Polymorphisms of the androgen receptor gene associate with fatness , uterus and ovary measurements in the pig

Quantitative trait loci, QTL, for fatness and carcass traits in pigs have been recently mapped to the X chromosome approximately at the position where the androgen receptor gene (AR) is localized. The AR acts as a nuclear transcription factor regulating expression of a number of androgenic response genes during various stages of sexual development. The present study aimed to analyze association of the AR genotype on traits related to fatness and phenotype of primary female sexual organs, ovary and uterus. Animals of a F2 resource population based on Duroc and Berlin Miniature pig were genotyped at a multi-allelic microsatellite marker (CCTTT)n situated in the 5 ́ untranslated region and a bi-allelic CAG-insertion/deletion polymorphism (CAGINDEL) within exon 1 of the AR. Association analysis showed that the AR genotype both at the CAG-INDEL and microsatellite affect almost all fatness traits measured. The D allele inherited from Duroc was associated with the decrease of fat thickness. The AR genotypes also affect uterus and ovary measurements. The pigs homozygous for D allele were likely to have a lighter uterus, shorter uterus horns, oviducts and smaller ovaries than pigs homozygous for the M allele. Our results confirm the previously reported QTL for fatness traits and provide evidence for a QTL affecting dimensions of uterus and ovary on the X chromosome. The AR is a positional functional candidate gene for both trait complexes.


Introduction
The androgen receptor is classically activated by the binding of androgens.Two most important androgens are testosterone and 5α-dihydrotestosterone.After liganddependent activation the hormone-receptor complexes bind to hormone response elements (HRE) of target genes and interact with other basal transcription factors, coactivators and co-repressors to up or down regulate transcription of the target genes.A number of genes participating in cell metabolism, activity and response have been shown to be regulated by androgens (NANTERMET et al., 2004;NELSON et al., 2002).The AR has been shown to be expressed in diverse tissues like prostate, testis, ovary, uterus, mammary gland, anterior pituitary, thyroid, adrenal cortex, liver, muscle as well as bone during stages suggesting its importance for normal development and function of those tissues (PELLETIER, 2000;SLOMCZYNSKA et al., 2001;TRIPEPI et al., 2000).Mutations of the AR associate with various human disorders like male infertility, spinal and bulbar muscular atrophy (SBMA), breast, and prostate cancers (RIS-STALPERS et al., 1990;McPHAUL et al., 1993;BRINKMANN, 2001).The AR effects on development of reproductive organs/tissues like follicle and testis have been reported in the pig (SLOMCZYNSKA et al., 2001;TRIPEPI et al., 2000).Quantitative trait loci, QTL, for various traits like back fat, abdominal fat, intramuscular fat content, and carcass traits were mapped to the porcine X chromosome in close vicinity to AR (KNOTT et al., 1998;ROHRER and KEELE, 1998;HARLIZIUS et al., 2000;ROHRER, 2000;MALEK et al., 2001a,b).Hence, the AR is a functional and positional candidate gene for production and reproduction traits in pigs.Previously, polymorphisms have been reported in coding and promoter regions of the porcine AR gene (TRAKOOLJUL et al., 2004).Moreover, an association between the AR genotype and its mRNA transcript level has been observed.In the present study, the AR genotype at the CAG-insertion/deletion (INDEL) in the exon 1 and the CCTTT(n) microsatellite in 5´-untranslated region were analyzed for associations on production and reproduction traits like fat thickness, uterus and ovary dimensions.

Animals and phenotype records
Analyses were done in a three-generation porcine F2 resource population (DUMI resource population) that was established by reciprocal crossbreeding of Duroc and Berlin Miniature Pig breeds (HARDGE et al., 1999).F2 piglets were weaned at about 6 weeks of age and kept in flat decks until day 100 and subsequently in single pens until slaughter at 200 days of age.The following traits related to fatness were recorded: average back fat, fat depth at shoulder, fat depth at 10 th rib, loin fat depth, and fat depth at side.Dimensions of reproductive organs, vagina, cervix, uterus and ovary were recorded at slaughter.All sows were observed for estrous cycle (1) or non estrous (0).

Genotype analysis
Genomic DNA was isolated from tail cuts by standard procedure of proteinase K digestion and phenol-chloroform extraction.Genotyping of the CCTTT(n) microsatellite at 5´-UTR of the pig AR was performed as previously described (TRAKOOLJUL et al., 2000).The CAG-INDEL genotyping was based on PCR amplification length polymorphism using primers: forward 5´-agctgctccaccgatcttaaag-3´ and reward 5´-cttacacaattccttggcgctg-3´.The PCR was prepared in a total volume of 10 µl containing 1 x PCR buffer, 0.25 mM dNTPs, 0.25 µM of each primer, 0.25 U of Taq polymerase, and 25 ng of DNA.Touch-down step PCR was performed in Thermocycler (MJ research) programmed at 94 °C for 3 min, 10 x [94 °C 30 sec, 63 -58 °C (-0.5 °C/ cycle) 30 sec, 72 °C 1 min], 24 x [94 °C 30 sec, 58 °C 30 sec, 72 °C 1 min] with additional extension of 72 °C for 5 min.Amplified fragments were separated on 8 % denatured polyacrylamide gel, run in 1 x TBE buffer at 50 watts for 90 min using vertical electrophoresis system.Mendelian inconsistency of genotype in the pedigree was eliminated using PedCheck program (O´ CONNELL and WEEKS 1998).
Association analysis Statistic analyses were examined with SAS for Windows version 8.2, using the procedure PROC MIXED.Association effects of the AR (CAG-INDEL) and microsatellite (CCTTTn) genotypes on fatness traits were analyzed with the statistical model (1) including fixed effects of parity and genotype and random effects of family (a combination of F1 boar x F1 sow), separately for the sexes.Individual slaughter weight was in the model 1 as covariate.For uterus measurements, gilt status (0 = noncyclic or 1 = cyclic) and genotype were included as fixed effects where as family as random effects in the model (2).Results are presented as least square means (LSM) and standard errors of least square means (SE).All multiple comparisons were performed with adjustments according to Tukey-Kramer.Statistical models used for the analyses are as follows: Model (1) for fatness traits: y ijkl = µ + parity i + genotype j + family k + b(Slaughter weight) ijkl +e ijkl where: parity (fixed effect) i = 1-5 genotype (fixed effect) j = 1-3 (CAG-INDEL) or 1-5 (microsatellite) for females or j = 1-2 (CAG-INDEL) or 1-4 (microsatellite) for males family (random effect) k = 21 slaughter weight as covariate Model ( 2) for uterus and ovary measurements:

Allele segregation
Genotyping of the DUMI animals for the CAG-INDEL showed two segregating alleles with allele size of 213 and 216 bp, namely "M" and "D" allele, respectively.This CAG repeat codes for polyglutamine, which is corresponding to amino acid position 179 to 188 of the pig AR.The D allele predicts for 10 glutamine residues and is a common allele found in Duroc and other breeds.The M allele codes for 9 glutamine residues and is an alternative deleted allele detected in Berlin Miniature pig.At the CCTTT(n) microsatellite, four segregating alleles with 188, 203, 204 and 228 bp were observed (Figure 1).Sequencing and genotyping of the DUMI founders revealed four haplotypic combinations: M-204, M-228, D-188, and D-203.Since the Duroc and Berlin Miniature pig founders were fixed for different alleles at the CAG-INDEL, all F1 sows were heterozygous and the genotypic ratio of their F2 followed X-linked inheritance pattern.Association analysis -fatness traits Association analyses for fatness traits are shown in Table 1.The results showed that the AR genotype at the CAG-INDEL was associated with all fatness traits measured.However, significant differences of LSM were found in males but not in females.The D allele from Duroc was associated with low fatness while the M allele from the Berlin miniature pig with high fatness.Average back fat thickness of male pigs carrying the M allele was 3.22 mm higher than of pigs with D allele.No significant effect was found between the microsatellite genotype and fatness traits in females.However, pigs with the 188/188 and 188/203 genotypes, i.e. with both alleles from Duroc, tend to deposit less fat than other genotypes.Pigs with genotype with both alleles from Berlin Miniature pig were dropped from comparisons because of small number of animals available.Allele effects on fatness were more pronounced in males than in females.Male pigs with genotype 204/0 were significantly different from pigs with genotype 203/0 or 188/0 in all fat traits except fat depth at side as shown in Table 1.a,b,c,d,e indicate significant differences, ns = not significant differences, P-value is adjusted according to Tukey-Kramer Association analysis -uterus and ovary measurements Association analyses showed both the CAG-INDEL and microsatellite genotypes affect uterus measurements.The homozygous DD pigs had significantly higher weight of vagina and cervix than homozygous MM pigs (p < 0.001).However, the pigs homozygous for MM genotype were likely to have a heavier uterus and longer uterus horns than homozygous DD pigs.Homozygous MM pigs had longer oviducts and larger ovaries (left and right) than homozygous DD pigs (Table 2).Similarly, associations between the microsatellite and uterus and ovary measurements were found significant.The pigs carrying genotype 204/228 with both alleles inherited from Berlin Miniature pig were likely to have a heavier uterus body, longer uterus horns and larger ovaries than pigs with the genotype 188/188 with both alleles from Duroc.Moreover, pigs with the genotype 203/228 and 203/204 showed larger uterine horns compared to pigs with the genotype 188/188 (Table 3).1.77 ± 0. 14 a,c,d,e 1.93 ± 0.10 b,c 1.47 ± 0. 09 d,f 1.44 ± 0.09 e,f 0.02 -0.007 a,b,c,d,e,f,g,h indicate significant differences, ns = not significant differences, P-value is adjusted according to Tukey-Kramer

Discussion
The present study shows strong associations between the AR genotype and fat traits within the DUMI resource population.The results are consistent with the X-linked QTL for fatness traits previously reported in the pig (KNOTT et al., 1998;HARLIZIUS et al., 2000;ROHRER, 2000).As expected, the "M" allele from Berlin Miniature pig associates with high fat deposition.Like observed in other resource populations the alleles originated from Asian breeds (the Berlin Miniature Pig was breed from German Landrace, Saddleback and Vietnamese Potbelly Pig) increase back fat except those on chromosome 7.Moreover, dose dependent responses of the M allele on fatness were likely to be present in females but not statistically significant.A small number of homozygous MM may be a cause.Since the AR is located on the X chromosome, on the other side, deviations from a random X-inactivation in heterozygous females may be occurred and this can hide the allele effects (GARTLER and GOLDMAN, 2001;BROWN, 2001).There are evidences suggesting that the AR has direct and/or indirect effects on adipose tissue and hence fat depositions.It has been found that plasma testosterone level associated with back fat thickness and weight gain in line selected boars (ANDRESEN, 1976).The AR is highly expressed in preadipocytes and mature adipocytes (DIEUDONNE et al., 1998).It has been shown that the AR can directly determine the mesenchymal precursor cell into adipogenic lineage (in vitro) by acting at sites of myogenic and adipogenic differentiation pathways (SINGH et al., 2003).De PERGOLA (2000) reported that testosterone inhibits differentiation of adipocyte precursor cells, suppresses lipid uptake and lipoprotein lipase activity in adipocytes, and up-regulates the number of β-adrenergic receptors.Androgens may also affect fat deposition by inhibiting expression of peroxisomal proliferators-activated receptor γ 2 (PPARγ2) and CCAAT/enhancer binding protein α (C/EBPα) which are transcription factors regulating adipogenic differentiation.Moreover, the effects of the AR on fat traits may be mediated through leptin pathway influencing feed intake and energy balance since testosterone can partly regulates leptin expression and secretion in adipocytes (WAUTERS et al., 2000).
QTL for reproduction traits, including traits related to litter size, reproductive organs measurements, puberty and gestation development have been assigned to several autosomes, however there is only one QTL for FSH plasma concentration on SSCX (ROHRER et al., 2001).Here we provide evidence for a QTL for uterus and ovary dimension on the X chromosome.The present results showed that the AR genotypes affect uterus and ovary measurements, which are important components of reproductive traits in pigs (FORD et al., 2002).Not surprising that alleles from the Berlin Miniature pig associate with long horns of uterus and large ovary weight since genetics of Asian-related breeds is generally expected to have positive influence on prolificacy.Size of vagina and cervix are not affected by AR genotype or show inverse effect of Duroc or Miniature pig originated alleles, respectively.Variation in cervix and vagina size parallels variation in body size to a higher extent than size of uterus and ovary that are of higher functional importance with regard to fertility.Similar results have been observed in a study comparing uterine dimension of Large White and Meishan gilts at the same physiological stages (BAZER et al., 1988).Effects of breeds influencing length of uterine horns were reported in the study of pure breed Large White, Yorkshire and crossbred (ISLER et al., 2002).The AR could have diverse effects depending on stages of developments since it functions as nuclear transcription factors determining up-or down-regulation of genes in response to sex hormones.It has been shown that the female reproductive tract is under negative control of androgens (DREWS et al., 2002).On the other hand, the androgen receptor together with other nuclear receptors cooperatively plays important roles in functional phases of uterus.For example, estrogen administration positively regulates and restores the AR mRNA expression in mouse uterus and vagina after ovarectomy (PELLETIER et al., 2004).The AR is detected in nuclei of most cell types of pig uterus including epithelial, stroma, myometrial, and endothelial cells (CARDENAS and POPE, 2003).It presents mainly in glandular and luminal epithelia and remains unchanged throughout the estrous cycle and early pregnancy.Besides, the AR expresses abundantly in granulosa cells of growing follicles and promotes follicle development (WEIL et al., 1998).These indicate that the AR interacts with other genes at physiological levels to determine sex organ development and hence potentially influences various fertility related traits.
The present results show strong associations between the AR genotypes and fatness traits, length of uterus horns and weight of ovary in pigs.Taking into account linkage disequilibrium that is expected in an intercross of divergent breeds the present results confirm the previously reported QTL for fatness traits on SSCX and for the first time provide evidence for a QTL affecting dimensions of uterus and ovary on the X chromosome in the pig.The study promotes AR as a positional functional candidate gene for traits related to obesity and reproduction.Our previous study showed a number of polymorphisms within the gene and SNPs in the promoter.These polymorphisms are likely to affect gene function and expression and consequently to influence biochemical pathways and physiology of associated traits.Further studiesin vivo and in vitro -will elucidate and prove whether AR represents the QTL and the polymorphisms detected represent the genomic variations underlying the effects on phenotypic variation.

Table 2
Effects of the CAG-INDEL genotype of AR on uterus and ovary measures (LSM ± SE) (Effekte der CAG-INDEL Genotypen am AR auf Uterus-und Ovarmaße) a,b indicate significant differences, ns = not significant differences, P-value is adjusted according to Tukey-Kramer

Table 3
Effects of microsatellite genotype of AR on uterus and ovary measures (LSM ± SE) (Effekte der Mikrosatelliten Genotypen am AR auf Uterus-und Ovarmaße)