CTLA-4 gene polymorphisms are associated with obesity in Turner Syndrome.

Turner syndrome (TS) is characterized by a set of clinical conditions, including autoimmune/inflammatory diseases and infectious conditions, that can compromise a patient's quality of life. Here we assessed polymorphisms in CTLA-4 +49A/G (rs231775), PTPN22 +1858G/A (rs2476601), and MBL2 -550 (H/L) (rs11003125), -221(X/Y) (rs7096206) and exon 1 (A/O) in women from northeastern Brazil to determine whether polymorphisms within these key immune response genes confer differential susceptibility to clinical conditions in TS. A case-control genetic association study was performed, including 86 female TS patients and 179 healthy women. An association was observed for the A/G genotype of CTLA-4 +49A/G in TS patients (p=0.043, odds ratio [OR]=0.54). In addition, an association between the CTLA-4 G/G genotype and obesity was detected in TS patients (p=0.02, OR=6.04). Regarding, the -550(H/L) polymorphism in the MBL2 promoter, the frequency of the H/L genotype was significantly higher in the TS group than healthy controls (p=0.01, OR=1.96). The H/H genotype indicated a protective effect in TS patients (p=0.01, OR=0.23). No differences were observed in the distribution of -221(X/Y), MBL2 exon 1 variants, and PTPN22 +1858G/A in any assessed groups. CTLA-4 variants are potentially involved in obesity in this cohort of TS patients from northeastern Brazil.

Introduction

Turner syndrome (TS) is one of the most common chromosomal abnormalities in humans and is characterized by the presence of one X chromosome and total or partial loss of the second sex chromosome. TS is estimated to affect 1 in every 2500 live female births (Stochholm ). Individuals with TS exhibit a set of phenotypic features, including short stature and gonadal dysgenesis. Other clinical conditions, such as osteoporosis, dyslipidemia, obesity and congenital malformations are also observed (Ostberg ; Carvalho ; Bispo ; Ríos Orbañanos ).

Some studies have reported increased levels of autoantibodies (anti-thyroid peroxidase and anti-glutamic-aciddecarboxylase) in TS patients and an increased risk of developing a range of autoimmune diseases, such as Hashimoto’s thyroiditis, type I diabetes mellitus, celiac disease, Crohn’s disease, ulcerative colitis, juvenile rheumatoid arthritis, Addison’s disease, autoimmune hepatitis, psoriasis, vitiligo, and alopecia (Mortensen ; Bianco ; Jørgensen ; Bakalov ). In addition, ovarian insufficiency and absence of a second normal X chromosome are linked to an increased risk of autoimmune disorders in these patients. However, the underlying pathophysiological mechanisms related to the immune unbalance remain to be fully elucidated (Mortensen ; Bakalov ).

A wide variety of autoimmune/inflammatory diseases and infectious conditions have been associated with a set of genes related to immune regulation, including the tyrosine-protein phosphatase non-receptor type 22 gene (PTPN22), cytotoxic T-lymphocyte-associated protein 4 gene (CTLA4), and mannose-binding lectin (MBL2) (Bottini ; Bevilacqua Filho ; Katkam ). Polymorphisms within these three genes have been analyzed due to their importance in immune balance and homeostasis within the body. Even though a body of evidence indicates immune deregulation processes in TS, only PTPN22 rs2476601 has been assessed in Brazilian TS patients from São Paulo (Southeast region) (Bianco ). Furthermore, studies evaluating the role of MBL2 and CTLA-4 polymorphisms in TS and its association with clinical features are lacking.

To understand the role of these key genes in immune misbalance and its consequences, we assessed whether PTPN22, CTLA-4, and MBL2 polymorphisms confer susceptibility to autoimmune conditions or other inflammation-related features in TS patients from Northeast Brazil.

Material and Methods

Patients and controls

This study included 86 patients with cytogenetic diagnosis of TS, who attended at Medical Genetics Service of Institute of Integral Medicine Professor Fernando Figueira and at Pediatric Endocrinology Service of Clinical Hospital of Federal University of Pernambuco. It was proposed as a pilot study. At time of TS diagnosis, patients mean age was 11.48 years old (SD ± 7.52 years old), ranging from 0.1 to 33 years. Clinical data shown in the Table 1 were obtained from medical records of each patient. The control group included 179 healthy women from the same geographical region. Their mean age was 34.62 years old (SD ± 13.4 years old), ranging from 8 to 72 years. Exclusion criteria for the control group included an individuals’ history of autoimmune and inflammatory chronic disease, also in close relatives such as parents. All individuals (or their legal responsible) included in this research signed an informed consent term, which followed the Declaration of Helsinki guidelines and presented the approval number from local Ethics Committee (Record: CEP/IMIP N° 802/06; CEP/CCS/UFPE N° 493/11).

Table 1

Clinical characterization of all Turner syndrome patients enrolled in our study.

Clinical characteristics	N
Short stature	69
Skeletal abnormalities	72 a
Osteopenia/osteoporosis	4
Sexual infantilism	100 b
Primary amenorrhea	31
Obesity	9
Dyslipidemia	5
Autoimmune thyroid disease	11
Alopecia	2
Inflammatory diseases	9
Infectious diseases	6
Neurological disease	23
Cardiovascular disease	17
Renal malformations	10
Skin diseases	17
Edema	35
Mammary hypertelorism	22
Hearing impairment	7
Ear malformations	21
Muscle hypotonia	3
Arched palate	7
Nails malformations	33
Eyes anatomic alterations	13
Visual impairment	2
Short and webbed neck	35
Low posterior hairline	21
Skin redundancy in the neck	18

Main clinical conditions are cubitus valgus, pectus scavatum and genu valgus;

Absent mammary development, absent axillary and pubic hair, hypoplastic uterus and absent ovaries;

Karyotyping

Clinical diagnosis of 86 TS patients was confirmed by chromosome analysis in peripheral blood leucocyte. Karyotypes found were as follows: 45,X (n=47, 54.65%); 45,X/46,X,i(Xq) (n=15, 17.44%); 46,X,i(Xq) (n=4, 4.65%); 45,X/46,XY (n=4, 4.65%); 45,X/46,X,r(X) (n=3, 3.49); other chromosomal constitutions summed 13 individuals (15.12%).

DNA extraction and genotyping

Genomic DNA was extracted from whole blood using IllustraTM Blood GenomicPrep Mini Spin Kit (GE Healthcare) according to manufacturer’s instructions. SNPs selection was based on minimum allele frequency (MAF) of 10% and/or SNP consequence/function upon gene action. A total of five SNPs were selected distributed as follows PTPN22 + 1858G/A (rs2476601) at codon 620, CTLA-4 +49A/G (rs231775) within codon 17 in the first exon and MBL2 promoter region -550(H/L) (rs11003125), -221(X/Y) (rs7096206). Genotyping was performed using TaqMan SNP genotyping assays and Taqman Universal Master Mix (Applied Biosystems®, CA) according to manufacturer instructions. SNP assessment within MBL2 exon 1 (A/O) was performed using SYBR Green (Qiagen, Hilden, Germany) as previously described (Hladnik ). All Sybr Green endpoint PCRs were performed, including all three possible genotypes as positive controls, in a Rotor-Gene 6000 TM apparatus (Corbett Research Mortlake, Sydney, Australia). Ten randomly chosen MBL2 genotyped samples were sequenced in order to double-check the Melting Temperature assay (MTA) results. We found 100% concordance between sequenced samples and the MTA results.

Statistical analysis

Statistical analyses were carried out using SNPStats available at http://bioinfo.iconcologia.net/SNPstats_web and R software (https://www.r-project.org/). Hardy–Weinberg equilibrium was tested for each polymorphism by comparing observed with expected frequencies using chi-square (χ2) tests. Differences in allele and genotype frequencies from each studied polymorphism in patients and controls were assessed using χ2 or Fisher’s exact test. Odds Ratio (OR) and 95% Confidence Intervals (CI) were also calculated and a p-value < 0.05 was considered statistically significant. Combined alleles for the PTPN22 and CTLA-4 genes, haplotypes for the MBL2 gene and a possible association of these combined alleles and haplotype with clinical conditions in TS patients were also assessed. Combined genotypes for the MBL2 gene were assessed by Arlequin version 3.1 software (Excoffier ). Fisher’s exact test was performed to evaluate difference between combined genotypes in case-control and associations with clinical data in TS group. We compared genotype and allele distribution for all SNPs assessed in this study in TS patients and control group. Posteriorly, we evaluated a possible association of all polymorphic variants with differential presence of autoimmune diseases such as autoimmune thyroid disease and alopecia and other clinical features as follows: obesity, dyslipidemia, inflammatory and infectious conditions in TS patients. Post-hoc (goodness of fit χ2 tests) power analysis was performed with the G* Power software (version 3.1.9.2, available at http://www.gpower.hhu.de/), with α error probability of 0.05

Results

PTPN22 and CTLA-4 gene polymorphisms

The allele and genotype distributions of PTPN22 rs2476601 (G > A) and CTLA-4 rs231775 (A > G) among TS patients and healthy controls are summarized in Table 2. Conformity to Hardy-Weinberg equilibrium (HWE; p > 0.05) was observed in both SNP distributions, and no significant differences were found in the allele and genotype frequencies of these variants in both groups (Table 2).

Table 2

Genotype and allele distribution of PTPN22 and CTLA-4 gene polymorphisms in TS and controls group.

Polymorphism	Patients N (%)	Controls N (%)	Odds ratio (95% CI)	p-value
PTPN22 rs2476601
Allele	172	358
G	170 (99%)	345 (96%)	1.00
A	2 (1%)	13 (4%)	0.31 (0.03-1.40)	0.16
Genotype	86	179
GG	84 (97.7%)	166 (92.7%)	1.00
GA	2 (2.3%)	13 (7.3%)	0.30 (0.03-1.39)	0.15
AA	0 (0.0%)	0 (0.0%)	0 (0 inf)	1.00
CTLA-4 rs231775
Allele	172	340
A	114 (66%)	203 (60%)	1.00
G	58 (34%)	137 (40%)	0.75 (0.50-1.12)	0.17
Genotype	86	170
AA	41 (47.7%)	59 (34.7%)	1.00
AG	32 (37.2%)	85 (50%)	0.54 (0.29 - 0.99)	0.043*
GG	13 (15.1%)	26 (15.3%)	0.72 (0.30-1.66)	0.44

Significant difference using Fisher’s exact test.

We did not detect the homozygous A/A genotype of PTPN22 rs2476601 (G > A) in any of the studied groups. A lower frequency of the A allele and G/A genotype was observed in both TS patients and controls (Table 2). No significant association was identified between the assessed SNPs and the presence of any clinical conditions in women with TS (p > 0.05; Fisher’s exact test) (Table S1). The power was 80.7% (α-error = 5% confidence) to detect a medium effect size (w=0.3) for PTPN22 genotypes in both patients and controls.

Concerning CTLA-4 rs231775, significantly different distributions of genotype frequencies were observed in TS patients compared to the control group. An association was detected for the A/G genotype (p=0.043, OR=0.54), indicating a differential distribution for this SNP in TS patients. The power was 99% (α-error = 5% confidence) to detect a medium effect size (w=0.3) for CTLA4 genotypes comparing patients and controls. Furthermore, when assessing the clinical features of TS and the allele and genotype distribution, we detected an association between the CTLA-4 allele (recessive model: A/A—A/G vs. G/G) and obesity in TS patients (p=0.02, 95% CI 1.37-26.75, OR=6.04) (Table 3, Table S2).

Table 3

Genotype distribution of CTLA-4 gene polymorphisms in TS group.

Model	Polymorphism	Obesity N (%)	Non-obesity N (%)	Odds ratio (95% CI)	p-value
Recessive	CTLA-4
	rs231775
	Genotype
	AA- AG	5 (55.6%)	68 (88.3%)	1.00
	GG	4 (44.4%)	9 (11.7%)	6.04 (1.37 - 26.75)	0.023

Combined alleles for PTPN22 and CTLA-4 genes

The combined alleles of PTPN22 and CTLA-4 and frequencies for both groups are given in Table 5. No significant differences were observed. Furthermore, no significant association was established between the combined alleles and clinical status of TS patients (Table S3).

Table 5

Analyses of combined alleles of PTPN22 and CTLA4 genes, haplotypes and genotypes of -550 and -221 promoter region and exon 1 of MBL2 gene in TS patients and controls.

PTPN22	CTLA4	Frequencies in TS	Frequencies in controls	p-value	OR (95% C.I.)
G	A	0.6512	0.5697	Reference	1.00
G	G	0.3372	0.3939	0.16	0.76 (0.52-1.11)
A	A	0.0116	0.027	0.25	0.39 (0.08 - 1.93)
A	G	0	0.0093	1	0.00 (-Inf - Inf)
Global haplotype association p-value: 0.14
MBL producers and haplotype		Frequencies in TS	Frequencies in controls	p-value	OR (95% C.I.)
High MBL producers
LYA		0.3827	0.4272	Reference	1.00
HYA		0.2586	0.2694	0.89	1.04 (0.62 1.73)
Low MBL producers
LXA		0.1453	0.1467	0.69	1.12 (0.64 1.94)
Deficient MBL producers
LYO		0.1464	0.0861	0.13	1.75 (0.85 3.60)
HYO		0.0669	0.0706	0.84	1.10 (0.45 2.67)
Global haplotype association p-value: 0.63
MBL producers and genotypes		TS group (n = 65)	Control (n = 150)	p-value	OR (95%C.I.)
High MBL producers
HYA/HYA
HYA/LYA		27 (0.42)	76 (0.51)	Reference	Reference
LYA/LYA
Low MBL producers
LXA/LXA
LYA/LXA
HYA/LXA
HYA/HYO		32 (0.49)	63 (0.42)	0.278	1.42(0.74-2.76)
HYA/LYO
LYA/LYO
Deficient MBL producers
HYO/LXA
HYO/LYO
LYO/LXA		6 (0.09)	11 (0.07)	0.558	1.53(0.42-5.06)
LYO/LYO
HYO/HYO

MBL2 gene polymorphisms

The MBL2 genotype and allele distributions in TS patients and controls are given in Table 4. All allelic and genotypic frequencies of MBL2 polymorphisms were in HWE in the TS and control groups, except for rs11003125 (-550 H/L) in the TS group (p < 0.05). The H/L genotype frequency was significantly higher in TS patients than controls (overdominant model: H/L vs. L/L—H/H: p=0.01, OR=1.96, 95% CI 1.11-3.50). Furthermore, the H/H genotype indicated a protective effect in TS patients compared to healthy controls (recessive model: H/H vs. L/L—H/L: p=0.01, OR=0.23, 95% CI 0.04-0.83). Again, this result cannot be fully explained due to an absence of HWE in this specific group. The power was 99.9% (5% confidence) to detect a medium effect size (w=0.3) for -550 MBL2 genotypes in both groups.

Table 4

Genotype and allele distribution of MBL2 gene polymorphisms in TS and controls group.

Inheritance Model	Polymorphism	Patients N (%)	Controls N (%)	Odds ratio (95% CI)	p-value
	MBL2 rs11003125
	Allele	172	300
	L	116 (67.0%)	198 (66.0%)	1.00
	H	56 (33.0%)	102 (34.0%)	0.93 (0.61- 1.42)	0.76
Recessive	Genotype	86	150
	L/L—H/L	83 (96.5%)	130 (86.7%)	1.00
	H/H	3 (3.5%)	20 (13.3%)	0.23 (0.04-0.83)	0.01 *
Overdominant	Genotype	86	150
	L/L—H/H	36 (41.9%)	88 (58.7%)	1.00
	H/L	50 (58.1%)	62 (41.3%)	1.96 (1.11-3.50)	0.01 *
	MBL2 rs7096206
	Allele	172	300
	Y	147 (85%)	256 (85%)	1.00
	X	25 (15%)	44 (15%)	0.98(0.55-1.73)	1.0
Recessive	Genotype	86	150
	Y/Y—X/Y	84 (97.7%)	144 (96.0%)	1.00
	X/X	2 (2.3%)	6 (4.0%)	0.57(0.05-3.29)	0.7
	MBL2 Exon 1
	Allele	126	300
	A	98 (78%)	253 (84%)	1.00
	O	28 (22%)	47 (16%)	1.53(0.87-2.66)	0.12
Dominant	Genotype	63	150
	A/A	39 (61.9%)	108 (72%)	1.00
	A/O - O/O	24 (38.1%)	42 (28%)	1.57(0.80-3.06)	0.14

Significant difference using Fisher’s exact test.

Regarding MBL2 rs7096206 (-221 X/Y), no differences in allele or genotype frequencies were observed between TS patients and controls. The power was 30.5% (5% confidence) to detect a medium effect size (w=0.3) for -221 MBL2 genotypes when comparing patients and controls. Furthermore, no significant differences were observed in the genotype and allele frequencies of exon 1 variants between TS patients and controls. The power was 91.2% (5% confidence) to detect a medium effect size (w=0.3) for exon 1 MBL2 genotypes when comparing TS patients and controls. In patients with TS, no SNP in the MBL2 gene or promotor region was associated with clinical characteristics (Table S4).

Haplotypes and combined genotypes of MBL2 gene -550, -221, and exon 1 variants

The frequencies of haplotypes and combined genotypes originating from linkage disequilibrium between the MBL2 -550 and -221 promoter region and exon 1 polymorphisms are given in Table 5. Haplotypes were combined in different groups; haplotypes associated with high expression of MBL (LYA, HYA), low production of MBL (LXA), and deficient expression of MBL (LYO, HYO). No significant differences were found in haplotype frequencies between the analyzed groups. Furthermore, no significant association was established among haplotypes and the clinical data of TS patients (Table S5).

Genotypes were classified as high (HYA/HYA, HYA/LYA, and LYA/LYA); low (LXA/LXA, LYA/LXA, HYA/LXA, HYA/HYO, HYA/LYO, and LYA/LYO); and deficient producers of MBL (HYO/HYO, HYO/LXA, HYO/LYO, LYO/LXA, and LYO/LYO). Significant differences were not found between the evaluated groups. Moreover, no significant difference was observed among combined genotypes and the clinical data of TS patients (Table S6).

Discussion

To date, only a few assays have been performed involving genes linked to innate and adaptive immunity in patients with TS, even though the immune response seems impaired in these patients (Bianco , 2012).

We included PTPN22 in our analyses due to its importance in the host immune system; this gene encodes LYP, an important negative regulator of T cell activation (Bottini ). Our results indicate an absence of an association between the selected PTPN22 SNP and autoimmunity, inflammatory, and infectious conditions in TS women.

Our results differ from Bianco , who found an association between the same SNP (PTPN22 rs2476601) and the development of autoimmune diseases in TS in another cohort from Southeast Brazil. The frequency of the A allele in women with TS (1.1%) and healthy controls (3.6%), as well as the heterozygote genotype in both groups (2.3% and 7.3%, respectively), was lower in our study than that of Bianco . In their study, the frequency of the A allele and heterozygote genotype were 18.3% and 28.2% in the TS group, and 9.2% and 16.1% in controls, respectively. A disease-associated homozygote genotype was present in 4.2% of patients and 1.1% of controls.

This difference between studies could be due to variation in allele frequencies of some disease-associated SNPs in different ethnic groups (Mori ), because the Brazilian population exhibits variety in allele distribution. Geographic distributions in Brazil exhibit ethnic disparities, mainly due to the genetic burden of heterogeneous colonization sources (Pena ; Coelho ).

We also evaluated PTPN22 rs2476601 (G > A) and other clinical features in TS patients, such as obesity, as well as dyslipidemia, given its role in modulating inflammatory conditions. As obesity is a disease characterized by chronic mild inflammation, the concentration of acute phase proteins and cytokines associated with inflammation are higher in obese individuals compared to normal weight individuals (Trayhurn, 2007). However, our analysis did not indicate an association between these clinical manifestations in the TS group, which is similar to a previous study by Salinas-Santander , in which PTPN22 +1858G/A was not associated with differential susceptibility to overweight and the development of obesity in adolescents.

The CTLA-4 +49A/G polymorphism at exon 1 is involved in the negative regulation of T cells. Our assay showed an association between the rs231775 G/G genotype and obesity in the TS group (p=0.02).

The presence of the CTLA-4 +49A/G variant has been associated with different diseases, and G/G individuals may possess CTLA-4 protein with a weak suppression function compared to individuals with the A/A genotype (Chistiakov and Turakulov, 2003). Therefore, increased T-cell activation due to this reduced inhibitory signal to T cells would be associated with the pathogenesis of several autoimmune/inflammatory diseases, including obesity, as observed with TS patients in the present study. The adipocytes of obese individuals with TS express fewer anti-inflammatory elements and high amounts of pro-inflammatory factors, leading to a misplaced response in immune cells (Ostberg ; Bakalov ).

MBL2 polymorphisms and their influence on serum protein levels have been evaluated extensively and found to be associated with recurrent and severe infections (Sumiya ; Summerfield ), such as tuberculosis (da Cruz ) and autoimmune diseases, including celiac disease (Boniotto ), systemic lupus erythematosus (Lee ), Sjögren’s syndrome (Tsutsumi ), and autoimmune thyroid disease (AITD) (Bevilacqua Filho ).

Polymorphisms at exon 1 and the promotor region of MBL2 were evaluated in our study due to their role in innate immunity and modulation of the inflammatory response. A relationship between clinical parameters and this genetic variant has not been assessed previously in TS, making our study the first to be performed in women with TS.

No association was found between the MBL2 -221 (X/Y allele) polymorphism and different clinical conditions in TS patients. This MBL2 promoter variation has a significant down-regulating effect on the serum MBL concentration, leading to ineffective clearance of apoptotic cells and the spread of self-antigens, permitting an immune response toward autoimmunity and tissue damage (Bouwman ; Araujo ). The Y variant is associated with high serum MBL expression (Madsen ) and has been involved in susceptibility to the development of different diseases (Lee ; da Cruz ).

On the other hand, our results revealed significant differences regarding the -550 (H/L allele) promoter polymorphism between TS patients and controls (p < 0.05), revealing a different distribution in both groups.

Notably, in our study population, the -550 H/L MBL2 variant was not in HWE, though the power analysis excluded type I and II statistical error. Therefore, we suggest that the TS condition acts upon the allelic distribution, causing deviation from HWE. Although differences in the MBL2 polymorphism distribution have been detected, no significant association was found regarding the MBL2 -550 (H/L allele) variant and clinical data of the TS group.

In summary, even though our study is a pilot study, due to limited number of TS patients and controls included, our results indicate a differential distribution for some polymorphisms within key inflammation-regulating genes in TS patients. The understanding of how key immune genes and its variants are related in TS might help in future therapy strategies. These findings may open up a new potential line of research to improve life’s quality in these individuals.