Dysfunction of intraflagellar transport-A causes hyperphagia-induced obesity and metabolic syndromeDamon T. Jacobs1, Luciane M. Silva1, Bailey A. Allard1, Michael P. Schonfeld1, Anindita Chatterjee1,George C. Talbott2, David R. Beier2,3 and Pamela V. Tran1,2,*
ABSTRACTPrimary cilia extend from the plasma membrane of most vertebratecells and mediate signaling pathways. Ciliary dysfunction underliesciliopathies, which are genetic syndromes that manifest multipleclinical features, including renal cystic disease and obesity. THM1(also termed TTC21B or IFT139) encodes a component of theintraflagellar transport-A complex and mutations in THM1 have beenidentified in 5% of individuals with ciliopathies. Consistent with this,deletion of murine Thm1 during late embryonic development resultsin cystic kidney disease. Here, we report that deletion of murine Thm1during adulthood results in obesity, diabetes, hypertension and fattyliver disease, with gender differences in susceptibility to weight gainand metabolic dysfunction. Pair-feeding of Thm1 conditional knock-out mice relative to control littermates prevented the obesity andrelated disorders, indicating that hyperphagia caused the obesephenotype. Thm1 ablation resulted in increased localization ofadenylyl cyclase III in primary cilia that were shortened, withbulbous distal tips on neurons of the hypothalamic arcuate nucleus,an integrative center for signals that regulate feeding and activity. Inpre-obese Thm1 conditional knock-out mice, expression ofanorexogenic pro-opiomelanocortin (Pomc) was decreased by 50%in the arcuate nucleus, which likely caused the hyperphagia. FastingofThm1 conditional knock-outmice did not alterPomc nor orexogenicagouti-related neuropeptide (Agrp) expression, suggesting impairedsensing of changes in peripheral signals. Together, these dataindicate that the Thm1-mutant ciliary defect diminishes sensitivity tofeeding signals, which alters appetite regulation and leads tohyperphagia, obesity and metabolic disease.
KEY WORDS: Primary cilia, IFT complex A, POMC, Obesity mousemodel
INTRODUCTIONObesity is a global epidemicwith significant morbidity andmortality.Obesity often leads to metabolic syndrome, a combination of adversehealth conditions that includes dyslipidemia, hypertension, glucoseintolerance and insulin resistance (O’Neill and O’Driscoll, 2014).
These increase risk for diabetes mellitus type 2, cardiovasculardisease and non-alcoholic fatty liver disease, for which treatments areinvasive and largely ineffective (Shin et al., 2013). Despite extensiveinvestigations, much is still unknown regarding the molecularmechanisms underlying onset of obesity and associated metabolicdisorders.
Obesity arises when caloric intake exceeds caloric expenditure.This energy balance is controlled by neural circuitry that initiates inthe hypothalamic arcuate nucleus (ARC), a central processing centerfor signals that regulate feeding and activity. In the ARC, neuronsexpressing pro-opiomelanocortin (POMC) and agouti-relatedpeptide/neuropeptide Y (AgRP/NPY) are two distinct neuronpopulations that respond to signals emanating from peripheraltissues (Sohn et al., 2013). In response to feeding, satiety signals,such as leptin or insulin, are released into the bloodstream byadipose tissue and the pancreas. Upon reaching the POMC-expressing neurons of the ARC, these elicit a response to stopfood-seeking behavior and increase physical activity (Millington,2007). In contrast, fasting signals such as ghrelin, which is secretedby an empty stomach, signal to the AgRP/NPY-expressing neuronsto elicit a food-seeking response (Liu et al., 2012). In the satiatedstate, leptin further inhibits AgRP/NPY-expressing neurons toenhance the satiety signal. Deficiency of leptin or of the leptinreceptor in the ob/ob or db/db mouse models, respectively,dysregulates the feeding/activity signaling axis resulting inexcessive food intake (hyperphagia) and obesity (Islam, 2013).
Ciliopathies are genetic syndromes that link hyperphagia andobesity to dysfunction of the primary cilium, an antenna-likesensory organelle that regulates signaling pathways and is present onalmost all vertebrate cells (Berbari et al., 2009). Within the cilium,protein complexes carry cargoes of structural or signaling proteinsbidirectionally along microtubular tracks in a process termedintraflagellar transport (IFT). The IFT machinery comprisesIFT-B and IFT-A protein complexes, which are transported bykinesin and cytoplasmic dynein motors. Ciliopathies affect multipleorgans and clinical features can include cystic disease of the kidney,liver and pancreas, retinal degeneration, facial anomalies, mentalretardation and polydactyly. In two ciliopathies, Bardet–Beidlsyndrome (BBS) and Alström syndrome, obesity also presents as amajor clinical feature (Girard and Petrovsky, 2011; Quinlan et al.,2008). BBS results from mutations of at least 20 genes (Lindstrandet al., 2014), which encode products that facilitate or assemble into aprotein complex, the BBSome, which transports cargo to membranecompartments and within the ciliary membrane (Guo andRahmouni, 2011). Alström syndrome results from mutations of asingle gene, ALMS1, whose gene product localizes to the basal body(Girard and Petrovsky, 2011). In mice, loss of Bbs2, Bbs4 and Bbs6causes hyperphagia, obesity and hyperleptinemia (Rahmouni et al.,2008), and hypothalamic neurons of Bbs4−/− mice lack ciliarylocalization of appetite-regulating G-protein coupled receptorsReceived 1 April 2016; Accepted 29 April 2016
1Department of Anatomy and Cell Biology and The Kidney Institute, University ofKansas Medical Center, Kansas City, KS 66160, USA. 2Genetics Division, Brighamand Women’s Hospital, Harvard Medical School, Boston, MA 02115, USA. 3Centerfor Developmental Biology and Regenerative Medicine, Seattle Children’sResearch Institute, Seattle, WA 98105, USA.
This is an Open Access article distributed under the terms of the Creative Commons AttributionLicense (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use,distribution and reproduction in any medium provided that the original work is properly attributed.
(Berbari et al., 2008). Alms1-mutant mice are hyperphagic andobese (Arsov et al., 2006; Collin et al., 2005), and foz/foz mice,which harbor a truncating mutation in Alms1, manifest a progressiveloss of neuronal cilia in the hypothalamus (Heydet et al., 2013).Further, ablation of the complex B genes, Ift88 or Kif3a, eitherduring adulthood or specifically in POMC-expressing cells of mice,causes hyperphagia, obesity and hyperleptinemia (Davenport et al.,2007). These data underscore the importance of neuronal primarycilia in regulating the feeding/activity signaling axis. Finally,hypomorphic Rpgrip1l+/− mutants of the transition zone, whichregulates entry of proteins into the cilium, are also hyperphagic andobese (Stratigopoulos et al., 2014).A role for IFT complex A in regulating energy homeostasis has
not been reported. Previously, we identified THM1 as an IFTcomplex A protein (Tran et al., 2008), and pathogenic alleles ofTHM1 have been identified in 5% of individuals with ciliopathies,including BBS, for which obesity is a major clinical component(Davis et al., 2011). Loss of THM1 impairs retrograde IFT, causingshortened primary cilia with bulbous distal tips where proteinparticles accumulate (Tran et al., 2008). In mouse, early embryonicloss of Thm1 causes polydactyly, craniofacial and neural tubedefects and perinatal lethality (Tran et al., 2008), whereas deletionof Thm1 during late embryogenesis causes cystic kidney disease
(Tran et al., 2014). Collectively, these mouse mutants demonstratemany of the clinical features of ciliopathies. Because obesity is aprimary clinical feature of BBS, we examined whether deletion ofmurine Thm1 also causes obesity and affects neuronal signaling inthe ARC, misregulating energy homeostasis.
RESULTSThm1 conditional knock-out mice become obeseWe deleted Thm1 at 5 weeks of age using a tamoxifen-inducible Crerecombinase driven by the Rosa26 locus, then monitored Thm1conditional knock-out (cko) mice over a 13-week period. Threeweeks following gene deletion, Thm1-cko mice showedsignificantly increased body weight relative to wild-type control(WT) littermates. By thirteen weeks post-Thm1 ablation, Thm1-ckofemales and males showed 1.8- and 1.4-fold higher body weights,respectively than their WT littermates (Fig. 1A-C). Adipose depotsof Thm1-cko mice were significantly larger (Fig. S1; Fig. 1D,E),and Thm1-cko females and males exhibited ∼3.7- and 2.0-foldhigher percentage body fat than female and male WT littermates,respectively (Fig. 1F). Histological analysis of peri-renal, whiteadipose tissue revealed that Thm1-cko adipocytes were enlarged∼threefold in diameter, corresponding to a 45-fold increase involume (Fig. 1G, upper panels). Analysis of WT subcutaneous
Fig. 1. Loss of Thm1 causes obesity. (A) Ad libitum-fed WT and Thm1-cko mice 13 weeks post-tamoxifen injection. A conditional allele of Thm1 wasdeleted using a tamoxifen-inducible Cre recombinase at 5 weeks of age. (B,C) Weekly body weight measurements over a 13-week period. Data points representmeans±s.e.m. Two-tailed unequal variance t-test. *P<0.05. N=9 WT; N=9 Thm1-cko each for male and female. (D,E) Adipose tissue weights at 13 weeks post-tamoxifen injection. Error bars represent means±s.e.m. *P<0.05. (F) Percent body fat. Isolated adipose depots were weighed and summed, then divided overbody weight. Error bars represent means±s.e.m. *P<0.05. (G) H&E staining of peri-renal fat and brown fat. Scale bars: 100 µm.
790
RESEARCH ARTICLE Disease Models & Mechanisms (2016) 9, 789-798 doi:10.1242/dmm.025791
scapular brown adipose tissue (BAT) showed numerous smalllocules containing lipid droplets. In contrast, Thm1-cko BATshowed loss of lipid locules and increased lipid droplet size,characteristic of obesity (Fig. 1G, lower panels).To confirm genomic recombination of Thm1, mice harboring a
tdTomato;EGFP reporter, which fluoresces tdTomato in non-recombined tissue and EGFP in recombined tissue, were bred intothe Thm1fl/fl line, which was subsequently crossed to Thm1aln/+;Cre+mice. The aln allele is a null allele. Fluorescence analysis of arcuatenucleus, hypothalamus, skeletal muscle and white and brown adiposetissues of progeny that were tamoxifen-injected and harboring aCre+allele fluoresced green, indicating recombination (Fig. S2). Varyingexpression of tdTomato was also observed in these tissues, indicatingrecombination was partial. Skeletal muscle showed the most effectiverecombination. Western blot analysis was also performed on proteinextracts of the arcuate nucleus, hypothalamus, skeletal muscle andwhite and brown adipose tissues. Thm1-cko tissues showed 35-45%less protein relative to control Thm1fl/+;Cre+ extracts (Fig. S2).
Thm1-cko mice develop metabolic syndromeTo determine whether the Thm1-cko obese phenotype wasaccompanied by metabolic abnormalities, we measured serumlevels of adipocyte-derived leptin and resistin, pancreatic β-cell-derived insulin and C-peptide, and glucose-dependentinsulinotropic peptide (GIP) in WT and Thm1-cko mice at13 weeks post-Thm1 deletion using a metabolic multiplex ELISA.Resistin has been linked to diabetes, cardiovascular disease andnon-alcoholic fatty liver disease and is proposed to modulatemetabolic and inflammatory pathways (Jamaluddin et al., 2012).Insulin promotes uptake of blood glucose into skeletal muscle,adipose tissue and liver, and like leptin, decreases appetite in thehypothalamus. Upon insulin formation, C-peptide is released as aby-product, and thus, C-peptide levels often reflect levels of insulinsynthesis (Landreh et al., 2013). Finally, GIP is released by cells ofthe gastrointestinal tract and stimulates insulin secretion in responseto glucose (Lynn et al., 2001). Leptin, insulin and C-peptide levelswere elevated in both Thm1-cko males and females, and resistin andGIP levels were also increased in Thm1-cko females (Fig. 2A-D).These data indicate disturbances in insulin metabolism andresistance to leptin and insulin in obese Thm1-cko mice.To examine glucose and insulin metabolism further, wemeasured
blood glucose levels at 0, 6 and 13 weeks post-tamoxifen injection.At 6 and 13 weeks post-tamoxifen injection, Thm1-cko miceshowed elevated glucose levels relative toWTmice (Fig. 2E,G). Wenext performed a glucose tolerance test (GTT), which measuresability to clear a glucose load from the bloodstream. An alteredresponse to a GTT can serve as an indicator of diabetes mellitus type2 (DM2), which is characterized by an inability to effectively andefficiently transfer glucose from the bloodstream into tissues viainsulin signaling (Muoio and Newgard, 2008). Following a bolusintraperitoneal (i.p.) injection of a 20% glucose solution at time 0(T0), blood glucose spiked at 30 min (T30) and gradually cleared tofasting levels by 2 h (T120) in WT and Thm1-cko mice (Fig. 2F,H).Thm1-cko females showed a similar glucose clearance as WTfemales. However, in Thm1-cko males, glucose levels continued torise until 1 hour post-injection (T60) and remained significantlyelevated at 2 h (T120) (Fig. 2H). This impaired response suggests adiabetic state in Thm1-cko males.In addition, we weighed internal organs and found that the livers
of Thm1-cko mice were significantly heavier than those ofWTmice(Fig. S3). Fatty deposits were visible on Thm1-cko livers at thewhole-mount level (data not shown), and histological analysis of
liver sections using H&E and Oil Red O staining confirmed amarked increase in size and number of lipid droplets in Thm1-ckoliver, indicative of fatty liver disease (Fig. 3).
Thm1-cko mice are hyperphagicTo determine the cause of weight gain in Thm1-cko mice, weexamined energy expenditure and food intake. Using a force plateactimeter, activity of WT and Thm1-cko mice were monitored atmid-day during 10-min intervals 1 week, 6 weeks and 13 weeksfollowing gene deletion. Total distance traveled, rate of movement,and focused stereotypy (repetitive movement) were not significantlydifferent between WT and Thm1-cko mice (Fig. S4). In contrast,measurements of daily food intake over the 13-week period revealedthat female and male Thm1-cko mice consumed ∼19% and 9%more diet than female and male WT mice, respectively. Todetermine whether hyperphagia caused the Thm1-cko weight gain,we pair-fed control and Thm1-cko mice for 13 weeks followingThm1 ablation. Throughout these 13 weeks, WT and Thm1-ckomice showed similar body weights (Fig. 4A-C), suggestinghyperphagia is a main contributor to the Thm1-cko obesephenotype. At the 13-week time point, we noted that pair-fedThm1-cko females did emerge slightly heavier than control females(Fig. 4B), with heavier white adipose tissues (Fig. 4D) and ∼2.3-fold higher percent body fat than control females (Fig. 4F). Pair-fedThm1-cko male mice did not exhibit elevated weight gain relative tocontrol males (Fig. 4C,E,F). These findings suggest an increasedpropensity in Thm1-cko females to increase adipose tissue weight.
In pair-fed Thm1-cko mice, blood glucose levels remainedsimilar to those of WT mice at 0, 6 and 13 weeks post-Thm1deletion. In response to a GTT, pair-fed female and male WT andThm1-cko mice showed similar glucose clearance (Fig. 5A,B).Additionally, organ weights, including that of liver, were similarbetween pair-fed WT and Thm1-cko mice (Fig. S5). Histologicalanalysis revealed normal liver morphology in pair-fed Thm1-ckomice (Fig. 5C). These data indicate that the DM2 and fatty liverdisease present in ad libitum-fed Thm1-cko mice wereconsequences of the obese phenotype.
Thm1-cko mice exhibit altered neuropeptide geneexpression in the hypothalamic arcuate nucleusIn response to signals such as leptin or insulin, POMC- and AgRP/NPY-expressing neurons of the ARC regulate feeding response(Belgardt et al., 2009). A satiated state causes upregulation ofPOMC expression, which attenuates food-seeking behavior andincreases physical activity. Conversely, fasting increases AgRPand NPY expression, which augments food-seeking behavior anddecreases physical activity. To determine if Thm1-cko mice exhibitsignaling defects in the ARC, we examined primary cilia andneuropeptide gene expression in the ARC two weeks followinggene deletion. At this time point, WT and Thm1-cko body weightswere similar (Fig. 1B,C), so that any observed differences in geneexpression between WT and Thm1-cko mice might suggestmechanisms that initiate the obese phenotype. To examineprimary cilia, sections of the hypothalamus were immunostainedfor the neural ciliary marker, adenylyl cyclase 3 (AC3, also knownas ADCY3) (Bishop et al., 2007). The ARC is situated around theventrolateral region of the third ventricle. Neuronal primary cilia inthis region in Thm1-cko mice were shortened with a bulbous distaltip and showed more intense expression of AC3, suggesting anaccumulation of proteins in the mutant cilia (Fig. 6). QuantitativePCR of RNA lysates from the ARC of WT and Thm1-cko micerevealed ∼50% lower POMC expression in Thm1-cko mice than in
791
RESEARCH ARTICLE Disease Models & Mechanisms (2016) 9, 789-798 doi:10.1242/dmm.025791
WT mice (Fig. 7A). Because POMC neuropeptides attenuate food-seeking behavior, this decrease in POMC might be a primary causeof the hyperphagia. In response to fasting,WTARCextracts showed
decreased POMC expression and increased AgRP and NPYexpression. In contrast, fasting of Thm1-cko mice increased NPYtranscript levels, but did not cause expected alterations of POMCand AgRP transcripts. These results suggest that loss of Thm1 causesmisregulation of the feeding/activity signaling axis in the ARC.
Primary cilia regulate hedgehog (Hh) and Wnt signalingpathways (Nozawa et al., 2013; Oh and Katsanis, 2013), andthese pathways have been implicated in various aspects of adiposebiology (Christodoulides et al., 2009; Cousin et al., 2007).Activation of Hh or Wnt signaling inhibits adipocytedifferentiation (Ross et al., 2000; Spinella-Jaegle et al., 2001),and activation of Wnt signaling in mature brown adipocytes drivestheir conversion to white adipocytes (Tseng et al., 2005). Yet a rolefor these developmental signaling pathways in regulating energyhomeostasis in the hypothalamus has not been investigated.Previously, we identified THM1 as a negative regulator of Hhsignaling during embryonic development, including in neuronaltissue (Tran et al., 2008). Thus, we queried whether Hh signaling
Fig. 3. Obese Thm1-cko mice show hepatic steatosis. H&E and Oil Red Ostaining of liver sections from WT and Thm1-cko mice at 13 weeks post-tamoxifen injection. Scale bar: 100 µm.
Fig. 2. Obese Thm1-cko mice show hyperleptinemia,hyperinsulinemia and diabetes. (A-D) Serum metabolitelevels at 13 weeks post-tamoxifen injection. (E,G) Bloodglucose levels at 0, 6 and 13 weeks post-tamoxifen injection.N=6 WT; N=6 Thm1-cko each for male and female.(F,H) Glucose tolerance test. Mice ranging from 13 to20 weeks post-Thm1 deletion were challenged with aglucose bolus (2 mg/g body weight) at T0 by i.p. injection.Blood glucose was monitored at 30-min intervals todetermine clearance rate. N=3 WT females; N=6 Thm1-ckofemales; N=4 WT males; N=6 Thm1-cko males. Data pointsrepresent means±s.e.m. Two-tailed unequal variance t-test.*P<0.05.
792
RESEARCH ARTICLE Disease Models & Mechanisms (2016) 9, 789-798 doi:10.1242/dmm.025791
Disea
seModels&Mechan
isms
was altered in the ARC of Thm1-cko mice two weeks followinggene deletion. Using qPCR, we examined levels of Hh targets andsignaling components, Patched1 (Ptc1), Ptc2, Gli1, Gli2, Gli3 andHhip. In both non-fasted and fasted states, we did not observesignificant differences in gene expression between WT and Thm1-cko mice (Fig. 7B).Because mouse embryonic fibroblasts lacking Thm1 show an
increased response to Wnt3a ligand (Tran et al., 2014), we furtherexamined expression of neural Wnt signaling targets Axin2, Lef1,NeuroD1, SP5 and Cacna1g in the ARC of Thm1-cko mice. In non-fasted states, gene expression levels between WT and Thm1-ckomice were similar (Fig. 7C). In response to fasting, SP5 andNeuroD1 expression increased in WT mice, suggesting canonicalWnt signaling might be involved in the fasting response. In contrast,fasting of Thm1-cko mice did not significantly increase SP5expression (P=0.088) nor alter NeuroD1 expression, supporting anaberrant response to fasting in Thm1-cko mice.In addition, we examined serum levels of leptin, insulin, C-peptide,
resistin and GIP in Thm1-cko mice at this time point; two weeksfollowing gene deletion and prior to weight gain. Levels of allmetabolites were similar between WT and Thm1-cko mice (Fig. S6),indicating that the metabolic disturbances observed in obese Thm1-cko mice follow the molecular changes that occur in the ARC.
DISCUSSIONIn this study, we demonstrate that deletion of Thm1, a component ofthe IFT-A complex, causes hyperphagia-induced obesity in mice.These findings add a new mutant class to the list of ciliary mouse
models of hyperphagia and obesity, which includes mutants of theBBS complex, of Alms1, of the IFT-B complex, and of the transitionzone (Arsov et al., 2006; Collin et al., 2005; Davenport et al., 2007;Rahmouni et al., 2008; Stratigopoulos et al., 2014). The Thm1-ckoobese phenotype further causes glucose intolerance and insulinresistance, which together indicatemetabolic syndrome. In the humanpopulation,metabolic syndromehas becomeepidemicworldwide anda predictor of DM2, cardiovascular disease, and non-alcoholic fattyliver disease (Shin et al., 2013). Obese Thm1-cko mice also developdiabetes and fatty liver disease, modeling the human condition.
Prior to the increased weight gain and elevated serum metabolitelevels in Thm1-cko mice, POMC mRNA levels in the ARC werereduced. This POMC reduction is likely a main cause of thehyperphagia. POMC-derived peptides attenuate food-seekingbehavior and POMC-null mice are hyperphagic and obese(Yaswen et al., 1999). Further, deletion of Ift88 in POMC-expressing cells in mice resulted in hyperphagia and obesity,demonstrating that ciliary function specifically of POMC-expressing neurons is crucial to the neuronal circuitry thatcontrols appetite (Davenport et al., 2007). Reduced POMC levelshave been reported also in obese Bbs2−/−, Bbs4−/− and Bbs6−/−
mice (Rahmouni et al., 2008), in obese Rpgrip1l+/− mutant mice(Stratigopoulos et al., 2014), and recently, in pre-obese Bbs1;CreLRb
cko mice, in which Bbs1 was deleted in leptin receptor (LRb)-expressing cells (Guo et al., 2016). In contrast, POMC levels werenot reduced in Ift88;CreLRb cko mice. These data suggest thatdeficiency of Bbs, Rpgrip1l and Thm1 might perturb similarsignaling pathways.
Fig. 4. Hyperphagia is a primarycause of obesity in Thm1-cko mice.(A) Pair-fed WT and Thm1-cko mice13 weeks post-tamoxifen injection.(B,C) Weekly body weights of pair-fedmice over a 13-week period. Datapoints represent means±s.e.m.(D,E) Adipose tissue weights of pair-fed mice at 13 weeks post-tamoxifeninjection. Error bars representmeans±s.e.m. Two-tailed unequalvariance t-test. *P<0.05. (F) Percentbody fat of pair-fed mice at 13 weekspost-tamoxifen injection. Error barsrepresent means±s.e.m. *P<0.05; N=7WT females; N=7 Thm1-cko females;N=6 WT males; N=8 Thm1-cko males.
793
RESEARCH ARTICLE Disease Models & Mechanisms (2016) 9, 789-798 doi:10.1242/dmm.025791
In Bbs2−/− and Bbs6−/−mutant mice, POMC-expressing neuronswere decreased by ∼20%, contributing to but not completelyaccounting for the 40% reduction in POMC transcript levels (Seoet al., 2009). Data in our lab show that the number of P-STAT3-expressing neurons, which comprise both POMC- and AgRP-expressing neurons (Ha et al., 2013), is similar between WT andThm1-cko mice, suggesting that the number of these neuronalpopulations is unaffected. This is likely because Thm1 was deletedin adulthood, whereas Bbs2 and Bbs4 were deleted from thebeginning of embryogenesis. Thus, we propose that the decrease inPOMC transcripts in Thm1-cko mice results from altered signalingand not from POMC neuron number. Deregulated POMCtranscription in Thm1-cko mice in response to fasting furthersupports altered signaling. The ciliary phenotypes of these mutantsdiffer; cilia are ablated in complex B Ift88 and Kif3a mutants(Davenport et al., 2007), unchanged at the structural level in Bbs2and Bbs4 mutants (Berbari et al., 2008), shortened with a bulb-likestructure at the distal tip in IFT complex A Thm1 mutants (Tranet al., 2008) and lengthened in transition zone Rpgrip1l mutants
(Gerhardt et al., 2015; Stratigopoulos et al., 2014). Collectively,these data suggest that regulation of POMC and of molecularcomponents of POMC-expressing cells constitutes an importantunderlying mechanism of obesity in these ciliary mutants.
In WT mice, fasting decreased POMC expression and increasedAgRP and NPY expression to increase food-seeking behavior. Incontrast, in Thm1-cko mice, fasting caused upregulation of NPY,whereas POMC and AgRP levels remained unchanged, suggestingthat the Thm1 ciliary defect impairs the sensing of changes inperipheral signals. Similar to Thm1-cko mice, fasted Bbs2−/−,Bbs4−/− and Rpgrip1l+/− mice showed an impaired POMCtranscriptional response, whereas in contrast to Thm1-cko mice,AgRP and NPY transcriptional responses were normal (Seo et al.,2009). Our findings in the Thm1-cko mice reflect differentialregulation of NPY and AgRP. Differential regulation of NPY andAgRP has also been observed in rats devoid of a functional leptinreceptor (Korner et al., 2001) and in fasted C57BL/6 mice that uponre-feeding showed restored POMC levels and reduced NPY levels,but unaltered AgRP expression (Swart et al., 2002).
Although leptin regulates POMC, NPY and AgRP expression, therole of primary cilia in leptin signaling has been controversial.Genetic deletion of Bbs2, Bbs4, Bbs6 or of Rpgrip1l, which trafficproteins to the cilium, and siRNA-mediated knock-down of Kif3aand Ift88 in the hypothalamic third ventricle of mice, impairedresponse of these animals to the appetite- and weight-reducing effectsof exogenous leptin (Han et al., 2014; Seo et al., 2009; Stratigopouloset al., 2014). These studies were performed on lean, calorie-restrictedor pre-obese mice that were not hyperleptinemic, suggesting ciliarydysfunction causes a primary defect in leptin signaling. In contrast,pre-obese Bbs4-null mice and both pre-obese and lean, calorie-restricted mice with a global Ift88 deficiency induced duringadulthood responded normally to an intraperitoneal injection ofleptin, suggesting leptin resistance is a secondary consequence of theobese phenotype (Berbari et al., 2013). A recent report might accountfor some of these discrepancies. Bbs1;CreLRbmice developedmorbidobesity as a result of hyperphagia and reduced energy expenditure, incontrast to Ift88;CreLRb mice, which showed mild weight gain thatwas not a result of hyperphagia (Guo et al., 2016). Further, calorie-restricted Bbs1;CreLRb mice showed leptin resistance, whereas Ift88;CreLRb mice did not. Knock-down of Bbs1, but not of Ift88, showedimpaired trafficking of the leptin receptor to the cell membrane invitro, suggesting that BBS1, but not IFT88, is involved inintracellular trafficking of the leptin receptor. These results furtherimply that BBS leptin resistance is independent of the primarycilium. The results of these collective studies underscore the need toinvestigate leptin sensitivity in other ciliary mouse mutants, includingthose of a different ciliary mutant class, like Thm1.
Fig. 6. Thm1-cko primary cilia are stunted with abulbous distal tip and show enrichment of AC3.Immunostaining for AC3 (ACIII) on sections of thearcuate nucleus (ARC). Inset shows highermagnification of dotted region of the ARC. 3V, thirdventricle.
Fig. 5. Pair-fed Thm1-cko mice show normal glucose metabolism andliver morphology. (A,B) Glucose tolerance test for pair-fed mice at 13 to20 weeks post-Thm1 deletion. N=4 WT; N≥4 Thm1-cko each for male andfemale. Data points represent means±s.e.m. Two-tailed unequal variancet-test. (C) H&E staining of liver sections from WT and Thm1-cko mice at13 weeks post-tamoxifen injection.
794
RESEARCH ARTICLE Disease Models & Mechanisms (2016) 9, 789-798 doi:10.1242/dmm.025791
Disea
seModels&Mechan
isms
In addition to leptin, other signals such as insulin, glucose andestrogen, regulate POMC, NPY and AgRP expression (Gao et al.,2007; Varela and Horvath, 2012). Six hours of re-feeding followinga 24-h fast in C57BL/6 mice restored POMC and NPY levels, butleptin administration at a dose that reduced food intake did not affectneuropeptide gene expression, suggesting that multiple signalswork in concert to effect changes in POMC, AgRP and NPY (Swartet al., 2002). It has been shown that during embryonic development,primary cilia regulate multiple signaling pathways. Similarly, it ispossible that multiple ciliary-mediated pathways converge onneuropeptide gene expression. Identification of these signals andpathways will be crucial to delineating the role of primary cilia inenergy homeostasis.The importance of primary cilia during development is well
recognized, but the role of primary cilia during tissue maintenanceis only beginning to be understood. This study and others (Berbari
et al., 2013; Davenport et al., 2007) demonstrate that primary ciliaare important in adult hypothalamic neurons. This contrasts with thekidney, where loss of primary cilia during kidney developmentcauses aggressive cystic kidney disease, but loss of primary ciliaonce the kidney has matured results in very slow, late-onset cystdevelopment initiating 6 months following gene deletion(Davenport et al., 2007). Consistent with this, Thm1-cko kidneysshowed a normal phenotype at the end of this study period 13 weeksfollowing Thm1 deletion at 5 weeks of age (data not shown). Thesensitivity of the adult hypothalamus to ciliary changes might beattributed to the plasticity required for regulating energyhomeostasis (Horvath, 2005).
We observed phenotypic differences between Thm1-ckofemales and males. Thm1-cko females showed a greater increasein adipose tissue weight, which was reflected also in higher levelsof leptin and resistin. Similarly, Bbs4−/− and Bbs1−/− female miceshow a more severe obese phenotype than their male counterparts(Eichers et al., 2006; Guo et al., 2016). Yet Thm1-cko males, andnot females, were diabetic at the end of the study period. Suchdifferences are consistent with a study that showed gender-specificmetabolic changes in ob/ob mice, with changes in ob/ob malesand females associated with insulin signaling and lipidmetabolism, respectively (Won et al., 2013). Additionally, pair-fed Thm1-cko females, but not pair-fed Thm1-cko males, wereheavier than control littermates at the end of the 13-week study,further showing an increased tendency in Thm1-cko females toincrease adipose tissue weight. Pair-feeding of Bbs mutant micesimilarly did not completely rescue the increase in fat depotsrelative to control littermates, suggesting reduced energyexpenditure (Guo et al., 2016; Rahmouni et al., 2008).Similarly, energy expenditure such as in the form of basalmetabolic rate or thermoregulation, which could not be measuredby the actimeter, might also be affected by deficiency of Thm1.
Finally, loss of THM1 resulted in shortened primary cilia withbulbous distal tips and increased ciliary localization of AC3. Thisphenotype, including protein accumulation in the mutant cilia, ischaracteristic of an IFT-A mutant phenotype (Tran et al., 2008). Yetthis enrichment ofAC3 inThm1-mutant cilia contrasts with the absentor decreased ciliary localization of AC3 in mouse embryonicfibroblasts of other complex A mutants, Ift144 and Ift122 (Liemet al., 2012) and also of transition zone mutants, Rpgrip1l(Stratigopoulos et al., 2014). Deletion of AC3 in mice causesobesity (Wang et al., 2009), whereas a gain-of-function mutation inAC3 protects mice from diet-induced obesity (Pitman et al., 2014).The contrasting effects of Thm1 deficiency versus Ift144 or Ift122deficiency on AC3 ciliary localization might reflect the uniquebiochemical functions of individual IFT proteins. Further, theincreased ciliary localization of AC3 in Thm1-cko neuronal ciliacoupled with the obese phenotype questions the functionality of theadenylyl cyclase, and implicates a role for THM1 in regulating AC3function.
In summary, the Thm1-cko mouse provides the first IFT complexA mouse model of hyperphagia and obesity. Prior to increasedweight gain, deficiency of Thm1 downregulates POMC expressionin the ARC and misregulates POMC and AgRP expression inresponse to fasting. Interestingly, the Thm1-mutant phenotypesuggests mechanisms similar to Bbs and transition zone mutantsand not to Ift-B mutants, suggesting differential roles of IFT-Band IFT-A complexes in regulating energy homeostasis. Futureinvestigations into the molecular mechanisms regulatingneuropeptide gene expression and the role of THM1 and primarycilia in different neuronal populations of the ARC will provide a
Fig. 7. Pre-obese Thm1-cko mice exhibit altered neuropeptide geneexpression in the hypothalamic arcuate nucleus. Gene expressionanalysis at 2 weeks post-Thm1 deletion on extracts of the arcuate nucleususing qPCR for (A) appetite-regulating neuropeptides, and (B) hedgehog and(C) Wnt target genes. Gene expression was normalized to β-actin expression.Error bars represent means±s.e.m. Two-tailed unequal variance t-test.*P≤0.05; **P≤0.005; N≥7 WT; N≥7 Thm1-cko each for fasted (F) and non-fasted mice.
795
RESEARCH ARTICLE Disease Models & Mechanisms (2016) 9, 789-798 doi:10.1242/dmm.025791
Disea
seModels&Mechan
isms
greater understanding of the neural circuitry that controls energyhomeostasis, which will offer potential therapeutic targets againsthyperphagia and obesity.
MATERIALS AND METHODSMiceThm1-cko mice were generated as previously described (Tran et al., 2014).Briefly, Thm1-cko mice were generated with one allele harboring the alnmutation, which results in a null allele, and a floxed allele containing LoxPsites flanking exon 4. Deletion of the floxed exon was carried out usingtamoxifen-inducible ROSA26CreERT mice (Jackson Laboratories, Stock004847). Cre recombinase expression was induced at 5 weeks of age by i.p.injection of 10 mg tamoxifen/40 g mouse weight. Tamoxifen (Sigma T5648)was suspended in corn oil (Sigma C8267) at 30 mg/ml. To examinerecombination in control (Thm1fl/+;CreERT+) or Thm1-cko (Thm1fl/aln;CreERT+) mice, Gt(ROSA)26Sortm4(ACTB-tdTomato,-EGFP)Luo/Gt(ROSA)26Sortm4(ACTB-tdTomato,-EGFP)Luo reporter mice (Jackson Laboratories, Stock007676), which express tdTomato in non-recombined tissue and EGFP inrecombined tissue, were bred into theThm1fl/fl lines.Thm1fl/fl;tdTom+ femaleswere then bred with Thm1aln/+;CreERT+ males to generate progenyexpressing the reporter. Tissue from Thm1fl/+;tdTom+ or Thm1fl/aln;tdTom+mice and from Thm1fl/+;CreERT+;tdTom+ or Thm1fl/aln;CreERT+;tdTom+were embedded in OCT compound, cryosectioned at 10-µm thicknesses, andviewed under a fluorescence microscope. Mice were on a mixed geneticbackground containing FVB, SV129 and C57BL/6J strains. All animalprocedures were conducted in accordance with KUMC-IACUC andAAALAC rules and regulations.
FeedingMice were fed a Breeder diet (PicoLab 5058). To measure daily dietconsumption, WT and Thm1-cko mice were housed individually. Pairfeeding was performed by monitoring the daily diet intake (by weight) ofcontrol mice and providing the same amount to Thm1-cko littermates.
Serum collectionSubmandibular blood (100-200 µl) was collected in a Microvette CB300zblood collection tube (Kent Scientific, Torrington, CN). Serum wasisolated by centrifugation for 10 min at 1800×g at 4°C using a tabletopcentrifuge (PrismR; C2500-R). Serum was collected and treated withprotease inhibitor cocktail (Sigma) and stored at −80°C until analysis.Serum was analyzed using a Milliplex MAP Mouse Metabolic HormoneMagnetic Bead Panel – Metabolism Multiplex assay (EMD Millipore;MMHMAG-44K).
Activity levelsActivity levels were monitored by placing the mice onto a BASi Force PlateActimeter (KUMCRodent Behavior Facility) that uses force transduction tomonitor and track locomotor activity. Mice were allowed free access to the1 m×1 m actimeter arena for 10 min during the light cycle. Force PlateActimeter activity software reports locomotion behaviors including distancetraveled, rate of movement, and focused stereotypy.
Glucose tolerance testWT and Thm1-cko mice were fasted for 8-12 h prior to determining thebaseline glucose level. A bolus (2 mg/kg) of 20% glucose in PBS solutionwas delivered via i.p. injection at time 0 (T0). Blood glucose level wasdetermined using a Contour blood glucose monitor (Model 7151H, BayerCorp.) and was sampled from the tail vein at 30-min intervals for 4 h.Glucose levels between WT and Thm1-cko mice were compared using aStudents’ t-test at each time interval.
qPCR of ARC RNA extractsWhole mouse brain was harvested, then further dissected to obtain thewholehypothalamus, and finally, the arcuate nucleus was isolated from thehypothalamus. Isolated tissues were immediately placed onto dry ice forstorage. Tissue was homogenized with a pestle (www.bioexpress.com;C-3260-1) and RNA was extracted using the RNeasy mini kit (Qiagen).
cDNA was generated using qScript cDNA Supermix (Quanta Biosciences,95048-500) and real-time PCRwas performed using qPCR PerfeCTa SYBRGreen FastMix (Quanta BioSciences, 95072-012) in a BioRad CFXConnect Real-Time System. Intron-spanning primers for qPCR assays weredesigned using the Roche Universal ProbeLibrary Assay Design Center andsynthesized by IDT Technologies. All primer sequences are listed inTable S1.
Tissue processingMice were anesthetized with an i.p. injection of ketamine and xylazine, andwere cardiac perfused. Perfusion was performed with 7 ml of phosphatebuffer saline (PBS) followed by 7 ml of 4% paraformaldehyde (PFA) inPBS at∼3.5 ml/min. Tissues were isolated and submersed in 4% PFA on icefor 2-12 h or in Bouin’s fixative (Poly Scientific, Bay Shore, NY) for >24 h.Bouin’s fixed tissues were dehydrated through an ethanol series, paraffin-embedded and sectioned at 7-µm thicknesses. Brain tissues were fixed in 4%PFA, infused with 30% sucrose, embedded in Tissue-Tek OCT compound(Sakura) and cryosectioned at 10-15-µm thicknesses.
HistologyTissues were stained with hematoxylin and eosin using a standard protocol.Liver tissues were stained using Oil Red O (ORO) stain (Sigma-Aldrich,O0625). An initial stock of 0.5% ORO in 2-propanol was diluted to 0.3%ORO in 60% isopropanol and filtered into a Coplin jar for staining. PFA-fixed sections were rinsed once with 60% isopropanol then immersed in0.3% ORO in 60% isopropanol for 12 min. Sections were rinsed with 60%isopropanol for 3-4 min to remove excess ORO. Nuclei were stained withhematoxylin solution (5 dips) and rinsed with distilled H2O for 3 min.Sections were mounted with 100 µl ImmunoHistoMount solution (Sigma,I1161). Staining was viewed and imaged using a Nikon 80i microscopeequipped with a Nikon DS-Fi1 camera.
ImmunofluorescenceFixed tissues were OCT-embedded and sectioned at 10-15 µm. Primaryantibodies were diluted into blocking buffer (2% BSA in PBS) and tissuesections were incubated at 4°C for 1-12 h. Primary antibody used wasanti-adenylate cyclase III (C-20; Santa Cruz, sc-588; 1:100 dilution)and secondary antibody used was Alexa Fluor 594 goat anti-rabbit(Thermo Fisher, A-11012; 1:1000 dilution). Tissues were labeled withDAPI (5 µg/ml in PBS) to visualize nuclei. Tissue sections were thenmounted with Fluoromount-G mounting media (Electron MicroscopyServices). Immunolabeled tissues were viewed and imaged using a LeicaTCS SPE confocal microscope configured on a DM550 Q uprightmicroscope.
StatisticsTwo-tailed unequal variance t-tests were performed. P-value <0.5 wasconsidered significant.
AcknowledgementsWe thank members of the KUMC Department of Anatomy and Cell Biology and ofthe Kidney Institute for helpful discussions as well as Jing Huang of the KUMCHistology Core andMichelleWinter of the Rodent Behavior Facility for their technicalassistance.
Competing interestsThe authors declare no competing or financial interests.
Author contributionsD.T.J., L.M.S., B.A.A., M.P.S., A.C., G.C.T. and P.V.T. performed the experiments.D.T.J., D.R.B. and P.V.T. designed the experiments and wrote the manuscript.
FundingThis work was supported by the National Institutes of Health [T32DK71496-6 A1, PI:Jared Grantham, University of Kansas Medical Center (KUMC), Kansas City, USA;P20 GM103418, PI: Douglas Wright, KUMC, to D.T.J.]; [R01HD36404 to D.R.B.];[R21DK088048; P20 GM104936-06, PI: Dale Abrahamson, KUMC; R01DK103033to P.V.T.].
796
RESEARCH ARTICLE Disease Models & Mechanisms (2016) 9, 789-798 doi:10.1242/dmm.025791
Supplementary informationSupplementary information available online athttp://dmm.biologists.org/lookup/doi/10.1242/dmm.025791.supplemental
ReferencesArsov, T., Silva, D. G., O’Bryan, M. K., Sainsbury, A., Lee, N. J., Kennedy,C., Manji, S. S. M., Nelms, K., Liu, C., Vinuesa, C. G. et al. (2006).Fat aussie–a new Alstrom syndrome mouse showing a critical role forALMS1 in obesity, diabetes, and spermatogenesis. Mol. Endocrinol. 20,1610-1622.
Belgardt, B. F., Okamura, T. and Bruning, J. C. (2009). Hormone and glucosesignalling in POMC and AgRP neurons. J. Physiol. 587, 5305-5314.
Berbari, N. F., Lewis, J. S., Bishop, G. A., Askwith, C. C. and Mykytyn, K.(2008). Bardet–Biedl syndrome proteins are required for the localization of Gprotein-coupled receptors to primary cilia. Proc. Natl. Acad. Sci. USA 105,4242-4246.
Berbari, N. F., O’Connor, A. K., Haycraft, C. J. and Yoder, B. K. (2009).The primary cilium as a complex signaling center. Curr. Biol. 19, R526-R535.
Berbari, N. F., Pasek, R. C., Malarkey, E. B., Yazdi, S. M. Z., McNair, A. D., Lewis,W. R., Nagy, T. R., Kesterson, R. A. and Yoder, B. K. (2013). Leptin resistance isa secondary consequence of the obesity in ciliopathy mutant mice. Proc. Natl.Acad. Sci. US A 110, 7796-7801.
Bishop, G. A., Berbari, N. F., Lewis, J. and Mykytyn, K. (2007). Type III adenylylcyclase localizes to primary cilia throughout the adult mouse brain. J. Comp.Neurol. 505, 562-571.
Christodoulides, C., Lagathu, C., Sethi, J. K. and Vidal-Puig, A. (2009).Adipogenesis and WNT signalling. Trends Endocrinol. Metab. 20, 16-24.
Collin, G. B., Cyr, E., Bronson, R., Marshall, J. D., Gifford, E. J., Hicks, W.,Murray, S. A., Zheng, Q. Y., Smith, R. S., Nishina, P. M. et al. (2005). Alms1-disrupted mice recapitulate human Alstrom syndrome. Hum. Mol. Genet. 14,2323-2333.
Cousin, W., Fontaine, C., Dani, C. and Peraldi, P. (2007). Hedgehog andadipogenesis: fat and fiction. Biochimie 89, 1447-1453.
Davenport, J. R., Watts, A. J., Roper, V. C., Croyle, M. J., van Groen, T., Wyss,J. M., Nagy, T. R., Kesterson, R. A. and Yoder, B. K. (2007). Disruption ofintraflagellar transport in adult mice leads to obesity and slow-onset cystic kidneydisease. Curr. Biol. 17, 1586-1594.
Davis, E. E., Zhang, Q., Liu, Q., Diplas, B. H., Davey, L. M., Hartley, J., Stoetzel,C., Szymanska, K., Ramaswami, G., Logan, C. V. et al. (2011). TTC21Bcontributes both causal and modifying alleles across the ciliopathy spectrum. Nat.Genet. 43, 189-196.
Eichers, E. R., Abd-El-Barr, M. M., Paylor, R., Lewis, R. A., Bi, W., Lin, X.,Meehan, T. P., Stockton, D. W., Wu, S. M., Lindsay, E. et al. (2006). Phenotypiccharacterization of Bbs4 null mice reveals age-dependent penetrance andvariable expressivity. Hum Genet. 120, 211-226.
Gao, Q., Mezei, G., Nie, Y., Rao, Y., Choi, C. S., Bechmann, I., Leranth, C., Toran-Allerand, D., Priest, C. A., Roberts, J. L. et al. (2007). Anorectic estrogenmimicsleptin’s effect on the rewiring of melanocortin cells and Stat3 signaling in obeseanimals. Nat. Med. 13, 89-94.
Gerhardt, C., Lier, J. M., Burmuhl, S., Struchtrup, A., Deutschmann, K., Vetter,M., Leu, T., Reeg, S., Grune, T. and Ruther, U. (2015). The transition zoneprotein Rpgrip1l regulates proteasomal activity at the primary cilium. J. Cell Biol.210, 115-133.
Girard, D. and Petrovsky, N. (2011). Alstrom syndrome: insights into thepathogenesis of metabolic disorders. Nat. Rev. Endocrinol. 7, 77-88.
Guo, D.-F. and Rahmouni, K. (2011). Molecular basis of the obesityassociated with Bardet-Biedl syndrome. Trends Endocrinol. Metab. 22,286-293.
Guo, D.-F., Cui, H., Zhang, Q., Morgan, D. A., Thedens, D. R., Nishimura, D.,Grobe, J. L., Sheffield, V. C. and Rahmouni, K. (2016). The BBSome controlsenergy homeostasis bymediating the transport of the leptin receptor to the plasmamembrane. PLoS Genet. 12, e1005890.
Ha, S., Baver, S., Huo, L., Gata, A., Hairston, J., Huntoon, N., Li, W., Zhang, T.,Benecchi, E. J., Ericsson, M. et al. (2013). Somato-dendritic localization andsignaling by leptin receptors in hypothalamic POMC and AgRP neurons. PLoSONE 8, e77622.
Han, Y. M., Kang, G. M., Byun, K., Ko, H. W., Kim, J., Shin, M.-S., Kim, H.-K., Gil,S. Y., Yu, J. H., Lee, B. et al. (2014). Leptin-promoted cilia assembly is critical fornormal energy balance. J. Clin. Invest. 124, 2193-2197.
Heydet, D., Chen, L. X., Larter, C. Z., Inglis, C., Silverman, M. A., Farrell, G. C.and Leroux, M. R. (2013). A truncating mutation of Alms1 reduces the number ofhypothalamic neuronal cilia in obese mice. Dev. Neurobiol. 73, 1-13.
Horvath, T. L. (2005). The hardship of obesity: a soft-wired hypothalamus. Nat.Neurosci. 8, 561-565.
Islam, M. S. (2013). Animal models of diabetic neuropathy: progress since 1960s.J. Diabetes Res. 2013, 149452.
Jamaluddin, M. S., Weakley, S. M., Yao, Q. and Chen, C. (2012). Resistin:functional roles and therapeutic considerations for cardiovascular disease.Br. J. Pharmacol. 165, 622-632.
Korner, J., Savontaus, E., Chua, S. C., Jr., Leibel, R. L. and Wardlaw, S. L.(2001). Leptin regulation of Agrp and Npy mRNA in the rat hypothalamus.J. Neuroendocrinol. 13, 959-966.
Landreh, M., Johansson, J. and Jornvall, H. (2013). C-peptide: a moleculebalancing insulin states in secretion and diabetes-associated depositoryconditions. Horm. Metab. Res. 45, 769-773.
Liem, K. F., Jr, Ashe, A., He, M., Satir, P., Moran, J., Beier, D., Wicking, C. andAnderson, K. V. (2012). The IFT-A complex regulates Shh signaling through ciliastructure and membrane protein trafficking. J. Cell Biol. 197, 789-800.
Lindstrand, A., Davis, E. E., Carvalho, C. M. B., Pehlivan, D., Willer, J. R., Tsai,I.-C., Ramanathan, S., Zuppan, C., Sabo, A., Muzny, D. et al. (2014). RecurrentCNVs and SNVs at the NPHP1 locus contribute pathogenic alleles to Bardet-Biedlsyndrome. Am. J. Hum. Genet. 94, 745-754.
Liu, T., Wang, Q., Berglund, E. D. and Tong, Q. (2012). Action of neurotransmitter:a key to unlock the AgRP neuron feeding circuit. Front Neurosci. 6, 200.
Lynn, F. C., Pamir, N., Ng, E. H. C., McIntosh, C. H. S., Kieffer, T. J. andPederson, R. A. (2001). Defective glucose-dependent insulinotropic polypeptidereceptor expression in diabetic fatty Zucker rats. Diabetes 50, 1004-1011.
Millington, G. W. M. (2007). The role of proopiomelanocortin (POMC) neurones infeeding behaviour. Nutr. Metab. 4, 18.
Muoio, D. M. and Newgard, C. B. (2008). Mechanisms of disease: molecular andmetabolic mechanisms of insulin resistance and beta-cell failure in type 2diabetes. Nat. Rev. Mol. Cell Biol. 9, 193-205.
Nozawa, Y. I., Lin, C. and Chuang, P.-T. (2013). Hedgehog signaling from theprimary cilium to the nucleus: an emerging picture of ciliary localization, traffickingand transduction. Curr. Opin. Genet. Dev. 23, 429-437.
Oh, E. C. and Katsanis, N. (2013). Context-dependent regulation of Wnt signalingthrough the primary cilium. J. Am. Soc. Nephrol. 24, 10-18.
O’Neill, S. and O’Driscoll, L. (2014). Metabolic syndrome: a closer look at thegrowing epidemic and its associated pathologies. Obes. Rev. 16, 1-12.
Pitman, J. L., Wheeler, M. C., Lloyd, D. J., Walker, J. R., Glynne, R. J. andGekakis, N. (2014). A gain-of-function mutation in adenylate cyclase 3 protectsmice from diet-induced obesity. PLoS ONE 9, e110226.
Quinlan, R. J., Tobin, J. L. and Beales, P. L. (2008). Modeling ciliopathies: primarycilia in development and disease. Curr. Top. Dev. Biol. 84, 249-310.
Rahmouni, K., Fath, M. A., Seo, S., Thedens, D. R., Berry, C. J., Weiss, R.,Nishimura, D. Y. and Sheffield, V. C. (2008). Leptin resistance contributes toobesity and hypertension in mouse models of Bardet-Biedl syndrome. J. Clin.Invest. 118, 1458-1467.
Ross, S. E., Hemati, N., Longo, K. A., Bennett, C. N., Lucas, P. C., Erickson, R. L.and MacDougald, O. A. (2000). Inhibition of adipogenesis by Wnt signaling.Science 289, 950-953.
Seo, S., Guo, D.-F., Bugge, K., Morgan, D. A., Rahmouni, K. and Sheffield, V. C.(2009). Requirement of Bardet-Biedl syndrome proteins for leptin receptorsignaling. Hum. Mol. Genet. 18, 1323-1331.
Shin, J.-A., Lee, J.-H., Lim, S.-Y., Ha, H.-S., Kwon, H.-S., Park, Y.-M., Lee, W.-C.,Kang, M.-I., Yim, H.-W., Yoon, K.-H. et al. (2013). Metabolic syndrome as apredictor of type 2 diabetes, and its clinical interpretations and usefulness.J. Diabetes Investig. 4, 334-343.
Sohn, J.-W., Elmquist, J. K. and Williams, K. W. (2013). Neuronal circuits thatregulate feeding behavior and metabolism. Trends Neurosci. 36, 504-512.
Spinella-Jaegle, S., Rawadi, G., Kawai, S., Gallea, S., Faucheu, C., Mollat, P.,Courtois, B., Bergaud, B., Ramez, V., Blanchet, A. M. et al. (2001). Sonichedgehog increases the commitment of pluripotent mesenchymal cells into theosteoblastic lineage and abolishes adipocytic differentiation. J. Cell Sci. 114,2085-2094.
Stratigopoulos, G., Martin Carli, J. F., O’Day, D. R., Wang, L., LeDuc, C. A.,Lanzano, P., Chung, W. K., Rosenbaum, M., Egli, D., Doherty, D. A. et al.(2014). Hypomorphism for RPGRIP1L, a ciliary gene vicinal to the FTO locus,causes increased adiposity in mice. Cell Metab. 19, 767-779.
Swart, I., Jahng, J.W., Overton, J. M. and Houpt, T. A. (2002). Hypothalamic NPY,AGRP, and POMC mRNA responses to leptin and refeeding in mice.Am. J. Physiol. Regul. Integr. Comp. Physiol. 283, R1020-R1026.
Tran, P. V., Haycraft, C. J., Besschetnova, T. Y., Turbe-Doan, A., Stottmann,R. W., Herron, B. J., Chesebro, A. L., Qiu, H., Scherz, P. J., Shah, J. V. et al.(2008). THM1 negatively modulates mouse sonic hedgehog signal transductionand affects retrograde intraflagellar transport in cilia. Nat. Genet. 40, 403-410.
Tran, P. V., Talbott, G. C., Turbe-Doan, A., Jacobs, D. T., Schonfeld, M. P., Silva,L. M., Chatterjee, A., Prysak, M., Allard, B. A. and Beier, D. R. (2014).Downregulating hedgehog signaling reduces renal cystogenic potential of mousemodels. J. Am. Soc. Nephrol. 25, 2201-2212.
Tseng, Y.-H., Butte, A. J., Kokkotou, E., Yechoor, V. K., Taniguchi, C. M.,Kriauciunas, K. M., Cypess, A. M., Niinobe, M., Yoshikawa, K., Patti,M. E. et al. (2005). Prediction of preadipocyte differentiation by geneexpression reveals role of insulin receptor substrates and necdin. Nat. Cell Biol.7, 601-611.
Varela, L. and Horvath, T. L. (2012). Leptin and insulin pathways in POMC andAgRP neurons that modulate energy balance and glucose homeostasis. EMBORep. 13, 1079-1086.
797
RESEARCH ARTICLE Disease Models & Mechanisms (2016) 9, 789-798 doi:10.1242/dmm.025791
Wang, Z., Li, V., Chan, G. C. K., Phan, T., Nudelman, A. S., Xia, Z. and Storm,D. R. (2009). Adult type 3 adenylyl cyclase-deficient mice are obese. PLoS ONE4, e6979.
Won, E.-Y., Yoon, M.-K., Kim, S.-W., Jung, Y., Bae, H.-W., Lee, D., Park, S. G.,Lee, C.-H., Hwang, G.-S. and Chi, S.-W. (2013). Gender-specific metabolomic
profiling of obesity in leptin-deficient ob/ob mice by 1H NMR spectroscopy. PLoSONE 8, e75998.
Yaswen, L., Diehl, N., Brennan, M. B. and Hochgeschwender, U. (1999). Obesityin the mouse model of pro-opiomelanocortin deficiency responds to peripheralmelanocortin. Nat. Med. 5, 1066-1070.
798
RESEARCH ARTICLE Disease Models & Mechanisms (2016) 9, 789-798 doi:10.1242/dmm.025791
