Introduction

Corticotropin-releasing factor (CRF) and its homologue, urocortin (Ucn), are expressed in hypothalamic regions implicated in energy homeostasis including the paraventricular nucleus. ¹ Both peptides decrease food intake when administered centrally in rodents,^2,3 whereas the CRF antagonist, hCRF (9–41), blocks acute stress-induced hypophagia.⁴ Furthermore, the effects of leptin on food intake are attenuated by coinfusion of hCRF (9–41), suggesting that CRF-like peptides are required to mediate the central effects of leptin.⁵ Levels of endogenous CRF and CRF-binding protein, which sequesters CRF and urocortin, are altered by changes in nutritional status brought about by short-term food restriction/repletion.¹ The effects of CRF and related peptides are mediated by two receptors, CRF-R1 and CRF-R2. CRF-R1 binds CRF and its known mammalian paralogue, urocortin, with comparable affinities, whereas CRF-R2 binds Ucn with a much higher affinity than CRF.⁶ Although gross abnormalities in feeding behaviour or body weight have not been observed in CRF-R1,^7,8 CRF-R2^9,10 or CRF-R1/CRF-R2 knockout mice,⁹ suggestion for a specific role of CRF-R2 in energy balance was provided by the recent identification of two CRF-R2-specific ligands, Ucn III and Ucn II,^11,12,13 which attenuate feeding in mice. Indeed, the expression of CRF-R2 mRNA in the ventromedial hypothalamus is positively correlated with plasma leptin concentrations, thus supporting the notion that leptin reduces food intake, at least in part, via a CRF-R2-mediated pathway.¹⁴ Finally, genome-wide linkage analyses have identified linkage between body mass index (BMI) and a region on chromosome 7p15 encompassing the CRF-R2 gene, suggesting that genetic variation at this locus may be associated with human obesity.¹⁵

The aim of the present study was to establish whether genetic variation in the human CRF-R1 and CRF-R2 genes contributed to severe early-onset human obesity. The coding regions of the CRF-R1 and CRF-R2 genes were examined in 51 unrelated subjects with severe early-onset obesity. Having detected common single-nucleotide polymorphisms in these genes, we performed association studies between genetic variants of the CRF-R1 and CRF-R2 genes and obesity-related phenotypes in subjects from a UK population-based cohort.

Experimental subjects and procedures

Subjects with severe obesity of early onset (<10 y of age) have been recruited to the UK Genetics of Obesity Study (GOOS) after multiregional ethics committee approval. In all, 51 subjects from this cohort, all UK Caucasians, were randomly selected (mean age 5.82.1 y) for screening of CRF-R1 and CRF-R2. BMI (weight in kg/height in m²) standard deviation scores (SDSs) were calculated using the UK 1990 growth reference data.¹⁶ The mean BMI SDS of subjects studied was 4.20.8.

The Ely Study is a prospective population-based cohort study of the aetiology and pathogenesis of Type II diabetes and related metabolic disorders in 1122 subjects, all Caucasians from the UK and aged between 40 and 65 y.¹⁷ Anthropometric and biochemical data were available in this study both at phase 1 and phase 2, 4.5 y later. Although there were no specific exclusion criteria, none of the subjects were taking medications known to affect body weight.

DNA isolation and mutational screening

Genomic DNA was isolated from whole blood using a QIAamp blood kit (Qiagen, London). The human CRF-R1 gene consists of 14 exons and has been localised to chromosome 17. We screened the entire coding region of CRF-R1, including all intron/exon boundaries with the exception of exon 1 for which the genomic sequence flanking the 33 bp of this exon was not available in the public database. The human CRF-R2 gene consists of 12 exons and has been localised to chromosome 7. The entire coding sequence of CRF-R2, including all intron/exon boundaries was screened. Oligonucleotide primers (sequences available upon request) flanking exons were designed and used in a PCR reaction performed with AmpliTaq Gold (Perkin-Elmer, UK) and carried out under standard conditions with 35 cycles of 95°C for 30 s, 55°C for 30 s and 72°C for 1 min. PCR products were subsequently analysed by denaturing high-performance liquid chromatography (DHPLC) (Transgenomic Ltd, UK). Bidirectional sequencing was performed on all PCR products that showed an aberrant DHPLC pattern using the described PCR primers at an annealing temperature of 52°C. Owing to the large size of CRF-R2 exon 2, PCR was performed as described above; however, the amplicons were analysed by direct nucleotide sequencing rather than DHPLC. Sequencing was carried out using Big Dye terminator chemistry (Perkin-Elmer, Foster City, CA, USA) on an ABI 377 automated DNA sequencer. Sequences were assembled and examined using Sequencher software (Gene Codes, Ann Arbor, MI, USA).

Genotyping

Primers with mutagenic base pairs were designed so that in the presence of either polymorphism a restriction site is introduced. A fragment of CRF-R1 exon 10, containing C861T was PCR amplified with primers (hCRF-R1-restriction fragment length polymorphism(RFLP)-Fwd, 5'-CCTTTCCTCTGTGGCCTTCTAGGC-3' hCRF-R1-RFLP-Rev, 5'-GGTTCTTACCAGCAGGACCAGG-3') designed to introduce an HhaI restriction site when the rare variant is present. The genomic region encompassing the CRF-R2 exon 10 Ser349Ser polymorphism was PCR amplified with primers (hCRF-R2-RFLP-Fwd, 5'-CATCACCTACATGCTCTTCTTCG-3'; hCRF-R2-RFLP-Rev, 5'-CCATCTGGTCACAGGCCCCACCG-3') designed to introduce an RsaI restriction site in the presence of the rare allele. PCR products were digested with 2 U of the appropriate restriction enzyme and electrophoresed on a 3.0% (wt/vol) agarose gel (Gibco BRL, Paisley, UK). Of note, the discrepancy observed between the number of subjects genotyped for the G1047A and C861T reflects either failure of the genotyping assay to provide a clear result or an insufficient DNA volume to type all the subjects for both sequence variants.

Statistical analyses

Association analyses between the genotype and the phenotype were performed using the MIXED procedure in the Statistical Analysis System (SAS for Windows 8. 2); this allows the incorporation of unbalanced repeated data, which considerably improves the power of the study. The allowance for unbalanced repeated data allows the inclusion of individuals who contributed one as well as two measures of the phenotypic variables. Means were adjusted for age and sex, and P-values are presented comparing the C/C (wild-type) genotype group and C861T carriers uncorrected for multiple testing.

Results

CRF-R1

All nonsynonymous variants found in both CRF-R1 and CRF-2R in the 51 severely obese subjects are listed in Table 1. A missense mutation (Val161Met) in CRF-R1 was identified in a single 6-y-old girl with hyperphagia and severe, early-onset obesity (BMI SDS=3.3), but not detected in 144 alleles from nonobese control subjects. However, the mutation did not segregate with obesity in extended family studies (data not shown).

Table 1 - Frequency of missense variants identified in the hCRFR-1 and hCRFR-2 genes in obese and lean control subjects.

Full table

A common silent polymorphism (C861T, Cys287Cys) was found in heterozygous form in 34 (10.3%) subjects. To determine whether C861T was associated with any obesity-associated traits, 512 individuals from the Isle of Ely Study were genotyped by a forced HhaI RFLP. The allele frequency of the C861T variant was 0.224 and the observed number of individuals with the three genotypes did not significantly deviate from Hardy–Weinberg predictions (P=0.6227). The relationships between C861T- and obesity-associated traits were examined using mixed model analysis incorporating phenotypic data from two independent measurements 4.5 y apart, adjusted for age and sex. Carriers of the rare allele had a significantly higher BMI (P=0.0083) than wild-type subjects (Table 2). No significant associations were found between the C861T genotype and waist–hip ratio, systolic or diastolic blood pressure or levels of fasting plasma insulin, fasting triglycerides or cholesterol.

Table 2 - Association studies of the CRF-R1 sequence variant, C861T (Cys287Cys), in subjects from the Isle of Ely Study.

Full table

CRF-R2

Three different nonsynonymous nucleotide substitutions in the CRF-R2 gene were identified in severely obese probands (Table 1). A mutation (Val411Met) was found in heterozygous form in a 5-y-old girl with hyperphagia and severe, early-onset obesity (BMI SDS=3.5), but not in 140 alleles from control subjects. The probands' mother and maternal grandfather, but no other family members carried the Val411Met mutation. Both the mother and grandfather were overweight (BMI SDS=1.6, BMI=28 kg/m² and BMI SDS=1.6, BMI=29 kg/m², respectively), and the mother gave a history of being obese since childhood (Figure 1). However, additional family members were equally or more obese and did not carry the mutation. Therefore, while the cosegregation studies do not exclude a pathological role for this variant, they do not provide strong support. Three subjects (5.9%) were found to be heterozygous for Glu220Asp and two subjects (3.9%) heterozygous for Val240Ile. Both these variants were also found in 1–3% of lean control subjects.

Figure 1.

Pedigree of subject with Val411Met mutation in CRF-R2. Males represented by squares, females by circles, filled in symbols represent subjects with a history of early-onset obesity (<10 y). N=normal allele, M=mutant allele, BMI SDS=BMI standard deviation score, proband indicated by an arrow.

Full figure and legend (19K)

One common silent polymorphism (G1047A, Ser349Ser) was found in six subjects from the obese cohort (11.8%). To determine whether the G1047A polymorphism was associated with obesity-associated traits, 604 individuals from the Isle of Ely Study were genotyped by a forced RsaI RFLP. The genotype frequency did not deviate significantly from Hardy–Weinberg expectations (P=0.1505). The determination of the relationships between genotype and obesity-related phenotypes revealed no statistically significant associations.

Discussion

To our knowledge, the current study represents the first genetic analysis of the CRF-R1 and CRF-R2 genes in relation to human obesity. One missense mutation (Val161Met) and one common polymorphism were identified (C861T; Cys287Cys) in CRF-R1. Although the missense mutation was not found in 72 nonobese control subjects, it did not cosegregate with obesity when family studies were undertaken. CRF-R1 is an alternatively spliced form of the receptor in which 29 amino acids of exon 6 are inserted into the first intracellular loop. Val161Met is found within this insertion. Previous investigations of the structure–function relationship of CRF-R1 have identified the N-terminal domain, the second extracellular domain and junction of the third extracellular domain and fifth transmembrane domain, as regions involved in receptor ligand binding and/or receptor activation.^18,19,20 Although CRF-R1 has been shown to bind CRF with two-fold lower affinity than CRF-R1, the ability of CRF to stimulate adenylate cyclase was reduced 100-fold with CRF-R1, suggesting that this loop participates in the interaction with G proteins.²¹ The physiological role of this splice variant is unclear and the functional significance of Val161Met remains to be determined.

Carriers of the rare allele at C861T had a significantly higher BMI than noncarriers (P<0.01), with no independent effects being observed on waist–hip ratio, fasting plasma insulin or blood pressure. This silent variant would not be expected to affect receptor structure but may influence receptor expression either directly or indirectly, through linkage disequilibrium with an as yet unidentified variant in the regulatory elements of the gene. Variation in CRF-R1 receptor expression levels could influence body weight through a number of mechanisms. Firstly, as CRF has central anorexigenic effects in the hypothalamus,² reduced receptor expression levels might directly influence the control of appetite and food intake. CRF-R1 receptors are also highly expressed on the corticotrophs of the anterior pituitary,²² and variation in expression levels may influence the production of ACTH and therefore plasma cortisol levels. As glucocorticoids have effects on adiposity through central and peripheral mechanisms,²³ this could also be a mechanism linking CRF-R1 expression and obesity. We do not have information on diurnal cortisol dynamics or daily cortisol production rates in the Isle of Ely study, and are therefore unable to address this question directly. However, it is worth noting that the C861T variant was not associated with any alteration in waist–hip ratio, a parameter that might be expected to be altered if the effect of the C861T allele was through increased ambient levels of glucocorticoids. It should be noted that the controlling elements of genes may have tissue-specific functions and it is plausible that a promoter or enhancer variant might affect the expression of the CRF-R1 gene in the hypothalamus, but not in the anterior pituitary corticotroph.²⁴

Mutational screening of CRF-R2 in subjects with severe obesity identified three heterozygous missense mutations (Glu220Asp, Val240Ile and Val411Met) and a common variant G1047A. Since two of the variants (Glu220Asp and Val240Ile) were identified in lean controls, they were not further investigated. Val411Met was identified in one severely obese proband and not in lean controls. Although Val411Met did not clearly cosegregate with obesity when family studies were undertaken, a contributory role of the mutation to the development of obesity in that family cannot be excluded. No significant associations between the common G1047A polymorphism- and obesity-related phenotypes were found in a UK population-based cohort.

In conclusion, our results suggest that mutations within the coding sequences of the CRF-R1 or CRF-R2 genes are unlikely to make a major contribution to severe early-onset obesity. The regulatory regions of these genes have not been clearly defined, thus pathogenic mutations in these regions cannot be excluded. The robustly significant association found between a CRF-R1 polymorphism and BMI in our UK Caucasian population should provide a stimulus for further examination of this gene in relation to BMI and obesity in other populations.

References

1. Cone RD. The corticotropin-releasing hormone system and feeding behavior—a complex web begins to unravel. Endocrinology 2000; 141: 2713–2714. | Article | PubMed | ISI | ChemPort |
2. Arase K, York DA, Shimizu H, Shargill N, Bray GA. Effects of corticotropin-releasing factor on food intake and brown adipose tissue thermogenesis in rats. Am J Physiol 1988; 255: E255–9. | PubMed | ISI | ChemPort |
3. Spina M, Merlo-Pich E, Chan RK, Basso AM, Rivier J, Vale W, Koob GF. Appetite-suppressing effects of urocortin, a CRF-related neuropeptide. Science 1996; 273: 1561–1564. | Article | PubMed | ISI | ChemPort |
4. Krahn DD, Gosnell BA, Grace M, Levine AS. CRF antagonist partially reverses CRF- and stress-induced effects on feeding. Brain Res Bull 1986; 17: 285–289. | Article | PubMed |
5. Gardner JD, Rothwell NJ, Luheshi GN. Leptin affects food intake via CRF-receptor-mediated pathways. Nat Neurosci 1998; 1: 103. | Article | PubMed | ISI | ChemPort |
6. Vaughan J, Donaldson C, Bittencourt J, Perrin MH, Lewis K, Sutton S, Chan R, Turnbull AV, Lovejoy D, Rivier C, Rivier J, Sawchenko PE, Vale W. Urocortin, a mammalian neuropeptide related to fish urotensin I and to corticotropin-releasing factor. Nature 1995; 378: 287–292. | Article | PubMed | ISI | ChemPort |
7. Smith GW, Aubry JM, Dellu F, Contarino A, Bilezikjian LM, Gold LH, Chen R, Marchuk Y, Hauser C, Bentley CA, Sawchenko PE, Koob GF, Vale W, Lee KF. Corticotropin releasing factor receptor 1-deficient mice display decreased anxiety, impaired stress response, and aberrant neuroendocrine development. Neuron 1998; 20: 1093–1102. | Article | PubMed | ISI | ChemPort |
8. Timpl P, Spanagel R, Sillaber I, Kresse A, Reul JM, Stalla GK, Blanquet V, Steckler T, Holsboer F, Wurst W. Impaired stress response and reduced anxiety in mice lacking a functional corticotropin-releasing hormone receptor 1. Nat Genet 1998; 19: 162–166. | Article | PubMed | ISI | ChemPort |
9. Bale TL, Contarino A, Smith GW, Chan R, Gold LH, Sawchenko PE, Koob GF, Vale WW, Lee KF. Mice deficient for corticotropin-releasing hormone receptor-2 display anxiety-like behaviour and are hypersensitive to stress. Nat Genet 2000; 24: 410–414. | Article | PubMed | ISI | ChemPort |
10. Coste SC, Kesterson RA, Heldwein KA, Stevens SL, Heard AD, Hollis JH, Murray SE, Hill JK, Pantely GA, Hohimer AR, Hatton DC, Phillips TJ, Finn DA, Low MJ, Rittenberg MB, Stenzel P, Stenzel-Poore MP. Abnormal adaptations to stress and impaired cardiovascular function in mice lacking corticotropin-releasing hormone receptor-2. Nat Genet 2000; 24: 403–409. | Article | PubMed | ISI | ChemPort |
11. Reyes TM, Lewis K, Perrin MH, Kunitake KS, Vaughan J, Arias CA, Hogenesch JB, Gulyas J, Rivier J, Vale WW, Sawchenko PE. Urocortin II: a member of the corticotropin-releasing factor (CRF) neuropeptide family that is selectively bound by type 2 CRF receptors. Proc Natl Acad Sci USA 2001; 98: 2843–2848. | Article | PubMed | ChemPort |
12. Lewis K, Li C, Perrin MH, Blount A, Kunitake K, Donaldson C, Vaughan J, Reyes TM, Gulyas J, Fischer W, Bilezikjian L, Rivier J, Sawchenko PE, Vale WW. Identification of urocortin III, an additional member of the corticotropin-releasing factor (CRF) family with high affinity for the CRF2 receptor. Proc Natl Acad Sci USA 2001; 98: 7570–7575. | Article | PubMed | ChemPort |
13. Hsu SY, Hsueh AJ. Human stresscopin and stresscopin-related peptide are selective ligands for the type 2 corticotropin-releasing hormone receptor. Nat Med 2001; 7: 605–611. | Article | PubMed | ISI | ChemPort |
14. Nishiyama M, Makino S, Asaba K, Hashimoto K. Leptin effects on the expression of type-2 CRH receptor mRNA in the ventromedial hypothalamus in the rat. J Neuroendocrinol 1999; 11: 307–314. | Article | PubMed |
15. Wu X, Cooper RS, Borecki I, Hanis C, Bray M, Lewis CE, Zhu X, Kan D, Luke A, Curb D. A combined analysis of genomewide linkage scans for body mass index from the National Heart, Lung, and Blood Institute Family Blood Pressure Program. Am J Hum Genet 2002; 70: 1247–1256. | Article | PubMed | ISI | ChemPort |
16. Cole TJ. Conditional reference charts to assess weight gain in British infants. Arch Dis Child 1995; 73: 8–16. | PubMed | ISI | ChemPort |
17. Wareham NJ, Byrne CD, Williams R, Day NE, Hales CN. Fasting proinsulin concentrations predict the development of type 2 diabetes. Diabetes Care 1999; 22: 262–270. | Article | PubMed | ISI | ChemPort |
18. Wille S, Sydow S, Palchaudhuri MR, Spiess J, Dautzenberg FM. Identification of amino acids in the N-terminal domain of corticotropin-releasing factor receptor 1 that are important determinants of high-affinity ligand binding. J Neurochem 1999; 72: 388–395. | Article | PubMed | ISI | ChemPort |
19. Liaw CW, Grigoriadis DE, Lorang MT, De Souza EB, Maki RA. Localization of agonist- and antagonist-binding domains of human corticotropin-releasing factor receptors. Mol Endocrinol 1997; 11: 2048–2053. | Article | PubMed | ChemPort |
20. Liaw CW, Grigoriadis DE, Lovenberg TW, De Souza EB, Maki RA. Localization of ligand-binding domains of human corticotropin-releasing factor receptor: a chimeric receptor approach. Mol Endocrinol 1997; 11: 980–985. | PubMed |
21. Xiong Y, Xie LY, Abou-Samra AB. Signaling properties of mouse and human corticotropin-releasing factor (CRF) receptors: decreased coupling efficiency of human type II CRF receptor. Endocrinology 1995; 136: 1828–1834. | PubMed |
22. Potter E, Sutton S, Donaldson C, Chen R, Perrin M, Lewis K, Sawchenko PE, Vale W. Distribution of corticotropin-releasing factor receptor mRNA expression in the rat brain and pituitary. Proc Natl Acad Sci USA 1994; 91: 8777–8781. | Article | PubMed | ChemPort |
23. Cavagnini F, Croci M, Putignano P, Petroni ML, Invitti C. Glucocorticoids and neuroendocrine function. Int J Obes Relat Metab Disord 2000; 24 (Suppl 2): S77–S79. | Article | PubMed |
24. Rubinstein M, Mortrud M, Liu B, Low MJ. Rat and mouse proopiomelanocortin gene sequences target tissue-specific expression to the pituitary gland but not to the hypothalamus of transgenic mice. Neuroendocrinology 1993; 58: 373–380. | PubMed |

Acknowledgements

This work was supported by the Wellcome Trust (NJW, ISF, SOR) and the MRC (SOR, NJW). We are grateful to all the physicians who have referred patients to the GOOS study, in particular, to Dr B Houston for help with clinical studies.