R�/� (4, 21, 22), EP2R�/�

(23–29), EP3R�/� (4, 5, 21), and EP4R�/� (30–32) mice. EP1R�/�

mice show a decreased aberrant foci formation to azoxymethane(33). EP2R�/� mice have impaired ovulation and fertilization,salt-sensitive hypertension, impaired vasodepressor response toPGE2, and a loss of bronchodilation (29). EP4R�/� mice have animpaired vasodepressor response to i.v. infusion of PGE2 anddecreased inflammation and bone resorption (34). EP3R�/� micehave an impaired febrile response (5), impaired duodenal bicar-bonate secretion (35), enhanced vasodepressor response to PGE2(36), and increased bleeding tendency (37). Effects of the loss ofEP3R were also observed on tumor-induced angiogenesis (38),colon cancer development (39), allergic inflammation (40), andinflammatory pain (41). PGE2 levels in these mice were not foundto be altered in these studies.

Although EP3R�/� mice have been well characterized, therehave not been any reports on obesity or altered feeding patterns inthese mice. PGE2 has been shown to have an effect on adipocytesby inhibiting lipolysis and stimulating the secretion of leptin, but thespecific prostanoid receptor subtype important for mediating theseeffects has remained unknown (42). The heterozygous COX2�/�

mice but not the COX1�/� or COX2�/� transgenic mice have been

shown to develop obesity (43), but no mechanistic explanationsinvolving altered PGE2 signaling have been developed.

Effects of prostaglandin D (PGD) and PGE2 on sleep archi-tecture have been studied extensively (cf. ref. 44), and sleeppromotion by PGE2 applied in the subarachnoid space wasreported by Ram et al. (45), but the repeated night activitydescribed in this report has not been observed.

EP3R is a G protein-coupled receptor (46) with several splicevariants (47), but the tissue-specific distribution of these variantsis unknown. The EP3R-like immunoreactivity is richly expressedin the rodent brain, with the highest density of EP3R receptorsin the different hypothalamic nuclei involved in thermoregula-tion and sleep regulation (48–50). We have shown that EP3Rmediates the effects of PGE2, a potent pyrogen, on the thermo-sensitivity of anterior hypothalamic neurons (51). EP3R-likeimmunoreactivity is also abundant in monoaminergic nuclei suchas the raphe nuclei and locus ceruleus (48), where this prosta-noid receptor may affect appetite and feeding through modu-lation of the serotonergic and noradrenergic signaling.

We show here that EP3R�/� mice exhibit an obese phenotypewith a ‘‘night’’ eating component that is demonstrated by increasedfeeding during the light cycle of the day. The obese phenotype inEP3R�/� mice is congruent with previous observations that PGE2stimulated leptin release, affected lipolysis, and had an effect on theoverall endocrine state of the organism (43, 52). More broadly, thisfinding contributes to the increased understanding of the couplingbetween obesity and inflammatory signaling (53–55).

ResultsEP3R�/� Mice Showed Increased Body Weight Caused by Increased FatStorage. Male EP3R�/� mice fed ad libitum on an 11% fat dietexhibited an increase in body weight that resulted in an obesephenotype (Fig. 1A). EP3R�/� body weights showed a gradualincrease, with continuous accumulation up to 20% above that ofWT littermate controls by week 20 (t � 4.41; P � 0.001) andcontinuing thereafter up to a recorded 30% increase by week 40(t � 4.7; P � 0.006) (Fig. 1B). Body weight gain was primarilyattributable to increased adipose tissue deposition (Fig. 1C). Com-parison of the body composition of 3- and 6-month-old micerevealed that the early deposition of intraabdominal fat pads(gonadal, retroperitoneal, and mesenteric) in EP3R�/� mice at �3months was followed by significant deposition of fat in adjacenttissues that include the liver and subcutaneous fat pads (inguinaland the groin) in EP3R�/� mice at �6 months (Fig. 1D). At 3 and6 months of age, abdominal fat accounted for 6.62 � 0.47% and8.59 � 0.25% of the total body weight in EP3R�/� mice comparedwith the 3.37 � 0.69% and 3.46 � 0.16% in the WT mice,

Author contributions: M.S.-A., I.V.T., C.N.D., B.C., and T.B. designed research; M.S.-A., I.K.,S.E.B., I.V.T., C.N.D., and B.C. performed research; I.K., S.E.B., I.V.T., and C.N.D. contributednew reagents/analytic tools; M.S.-A., I.K., I.V.T., C.N.D., B.C., and T.B. analyzed data; andM.S.-A. and T.B. wrote the paper.

The authors declare no conflict of interest.

Freely available online through the PNAS open access option.

Abbreviation: PGE2, prostaglandin E2.

*To whom correspondence should be addressed. E-mail: tbartfai@scripps.edu.

© 2007 by The National Academy of Sciences of the USA

www.pnas.org�cgi�doi�10.1073�pnas.0611209104 PNAS � February 20, 2007 � vol. 104 � no. 8 � 3009–3014

PHYS

IOLO

GY

Dow

nloa

ded

by g

uest

on

Apr

il 30

, 202

0 D

ownl

oade

d by

gue

st o

n A

pril

30, 2

020

Dow

nloa

ded

by g

uest

on

Apr

il 30

, 202

0

respectively (Fig. 1C). Subcutaneous fat pads at both ages ac-counted for 2.49 � 0.33% and 5.2 � 0.28% of the total body weightin EP3R�/� mice compared with the 1.99 � 0.32% and 1.47 �0.12% in the WT mice, respectively (Fig. 1D). No significantdifferences in the brown adipose tissue percentage of body weightwere seen between EP3R�/� and WT mice.

EP3R�/� Mice Showed High Serum Levels of Insulin and Leptin at 3 and6 Months of Age. Measurements of circulating levels of insulin andleptin demonstrated that EP3R�/� mice had significantly elevatedlevels of both hormones. At 3 and 6 months of age, insulin levelsaccounted for 12.30 � 1.37 ng/ml and 14.50 � 1.37 ng/ml inEP3R�/� mice compared with 6.95 � 1.44 ng/ml and 11.01 � 1.96ng/ml in the WT mice during the light cycle (insulin, t � 2.3, P �0.042, EP3R�/� mice vs. WT mice at 3 months of age; insulin, t �1.4, P � 0.187, EP3R�/� mice vs. WT mice at 6 months of age);insulin levels accounted for 15.43 � 1.5 ng/ml and 16.49 � 0.65ng/ml in EP3R�/� mice compared with 11.01 � 1.96 ng/ml and6.18 � 3.91 ng/ml in the WT mice during the dark cycle (insulin, t �2.74, P � 0.033, EP3R�/� mice vs. WT mice at 3 months of age;insulin, t � 4.7, P � 0.003, EP3R�/� mice vs. WT mice at 6 monthsof age) (Fig. 2 Upper).

Leptin levels were higher in EP3R�/� mice compared withWT mice during both the light and dark cycles. At 3 and 6 monthsof age, leptin levels were 19.61 � 2.79 ng/ml and 22.94 � 2.43ng/ml in EP3R�/� mice compared with 10.18 � 1.45 ng/ml and7.17 � 1.03 ng/ml in the WT mice during the light cycle (leptin,t � 3.0, P � 0.013, EP3R�/� mice vs. WT mice at 3 months ofage; leptin, t � 4.8, P � 0.0019, EP3R�/� mice vs. WT mice at6 months of age); leptin levels were 21.27 � 0.95 ng/ml and25.07 � 1.47 ng/ml in EP3R�/� mice compared with 7.16 � 1.03ng/ml and 2.78 � 0.11 ng/ml in the WT mice during the dark cycle(leptin, t � 9.1, P � 0.0001, EP3R�/� mice vs. WT mice at 3months of age; leptin, t � 10.26, P � 0.0001, EP3R�/� mice vs.WT mice at 6 months of age) (Fig. 2 Lower).

EP3R�/� Mice Showed Impaired Glucose Tolerance and Insulin Resis-tance. At 3 months of age, the increased adiposity in EP3R�/�

mice was accompanied by abnormalities in glucose metabolism.Even though mice were fasted for 24 h before the glucose-tolerance test, basal levels of glucose in EP3R�/� mice weresignificantly higher compared with WT mice (EP3R�/�,149.25 � 15.91 mg/dl; WT, 103.01 � 10.1 mg/dl; t � 2.45; P �0.04) (Fig. 3A). After the mice were challenged with a 1.5 mg ofglucose per gram of body weight, serum glucose concentrationswere similarly elevated at 15 min in EP3R�/� and WT mice, butserum glucose concentrations reached higher levels in EP3R�/�

mice at 30 min (445 � 21.08 mg/dl) than in the WT mice (381 �22.54 mg/dl) (t � 2.08; P � 0.04). At 60 min after the glucose

Fig. 1. EP3R�/� mice show an increase in body weightthat results in an obese phenotype. (A) A male EP3R�/�

mouse (Right) and a WT littermate (Left) at 30 weeks ofage. (B) Measurement of body weight at different timepoints shows that overweight becomes significant onEP3R�/� mice after week 12 (t � 3.17; P � 0.048). Evalu-ation of body weight at different time points shows aconsistent increment of body weight in EP3R�/� mice atweeks 16 (t � 5.29; P � 0.0001), 20 (t � 4.41; P � 0.001),27 (t � 5.97; P � 0.0001), 31 (t � 6.36; P � 0.0001), and 40(t � 4.7; P � 0.0006). After week 30, weight differencesare �30% of the body weight of corresponding WTlittermate controls. (C) Early deposition of intraabdomi-nal and subcutaneous fat pads in EP3R�/� mice at 3months of age becomes an important contributor tobody weight gain (t � 3.79; P � 0.0053); at 6 months ofage, the contribution becomes higher (t � 18.36; P �0.0001). (D)At3months, intraabdominal fatpadsarethemain site of fat accumulation in EP3R�/� mice (t � 4.35;P � 0.0024). At 6 months, the intraabdominal fat pads(t � 19.55; P � 0.0001), liver (t � 4.68; P � 0.0016), andsubcutaneous fat pads (t � 12.09; P � 0.0001) becomeimportant sources of fat deposits in EP3R�/� mice. Aster-isks indicate significant differences at each given point.For P values, see above.

Fig. 2. Endocrine changes in EP3R�/� mice vs. WT littermates at 3 and 6 monthsof age. (Upper) Plasma insulin levels. (Lower) Plasma leptin levels. Asterisksindicate significant differences at each given point. For P values, see text.

3010 � www.pnas.org�cgi�doi�10.1073�pnas.0611209104 Sanchez-Alavez et al.

Dow

nloa

ded

by g

uest

on

Apr

il 30

, 202

0

challenge, glucose levels were similar in both EP3R�/� and WTmice, but higher levels in EP3R�/� mice were reached again at120 min (EP3R�/�, 242.25 � 19.55 mg/dl; WT, 136.25 � 20.42mg/dl; t � 3.78; P � 0.0091) and 180 min (EP3R�/�, 157.5 � 6.38mg/dl; WT, 119.24 � 3.72 mg/dl; t � 5.17; P � 0.0021) (Fig. 3A).

At 6 months of age, abnormalities in glucose metabolism weresignificantly higher in EP3R�/� mice than at 3 months of age(Fig. 3B). Basal levels of glucose in EP3R�/� mice were signif-icantly higher compared with WT mice (EP3R�/�, 196.4 � 22.84mg/dl; WT, 105.82 � 6.17 mg/dl; t � 4.16; P � 0.0024). After aglucose challenge, EP3R�/� mice showed higher glucose levelsat 15 min (465.2 � 22.05 mg/dl) than did WT mice (382.5 � 22.31mg/dl) (t � 2.6; P � 0.0283). The same was true at 30 min(EP3R�/�, 487.8 � 11.7 mg/dl; WT, 404.83 � 33.35 mg/dl; t �2.26; P � 0.049), 60 min (EP3R�/�, 420.6 � 23.54 mg/dl; WT,282 � 22.15 mg/dl; t � 4.27; P � 0.0021), 120 min (EP3R�/�,286.1 � 25.93 mg/dl; WT, 233.65 � 16.75 mg/dl; t � 1.7; P �0.1135), and 180 min (EP3R�/�, 184.6 � 15.14 mg/dl; WT,146.5 � 9.21 mg/dl; t � 2.24; P � 0.048) (Fig. 3B).

We determined the ability of insulin to acutely stimulateglucose disposal or clearance by performing an acute insulinchallenge in EP3R�/� and WT mice at both 3 and 6 months ofage. At 3 months of age, the ability of insulin to acutely stimulateglucose disposal in EP3R�/� mice was significantly blunted at 15min, indicating a short-term decrease in insulin sensitivity(EP3R�/�, 158.3 � 10.59 mg/dl; WT, 125.8 � 11.79 mg/dl, whichcorresponded to a 84.36% and 63.63% reduction in baselinevalues of EP3R�/� and WT, respectively; P � 0.0246) (Fig. 3 C

and C�). At 6 months of age, basal levels of glucose were higher inEP3R�/� mice (EP3R�/�, 287.25 � 48.32 mg/dl; WT, 173.02 �17.16 mg/dl), and acute insulin challenge was unable to stimulateglucose disposal in EP3R�/� mice at 15 min (EP3R�/�, 265 �53.71 mg/dl; WT, 125.3 � 16.04 mg/dl, which corresponded to a90.45% and 73.01% reduction in baseline values of EP3R�/� andWT, respectively; P � 0.046). Even though values at 30 min(EP3R�/�, 280.12 � 48.5 mg/dl; WT, 148.02 � 13.21 mg/dl), 60min (EP3R�/�, 274.07 � 41.69 mg/dl; WT, 156.23 � 22.43mg/dl), or 120 min (EP3R�/�, 290.94 � 44.14 mg/dl; WT,185.16 � 28.89 mg/dl) were significantly higher in EP3R�/� mice,comparisons of the percentage of change from their own base-line were not significant (Fig. 3 D and D�).

EP3R�/� Mice Exhibited Increased Motor Activity and Core BodyTemperature. Radiotelemetric evaluation of motor activity indicatesthat EP3R�/� mice displayed peaks of increased motor activityduring the light part of the day, when mice usually spend most oftheir time sleeping. Increased nocturnal activity was recordedcontinuously over 21 days in six EP3R�/� mice and six WTlittermates. Continuous motor activity profile during 21 days isshown (Fig. 4A), and this profile demonstrates that the peaks ofnocturnal activity in EP3�/� mice are not episodic but instead arerecurrent daily events occurring with a frequency of between twoand four episodes per night (Fig. 4 A and C). The increase in motoractivity during the light cycle in EP3R�/� mice compared with WTmice was punctuated and separated by phases in which the motoractivity of the EP3R�/� mice was indistinguishable from that of theWT mice (Fig. 4 B and C). However, cumulative analysis demon-strated that these peaks contributed to an overall 60.3% increasedmotor activity in EP3R�/� mice. Increased motor activity inEP3R�/� mice was also observed during the dark cycle, and thisincrease accounted for 24.73% of the motor activity (Fig. 4B). Corebody temperature of EP3R�/� mice was slightly elevated during thecorresponding peaks of increased motor activity (see Fig. 4D).

EP3R�/� Mice Showed Increased Food Consumption. Measurement offood consumption in EP3R�/� and WT mice demonstrated thatEP3R�/� ate significantly more food than the WT littermates atboth ages. At 3 months of age, hourly food intake was similarbetween EP3R�/� and WT mice during the light cycle. However,EP3R�/� mice during the dark period showed an increase in foodintake characterized by a continuous feeding (monophasic),whereas WT littermate controls showed an increase in feedingfollowed by a period of temporary decrease and then by a secondincrease in feeding, making the pattern biphasic (Fig. 5A). As aconsequence, statistical differences are only observed in the darkperiod at 0:00 and 1:00 a.m. (0:00: EP3R�/�, 0.338 � 0.02 g; WT,0.236 � 0.03 g; t � 2.35; P � 0.0466; 1:00: EP3R�/�, 0.298 � 0.03 g;WT, 0.184 � 0.03 g; t � 2.33; P � 0.0481). At 6 months of age, theincrease in feeding strongly correlated with the peaks of activity andelevated temperature during both the light/inactive and the dark/active part of the day (Fig. 5B). Cumulative food intake was higherin EP3R�/� mice in both periods. During the light phase, significantdifferences were at 7:00 a.m. (EP3R�/�, 0.19 � 0.03 g; WT, 0.09 �0.02 g; t � 2.25; P � 0.0379), 9:00 a.m. (EP3R�/�, 0.17 � 0.04 g; WT,0.05 � 0.01 g; t � 2.51; P � 0.022), 15:00 p.m. (EP3R�/�, 0.16 �0.02 g; WT, 0.04 � 0.02 g; t � 3.27; P � 0.004), and 17:00 p.m.(EP3R�/�, 0.17 � 0.06 g; WT, 0.03 � 0.01 g; t � 2.14; P � 0.047).During the dark phase, significant differences were at 0:00 a.m.(EP3R�/�, 0.36 � 0.03 g; WT, 0.17 � 0.04 g; t � 3.08; P � 0.006)and 1:00 a.m. (EP3R�/�, 0.27 � 0.03 g; WT, 0.15 � 0.03 g; t � 2.2;P � 0.041).

Both EP3R�/� and WT mice underwent food deprivation for24 h, and subsequent food intake was measured every 12 h. Baselinemeasurement for 3 days confirmed differences in food intake. Fooddeprivation increased food intake in the next dark and light periodin both strains, and full recovery was reached on day 4 (Fig. 5B).

Fig. 3. EP3R�/� mice showed impaired gluclose tolerance and insulin resis-tance. (A and B) Glucose-tolerance test. Mice were challenged with 1.5 mg ofglucose per gram of body weight, and glucose concentrations were measuredbefore glucose challenge and after 15, 30, 60, 120, and 180 min. (C and D)Percentage of changes from baseline on the ability of insulin to acutelystimulate glucose disposal or clearance by performing an acute insulin chal-lenge in EP3R�/� and WT mice at both ages. (C� and D�) The raw correspondentblood glucose values (in milligrams per deciliter) used for C and D. Asterisksindicate significant differences at each given point. For P values, see text.

Sanchez-Alavez et al. PNAS � February 20, 2007 � vol. 104 � no. 8 � 3011

PHYS

IOLO

GY

Dow

nloa

ded

by g

uest

on

Apr

il 30

, 202

0

DiscussionThe observations reported here that EP3R�/� mice have anobese phenotype include measurements of body weight, as wellas endocrine parameters known to increase with obesity, such asinsulin and leptin plasma levels. Although it has been reportedthat PGE2 stimulates leptin release from adipocytes (52), weshow that the lack of EP3R signaling in EP3R�/� mice does notprevent the large increase in leptin levels (Fig. 2 Lower), whichis commensurate with the increased fat deposits (Fig. 1C). Thus,PGE2 stimulation of leptin release does not exclusively dependon signaling through the EP3 prostanoid receptor subtype.

The levels of PGE2 in obese humans are elevated (56). It will beinteresting to test the hypothesis that such elevation could beattributable to disruption of PGE2-negative feedback throughEP3R.

EP3R receptors are broadly expressed in the periphery and inthe brain. Our study did not determine the site(s) of action thatcontributes to the obese phenotype of EP3R�/� mice, becauseboth peripheral and central expression of EP3R is absent in thenull mouse. The exact localization of the prostanoid receptor(s)-mediated febrile effects by transgenic and lentiviral techniquesis currently being studied by Lazarus (57), and similar techniqueswill be needed to precisely determine the site(s) at which the lackof PGE2 signaling via EP3R contributes to this obese phenotype.Although it is possible to speculate that obesity in EP3R�/� mice

may be caused in part by the lack of hypothalamic PGE2signaling, numerous peripheral actions of PGE2 are also medi-ated through this prostanoid receptor subtype (34).

The finding of increased feeding coupled with nocturnal motoractivity in EP3R�/� mice compared with WT littermates suggeststhat EP3R�/� mice do not stabilize sleep and may wake up moreeasily. The role of PGE2 as a somnogenic agent alongside prosta-glandin D in sleep has been reported (45). Our data suggest thatEP3R may actually mediate a part of this response.

The most likely brain–EP3R-linked response is the febrileresponse. It has been shown that EP3R�/� mice do not mountfever in response to the pyrogens IL-1� and LPS, suggesting thatthe PGE2 action in the anterior hypothalamus, raphe pallidus,and other sites involved in the generation of the fever responseinvolves EP3R (5). The basal temperature regulation may alsoinvolve EP3R-mediated effects of PGE2, and we observed aslightly elevated core body temperature in freely moving adlibitum-fed EP3R�/� mice. The increased core body tempera-ture and motor activity would be predicted to lead to a decreasein body weight because energy demands are increased whenhigher body temperature needs to be maintained, but it appearsas though the increase in food consumption/energy intake has amore pronounced effect, thus resulting in body weight gain.

The onset of obesity does not occur late in development inEP3R�/� mice and does not seem to require additional age-dependent factors to come into play. It is important to note that

Fig. 4. Circadian rhythm profile for WT littermates and EP3R�/� mice. Although diurnal distribution of motor activity follows the light–dark cycle, EP3R�/� miceshow bouts of increased activity during the light cycle that are associated with grooming and eating behavior. Those bouts of activity are irregular and betterdetected during the resting phase (see �). (B) Continuous recording of core body temperature (CBT) and motor activity (MA) during 5 days at normothermicconditions (room temperature was 30°C) confirms that EP3R�/� and WT mice are nocturnal and that they follow the low activity–high resting (light cycle) andhigh activity–low resting (dark cycle) pattern. (C) Averaged data indicate that EP3R�/� mice have an increase in motor activity characterized by bouts of activitythat increases the core body temperature (see arrows). The increase in motor activity is associated with grooming and eating behavior. (D) Cumulative dataconfirm that EP3R�/� mice are more active during the light period. *, P � 0.032.

3012 � www.pnas.org�cgi�doi�10.1073�pnas.0611209104 Sanchez-Alavez et al.

Dow

nloa

ded

by g

uest

on

Apr

il 30

, 202

0

the obesity in EP3R�/� mice does not result from the high-fatdiet that is now commonly used to achieve an obese state inexperimental animals; rather, this obesity in EP3R�/� miceoccurs on standard chow and results from increased feeding inthe absence of a commensurate increase in energy expenditure.This increase in energy intake relative to energy expenditure isoften deemed to be an important factor in the etiology ofcommon forms of human obesity. In addition, it is worth notingthat EP3R�/� mice show another common feature of weight-gain scenarios in humans: night eating.

It is important to note that the lack of the full spectrum ofinflammatory signaling has been shown recently to lead toobesity in a multitude of transgenic models including IL-1R1�/�

(58), IL-1��/�/IL-6�/� double knockout (59), and IL-18�/� (60)mice. These observations, together with those in the EP3R�/�

mice described here, suggest that inf lammatory moleculesand/or signaling may be required for keeping body weighthomeostasis. The exact mechanism leading to obesity in thesemodels is not known, although effects of cytokines on insulinresistance and insulin receptor substrate expression (61) andphosphorylation have been proposed (62). It will be importantto determine whether those mechanisms are distinct or similarto those leading to obesity in EP3R�/� mice. Full pharmaco-logical characterization of the contribution of EP3 mediated-signaling in obesity and other phenotypes awaits the introductionof an EP3 prostanoid receptor-specific antagonist, in the samemanner that EP1 antagonist-medicated impulsive behaviorserved to verify the involvement of that receptor subtype in thebehavioral effects (63).

In summary, mice that are null for the prostanoid receptor EP3display increased feeding throughout the day along with additionalfeeding activity peaks during the night (or light period), resulting inobesity. These mice may represent an important obesity model.

Materials and MethodsAnimals. All procedures were approved by the InstitutionalAnimal Care and Use Committee of the Scripps Research

Institute and were carried out on male EP3R�/� mice back-crossed to C57BL/6 background over more than eight genera-tions and on WT littermates.

Food-Intake Measurements. Mice were fed ad libitum with mousebreeder diet (S-2335 Mouse Breeder, gross energy kcal (1 kcal �4.18 kJ)/g 4.39, protein % 17.50, fat % 11.72, fiber % 3.36; HarlanTeklad, Madison, WI) and separated in two groups on each age(n � 10 each group). Food intake was monitored every hour during24 h. For daily consumption, food and body weight were monitoredtwice per day at the onset of the dark and light period (6:00 a.m. and6:00 p.m.) for 9 days. Mice were deprived of food for 24 h on day4. Observation of food consumption was evaluated for 5 additionaldays after the food deprivation. Body weight was normalized formetabolic demands of body mass according to Kleiber’s function (gweight loss/g baseline weight0.75).

Body Weight and Fat Distribution. At the end of this period, micewere anesthetized, and intraabdominal fat pads (gonadal, retroper-itoneal, and mesenteric), liver, subcutaneous fat pad (inguinal andthe groin), and brown adipose tissue were dissected and weighed.

Glucose-Tolerance Tests. A glucose-tolerance test was performedat the onset of the light cycle (6:00 a.m.). Mice were weighed andfasted for 24 h before the glucose-tolerance test. Access todrinking water was allowed during this period. On the day of thetest, baseline glucose levels and body weight were determinedbefore challenge with a glucose load of 1.5 mg of glucose pergram of body weight (D-glucose, anhydrous; Sigma–Aldrich, St.Louis, MO) dissolved in sterile distilled water (0.75 g of D-glucose, anhydrous in 10 ml of sterile water). The mouse wasrestrained by holding the excess skin at the base of the neckbetween the technician’s thumb and forefinger. The mouse’s tailwas left hanging out and placed on a glass slide, and a segmentof �1 mm in length was cut off the tip of the tail by using a sharprazor blade. A small drop (�5 �l) of blood was placed on the

Fig. 5. EP3R�/� mice showed increased food consumption. (A and B) Hourly food-intake events as a function of time (A) and effects of 24 h of fasting inEP3R�/� and WT littermates (B). In A, at 3 months of age (Upper), EP3R�/� mice show an increase in food intake characterized by continuous feeding during thedark period (see arrows), whereas WT littermate controls show a similar increase in feeding followed by a period of temporary decrease and then by a secondincrease in food consumption in the same period. At 6 months of age (Lower), the increase in feeding in the EP3R�/� mice is extended to the light cycle (see arrows;P values are given in the text). In B, food intake was measured every 12 h. At 3 months of age (Upper), baseline measurement for 3 days confirmed that EP3R�/�

mice ate more during the dark period (*, P � 0.05). At 6 months of age (Lower), cumulative data indicate that EP3R�/� mice ate more during both light and darkcycles (*, P � 0.05). At both ages, food deprivation increases food intake in the next dark and light period in both strains and returns to the baseline patternafter day 4.

Sanchez-Alavez et al. PNAS � February 20, 2007 � vol. 104 � no. 8 � 3013

PHYS

IOLO

GY

Dow

nloa

ded

by g

uest

on

Apr

il 30

, 202

0

glucometer test strip (Home Diagnostics, Fort Lauderdale, FL).After a 5-second developing time, the baseline blood glucosevalue was recorded (in mg per deciliter), and the mouse wasreturned to his home cage. After the baseline glucose measure-ment, the mouse was injected i.p. by using a 1-ml syringe and a27-gauge needle. The time of the injection was noted, and 15-,30-, 60-, and 120-min postinjection blood glucose measurementswere performed again. It was sometimes necessary to remove ascab that formed at the initial tail-cut site to collect the secondblood sample (n � 6 for each group). After the end of the study,mice were returned, and food and water was provided at libitum.

Insulin-Resistance Test. The effects of insulin injection were as-sessed in nonfasted male mice. Similar to the glucose-tolerancetest, blood was withdrawn from the tail without anesthesia beforeadministration of human insulin (1 unit/kg, i.p.; Sigma–Aldrich).Samples were collected 15, 30, 60, and 120 min after the insulinchallenge. Blood glucose levels were determined by a bloodglucose meter (Home Diagnostics) (n � 6 for each group).

Plasma Levels of Leptin and Insulin. Plasma leptin and insulin levelswere determined at 6:00 a.m. and 6:00 p.m., at the onset of the lightand dark cycle, respectively. Mice were euthanized with isoflorane(5%), decapitated, and bled into EDTA-coated tubes. Blood wascentrifuged at 10,000 � g for 10 min at 4°C, and supernatants weretaken and stored at �70°C until further analysis. Plasma leptin wasmeasured by using a Mouse Leptin ELISA kit 96-well plate (catalogno. EZML-82K; Millipore, Billerica, MA) used for a nonradioac-tive quantification of leptin, and values were collected and aver-

aged � SEM. Mouse insulin was determined by RIA using a RatInsulin RIA kit (250 tubes; catalog no. RI-13K; Millipore) accord-ing to manufacturer’s instructions (n � 6 for each group).

Telemetry Device Implant. EP3R�/� and WT littermate male micewere anesthetized with isoflorane (induction, 3–5%; mainte-nance, 0.9–1.5%) and implanted with radio telemetry devices(TA10TA-F20; Data Sciences, Inc., St. Paul, MN) into theperitoneal cavity for core body temperature measurement. Micewere allowed to recover for 2 weeks and were then submitted forfreely moving recording (n � 10 for each group). Mice weremaintained in a temperature-controlled room (25°C) on a 12-hlight–dark cycle (light on at 6 a.m.). Core body temperature andmotor activity sensors were located in the transmitter implant.The cages were positioned onto the receiver plates. Radio signalsfrom the core body temperature and motor activity of eachanimal (number of horizontal movements) were continuouslymonitored with a fully automated data-acquisition system(Dataquest A.R.T.; Data Sciences, Inc.).

Data Analysis. Data were grouped and analyzed by using the pairedt test or ANOVA with repeated measures followed by post hocNewman–Keuls test. All results are expressed as means � SE.Metabolic efficiency was calculated as the energy intake divided bythe body weight gain over a certain period. Linear relationshipswere estimated by using Pearson’s moment correlation coefficient.

We thank Professor Jerold Chun for insightful suggestions. This workwas supported by funds from the Skaggs Institute of Chemical Biologyat The Scripps Research Institute.

1. Alheim K, Bartfai T (1998) Ann N Y Acad Sci 840:51–58.2. Ivanov AI, Romanovsky AA (2004) Front Biosci 9:1977–1993.3. Oka T (2004) Front Biosci 1:3046–3057.4. Oka T, Oka K, Kobayashi T, Sugimoto Y, Ichikawa A, Ushikubi F, Narumiya S, Saper CB

(2003) J Physiol (London) 551:945–954.5. Ushikubi F, Segi E, Sugimoto Y, Murata T, Matsuoka T, Kobayashi T, Hizaki H, Tuboi K,

Katsuyama M, Ichikawa A, et al. (1998) Nature 395:281–284.6. Walch L, Clavarino E, Morris PL (2003) Endocrinology 144:1284–1291.7. Higgs GA (1986) Prog Lipid Res 25:555–561.8. Hoozemans JJ, Veerhuis R, Janssen I, Rozemuller AJ, Eikelenboom P (2001) Exp Gerontol

36:559–570.9. Pinteaux E, Parker LC, Rothwell NJ, Luheshi GN (2002) J Neurochem 83:754–763.

10. Dinarello CA (2004) J Endotoxin Res 10:201–222.11. Saper CB (1998) Ann NY Acad Sci 856:90–94.12. Coelho MM, Luheshi G, Hopkins SJ, Pela IR, Rothwell NJ (1995) Am J Physiol 269:R527–R535.13. Derijk R, Van Rooijen N, Tilders FJ, Besedovsky HO, Del Rey A, Berkenbosch F (1991)

Endocrinology 129:330–338.14. Lenczowski MJ, Bluthe RM, Roth J, Rees GS, Rushforth DA, van Dam AM, Tilders FJ,

Dantzer R, Rothwell NJ, Luheshi GN (1999) Am J Physiol 276:R652–R658.15. Murakami N, Watanabe T (1989) Brain Res 478:171–174.16. Arnold J, Little RA, Rothwell NJ (1989) J Appl Physiol 66:1970–1975.17. Jepson MM, Millward DJ, Rothwell NJ, Stock MJ (1988) Am J Physiol 255:E617–E620.18. Xie Y, Kang X, Ackerman WEt, Belury MA, Koster C, Rovin BH, Landon MB, Kniss DA

(2006) Diabetes Obes Metab 8:83–93.19. Bos CL, Richel DJ, Ritsema T, Peppelenbosch MP, Versteeg HH (2004) Int J Biochem Cell

Biol 36:1187–1205.20. Narumiya S, Sugimoto Y, Ushikubi F (1999) Physiol Rev 79:1193–1226.21. Matsuoka Y, Furuyashiki T, Bito H, Ushikubi F, Tanaka Y, Kobayashi T, Muro S, Satoh

N, Kayahara T, Higashi M, et al. (2003) Proc Natl Acad Sci USA 100:4132–4137.22. Takeuchi O, Hoshino K, Kawai T, Sanjo H, Takada H, Ogawa T, Takeda K, Akira S (1999)

Immunity 11:443–451.23. Audoly LP, Tilley SL, Goulet J, Key M, Nguyen M, Stock JL, McNeish JD, Koller BH,

Coffman TM (1999) Am J Physiol 277:H924–H930.24. Stock JL, Shinjo K, Burkhardt J, Roach M, Taniguchi K, Ishikawa T, Kim HS, Flannery PJ,

Coffman TM, McNeish JD, et al. (2001) J Clin Invest 107:325–331.25. Tilley SL, Audoly LP, Hicks EH, Kim HS, Flannery PJ, Coffman TM, Koller BH (1999)

J Clin Invest 103:1539–1545.26. Hartney JM, Coggins KG, Tilley SL, Jania LA, Lovgren AK, Audoly LP, Koller BH (2006)

Am J Physiol 290:L105–L113.27. McCoy JM, Wicks JR, Audoly LP (2002) J Clin Invest 110:651–658.28. Li X, Okada Y, Pilbeam CC, Lorenzo JA, Kennedy CR, Breyer RM, Raisz LG (2000)

Endocrinology 141:2054–2061.29. Kennedy CR, Zhang Y, Brandon S, Guan Y, Coffee K, Funk CD, Magnuson MA, Oates

JA, Breyer MD, Breyer RM (1999) Nat Med 5:217–220.30. Mori K, Tanaka I, Kotani M, Miyaoka F, Sando T, Muro S, Sasaki Y, Nakagawa O, Ogawa

Y, Usui T, et al. (1996) J Mol Med 74:333–336.31. Sakamoto A, Matsumura J, Mii S, Gotoh Y, Ogawa R (2004) Shock 22:76–81.32. Sakuma Y, Tanaka K, Suda M, Komatsu Y, Yasoda A, Miura M, Ozasa A, Narumiya S,

Sugimoto Y, Ichikawa A, et al. (2000) Infect Immun 68:6819–6825.33. Watanabe K, Kawamori T, Nakatsugi S, Ohta T, Ohuchida S, Yamamoto H, Maruyama T,

Kondo K, Ushikubi F, Narumiya S, et al. (1999) Cancer Res 59:5093–5096.

34. Narumiya S, FitzGerald GA (2001) J Clin Invest 108:25–30.35. Takeuchi K, Ukawa H, Kato S, Furukawa O, Araki H, Sugimoto Y, Ichikawa A, Ushikubi

F, Narumiya S (1999) Gastroenterology 117:1128–1135.36. Audoly LP, Ma L, Feoktistov I, de Foe SK, Breyer MD, Breyer RM (1999) J Pharmacol Exp

Ther 289:140–148.37. Ma H, Hara A, Xiao CY, Okada Y, Takahata O, Nakaya K, Sugimoto Y, Ichikawa A,

Narumiya S, Ushikubi F (2001) Circulation 104:1176–1180.38. Amano H, Hayashi I, Endo H, Kitasato H, Yamashina S, Maruyama T, Kobayashi M, Satoh

K, Narita M, Sugimoto Y, et al. (2003) J Exp Med 197:221–232.39. Shoji Y, Takahashi M, Kitamura T, Watanabe K, Kawamori T, Maruyama T, Sugimoto Y,

Negishi M, Narumiya S, Sugimura T, et al. (2004) Gut 53:1151–1158.40. Kunikata T, Yamane H, Segi E, Matsuoka T, Sugimoto Y, Tanaka S, Tanaka H, Nagai H,

Ichikawa A, Narumiya S (2005) Nat Immunol 6:524–531.41. Minami T, Nakano H, Kobayashi T, Sugimoto Y, Ushikubi F, Ichikawa A, Narumiya S, Ito

S (2001) Br J Pharmacol 133:438–444.42. Fain JN, Leffler CW, Bahouth SW (2000) J Lipid Res 41:1689–1694.43. Fain JN, Ballou LR, Bahouth SW (2001) Prostaglandins Other Lipid Mediat 65:199–209.44. Hayaishi O, Matsumura H (1995) Adv Neuroimmunol 5:211–216.45. Ram A, Pandey HP, Matsumura H, Kasahara-Orita K, Nakajima T, Takahata R, Satoh S,

Terao A, Hayaishi O (1997) Brain Res 751:81–89.46. Yang J, Xia M, Goetzl EJ, An S (1994) Biochem Biophys Research Commun 198:999–1006.47. Kotani M, Tanaka I, Ogawa Y, Usui T, Tamura N, Mori K, Narumiya S, Yoshimi T, Nakao

K (1997) Genomics 40:425–434.48. Nakamura K, Kaneko T, Yamashita Y, Hasegawa H, Katoh H, Negishi M (2000) J Comp

Neurol 421:543–569.49. Ek M, Arias C, Sawchenko P, Ericsson-Dahlstrand A (2000) J Comp Neurol 428:5–20.50. Yoshida K, Nakamura K, Matsumura K, Kanosue K, Konig M, Thiel HJ, Boldogkoi Z, Toth

I, Roth J, Gerstberger R, et al. (2003) Eur J Neurosci 18:1848–1860.51. Tabarean IV, Behrens MM, Bartfai T, Korn H (2004) Proc Natl Acad Sci USA 101:2590–

2595.52. Fain JN, Leffler CW, Bahouth SW, Rice AM, Rivkees SA (2000) Prostaglandins Other Lipid

Mediat 62:343–350.53. Fantuzzi G (2005) J Allergy Clin Immunol 115:911–919.54. Bastard JP, Maachi M, Lagathu C, Kim MJ, Caron M, Vidal H, Capeau J, Feve B (2006)

Eur Cytokine Netw 17:4–12.55. Wisse BE (2004) J Am Soc Nephrol 15:2792–2800.56. Fain JN, Kanu A, Bahouth SW, Cowan GS, Jr., Hiler ML, Leffler CW (2002) Prostaglandins

Leukot Essent Fatty Acids 67:467–473.57. Lazarus M (2006) Mol Nutr Food Res 50:451–455.58. Garcia MC, Wernstedt I, Berndtsson A, Enge M, Bell M, Hultgren O, Horn M, Ahren B,

Enerback S, Ohlsson C, et al. (2006) Diabetes 55:1205–1213.59. Chida D, Osaka T, Hashimoto O, Iwakura Y (2006) Diabetes 55:971–977.60. Netea MG, Joosten LA, Lewis E, Jensen DR, Voshol PJ, Kullberg BJ, Tack CJ, van Krieken

H, Kim SH, Stalenhoef AF, et al. (2006) Nat Med 12:650–656.61. Jager J, Gremeaux T, Cormont M, Le Marchand-Brustel Y, Tanti JF (2007) Endocrinology

148:241–251.62. Kim JA, Yeh DC, Ver M, Li Y, Carranza A, Conrads TP, Veenstra TD, Harrington MA,

Quon MJ (2005) J Biol Chem 280:23173–23183.63. Matsuoka Y, Furuyashiki T, Yamada K, Nagai T, Bito H, Tanaka Y, Kitaoka S, Ushikubi

F, Nabeshima T, Narumiya S (2005) Proc Natl Acad Sci USA 102:16066–16071.

3014 � www.pnas.org�cgi�doi�10.1073�pnas.0611209104 Sanchez-Alavez et al.

Dow

nloa

ded

by g

uest

on

Apr

il 30

, 202

0

Corrections

BIOCHEMISTRY. For the article ‘‘The crystal structure of twomacrolide glycosyltransferases provides a blueprint for host cellantibiotic immunity,’’ by David N. Bolam, Shirley Roberts, MarkR. Proctor, Johan P. Turkenburg, Eleanor J. Dodson, CarlosMartinez-Fleites, Min Yang, Benjamin G. Davis, Gideon J.Davies, and Harry J. Gilbert, which appeared in issue 13, March27, 2007, of Proc Natl Acad Sci USA (104:5336–5341; firstpublished March 21, 2007; 10.1073�pnas.0607897104), the au-thors note that, in addition to writing the paper, Benjamin G.Davis should be credited with designing the research andanalyzing the data. The corrected author contributions footnoteappears below.

Author contributions: D.N.B., S.R., and M.R.P. contributedequally to this work; D.N.B., M.R.P., B.G.D., G.J.D., and H.J.G.designed research; D.N.B., S.R., and M.R.P. performed re-search; D.N.B., M.R.P., J.P.T., E.J.D., C.M.-F., M.Y., B.G.D.,and H.J.G. analyzed data; and B.G.D., G.J.D., and H.J.G. wrotethe paper.

www.pnas.org�cgi�doi�10.1073�pnas.0704090104

ENVIRONMENTAL SCIENCES. For the article ‘‘Combined climate andcarbon-cycle effects of large-scale deforestation,’’ by G. Bala, K.Caldeira, M. Wickett, T. J. Phillips, D. B. Lobell, C. Delire, andA. Mirin, which appeared in issue 16, April 17, 2007, of Proc NatlAcad Sci USA (104:6550–6555; first published April 9, 2007;10.1073�pnas.0608998104), the authors note the following. Onpage 6550, left column, in paragraph 1 of the main text, line 7,the phrase ‘‘However, deforestation also exerts a cooling influ-ence by (iv) decreasing the surface albedo’’ should instead read:‘‘However, deforestation also exerts a cooling influence by (iv)increasing the surface albedo.’’ Also on page 6550, right column,in Results, paragraph 2, line 1, the phrase ‘‘the atmospheric CO2concentration is higher by 299, 110, and 5 ppmv’’ should insteadread: ‘‘the atmospheric CO2 concentration is higher by 199, 110,and 5 ppmv.’’ Lastly, in the Fig. 1 legend, line 8, the phrase ‘‘netcooling, near-zero temperature change, and net warming, re-spectively’’ should instead read: ‘‘net warming, near-zero tem-perature change, and net cooling, respectively.’’ These errors donot affect the conclusions of the article.

www.pnas.org�cgi�doi�10.1073�pnas.0704096104

GENETICS. For the article ‘‘Deletion of the orphan nuclear recep-tor COUP-TFII in uterus leads to placental deficiency,’’ byFabrice G. Petit, Soazik P. Jamin, Isao Kurihara, Richard R.Behringer, Francesco J. DeMayo, Ming-Jer Tsai, and Sophia Y.Tsai, which appeared in issue 15, April 10, 2007, of Proc NatlAcad Sci USA (104:6293–6298; first published April 2, 2007;10.1073�pnas.0702039104), the authors note that the e-mailaddress for corresponding author Fabrice G. Petit appearedincorrectly. The correct address is fabrice.petit@u-psud.fr. Theonline version has been corrected.

www.pnas.org�cgi�doi�10.1073�pnas.0703353104

PHYSIOLOGY. For the article ‘‘Night eating and obesity in theEP3R-deficient mouse,’’ by Manuel Sanchez-Alavez, IzabellaKlein, Sara E. Brownell, Iustin V. Tabarean, Christopher N.Davis, Bruno Conti, and Tamas Bartfai, which appeared in issue8, February 20, 2007, of Proc Natl Acad Sci USA (104:3009–3014;first published February 16, 2007; 10.1073�pnas.0611209104),the authors note that the following acknowledgment was inad-vertently omitted from the article: ‘‘We thank Professor ShuhNarumiya (Kyoto University, Kyoto, Japan) (4, 5, 21) for thegeneration and subsequent generous transfer of the EP3R�/�

mice.’’

www.pnas.org�cgi�doi�10.1073�pnas.0704094104

www.pnas.org PNAS � June 5, 2007 � vol. 104 � no. 23 � 9911

CORR

ECTI

ON

S