Neural dysregulation of peripheral insulin action andblood pressure by brain endoplasmic reticulum stressSudarshana Purkayasthaa, Hai Zhanga, Guo Zhanga, Zaghloul Ahmedb, Yi Wanga, and Dongsheng Caia,1

aDepartment of Molecular Pharmacology and Diabetes Research Center, Albert Einstein College of Medicine, Bronx, NY 10461; and bDepartment of PhysicalTherapy and Neuroscience Program, College of Staten Island/City University of New York, Staten Island, NY 10314

Edited by Marc R. Montminy, Salk Institute for Biological Studies, La Jolla, CA, and approved January 6, 2011 (received for review May 18, 2010)

Chronic endoplasmic reticulum (ER) stress was recently revealed toaffect hypothalamic neuroendocrine pathways that regulate feed-ing and body weight. However, it remains unexplored whetherbrain ER stress could use a neural route to rapidly cause the peri-pheral disorders that underlie the development of type 2 diabetes(T2D) and the metabolic syndrome. Using a pharmacologic modelthat delivered ER stress inducer thapsigargin into the brain, thisstudy demonstrated that a short-term brain ER stress over 3 dwas sufficient to induce glucose intolerance, systemic and hepaticinsulin resistance, and blood pressure (BP) increase. The collectionof these changes was accompanied by elevated sympathetic toneand prevented by sympathetic suppression. Molecular studiesrevealed that acute induction of metabolic disorders via brain ERstress was abrogated by NF-κB inhibition in the hypothalamus.Therapeutic experiments further revealed that acute inhibition ofbrain ER stress with tauroursodeoxycholic acid (TUDCA) partiallyreversed obesity-associated metabolic and blood pressure disor-ders. In conclusion, ER stress in the brain represents a mediator ofthe sympathetic disorders that underlie the development of insulinresistance syndrome and T2D.

During the recent two decades, the epidemic of type 2 diabetes(T2D) has reached an explosive scale in the United States

and many other developed countries. The risk factors for the de-velopment of T2D include a group of prognostic disorders knowncollectively as insulin resistance syndrome, which manifests fre-quently in the form of glucose intolerance, insulin resistance,dyslipidemia, and blood pressure (BP) increase in associationwith aging and obesity. As broadly documented in the literature(1–7), all of these disorders are characterized by the existenceof stress and inflammatory molecules in the circulation and var-ious tissues. Although it is still poorly understood how all thesepathophysiological changes are etiologically connected, a varietyof intracellular stresses have been proposed as primary patho-genic factors (8, 9). These advances have included the recentunderstanding on endoplasmic reticulum (ER) stress (10–12), aset of intracellular molecular responses when the ER fails toadapt to various physiological or pathological conditions thatchallenge the normal functions of ER. Under diabetes-proneenvironmental changes, induction of ER stress was reported tooccur in insulin-secreting pancreatic β-cells (13–15) and variousinsulin-responsive peripheral tissues (16–18), which togethercause the compromised regulation of glucose homeostasis by in-sulin. Most recently, chronic ER stress was revealed to occur inthe hypothalamus under conditions of nutritional excess andcause hypothalamic hormonal (leptin and insulin) defects thatpromote feeding and weight gain (19, 20). Such effects of chronicbrain ER stress are predicted to incur long-term pathologicalchanges that contribute to T2D in a manner which is secondaryto weight gain and obesity (19, 20).The central nervous system (CNS), in particular, the comprised

hypothalamus, provides key regulations on various metabolicprocesses of the body. Recent research has elucidated multipleendocrine pathways that mediate hypothalamic control of feedingand body weight (21–23). On the other hand, the CNS, includingthe hypothalamus, critically controls the autonomic nervous sys-tem outflow, which has acute effects on physiology, such as me-tabolism (24). The neural pathways of the CNS may explain the

recent observations that pharmacologic manipulation of the hy-pothalamus can rapidly alter glucose metabolism in the peripheraltissues without the involvement of feeding or body weight change(25–29). Thus, the development of T2D conceivably involves aneural mechanism that alters peripheral metabolism in a bodyweight (obesity)-independent manner; however, investigation ofsuch mediators in the CNS still remains insufficient. Notably,ER stress has been known as an inducer of various pathologicalchanges not only in peripheral metabolic tissues (10–12) but alsoin the brain (30). Very recently, ER stress in the hypothalamuswas shown to promote appetite and weight gain to contribute toobesity-associated diseases (19, 20). In light of other researchthat has revealed the action of the hypothalamus in acute regula-tion of peripheral metabolism (25–29), the present study in-vestigated whether brain ER stress could have an acute, neuraleffect on peripheral physiology to direct the development of T2Dand the metabolic syndrome.

ResultsPharmacologic Induction of Brain ER Stress in Mice. Development ofER stress involves multiple steps of intracellular molecularpathways, including a pathway directed by inositol-requiring en-zyme 1 (IRE-1) and X-box–binding protein 1 (XBP-1), a pathwaydirected by activating transcription factor 6 (ATF-6), and a path-way directed by protein kinase R-like ER kinase (PERK) andeukaryotic translation initiation factor 2α subunit (eIF2α). Be-cause of the complexity of ER stress cascades, genetic methodsthat chronically created ER stress in animals often introducedconfounding pathophysiological changes (31, 32). Alternatively,pharmacologic strategies are useful to acutely induce ER stressin cultured cells and animals. Thapsigargin (TG), a classical ERstress-inducing chemical, has been extensively used to rapidlyinduce tissue ER stress through peripheral injection (such as i.p.)in different animal species, including rodents (33–35). In thisresearch, we performed intrabrain injection of TG via cannulapreimplanted into the ventral third ventricle, which is anatomi-cally adjacent to the hypothalamus. Normal C57BL/6 mice re-ceived a single injection of TG (1.0 μg), and the hypothalamusand other brain regions were harvested at 0, 2, 4, and 8 h post-injection for Western blot analysis of two ER stress indicators,phosphorylated IRE-1α and phosphorylated PERK. Datarevealed that TG increased hypothalamic phosphorylation levelsof both IRE-1α and PERK between 2 and 4 h postinjection,whereas these effects diminished at 8 h postinjection (Fig. S1).ER stress induction in other brain regions ranged from modestto negligible levels. In sum, TG injection via the ventral thirdventricle can induce a short-term ER stress in the brain and of-fers a pharmacologic model to study the roles of short-term ERstress that are generated in the brain.

Author contributions: D.C. designed research; S.P., H.Z., G.Z., Z.A., and Y.W. performedresearch; S.P., H.Z., G.Z., Z.A., and D.C. analyzed data; and S.P. and D.C. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.1To whom correspondence should be addressed. E-mail: dongsheng.cai@einstein.yu.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1006875108/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1006875108 PNAS | February 15, 2011 | vol. 108 | no. 7 | 2939–2944

MED

ICALSC

IENCE

S

Dow

nloa

ded

by g

uest

on

Oct

ober

1, 2

020

Acute Induction of Systemic Insulin Resistance by Brain ER Stress.Recent literature has reported that hypothalamic ER stress caninterfere with the feeding-inhibitory action of leptin (19, 20).Herein, we explored whether short-term brain ER stress inducedby TG could affect feeding and body weight of mice. Throughpreimplanted brain cannula, normal chow-fed C57BL/6 mice re-ceived 3 consecutive d of TG injections at the dose of 0, 0.3, 0.6,or 1.0 μg/d. Food intake levels of injected mice were monitored at2, 4, 16, and 24 h postinjection on each day. Compared with ve-hicle injection, only the highest dose of TG (1.0 μg) resulted inpromotion of food intake between 2 and 4 h postinjection (Fig.S2A). The ∼16- to 24-h food intake was, however, comparablebetween TG- and vehicle-treated mice, which was consistent withthe time course of ER stress induction by the chemical injection(Fig. S1). In line with the 24-h feeding, the protocol of 3-d TGtreatment did not significantly affect body weight of the mice(Fig. S2B). Also, adiposity was similar between TG- and vehicle-treated mice, assessed with MRI scanning (fat mass: 3.8 ± 0.3 gvs. 3.0 ± 0.5 g, P= 0.3, and lean mass: 21.2 ± 0.5 g vs. 20.9 ± 1.2 g,P = 0.8, in vehicle- vs. TG-treated mice). Thus, the previouslyproposed influence of brain ER stress on feeding and body weightimbalance (19, 20) should require accumulative effects of chronicbrain ER stress, which was not offered by the acute, short-termbrain ER stress in this experiment. In recognition of recent lit-erature that demonstrated that the brain control of body weightand systemic glucose metabolism is dissectible (25–29), we ex-amined whether short-term ER stress in the brain could rapidlyaffect peripheral glucose metabolism. Mice that received 3 d ofTG treatment at various doses were subjected to glucose tolerancetest (GTT). As shown in Fig. 1 A and B, there was a significantinduction of glucose intolerance in the mice that received 3 d ofTG injections at the dose of 0.6 or 1.0 μg/d. Insulin tolerance test(ITT) was also performed, and data revealed that TG treatment atthe dose of 0.6 or 1.0 μg/d significantly blunted the hypoglycemiceffect of insulin injection (Fig. 1 C and D). In addition, we mea-sured blood insulin levels of themice and found that TG treatment(0.6 or 1.0 μg/d) significantly increased fasting insulin levels (Fig.1E). Together, these data suggest that brain ER stress acutelycauses systemic insulin resistance and glucose intolerance.

Brain ER Stress Impairs Hepatic Insulin Signaling. The neural con-nection between the hypothalamus and the liver plays an im-portant role in the CNS control of systemic glucose homeostasis(24, 26). Suggested by this knowledge, we tested whether short-term TG treatment could be sufficient to affect hepatic insulinsignaling. In the experiment, mice received 3 d of TG injectionsvia third-ventricle cannula and were subjected to an i.v. injectionof insulin, and liver tissues were collected for Western blotanalysis of insulin signaling cascade at various molecular levels.As shown in Fig. 1 F and G, TG injection (1.0 μg/d) significantlyimpaired insulin-induced tyrosine phosphorylation of insulin re-ceptor (IR) and IR substrate 2 (IRS2) in the liver. Because of thereduced tyrosine phosphorylation, the recruitment of PI3-kinaseregulatory subunit p85 by IRS2 significantly decreased (Fig. 1 Fand G). As a result, Akt phosphorylation, which represents a keyfunctional step downstream of PI3-kinase, was poorly inducedby insulin (Fig. 1 F and G). Furthermore, because hepatic insulinsignaling is known to suppress gene expression of gluconeogenicenzymes, we tested whether the induction of hepatic insulin re-sistance by brain ER stress could be accompanied by elevated geneexpression of gluconeogenic enzymes, including phosphoenolpyr-uvate carboxykinase (pepck) and glucose-6-phosphatase (g6pase).Using real-time RT-PCR, we observed that mRNA levels ofpepck and g6pase significantly increased in mice treated with TGcompared with the control mice (Fig. 1H). Altogether, thesefindings suggest that brain ER stress can mediate systemic insulinresistance by inhibiting hepatic insulin signaling, and this effect isdissociable from obesity.

Brain ER Stress Elevates BP Levels. BP increase represents anothercomponent of insulin resistance syndrome. To more extensively

examine the role of brain ER stress in insulin resistance syn-drome, the experimental model established in Fig. 1 was used totest whether brain ER stress could acutely affect BP. NormalC57BL/6 mice received third-ventricle cannulation followed byradio-transmitter probe implantation for continuous BP moni-toring. After postsurgery recovery, mice were adapted to third-ventricle injection via preimplanted cannula until the BP levelswere not affected by the injection procedure per se. Sub-sequently, mice received daily injections of TG at the dose of 0,0.3, 0.6, or 1.0 μg/d for 3 consecutive d, and BP monitoringcontinued throughout the 3-d treatment period. As shown in Fig.2 A–E, TG treatment increased BP levels of the mice in a dose-dependent manner. When the injection dose increased to 1.0 μg/d, there was a profound increase of BP, with systolic, diastolic,and mean BP values elevated by 29.84, 24.77, and 27.46 mmHg,respectively. Such effects tended to be detected on day 2 but noton day 1. In summary, short-term induction of brain ER stress inmice was sufficient to induce BP elevation, another pathologicalaspect of the insulin resistance syndrome that underlies T2D.

Brain ER Stress Up-Regulates the Sympathetic Tone in the Periphery.The induction of both glucose intolerance and BP elevation byshort-term brain ER stress may suggest a rapid neural modula-

A B

C D

EF

G H

Fig. 1. Short-term brain ER stress causes systemic insulin resistance. (A–E)C57BL/6 mice received daily third-ventricle injections of TG at the doses of0.3, 0.6, or 1.0 μg/d for 3 consecutive d. Vehicle injection was used as neg-ative control and is labeled as 0 in the bar graphs. Mice were fasted over-night before receiving the final injection on day 3 and were examined at 2 hpostinjection for GTT (A and B), ITT (C and D), or plasma insulin levels (E).Area under the curve (AUC) data for GTT (B) and ITT (D) were calculated. ITTdata are presented as the percentages of time-course blood glucose levelsover the baseline level (C) and the percentage changes of AUC over thecontrol (vehicle injection) group (D). *P < 0.05; **P < 0.01; n = 8–9 mice pergroup. (F–H) Mice injected with TG (1.0 μg/d) or vehicle for 3 consecutived were anesthetized and challenged with insulin (INS) (5.0 units/kg) or salinefor 3 min via the inferior vena cava. Liver samples were rapidly harvestedand analyzed for insulin signaling with immunoprecipitation (IP) andWestern blots. (G) Tyrosine phosphorylation (p-Tyr) of IRβ and IRS2, thebinding of p85 to IRS2, and phosphorylated Akt (p-Akt) were quantitativelynormalized by the total protein levels (Total) of IRβ, IRS2, p85, and Akt, re-spectively. β-actin (β-act) was used as an internal control. (H) Liver tissueswere harvested from mice that received 3-d TG vs. vehicle treatment andanalyzed for mRNA levels of gluconeogenic enzymes pepck and g6pase withreal-time RT-PCR. *P < 0.05; **P < 0.01; n = 4–6 mice per group; AU, arbitraryunit. (Error bars reflect mean ± SEM.)

2940 | www.pnas.org/cgi/doi/10.1073/pnas.1006875108 Purkayastha et al.

Dow

nloa

ded

by g

uest

on

Oct

ober

1, 2

020

tion of the involved physiology. Hence, we decided to testwhether the metabolic effects of short-term brain ER stress couldbe driven by a shift of the sympathetic nervous system activitytoward sympathetic excitation. C57BL/6 mice received an intra-ventral third-ventricle injection of TG (0.3 μg/d) via preimplantedcannula for 3 consecutive d, and at 2 h after the last TG injection,the spontaneous renal sympathetic nerve activities were recor-ded. It was observed that the TG treatment led to a significantlyhigher mean frequency of firing rate (mean ± SD: 144.0 ± 15.4spikes per s) compared with vehicle-treated controls (mean ±SD: 129.0 ± 15.9 spikes per s) (P < 0.001; Fig. 3 A and B). Also,the spike amplitude was significantly larger in TG-treated ani-mals (mean ± SD: 46.7 ± 26.8 μV) than controls (mean ± SD:29.0 ± 17.4 μV) (P < 0.001; Fig. 3 A and C). In addition tothis electrophysiological study, we measured the plasma levels ofnorepinephrine, which is a known biochemical indicator of thesympathetic activity. As shown in Fig. 3D, the 3-d TG treatmentincreased plasma norepinephrine levels by nearly threefold.Thus, both electrophysiological and biochemical data suggestthat brain ER stress up-regulates the peripheral sympathetictone, which may represent an underlying basis for the metabolicand BP-raising effects of short-term brain ER stress.

Adrenergic Blocker Reverses the Systemic Effects of Brain ER Stress.We subsequently examined whether sympathetic up-regulationindeed represents a mechanism for the metabolic and BP-raisingeffects of brain ER stress. In the literature, the α-adrenergicblocker prazosin has been used to assess the sympathetic controlsof glucose metabolism (36) and BP (37). Thus, we tested if thischemical could abrogate the induction of glucose intoleranceand BP elevation by brain ER stress. C57BL/6 mice receivedthird-ventricle injections of TG (1.0 μg/d) for 3 consecutive d,and prazosin (1 mg/kg, i.p.) was administrated before the finalTG injection on day 3. Mice were subjected to GTT using thesame protocol in Fig. 1 A and B. Data revealed that prazosin

completely prevented TG from causing glucose intolerance (Fig.3E). Consistently, in the presence of prazosin, TG failed to in-crease the plasma insulin levels in the mice (Fig. 3F). Both setsof data confirmed that prazosin indeed abrogated the effect ofTG treatment in causing systemic insulin resistance. In addition,telemetric BP monitoring revealed that prazosin significantlyattenuated the hypertensive effect of TG treatment (Fig. 3 G–I).In sum, sympathetic up-regulation in the periphery underlies themetabolic and BP-raising effects of brain ER stress.

Activation of Hypothalamic NF-κB by Brain ER Stress. Numerousreports have demonstrated that the induction and pathologicalactions of ER stress are associated with inflammation (10, 11,38). Our recent work has linked ER stress to inflammatory nu-clear transcription factor NF-κB in the hypothalamus and pro-posed this link as a potential basis for neural inflammation that isassociated with obesity-induced T2D (19). In this study, we wereinterested in profiling how NF-κB in the hypothalamus mightreact to the pharmacologic induction of ER stress. NormalC57BL/6 mice received a single dose of TG injection (1.0 μg)via preimplanted third-ventricle cannula, and hypothalami were

A

B C

D E

Fig. 2. Short-term brain ER stress elevates BP levels in mice. Mice with third-ventricle cannulation were implanted with BP radio transmitter in the ca-rotid artery for continuous telemetric monitoring of BP and heart rates (HR).TG at the doses of 0.3, 0.6, or 1.0 μg/d or control vehicle (labeled as 0 in thebar graphs) was injected for 3 consecutive d. BP and HR of mice wereobtained before the 3-d treatment period to obtain the baseline levels(labeled as Basal) and continuously monitored during the 3-d treatmentperiod. (A) Representative profiles of BP in response to 1.0 μg of TG (Right)vs. vehicle (Left) injection. (B–E) Data represent average BP and HR valuesover a 2-h period after the final injection on day 3. SBP, systolic BP; DBP,diastolic BP; MBP, mean BP. *P < 0.05; **P < 0.01; n = 4 mice per group. (Errorbars reflect mean ± SEM.)

A

B

C

D

G

E F

H I

Fig. 3. Sympathetic activation in mice with short-term brain ER stress. (A–D)Mice received daily injections of TG (0.3 μg/d) or vehicle for 3 consecutive d.(A–C) At 2 h after the final injection on day 3, animals were anesthetized,and RSNA was recorded (A). n = 3–4 mice per group. The frequency of firing(B) and the spike amplitude (C) were both significantly different betweenTG- and vehicle-treated groups (P < 0.001). (D) Plasma norepinephrine (NE)concentrations of mice after 3-d TG (1.0 μg) or vehicle treatment. *P < 0.05;n = 7 mice per group. (E and F) C57BL/6 mice received daily third-ventricleinjections of TG vs. vehicle (Veh) for 3 consecutive d, and the final injectionwas provided to overnight-fasted mice in combination with i.p. injection ofprazosin (Prz, 1 mg/kg). At 2 h postinjection, mice were subjected to GTT (E)or blood sampling to measure plasma insulin concentrations (F). Data pre-sented for GTT are the AUC values. NS, not significant; n = 8–9 mice pergroup. (G–I) Mice preimplanted with third-ventricle cannula and artery BPtelemetric probes received daily third-ventricle injections of TG vs. vehicle(Veh) for 3 consecutive d, and the final injection was provided in combina-tion with (+) or without (−) i.p. injection of prazosin (Prz, 1 mg/kg). Datarepresent average BP values over a 2-h period after the final injection. SBP,systolic BP; DBP, diastolic BP; MBP, mean BP. *P < 0.05; n = 4 mice per group.(Error bars reflect mean ± SEM.)

Purkayastha et al. PNAS | February 15, 2011 | vol. 108 | no. 7 | 2941

MED

ICALSC

IENCE

S

Dow

nloa

ded

by g

uest

on

Oct

ober

1, 2

020

harvested at 0, 2, 4, and 8 h after TG injection for Western blotanalysis of NF-κB signaling. As shown in Fig. 4 A and B, phos-phorylation levels of NF-κB subunit RelA significantly increasedat 2 and 4 h after TG injection. Given that IκBα phosphorylationand its subsequent degradation are required for the activation ofthe classical NF-κB pathway, we also examined the phosphory-lation and protein levels of IκBα. It was revealed that IκBαphosphorylation increased while its total protein levels decreasedin the hypothalamus during the time course of 2–4 h after TGinjection (Fig. 4 A and B). The effects of TG injection on NF-κBsignaling substantially decreased at 8 h after TG injection (Fig. 4A and B), which correlated with the time course of brain ERstress after TG injection (Fig. S1). Thus, brain ER stress inducedby TG activated NF-κB in the hypothalamus. Because a singleinjection of TG led to hypothalamic NF-κB activation, NF-κBmight be mechanistically involved in the disease action of short-term brain ER stress.

NF-κB Inhibition Reverses the Systemic Effects of Brain ER Stress. Toexamine whether hypothalamic NF-κB could mediate the diseaseoutcomes of brain ER stress, we used our established viral in-jection approach (19) to generate mice with virus-mediated NF-κB inhibition in the hypothalamic arcuate nucleus. Dominant-negative IκBα (DNIκBα), which has been established as NF-κB–specific inhibitor (39, 40), was used to inhibit the NF-κB activity.Normal C57BL/6 mice received intra-arcuate injections of ade-noviruses expressing HA-tagged DNIκBα or GFP, immediatelyfollowed by third-ventricle implantation of cannula. As verifiedin Fig. 4C, virus-mediated exogenous gene expression of HA-tagged DNIκBα or GFP was induced specifically in the arcuatenucleus. We further used Western blot analysis to evaluate hy-pothalamic ER stress and NF-κB activities of these mice inresponse to third-ventricle TG injection. In GFP adenovirus-injected control mice, TG injection markedly induced hypotha-lamic ER stress and NF-κB activation (Fig. S3). The profoundinduction of ER stress and NF-κB activation was primarily at-tributed to the pharmacologic effect of TG because the non-specific effect caused by adenoviral injection procedure per sewas small (Fig. S3). However, arcuate delivery of DNIκBα pre-vented the up-regulation of hypothalamic NF-κB by TG (Fig. S3).GFP vs. DNIκBα adenovirus-injected mice received 3-d treatmentof TG (1.0 μg/d) or vehicle via preimplanted third-ventricle can-nula as described in Fig. 1 and were subsequently subjected toGTT. TG treatment impaired glucose tolerance in GFP adeno-virus-injected mice (Fig. 4 D and E), which was similar to theobservation in Fig. 1 A and B. In contrast, DNIκBα expression inthe arcuate nucleus prevented the induction of glucose in-tolerance by TG treatment (Fig. 4 D and E). These protectiveeffects were unrelated to the body weight of these mice (Fig. S4).Additional experiments using telemetry showed that TG signif-icantly increased BP levels in GFP adenovirus-injected mice butnot in DNIκBα adenovirus-injected mice (Fig. 4F). These resultstogether suggested that hypothalamic NF-κB worked as a medi-ator of the peripheral effects of short-term brain ER stress.

Inhibition of Brain ER Stress Acutely Reverses Insulin ResistanceSyndrome. Finally, we explored whether acute suppression ofbrain ER stress could rapidly reverse obesity-associated glucoseand BP disorders, given that obesity induced through chronichigh-fat diet (HFD) feeding was accompanied by the inductionof brain/hypothalamic ER stress (Fig. S5). In the experiment,C57BL/6 mice with HFD-induced obesity received two sepa-rate third-ventricle injections of the ER stress inhibitorTUDCA (1 μg) or vehicle, given at the beginning and the end ofan overnight fasting, respectively. At 2 h after the second in-jection, mice were subjected to metabolic profiling. As predicted,HFD-fed mice displayed glucose intolerance, insulin insensitivity,and hyperinsulinemia (Fig. 5 A–E). However, TUDCA treatmentsignificantly, although partially, reversed these metabolic dis-orders (Fig. 5 C–G). Molecular analysis further demonstratedthat TUDCA treatment partially improved hepatic insulin sig-

naling (Fig. 5 F and G) and significantly reduced the mRNAlevels of hepatic gluconeogenic enzymes (Fig. 5H) in HFD-fedmice. Subgroups of the mice were subjected to telemetry, anddata revealed that TUDCA treatment significantly decreasedsystolic, diastolic, and mean BP levels of HFD-fed mice (Fig. 5I–K). Notably, because TUDCA treatment was provided onlyduring an overnight fasting period in these experiments, the in-volvement of food intake was excluded. Also, the duration ofthe overnight period was too short to cause body weightchanges between TUCDA-injected and vehicle-injected mice.In this context, the therapeutic effects of TUDCA further con-solidated the finding that brain ER stress can rapidly induceneural dysregulation to mediate the development of T2D andrelated diseases.

DiscussionER stress can be induced by aberrant changes in various in-tracellular processes, such as protein folding/modification, redoxbalance, energy consumption, and Ca2+ regulation. The patho-genic roles of chronic ER stress have been identified in neurode-generative diseases that are characterized by excessive intracellularaccumulation and aggregation of misfolded proteins, as seen in

A B

C

D E F

Fig. 4. ER stress activates hypothalamic NF-κB to affect peripheral physiol-ogy. (A and B) C57BL/6 mice received a single injection of TG (1.0 μg) viapreimplanted cannula. Hypothalami were harvested for Western blot anal-ysis of NF-κB signaling at the indicated time points after TG injection.Phosphorylated RelA (p-RelA), phosphorylated IκBα (p-IκBα), and total pro-tein levels of IκBα were measured to reflect NF-κB activities. Vehicle injectionwas used as the negative control (labeled as 0 h). β-actin (β-act) was used asan internal control. (B) p-RelA and p-IκBα levels were normalized by the totalprotein levels of RelA and β-actin, respectively. *P < 0.05; **P < 0.01; n = 4–6mice per group; AU, arbitrary unit. (C) Mice received intra-arcuate injectionsof adenoviruses expressing either HA-tagged DNIκBα or GFP. At 1 wk post-injection, hypothalamic sections were immunostained for GFP (Upper) andHA (Lower) to verify the site-specific gene delivery. Nuclear staining withDAPI (blue) labels all of the cells in the sections and is merged with GFP(green) or HA (red) staining. Arc, arcuate nucleus; 3V, third ventricle. (Scalebar = 50 μm.) (D and E) Intra-Arc DNIκBα or GFP adenovirus-injected micereceived daily injections of TG (1 μg/d) or vehicle (Veh) via third-ventriclecannula for 3 consecutive d. At 2 h after the final injection, mice were ex-amined for GTT. (E) AUC data for GTT. *P < 0.05; n = 6–10 mice per group. (F)DNIκBα or GFP adenovirus-injected mice were implanted with third ventriclecannula and intra arterial telemetric BP probe. BP levels were monitoreddaily for 3 d before and after TG vs. vehicle injection. Data presented areaverage mean BP (MBP) values over a 2-h period after the final injection onday 3. *P < 0.05; n = 4 mice per group. (Error bars reflect mean ± SEM.)

2942 | www.pnas.org/cgi/doi/10.1073/pnas.1006875108 Purkayastha et al.

Dow

nloa

ded

by g

uest

on

Oct

ober

1, 2

020

Alzheimer’s disease, Parkinson disease, Huntington disease, andamyotrophic lateral sclerosis (30, 41, 42). Unexpectedly, recentstudies using several genetic models of ER stress, including PERKknockout mice (32), eIF2α knock-in mice (31), and XBP-1knockout mice (17), revealed that these mice manifested diabetesor prediabetes symptoms that were related to defects in insulinsecretion and actions caused by local ER stress. Classically, neu-rological diseases and metabolic diseases (such as T2D) involvedistinct biological/physiological disciplines, but recent attention tothe epidemic coexistence of neurodegenerative diseases and T2Dhas challenged this traditional view (43). Along with this knowl-edge, the CNS was recently recognized as a critical regulator of notonly feeding/body weight but also peripheral insulin–glucose ho-

meostasis (25–29). Using pharmacologic approaches, the currentstudy has revealed that short-term brain ER stress up-regulates thesympathetic pathway to rapidly cause several key components ofthe so-called metabolic syndrome, including glucose intolerance,insulin resistance, and BP increase. Compared with genetic ERstress models, our pharmacologic approach dissected the physio-logical outcomes of brain ER stress without the impact of chronicfeeding or body weight changes. Hence, this study reveals a para-digm by which brain ER stress uses the sympathetic route toacutely and directly mediate a cluster of peripheral metabolicdisorders associated with T2D and related problems (Fig. 6). Thisfinding complements the recent research which has proposed thatchronic brain ER stress causes neuroendocrine dysfunctions (suchas central leptin and insulin resistance) to mediate eating and bodyweight disorders (19, 20).Although the pharmacologic approach that delivered ER stress

modulators via the brain ventricle should have broad effectsin various brain regions, the involvement of the hypothalamusdeserves attention. Indeed, the pharmacologic induction of ERstress was prominent in the hypothalamus, and the disease effectsof brain ER stress were substantially reversible by hypothalamicNF-κB inhibition. Further, the ER stress–NF-κB connectionrevealed by the acute models in this study agrees with the knowl-edge that NF-κB can work as a downstream player of chronic ERstress (19, 44). All these understandings together can highlight thesusceptibility of the hypothalamus to stress and inflammation.Nevertheless, it remains an important topic regarding whether andhow the neural mechanisms of T2D and related disease requireER stress in other brain regions, in particular the regions thatclosely interact with the sympathetic nervous system. Answeringthese questions would call for further investigations.

Materials and MethodsAnimal Manipulations and Analyses. Adult male C57BL/6 mice (Jackson Labo-ratory) were implantedwith a guide cannula into the third ventricle followingthe method described previously (19). TG and TUDCA were injected at a vol-ume of 2 μL through the preimplanted cannula. Prazosin was injected intra-peritonially. Intra-arcuate injections of Ad5-CMV–driven DNIκBα and Ad5-CMV–driven GFP were performed according to the technique established inour previous research (19). GTT, ITT, and measurements of body weight andfood intake were performed at the indicated time points. Plasma insulin andnorepinephrine were measured by ELISA. Transcardial perfusion, brain sec-

A B

C

F

H

D E

G

I J K

Fig. 5. Acute reversal of obesity-related disorders by inhibiting brain ERstress. Male C57BL/6 mice were maintained on normal chow (Chw) or HFDfor 5 mo, since 2 mo of age, and then implanted with third-ventricle cannula.After 2 wk of postoperative recovery, mice received third-ventricle injectionof 1.0 μg of TUDCA (TU) or vehicle (Veh) on the night of day 1, followed byovernight fasting and a second injection of TUDCA or vehicle in the morningof day 2. (A–E) Mice at 2 h postinjection on day 2 were separately subjectedto GTT (A and B), ITT (C and D), and plasma insulin measurement (E). AUCdata of GTT (B) and ITT (D) are presented. ITT data are presented as thepercentages of time-course blood glucose over the baseline levels (C) andthe percentages of AUC values of treatment groups over the control group(D). *P < 0.05; **P < 0.01; n = 8–9 mice per group. (F and G) Mice at 2 hpostinjection on day 2 were anesthetized and challenged with insulin (INS;5.0 units/kg) or saline for 3 min via abdominal aorta injection. Liver sampleswere rapidly harvested and analyzed for insulin signaling with immuno-precipitation (IP) and Western blotting. (G) Phosphorylated IRβ (p-IRβ),phosphorylated IRS2 (p-IRS2), binding of p85 to IRS2, and phosphorylatedAkt (p-Akt) were quantitatively normalized by the total protein levels (Total)of IRβ, IRS2, p85, and Akt, respectively. β-actin (β-act) was used as an internalcontrol. **P < 0.01; n = 4–6 mice per group; AU, arbitrary unit. (H) Livertissues were harvested from chow-fed versus HFD-fed mice after 2-d TUDCAor vehicle treatment and analyzed for mRNA levels of pepck and g6pase.*P < 0.05; n = 4–6 mice per group; AU, arbitrary unit. (I–K) Mice were con-tinuously monitored for BP and heart rates (HR) via preimplanted artery BPtelemetric probes. Data presented represent average BP values over a 2-hperiod after TUDCA or vehicle injection on day 2. SBP, systolic BP; DBP, di-astolic BP; MBP, mean BP. *P < 0.05; **P < 0.01; n = 4 mice per group. (Errorbars reflect mean ± SEM.)

Fig. 6. The models of acute versus chronic brain ER stress in obesity/T2Dsyndrome. Pathological conditions associated with obesity and T2D induceboth acute and chronic ER stress in the brain. Acute brain ER stress rapidly up-regulates the sympathetic nervous system activity to cause peripheral insulinresistance, glucose intolerance, and related BP dysregulation. Chronic brainER stress, on the other hand, disrupts the hypothalamic neuroendocrinefunctions to cause eating disorder, overweight, and obesity. These neuraland neuroendocrine processes are physiologically interconnected and areaffecting each other at multiple levels; the combined effects of these twoprocesses represent a significant CNS basis for body weight-independent andbody weight-dependent development of T2D and related problems.

Purkayastha et al. PNAS | February 15, 2011 | vol. 108 | no. 7 | 2943

MED

ICALSC

IENCE

S

Dow

nloa

ded

by g

uest

on

Oct

ober

1, 2

020

tioning, immunostaining, and imaging were performed according to thepreviously established protocol (19). For details see SI Materials andMethods.

Telemetric Measurement of BP. BP of conscious mice was recorded with a ra-diotelemetrymonitoring system (DataSciences International) bypreimplantinga pressure sensor in the carotid artery following standard procedures (45). Forradiotelemetric probe (model TA11PA-C10; DSI) implantation, mice wereanesthetized, and the left common carotid arterywas separated. After ligatingone end of the artery (right below the carotid bifurcation) with 4-0 suture(Ethicon) and occluding the distal end with a microclip, a small incision wasmade near the proximal ligated end, and the pressure transmission catheterwas guided into the artery and secured optimally in place with sutures. Theradio telemetric transmitter attached to the catheter was passed s.c. andinserted into a s.c. pocket formed by a blunt dissection in the right flank. Micewere allowed 1–2 wk for postsurgical recovery. BP levels were continuouslyrecorded at a sampling rate of 2,000 Hz over a 300-s segment duration.

Measurement of Renal Sympathetic Nerve Activity (RSNA). Left renal nerveswere isolated in anesthetized mice and RSNA was recorded by pure iridiummicroelectrode. Amplified, filtered and digitized signals were fed into a

PowerLab data acquisition system and LabChart 7 software (AD Instruments)for acquisition and data analysis. For details, see SI Materials and Methods.

Immunoprecipitation, Western Blotting, and Real-Time RT-PCR. Liver tissuescollected from anesthetized mice with 3-min insulin (5 units/kg) or vehiclestimulation via inferior vena cava, were used for immunoprecipitation andWestern blotting following the method previously described (19, 40). ForReal-time RT-PCR, cDNA synthesized from total RNA was PCR amplified andquantified with SYBR Green PCR MasterMix (Applied Biosystems). For details,see SI Materials and Methods.

Statistical Analyses. Student’s t tests were used for comparisons involving onlytwo groups. ANOVA and appropriate post hoc analyses were used for com-parisons involving more than two groups. Data were presented as mean ±SEM. P < 0.05 was considered significant.

ACKNOWLEDGMENTS. This study was supported by National Institutes ofHealth Grants R01 DK078750 and R01 AG031774 and American DiabetesAssociation Junior Faculty Award 1-07-JF-09 (all to D.C.).

1. Ruan H, Lodish HF (2004) Regulation of insulin sensitivity by adipose tissue-derivedhormones and inflammatory cytokines. Curr Opin Lipidol 15:297–302.

2. Hotamisligil GS (2006) Inflammation and metabolic disorders. Nature 444:860–867.3. Lehrke M, Lazar MA (2004) Inflamed about obesity. Nat Med 10:126–127.4. Shoelson SE, Goldfine AB (2009) Getting away from glucose: Fanning the flames of

obesity-induced inflammation. Nat Med 15:373–374.5. Cai D (2009) NFκB-mediated metabolic inflammation in peripheral tissues versus

central nervous system. Cell Cycle 8:2542–2548.6. Schenk S, Saberi M, Olefsky JM (2008) Insulin sensitivity: Modulation by nutrients and

inflammation. J Clin Invest 118:2992–3002.7. Petersen KF, ShulmanGI (2006) Etiology of insulin resistance.Am JMed 119 (5 Suppl 1):

S10–S16.8. Cifuentes ME, Pagano PJ (2006) Targeting reactive oxygen species in hypertension.

Curr Opin Nephrol Hypertens 15:179–186.9. Frisard M, Ravussin E (2006) Energy metabolism and oxidative stress: Impact on the

metabolic syndrome and the aging process. Endocrine 29:27–32.10. Scheuner D, Kaufman RJ (2008) The unfolded protein response: A pathway that links

insulin demand with β-cell failure and diabetes. Endocr Rev 29:317–333.11. Hotamisligil GS (2010) Endoplasmic reticulum stress and the inflammatory basis of

metabolic disease. Cell 140:900–917.12. Glimcher LH, Lee AH (2009) From sugar to fat: How the transcription factor XBP1

regulates hepatic lipogenesis. Ann N Y Acad Sci 1173 (Suppl 1):E2–E9.13. Back SH, et al. (2009) Translation attenuation through eIF2α phosphorylation

prevents oxidative stress and maintains the differentiated state in β cells. Cell Metab10:13–26.

14. Huang CJ, et al. (2007) High expression rates of human islet amyloid polypeptideinduce endoplasmic reticulum stress mediated β-cell apoptosis, a characteristic ofhumans with type 2 but not type 1 diabetes. Diabetes 56:2016–2027.

15. Sachdeva MM, et al. (2009) Pdx1 (MODY4) regulates pancreatic β cell susceptibility toER stress. Proc Natl Acad Sci USA 106:19090–19095.

16. Ozcan U, et al. (2006) Chemical chaperones reduce ER stress and restore glucosehomeostasis in a mouse model of type 2 diabetes. Science 313:1137–1140.

17. Ozcan U, et al. (2004) Endoplasmic reticulum stress links obesity, insulin action, andtype 2 diabetes. Science 306:457–461.

18. Wang Y, Vera L, Fischer WH, Montminy M (2009) The CREB coactivator CRTC2 linkshepatic ER stress and fasting gluconeogenesis. Nature 460:534–537.

19. Zhang X, et al. (2008) Hypothalamic IKKβ/NF-κB and ER stress link overnutrition toenergy imbalance and obesity. Cell 135:61–73.

20. Ozcan L, et al. (2009) Endoplasmic reticulum stress plays a central role in developmentof leptin resistance. Cell Metab 9:35–51.

21. Ahima RS, Flier JS (2000) Leptin. Annu Rev Physiol 62:413–437.22. Morton GJ, Cummings DE, Baskin DG, Barsh GS, Schwartz MW (2006) Central nervous

system control of food intake and body weight. Nature 443:289–295.23. Myers MG, Cowley MA, Münzberg H (2008) Mechanisms of leptin action and leptin

resistance. Annu Rev Physiol 70:537–556.24. Klieverik LP, et al. (2009) Thyroid hormone modulates glucose production via

a sympathetic pathway from the hypothalamic paraventricular nucleus to the liver.Proc Natl Acad Sci USA 106:5966–5971.

25. Pocai A, et al. (2005) Hypothalamic K(ATP) channels control hepatic glucose

production. Nature 434:1026–1031.26. Lam TK, et al. (2005) Hypothalamic sensing of circulating fatty acids is required for

glucose homeostasis. Nat Med 11:320–327.27. Obici S, Zhang BB, Karkanias G, Rossetti L (2002) Hypothalamic insulin signaling is

required for inhibition of glucose production. Nat Med 8:1376–1382.28. Morton GJ (2007) Hypothalamic leptin regulation of energy homeostasis and glucose

metabolism. J Physiol 583:437–443.29. Gelling RW, et al. (2006) Insulin action in the brain contributes to glucose lowering

during insulin treatment of diabetes. Cell Metab 3:67–73.30. Hetz C, et al. (2009) XBP-1 deficiency in the nervous system protects against

amyotrophic lateral sclerosis by increasing autophagy. Genes Dev 23:2294–2306.31. Scheuner D, et al. (2001) Translational control is required for the unfolded protein

response and in vivo glucose homeostasis. Mol Cell 7:1165–1176.32. Harding HP, et al. (2001) Diabetes mellitus and exocrine pancreatic dysfunction in

perk-/- mice reveals a role for translational control in secretory cell survival.Mol Cell 7:

1153–1163.33. Chae HJ, et al. (2004) BI-1 regulates an apoptosis pathway linked to endoplasmic

reticulum stress. Mol Cell 15:355–366.34. Timmins JM, et al. (2009) Calcium/calmodulin-dependent protein kinase II links ER

stress with Fas and mitochondrial apoptosis pathways. J Clin Invest 119:2925–2941.35. Yan W, et al. (2002) Control of PERK eIF2α kinase activity by the endoplasmic

reticulum stress-induced molecular chaperone P58IPK. Proc Natl Acad Sci USA 99:

15920–15925.36. Ross H, Moser M (1993) Prazosin, diuretics, and glucose intolerance. Ann Intern Med

119:859–860.37. Khoury AF, Kaplan NM (1991) α-Blocker therapy of hypertension: An unfulfilled

promise. JAMA 266:394–398.38. Martinon F, Chen X, Lee AH, Glimcher LH (2010) TLR activation of the transcription

factor XBP1 regulates innate immune responses in macrophages. Nat Immunol 11:

411–418.39. Cai D, et al. (2004) IKKβ/NF-κB activation causes severe muscle wasting in mice. Cell

119:285–298.40. Cai D, et al. (2005) Local and systemic insulin resistance resulting from hepatic

activation of IKK-β and NF-κB. Nat Med 11:183–190.41. Nakagawa T, et al. (2000) Caspase-12 mediates endoplasmic-reticulum-specific

apoptosis and cytotoxicity by amyloid-β. Nature 403:98–103.42. Oyadomari S, Mori M (2004) Roles of CHOP/GADD153 in endoplasmic reticulum stress.

Cell Death Differ 11:381–389.43. Janson J, et al. (2004) Increased risk of type 2 diabetes in Alzheimer disease. Diabetes

53:474–481.44. Ma Y, Hendershot LM (2004) ER chaperone functions during normal and stress

conditions. J Chem Neuroanat 28:51–65.45. Farah VM, Joaquim LF, Bernatova I, Morris M (2004) Acute and chronic stress

influence blood pressure variability in mice. Physiol Behav 83:135–142.

2944 | www.pnas.org/cgi/doi/10.1073/pnas.1006875108 Purkayastha et al.

Dow

nloa

ded

by g

uest

on

Oct

ober

1, 2

020