Variations in Peptide YY and Y2 Receptor Genes Are Associated With Severe Obesity in Pima Indian Men

  1. Lijun Ma,
  2. P. Antonio Tataranni,
  3. Robert L. Hanson,
  4. Aniello M. Infante,
  5. Sayuko Kobes,
  6. Clifton Bogardus and
  7. Leslie J. Baier
  1. From the Diabetes Molecular Genetics Section, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Department of Health and Human Services, Phoenix, Arizona
  1. Address correspondence and reprint requests to Leslie J. Baier PhD, Diabetes Molecular Genetics Section, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, 4212 North 16th St., Phoenix, AZ 85016, E-mail: lbaier{at}


Peptide YY (PYY) and Y2 receptor (Y2R) may be important in the central regulation of body weight and food intake. To determine whether genetic variation in PYY and/or Y2R may contribute to morbid obesity in humans, these genes were sequenced in 83 extremely obese Pima Indians (BMI ≥50 kg/m2). Sequencing of PYY identified three single nucleotide polymorphsms (SNPs) in the untranslated region. Sequencing of the Y2R coding region identified one missense (Ala172Thr) substitution and two silent substitutions. Eight additional SNPs in the 5′ untranslated region of Y2R were identified from public databases. These SNPs were genotyped in 489 full-heritage adult Pimas (362 severely obese and 127 nondiabetic, nonobese subjects), who are not first-degree relatives, for association analysis. The PYY variants were not associated with obesity, whereas four variants from two haplotype blocks in Y2R were marginally associated (P = 0.054–0.067) with obesity. However, if the analysis was restricted to men (n = 167, 100 obese and 67 lean), the PYY variants and two SNPs in Y2R that were in complete linkage disequilibrium were significantly associated with severe obesity (P = 0.001 and P = 0.002, respectively). Our data suggest that the PYY-Y2R pathway may influence body weight through a sex-specific mechanism, but this finding requires confirmation in other populations.

Peptide YY (PYY), a member of the neuropeptide Y (NPY) family, is a 36–amino acid peptide secreted from the L-cells of the gastrointestinal tract in response to food intake (1,2). It has been suggested that PYY plays a role in the development of obesity in rodents, but this remains controversial. Peripheral administration of PYY was first reported to decrease food intake in rodents in 1993 (3). PYY3-36, the active form of PYY, also markedly inhibits food intake in rodents (2), and it has been shown to cross the blood-brain barrier and act on the arcuate nucleus of the hyperthalamus (4,5). PYY3-36 has a high affinity for the Y2 receptor (Y2R) (2,6), and animal studies have shown that PYY3-36 inhibits hypothalamic NPY-expressing neurons by binding to Y2R (7). Mice lacking the Y2R gene have increased body weight, food intake, and fat deposition (8). In contrast to these studies, Tschöp et al. (9) recently reported that PYY3-36 does not decrease food intake in rodents.

It has also been suggested that the PYY/Y2R pathway plays a role in human obesity. Peripheral administration of PYY3-36 was reported to inhibit food intake in humans (10), and plasma PYY concentrations were decreased in obese compared with lean subjects (10). In addition, two synonymous amino acid substitutions have been identified in Y2R that are associated with obesity in British Caucasian men (11).

The Pima Indians of Arizona have a high prevalence of severe obesity, and BMI in this population is highly heritable (12,13). As part of our genetic studies to identify variation in genes that affect body weight in this Native-American population, PYY and Y2R were analyzed as obesity candidate genes. These genes were sequenced to detect variation in a group of 83 extremely obese Pima Indians, and variants were analyzed for association with obesity in a case/control study.

The human PYY gene (NCBI accession no. NM_004160) on chromosome 17q21.1 is composed of four exons and three introns that span 1.2 kb. Three variants were detected by sequencing the PYY gene in 83 extremely obese Pima Indian subjects (BMI from 50.5 to 79.6 kg/m2) (Table 1). Two novel variants were positioned in the 5′ upstream flanking region, a C/T at −1746 bp (single nucleotide polymorphism [SNP]1) and a G/C at −1653 bp (SNP2), relative to the transcriptional start site. These two variants were in complete genotypic concordance among the 83 subjects and had a major allele frequency of 0.55. The third variant was a previously identified C/T in intron 3 (SNP3; rs162430), which was in tight linkage disequilibrium with SNP1/2 (D′ = 0.9999, Δ2 = 0.9999). SNP1 (as the representative SNP for SNPs 1 and 2; SNP1/2) and SNP3 were genotyped in the case/control group of 489 full heritage, non–first-degree–related adult Pima Indians (Table 2) for association analysis. Genotypes of SNP1 and SNP3 were in Hardy-Weinberg equilibrium when analyzed in the case, control, and combined groups, respectively (Table 1). The allele frequency of these two SNPs did not differ between the case and control group (Table 3). However, since several obesity genes in animal models have sex-specific effects, the analysis was repeated in men and women separately. In the gender-restricted analysis, SNP1/2 and SNP3 were associated with severe obesity in men (n = 167, 100 obese and 67 lean; P = 0.001) using a logistic regression full model (Table 3).

The human Y2R gene (NCBI accession no. U36269) on chromosome 4q31 consists of two exons that span 4.5 kb. Three variants were detected by sequencing the coding region of Y2R in the same 83 extremely obese subjects described above (Table 1). A G/A polymorphism, predicting an alanine-to-threonine substitution, was identified at codon 172 (Ala172Thr; SNP9). Two additional variants, both C/T predicting a silent (isoleucine) substitution, were identified at codons 195 and 312 (SNP10 and SNP11, respectively). These three coding SNPs and eight additional variants (SNPs 1–8, across 34 kb) (Table 1), located in the 5′ untranslated region that was identified in the dbSNP ( and/or Celera databases (, were genotyped in the 489 case/control subjects. SNP10 and SNP11 (SNP10/11) were found to be in perfect genotypic concordance among the 489 case/control samples and therefore provided identical information. Genotypes of these 11 SNPs were in Hardy-Weinberg equilibrium in the case, control, and combined groups, except for SNP10/11, which deviated modestly in the control group (P = 0.04) (Table 1). SNPs 6, 7, and 10/11 in Y2R were marginally associated (P = 0.054–0.067) with severe obesity (Table 3). However, when the analysis was restricted by gender, SNP10/11 was significantly associated with severe obesity in men (P = 0.002) (Table 3).

To better understand the underlying genetic models for these associations, logistic regression analyses were also applied to the entire case/control group, and the gender- restricted groups in different genetic models (Table 3). The PYY variants (SNP1/2 and SNP3) appear to function under a dominant model, where men with the major allele (both homozygotes and heterozygotes) were more prevalent in the obese group, and men homozygous for the minor allele tended to be in the lean group (P = 0.001–0.002, odds ratio [OR] >3.5). Variants in Y2R (SNPs 3, 6, and 7 in 5′ region, SNPs 10 and 11 in exon 2) were associated with severe obesity under a recessive model (Table 3). The strongest association in men was with SNP10/11, where men homozygous for the C allele tended to be in the lean group, and men homozygous or heterozygous for the T allele tended to be more prevalent in the obese group (P = 0.001, 0.33 < OR < 0.4).

A statistically significant interaction of genotype with sex was observed for several of these SNPs in both genes (PYY SNP1/2 and SNP3, P = 0.001 in dominant model; Y2R SNP10/11, P = 0.018 in recessive model) (online appendix Table 4 [available at]).

The genotypic information from 11 variants (SNPs 1–11) in Y2R was used to determine the haplotype structure of this gene (Fig. 1). Three haplotype blocks were identified. Y2R SNPs 2–4 represent the first haplotype block, SNPs 5–7 represent the second block, and SNPs 8–11 represent the third haplotype block. SNP9, the only coding missense substitution identified in Y2R, which was physically located in the third block, had a minor allele frequency that was too low for confident assignment to any haplotype block. In the association analysis, at least one variant in each of the three haplotype blocks of Y2R was modestly associated with severe obesity, but variants in the third haplotype block (SNP10/11) show a greater difference in men.

The finding of associations between variants in PYY and Y2R and severe obesity was unexpected in men only and may represent false-positive associations. However, a recent study by Hung et al. (11) on the PYY-Y2R pathway in British Caucasians also described an association between Y2R SNPs and human obesity with two SNPs identical to these significant SNPs detected in our current study (SNP10/11, rs1047214 [nt585C/T], and rs2880415 [nt936C/T]), and this association was only present in men. They found that men homozygous for the C allele of SNP rs1047214 had lower BMI than men homozygous or heterozygous for the T allele in British Caucasians (P = 0.017) (11), and rs2880415 was almost in complete genotypic concordance with rs1047214 (11). These findings are consistent with our results in Pima Indians. Therefore, it is possible that PYY or Y2R, or a different gene in the pathway that is activated by their interaction, is hormonally controlled. For example, NPY (a family member of PYY) expression is increased in male sheep in a testosterone-dependent manner (14), and testosterone also contributes to age-related changes of NPY expression (15). Since PYY belongs to the same family as NPY, it is possible that PYY also has a sex hormone–regulated element or other indirect hormone-controlled factor regulatory sites. Based on the sequence analysis of the PYY putative promoter by MatInspector (release professional 7.2.2,, PYY SNP1 (PYY-1746) is located in the anchor site of the ribonucleoprotein-associated zinc finger protein (MOK-2) binding region, and PYY SNP2 (PYY-1653) is located in the c-Myb/v-Myb core binding region. MOK-2 is preferentially expressed in testis tissue (16), providing a plausible mechanism for sex-specific expression of PYY. c-Myb is a transcription factor expressed in the hematopoietic system and the gastrointestinal tract that regulates the exquisite balance among cell division, differentiation, and survival (17). Androgen can influence the c-Myb expression level in mice (18); therefore, c-Myb binding could also potentially regulate PYY expression in a sex-dependent manner.

A third explanation for our association to be observed only in men is that more significant, female-specific obesity genes mask the contribution of PYY and Y2R to obesity in women. For example, female-specific effects on BMI and/or obesity-related traits have been reported for variants in several genes, including resistin (19), UCP3 (20), and FOXC2 (21,22).

As the present association was detected in a case/control study of extremely discordant individuals, the significance of these findings on the population basis is not yet clear. Further genetic analysis of PYY and Y2R in other populations should help clarify the biological importance of these genes in the development of severe human obesity.


The subjects are part of an ongoing longitudinal study of the etiology of obesity and type 2 diabetes among the Gila River Indian Community in central Arizona. Eighty-three extremely obese Pima subjects (37 men and 46 women, BMI 60.0 ± 6.1 kg/m2 [range 50.5–79.6 kg/m2], age 35.0 ± 9.5 years [means ± SD]), who were not first-degree relatives, were selected for sequence analysis of the PYY and Y2R genes. Unique variants in PYY and Y2R were genotyped in 489 full-heritage adult Pima Indians (who were not first-degree relatives) for association analyses, where 362 subjects were severely obese, as defined by a maximum BMI >45 kg/m2, and 127 subjects were nonobese control subjects as defined by BMI <30 kg/m2 at >35 years of age and not having diabetes. Diabetic status was determined by the criteria of the World Health Organization (23). The 83 obese subjects who were used for sequence analysis were part of the 362 obese subjects who were genotyped for association studies.

Sequence variant identification and genotyping.

Genomic DNA for sequencing and genotyping was obtained from peripheral lymphocytes. Sequence analysis of PYY included the entire four exons and three introns (1.2 kb), and 2 kb of the 5′ untranslated region. Sequence analysis of Y2R included only the coding region. The 5′ untranslated region of Y2R was not directly sequenced due to its large size (∼34 kb,, but all SNPs in this region that were available in the NCBI public database ( were genotyped. Sequencing of DNA from 83 subjects was done using Big Dye terminator (Applied Biosystems, Foster City, CA) on an automated DNA capillary sequencer (model 3730; Applied Biosystems). Variants identified by direct sequencing or by searching databases were genotyped in DNA from 489 subjects using the TaqMan Allelic Discrimination Assay (Applied BioSystems). The TaqMan genotyping reaction was amplified on a GeneAmp PCR system 9700 (95°C for 10 min, followed by 40 cycles of 95°C for 30 s and 60°C for 1 min 30 s), and fluorescence was detected on an ABI Prism 7700 (Applied BioSystems). Sequence information for all oligonucleotide primers and probes is available upon request.

Statistical analysis.

Statistical analyses were performed using SAS Institute (Cary, NC) software. Age and BMI are expressed as means ± SD. Differences of genotype frequencies were assessed with logistic regression between the lean and obese groups. For logistic regression modeling in the additive model, homozygotes for the major allele (aa) and heterozygotes (ab) and homozygotes for the minor allele (bb) were coded to a continuous numeric variable for genotype (0, 1, and 2). A dominant model was defined as contrasting genotypic groups aa + ab versus bb, and the recessive model was defined as contrasting genotypic groups aa versus ab + bb. Product terms were used to assess the interaction between genotype and sex. To examine pairwise linkage disequilibrium, haplotype frequencies were estimated with the EH program (Xie and Ott,, and these haplotype frequencies were used to calculate D-prime and Δ2. HaploView was used to place SNPs in haplotype blocks. The confidence interval method (24) was applied to define blocks ( P values <0.05 were considered statistically significance.

FIG 1. A:

D′ and Δ2 matrix for 11 SNPs identified in Y2R gene. D′ is defined as a measure of allelic association and Δ2 as a measure of concordance ( B: Haplotype structure and diversity of Y2R gene. Haplotype blocks and their frequencies were estimated using an accelerated EM algorithm. A, major allele; C, minor allele. ▾, tag SNPs that designate a parsimonious haplotype for each block. Thick line, frequency ≥10%; thin line, frequency between 1 and 10%.


Markers and map locations


Characteristics of Pima subjects in severely obese and nonobese case/control groups


Association of PYY and Y2R gene SNPs with severe obesity under logistic regression analyses


The authors thank Dr. Michal Prochazka for his critical reading of this manuscript.


  • Additional information for this article can be found in an online appendix at

    • Accepted February 7, 2005.
    • Received December 10, 2004.


| Table of Contents