Gene expression profiles post Roux-en-Ygastric bypassYuan Xu, MD, Eduardo J. B. Ramos, MD, Frank Middleton, PhD, Irina Romanova, PhD,Robert Quinn, DVM, Chung Chen, PhD, Undurti Das, MD, Akio Inui, MD,and Michael M. Meguid, MD, PhD, Syracuse, NY, and Kobe, Japan

Background. The hypothalamus is involved in regulation of food intake (FI) and fat deposition.Molecular mechanisms of weight loss after Roux-en-Y gastric bypass (RYGB) were studied by correlatingchanges in gene expression profiles in hypothalamic arcuate nucleus (ARC) and subcutaneousabdominal fat (SAF).Method. Diet-induced obese rats were divided into RYGB, sham-operated (SO-Obese), and sham-operated pair-fed (PF) groups. A non-obese group on a regular chow diet served as control. Body weight(BW) and FI were measured. Rats were killed 10 days after the operation. Plasma was analyzed forbiochemical indices, ARC and SAF were analyzed for gene expression profiles. Body SAF was alsoweighed. Data were analyzed by ANOVA and factor analysis.Results. BW and FI decreased in RYGB versus SO-Obese, as reflected by decreased SAF (53%). Genessimilarly expressed in ARC and SAF after RYGB were limited to several genes that predominantly relatedto metabolic pathways of carbohydrate, fat, neuropeptide, and cytokines. These expression profiles weresimilar to those seen in chow control and to those seen in a comparison of PF and SO-Obese.Conclusions. RYGB-induced weight loss is associated with changes in gene profile expressions that couldinfluence metabolic changes, contributing to weight loss. (Surgery 2004;136:246-52.)

From the Surgical Metabolism and Nutrition Laboratory, Neuroscience Program, Department of Surgery, andthe Department of Neuroscience and Physiology, SUNY Upstate Medical University, Syracuse, NY; theDepartment of Statistics, Management Information and Decision Sciences, Whitman School of Management,Syracuse University, Syracuse, NY; and the Division of Diabetes, Digestive and Kidney Diseases, Department ofClinical Molecular Medicine, Kobe University Graduate School of Medicine, Kobe, Japan

THE INCREASING PREVALENCE of obesity is due toincreased dietary intake and reduced physicalactivity.1 Obese patients are predisposed to type 2diabetes mellitus, hypertension, and coronaryheart disease (CHD).2 However, not all individualswho eat the same amount of food have the sameweight, suggesting that there are genetic dif-ferences that either predispose or prevent thedevelopment of obesity.3-4 Therefore, identificationof these genes and their actions is important tounderstand the pathophysiology of obesity.

Both hypothalamic neurons and adipose cellscontain the same genes except that their degree,whereby they express the product, may be differ-

Presented at the 65th Annual Meeting of the Society ofUniversity Surgeons, St. Louis, Missouri, February 11-14, 2004.

Supported in part by an educational grant from the Departmentof Surgery, the Hendrick’s Fund, and by material support fromEthicon, Cincinnati, Ohio, and Mead Johnson, Evansville, Ind.

Reprint requests: Michael M. Meguid, MD, PhD, Department ofSurgery, UniversityHospital, 750 EAdams St, Syracuse, NY 13210.

0039-6060/$ - see front matter

� 2004 Elsevier Inc. All rights reserved.

doi:10.1016/j.surg.2004.04.027

246 SURGERY

ent. Because the hypothalamus controls hunger,satiety, food intake (FI), and energy balance, andbecause adipose cells store the energy excess, wepostulated that a close interaction between themexists. It is likely that certain genes essential forenergy metabolism are similarly expressed inhypothalamic neurons and adipose cells. Wehypothesized that such common genes could beamong those that code for the metabolic processesleading to synthesis of peptides, neurotransmitters,and hormones such as leptin, adrenaline andnoradrenaline, cytokines, and glucocorticoids. Thispostulate can be verified by studying gene expres-sion profiles in specific hypothalamic food intake-regulating nuclei, such as the arcuate nucleus(ARC) and in fat deposits such as subcutaneousabdominal fat (SAF), in obese and non-obeseexperimental animals.

Roux-en-Y gastric bypass (RYGB) in morbidobesity produces prolonged and sustained weightloss.5 It is likely that the small gastric pouch and thereduction of nutrient absorption due to a longbypass of the small bowel contribute to weightloss. When this surgical technique was applied to

SurgeryVolume 136, Number 2

Xu et al 247

obese Zucker rats, they lost weight.6 However, thepossibility that the RYGB may reset the gene ex-pression profiles of FI-regulating peptides, neuro-transmitters, and cytokines in the hypothalamus,and in peripheral fat contributing to weight lossremains to be determined. We report the effect ofRYGB on changes in the pattern of gene expressionin ARC and in SAF to determine if there isa commonality between these 2 tissues in diet-induced obese rats after RYGB and in pair-fed (PF)diet restriction–induced weight loss compared tochow-fed control rats. These data serve to generatea hypothesis for future testing, rather than toprovide evidence regarding the mechanisms bywhich gastric bypass surgery influences weight loss.Furthermore, our results are placed into context byproviding some changes in blood parameter datapreviously described6,7 after RYGB. We also presentnew data on the effect of RYGB on FI, as measuredby meal size (MZ) and meal number (MN), and onthe decreases in SAF, and draw on the recentlyreported data bank of the complete genomepicture in rat obesity.8

METHODS

Approval for the experiments was obtained fromthe Committee for the Humane Use of Animals atSUNY Upstate Medical University. Animal care wasin accordance with the guidelines established bythe National Institutes of Health.

Twenty-four 3-week-old male Sprague-Dawleypups (Charles River, Wilmington, Mass), weighing54.9 ± 1.8 g, were housed in wire cages. Studysurroundings included 12-hour light-dark cycle(05:00-17:00); room temperature of 268C ± 18C;and 45% relative humidity. Diet-induced obesity(DIO) was achieved by a high-energy diet (HED,D12266; Research Diets, New Brunswick, NJ) anda highly palatable liquid diet (Boost Plus; MeadJohnson, Evansville, Ind). The HED diet consistedof 8% corn oil, 44% sweetened condensed milk,and 48% Purina rat chow. This diet provided 4.5kcal/g, of which 21% of the metabolizable energywas protein; 31%, fat; and 48%, carbohydrate; thelatter was 50% sucrose. Boost-Plus provided 1.5kcal/mL, of which 16.7% of the metabolizableenergy content was protein; 30.0%, fat; and 47.3%,carbohydrate. Rats had ad libitum HED pellets,Boost Plus, and tap water beginning at 3 to 4 weeksof age and ending at 11 to 12 weeks of age. Asa control, 3-week-old Sprague Dawley pups were fedregular chow diet (Rat Chow Diet # 5008; RalstonPurina, St Louis, Mo) for the same period (Chow-Control). After 7 weeks, the 24 DIO rats were

stratified based on the amount of body weight(BW) gained and then divided into 3 groups: (1)sham-operated ad lib (SO-Obese); (2) Roux-en-Ygastric bypass (RYGB); and (3) sham-operated pair-fed (PF). One week before the operation, the ratswere placed in individual metabolic cagesequipped with automated computerized rat eatermeter (ACREM) units, developed in this laboratoryto determine the microstructure of FI.9 It wasdesigned to continuously measure daily FI, MZ, andMN (ie, FI = MZ3MN), and thus determine whichcomponent of FI changed with RYGB, PF, and SO-Obese.

Rats were food deprived for 16 to 18 hoursbefore the operation and had either a RYGBprocedure or a celiotomy (sham operation). Theoperative procedures were previously docu-mented.6,7 Following the surgical procedure andpostoperative recovery, the rats were returned totheir individual ACREM cages. Boost-Plus andwater were provided for the first 3 days after theoperation, followed by an ad libitumHED diet for 7days. The PF group was restricted to consumptionof the average of the caloric intake combined fromliquid and solid diets consumed by the RYGBgroup. Daily FI, calorie intake, MZ, MN, and BWwere measured. Caloric intake was represented bythe total energy intake of HED pellets and by theBoost Plus consumed. After 10 days, the approxi-mate duration equivalent to 1 human year, the ratswere decapitated under isoflurane anesthesia. Theliver and SAF were dissected out and weighed, andthe differences between the groups calculated. Thebrain was rapidly removed and placed in dry ice,and the bilateral ARC was microdissected.10 Bloodwas obtained and analyzed for serum concentrationof glucose and triglyceride with the use ofenzymatic calorimetric kits (Sigma, St. Louis, Mo),and for serum insulin and leptin with the use ofenzyme immunoassay kits (DSL, Chicago, Ill). Theliver adipose content was determined as previouslydescribed.11

Qualitative changes in gene expression profilesin the ARC and SAF were performed as previouslydescribed.8 Briefly, the analysis of gene expressionpatterns was carried out with the use of the GeneChip Rat U34A Gene Chip (Affymetrix, SantaClara, Calif). To quantify gene expression, ARCand SAF were rapidly dissected from the humanelykilled animals, and RNA was extracted and purifiedfrom 3 of the animals in each group with the use ofthe RNeasy kit (Qiagen Inc, Valencia, Calif).Hybridizations of the best quality RNA (ie, largest28S:18S and 260:280 ratios) available from a singleanimal for both tissues in each treatment group

SurgeryAugust 2004

248 Xu et al

Table I. Changes in caloric intake, meal size, meal number, and body weight before and after the operation(mean ± SE).

Caloric intake(kcal/day)

Meal size(g/meal)

Mean number(meal/day)

Bodyweight (g)

Group Preop Postop Preop Postop Preop Postop Preop Postop

SO-Obese(n = 8)

129.7 ± 5.6 90.3 ± 5.2 1.3 ± 0.1 1.6 ± 0.1 12.3 ± 1.0 12.5 ± 0.7 492.1 ± 13.8 484.4 ± 15.4

RYGB(n = 6)

119.8 ± 10.0 31.4 ± 5.8 1.3 ± 0.1 0.9 ± 0.2* 13.1 ± 0.6 10.8 ± 0.8* 496.9 ± 16.3 421.7 ± 25.4*

PF(n = 7)

120.7 ± 6.2 48.4 ± 0.1 1.1 ± 0.1 — 12.3 ± 0.7 — 499.81 ± 14.93 436.1 ± 13.9

SO-Obese, Sham-operated obese; RYGB, Roux-en-Y gastric bypass; PF, sham-operated pair fed.*Versus SO-Obese, P < .05. Caloric intake was represented by the total energy intake of HED pellets and Boost Plus consumed.

were done. Thus, the 8 microarray data pointsrepresent 8 single microarray hybridizations. Tworounds of aRNA amplification were performed oneach sample. After purification of biotinylatedcRNA, 15 lg of this product was fragmentedrandomly into 35 to 200 bases (948C, 35 minutesin fragmentation buffer), added to a hybridizationcocktail that also contained a known concentrationof spiked in controls, heated (958C, 5 minutes),equilibrated (458C, 5 minutes), and then centri-fuged for 5 minutes. Samples were injected into theRat U34A Gene Chip array (Affymetrix) for in-cubation. Probe arrays were washed and stained onthe Affymetrix Fluidics Station 400 according to theEukGE-WS2 protocol. Fluorescent images werescanned and acquired with the use of the Hewlett-Packard G2500A Gene Array Scanner and GeneChip software (Affymetrix). Data obtained bymicroarray were analyzed with the use of GeneSpring (Silicon Genetics, Redwood City, Calif)software, with the Chow-Control group samplesfrom each tissue type serving as baseline controls.The expression level for each gene is thus thequotient of the level in the given sample divided bythe baseline sample as described in Results.

Differences in BW, caloric intake, MZ, and MNwere examined by using ANOVA and the non-parametric Mann-Whitney U and Student t test.Biochemical data were analyzed by using ANOVAand the Mann-Whitney U test. A P value < .05 wasconsidered statistically significant. Data are ex-pressed as mean ± SE.

RESULTS

Table I shows that preoperatively, FI, caloricintake, MZ, MN, and BW between the groups werenot statistically different. Postoperatively, in theRYGB compared with SO-Obese, caloric intake was

significantly decreased by 66% (31.4 ± 5.8 kcal/dayvs 90.3 ± 5.2 kcal/day; P < .05). The decrease incaloric intake was caused by a decrease in MZ,which was significantly reduced in the RYGB versusSO-Obese (0.9 ± 0.2 g/meal vs 1.6 ± 0.1 g/meal).MN was also decreased after the operation in theRYGB group versus SO-Obese (10.8 ± 0.8 meal/dayvs 12.5 ± 0.7 meal/day P < .05). BW significantlydecreased after RYGB versus SO-Obese (421.7 ±25.4 g vs 484.4 ± 15.4 g; P < .05). Mean weight lossafter the gastric bypass operation was 84.5 ± 11.5 g,representing a 16.1% BW loss. In the PF group, BWwas also significantly decreased compared to SO-Obese (436.1 ± 13.9 g vs 484.4 ± 15.4 g). Meanweight loss in PF group was 63.7 ± 5.4 g (12.7%).

Serum glucose concentration was significantlydecreased (P < .05) in the RYGB and PF versusSO-Obese (145.4 ± 10.5 mg/dL and 139.7 ± 5.4mg/dL vs 173.1 ± 8.1 mg/dL, respectively). Seruminsulin concentration was also significantly de-creased in the RYGB and PF groups versus SO-Obese (0.46 ± 0.04 ng/mL and 0.40 ± 0.12 ng/mLvs 0.85 ± 0.17 ng/mL, respectively). Serumconcentration of triglycerides was reduced in theRYGB and PF groups compared to SO-Obese (83.1± 6.7 mg/dL and 51.9 ± 5.0 mg/dL vs 126.1 ± 31.5mg/dL, respectively). Serum leptin concentrationwas decreased in both RYGB and PF groupsversus SO-Obese (452.4 ± 133.0 pg/mL and338.9 ± 61.7 pg/mL vs 944.7 ± 147.3 pg/mL,respectively). After RYGB, SAF was significantlydecreased in RYGB versus SO-Obese, as shown inFig 1. Liver fat content was also significantlydecreased in both RYGB and PF groups versusSO-Obese (44.0 ± 5.7 mg/g and 36.4 ± 7.6 mg/gvs 55.6 ± 3.8 mg/g, respectively).

The results of pattern analysis of the microarraydata comparing ARC and SAF are given in Fig 2and Table II. To obtain these lists, we searched for

SurgeryVolume 136, Number 2

Xu et al 249

all genes that displayed a highly specific pattern ofinterest in both tissue types—namely, an increase inSO-Obese samples versus Chow-Control that sub-sequently reverted to normal in the RYGB samplesmore than in the PF samples. Then, the expressiondata were examined to determine those genes withthe highest correlation to this multivariate patternin both tissues by using principal componentanalysis (PCA). Not surprisingly, the first principalcomponent (ie, ‘‘eigengene’’) described 90% of thevariance in the expression matrix for the top genesidentified. The expression patterns of the top 14

Table II. PCA reveals genes whose expressionpatterns correctly predict the treatment of bothhypothalamic and subcutaneous abdominal fattissues. The 4 treatment groups of the presentstudy include Chow-Control, RYGB, SO-Obese,and PF control. The PCA revealed a singularsolution (factor) that correctly described 90% ofthe variance across the treatment groups for thecombined tissue types. Listed below are some ofthe genes with the highest correlation to thepattern of interest (as shown in Fig 2). These genesare all increased in SO-Obese relative to Chow-Control rats and decreased by the RYGBprocedure.

Accession Correlation Gene description

X72792 0.9921 Alcohol dehydrogenaseX14318 0.9832 Rat mRNA for T-cell receptor

alpha chainM67465 0.9818 Hydroxy-delta-5-steroid

dehydrogenase, 3 beta andsteroid delta-isomerase

M84149 0.9738 Rat IgH chain VJ region mRNAAJ0111115 0.9672 Endothelial nitric oxide

synthase, 5’ region, partialD00036 0.9637 Phospholipase A2, group 1BM18349 0.9649 Rat leukocyte common antigenM23887 0.9608 Rat T-cell receptor beta-chain

mRNA V-regionAF055714 0.9576 Rattus norvegicus

hypertension-regulatedvascular factor

X60290 0.9567 Rattus norvegicusimmunoglobulin heavychain variable region

U25802 0.9563 Leuteinizing hormonesubunit beta

AA891764 0.9533 Low density lipoproteinreceptor-related protein 2

AA893618 0.9518 Glucocorticoid receptorAA998983 0.9514 Leptin receptor

PCA, Principal component analysis; O-Obese, sham-operated obese;RYGB, Roux-en-Y gastric bypass; PF, sham-operated pair fed; IgH VJ,immunoglobin H, variable J.

genes with the highest factor weights (correlation)on the first principal component are shown inTable II. These genes are increased in obese ratsrelative to Chow-Control and decreased afterRYGB. Inspection of the genes in this figureindicates transcriptional changes involving meta-bolic pathways, immune response, and inflamma-tion. In the ARC and SAF tissues in the obese rats,we noted an increased expression of genes foralcohol dehydrogenase, hydroxy-delta-5-steroid de-hydrogenase, and leptin, which are involved incarbohydrate, lipid, and protein metabolism.These expression levels were decreased afterRYGB and in diet-restricted PF rats. Similarly, theexpression of genes for T-cell receptor alpha chain,T-cell receptor beta chain, leukocyte commonantigen, prostaglandin F receptor, phospholipaseA2, and immunoglobulin heavy chain variableregion were increased in the obese rats andsuppressed after RYGB and PF, suggesting a rolefor cell and humoral immune responses in thepathobiology of obesity. The increased expressionof genes for hypertension-regulated vascular factor,low-density lipoprotein receptor, and endothelialnitric oxide synthase in the obese rats, and theirsuppression in RYGB and PF may explain the closerelationship seen between obesity, hypertension,and lipid abnormalities.

DISCUSSION

Using a rat model of RYGB in diet-inducedobese rats, we compared profiles of gene expres-sion changes that occur simultaneously in theprimary FI-regulating ARC of the hypothalamusand in SAF. We hypothesized that in weight gainthat occurs in obesity and in weight loss that occursafter RYGB or by pair-feeding, the same genesassociated with carbohydrate and fat metabolismare likely to be altered in ARC and SAF. Such anassociation would reflect a common link between

Fig 1. Subcutaneous abdominal fat weight after RYGB.*Versus SO-Obese and PF; yversus SO-Obese, PF, andRYGB.

SurgeryAugust 2004

250 Xu et al

Fig 2. Top 10 genes with the strongest correlation to the pattern of interest in the arcuate nucleus andsubcutaneous abdominal fat. All of these genes displayed a pattern of increasing in SO-Obese compared toChow-Control rats (Normal), then decreasing after RYGB to an extent that was generally greater than thedecrease in a calorie-restricted pair-fed (PF) rats. ADH3, alcohol dehydrogenase 3; Tcell rec, T-cell receptor;EST, expressed sequence tag; IgH VJ, immunoglobin H, variable J region; ATP3b4, ATPase, Ca++transporting, plasma membrane 4; CD28, cd28 antigen; NOS, nitric oxide synthase, endothelial; GOT2,glutamate oxaloacetate transaminase type 2, mitochondrial; PLA2, phospholipase A2, group 1b.

the brain and the gut. Our gene-array data have notbeen confirmed at the level of transcription. Ourdata serve to generate a hypothesis rather thanprovide evidence regarding the mechanisms bywhich gastric bypass surgery influences weight loss.A limitation of using Sprague-Dawley rats is that thevariation in the genetic pool is small compared tohumans, but one advantage is that it eliminates yetone more variable in this very complex situation.Consequently, this limitation needs to be con-sidered when interpreting our current data.

Morbid obesity treated by gastric bypass surgeryproducesweight loss.5 This is evident from the resultsof the present study. Caloric intake was significantlyreduced after RYGB via a reduction in MZ. Undernormal physiologic conditions, this would be com-pensated for by an increase in MN; this did notoccur.12 It is established that both MZ and MN areregulated separately in the brain. The reduced FIdecreased BW and SAF. In addition, other changesoccurred that include specific changes in theexpression profiles of several genes concerned withcontrol of FI. Increased expression of alcoholdehydrogenase, hydroxy-delta-5-steroid dehydroge-nase, glucocorticoid receptor, and leptin occurred inthe SO-Obese; their expression was suppressed afterRYGB in the ARC and the SAF, suggesting that there

are specific changes in the carbohydrate and lipidmetabolism with obesity and weight loss.

Alcohol dehydrogenase enzyme plays a majorrole in synthesis of fatty acids and triglycerides, andits genetic variation has an impact on CHD, which iscommonly associated with obesity.13 Hydroxy-delta-5-steroid dehydrogenase and luteinizing hormoneparticipate in the synthesis of pregnenolone, pro-gesterone, and other steroid hormones, thus havingroles in fertility and tissue-specific fat deposition.14

Their significance in obesity is emphasized byincreased incidence of pregnancy after weight lossafter RYGB.15 Similar decreases in these genes didnot occur with PF. The reason is not known but maybe related to the stress of the PF model.

Abdominal obesity is associated with metabolicabnormalities including hyperlipidemia, insulinresistance, and type 2 diabetes, which contributeto premature death from CHD. Glucocorticoidsregulate adipose tissue differentiation, function,and distribution; excessive adipose tissue causesabdominal obesity.16 The increase in the expres-sion of LDL receptor (Table II) is one of the lipidabnormalities that cause premature CHD. LDLreceptor is essential for the transport and metab-olism of LDL molecules. Glucocorticoid receptorand cortisol actions control insulin resistance,

SurgeryVolume 136, Number 2

Xu et al 251

carbohydrate, protein, and lipid metabolism.16,17

Excessive glucocorticoids cause hypertension, obe-sity, and glucose intolerance, which are commonfeatures in abdominal obesity and metabolicsyndrome.16,17 Enhanced production of gluco-corticoids specifically by the omental adipose tissueis the primary cause of abdominal obesity.17 Inaddition, glucocorticoids are anti-inflammatorybecause they suppress the production of pro-inflammatory cytokines, interleukin-6 (IL-6), andtumor necrosis factor-a (TNF-a).18

In obese subjects, the plasma concentrations ofIL-6 and TNF-a are elevated and, hence, obesity isconsidered to be an inflammatory condition.19,20

Previously, we observed that SAF and mesentericadipose tissues produced significantly higheramounts of IL-6 and TNF-a compared to control,while both plasma and adipose tissue concentra-tions of corticosterone was not significantly altered,tilting the balance more toward inflammation.9 Inthe present study, increase in the expression ofT-cell receptor alpha and beta chains was also notedin the SO-Obese rats, suggesting activation of Tcells, which secrete IL-6 and TNF-a, and supportingthe concept that inflammation plays a role in thepathobiology of obesity. Increased expression ofleptin receptor seen in the SO-Obese (presentstudy) expressed in both hypothalamic nuclei andadipose tissue also has proinflammatory actions.20

This suggests that the increase in glucocorticoidreceptor expression is probably compensatory innature, secondary to enhanced T-cell activation andincreased production of IL-6 and TNF-a seen inobesity. Increased expression of hypertension-regulated vascular factor occurs in the ARC andSAF tissues in SO-Obese rats. Furthermore, theN363S variant in the glucocorticoid receptor isassociated with overweight and CHD,21 similar toour findings in the SO-Obese rats. Thus, increasedexpression of hypertension-regulated vascular fac-tor and glucocorticoid receptor may explain theclose association between the obesity, hyperten-sion, and CHD reported in humans.

Hypothalamic neurotransmitters, neuropep-tides, and hormones such as neuropeptide Y,serotonin, dopamine, proopiomelanocortin, andmelanocortins control FI, hunger, appetite, andsatiety.22-25 Hence, one would expect an alterationin the expression of these genes in DIO, RYGB, andPF rats, especially when the expression profiles areperformed in the hypothalamic ARC. In the presentstudy, the emphasis has been on the genes that areexpressed in both the ARC and the SAF tissues. Thismay explain why altered expression of hypotha-lamic neuropeptides and hormones was not ob-

served since these genes are not expressed toa significant degree in the fat tissue.

Results of the present study showed that theincreased expression of the products of the variousgenes observed in the ARC and SAF tissues in theSO-Obese reverted to near normal after RYGB.Thus, beneficial effects of weight loss achieved bysurgery in obese rats can be related to decreasedexpression of the products of the various genesinvolved in carbohydrate, lipid, and protein meta-bolic pathways; immune response, and inflamma-tion. Hence, RYGB-induced weight loss is not justa mechanical event (due to physical restriction ofthe volume of the gastric pouch) but is associatedwith changes in the expression of genes at themolecular level, influencing metabolic changes. Itis also evident that food restriction, as occurs in PF,is not as successful in influencing changes in genesthat contribute to sustained weight loss, emphasiz-ing that gastric bypass and its physiologic sequelaeis more effective in reducing weight than dietrestriction in the morbidly obese.

REFERENCES

1. O’Brien PE, Dixon JB. The extent of the problem of obesity.Am J Surgery 2002;184:S4-6.

2. Mokdad AH, Bowman BA, Ford ES, Vinicor F, Marks JS,Koplan JP. The continuing epidemics of obesity anddiabetes in the United States. JAMA 2001;286:1195-2000.

3. Bouchard C, Perusse L, LeBlanc C, Tremblay A, TheriaultG. Inheritance of the amount and distribution of humanbody fat. Int J Obes 1988;12:205-15.

4. Allison DB, Kaprio J, Korkeila M, Koskenvuo M, Neale MC,Hayakaua K. The heritability of body mass index among aninternational sample of monozygotic twins reared apart. IntJ Obes 1996;20:501-6.

5. Brolin RE. Bariatric surgery and long-term control ofmorbid obesity. JAMA 2002;288:2793-6.

6. Xu Y, Ohinata K, Meguid M, Marx W, Tada T, Chen C.Gastric bypass model in the obese rat to study metabolicmechanisms of weight loss. J Surg Res 2002;107:56-63.

7. Meguid MM, Ramos EJB, Suzuki S, Xu Y, George ZM, DasUN, et al. A surgical rat model of human Roux-en-Y gastricbypass. J Gastrointest Surg. In press.

8. Middleton F, Ramos EJB, Xu Y, Diab H, Zhao X, Das UN,et al. Application of genomic technologies: DNA Micro-arrays and Metabolic Profiling of Obesity in the Hypothal-amus and in Subcutaneous Fat. Nutrition 2004;20:14-25.

9. Meguid MM, Kawashima Y, Campos ACL, Gelling P, Hill TW,Chen T-Y, et al. Automated computerized rat eater meter:Description and application. Physiol Behav 1990;48:759-63.

10. Palkovits M. Isolated removal of hypothalamic or otherbrain nuclei of the rat. Brain Res 1973;59:449-50.

11. Hartsook EW, Hershberger TV. A simplified method forsampling small animal carcasses for analyses. Proc Soc ExpBio Med 1963;113:973-7.

12. Meguid MM, Laviano A, Rossi-Fanelli F. Food intake equalsmeal size x meal number. Appetite 1998;31:404.

13. Hines LM, Stampfer MJ, Ma J, Gaziano JM, Ridker PM,Hankison SE, et al. Genetic variation in alcohol dehydro-

SurgeryAugust 2004

252 Xu et al

genase and the beneficial effect of moderate alcoholconsumption on myocardial infarction. New Engl J Med2001;344:549-55.

14. Toth MJ, Poehlman ET, Mathews DE, Techernof A, MacCossMJ. Effects of estradiol and progesterone on body compo-sition, protein synthesis, and lipoprotien lipase in rats. Am JPhysiol Endocrinol Metab 2001;280:E496-501.

15. Clark AM, Thornley B, Tomlinson L, Galletley C, NormanRJ. Weight loss in obese infertile women results inimprovement in reproductive outcome for all forms offertility treatment. Hum Reprod 1998;13:1502-5.

16. Glass AR, Burman KD, Dahms WT, Boehm TM. Endocrinefunction in human obesity. Metabolism 1981;30:89-104.

17. Bujalska IJ, Kumar S, Stewart PM. Does ventral obesityreflect ‘‘Cushing’s disease of the omentum’’? Lancet 1997;349:1210-3.

18. de Jong EC, Viera PL, Kalinski P, Kapsenberg ML. Cortico-steroids inhibit the production of inflammatory mediatorsin immature monocyte-derived DC and induce the de-velopment of tolerogenic DCs. J Leukocyte Biol 1999;66:201-4.

19. Ramos EJB, Xu Y, Romanova I, Middleton F, Chen C, QuinnR, et al. Is obesity an inflammatory disease? Surgery 2003;134:329-35.

20. Das UN. Is obesity an inflammatory condition? Nutrition2001;17:953-6.

21. Lin RCY, Wang XL, Morris BJ. Association of coronaryartery disease with glucocorticoid receptor N363S variant.Hypertension 2003;41:404-7.

22. Laviano A, Meguid MM. Serotonin and obesity. Curr MedChem-Central Nervous System Agents 2003;3:89-100.

23. Meguid MM, Fetissov SO, Varma M, Sato T, Zhang L,Laviano A. Hypothalamic dopamine and serotonin in theregulation of food intake. Nutrition 2000;16:843-57.

24. Cowley MA, Pronchuk N, Fan W, Dinulescu DM, ColmersWF, Cone RD. Integration of NPY, AGRP, and melanocortinsignals in the hypothalamic paraventricular nucleus: evi-dence of a cellular basis for the adipostat. Neuron 1999;24:155-63.

25. Schwartz MW, Woods SC, Porte D Jr, Seeley RJ, Baskin DG.Central nervous system control of food intake. Nature 2000;404:661-71.