Human Immunology 74 (2013) 348–352

Contents lists available at SciVerse ScienceDirect

www.ashi-hla.org

journal homepage: www.elsevier .com/ locate/humimm

Functional genetic variants of CTLA-4 and risk of tobacco-related oral carcinomain high-risk North Indian population

Vijaya Bharti a, Bidhu Kalyan Mohanti b, Satya Narayan Das a,⇑a Department of Biotechnology, All India Institute of Medical Sciences, Ansari Nagar, New Delhi 110029, Indiab Department of Radiation Oncology, BRA-IRCH, All India Institute of Medical Sciences, Ansari Nagar, New Delhi 110029, India

a r t i c l e i n f o

Article history:Received 19 September 2012Accepted 4 December 2012Available online 12 December 2012

0198-8859/$36.00 - see front matter � 2012 Americahttp://dx.doi.org/10.1016/j.humimm.2012.12.008

⇑ Corresponding author. Fax: +91 11 26589286.E-mail addresses: bharti_vijaya@yahoo.co.in (

rediffmail.com (B.K. Mohanti), satyandas@gmail.cosndas@aiims.ac.in (S.N. Das).

a b s t r a c t

Single nucleotide polymorphisms (SNPs) of the CTLA-4 gene have been implicated in susceptibility to dif-ferent cancer in different ethnic populations. We assessed the association of five SNPs [�1722C/T,�1661A/G and �318C/T in the promoter region49A/G in exon 1 and CT60A/G in the 30untranslated region(UTR)] with tobacco-related oral squamous cell carcinoma (OSCC) in North Indian subjects. We geno-typed 130 OSCC patients and 180 normal subjects by polymerase chain reaction- restriction fragmentlength polymorphism (PCR-RFLP) using BbvI, MseI, NcoI and BstEII restriction endonucleases. Among theseSNPs, �1722CC, �1661AG and CT60AA genotypes were more prevalent in OSCC patients as compared tocontrols and in the logistic regression analysis with odd ratio (OR) 2.85, 95% CI (0.69–11.68); OR 2.48, 95%CI (1.29–4.78) and OR 3.0, 95% CI (1.43–6.28) respectively, these genotypes showed strong associationwith OSCC risk. With higher prevalence in controls 49GG genotype and G allele (OR 0.57, 95% CI 0.40–0.81) appeared to be protective. Moreover, TACAG, TACGA and TATAG appeared as susceptible while TAC-GG and CACGG appeared as protective haplotypes. These results suggest significant risk modifying effectsof CTLA-4 �1722C/T, �1661A/G, �318T/C, CT60 A/G and 49A/G SNPs in tobacco-related OSCC in NorthIndian population.� 2012 American Society for Histocompatibility and Immunogenetics. Published by Elsevier Inc. All rights

reserved.

1. Introduction

Tobacco related oral squamous cell carcinoma (OSCC) is one ofthe common malignancies in Asian populations. In India it ac-counts for almost 40% of all cancers in Indian males [1]. These tu-mors remain localized for longer time and their metastatic spreadbeyond cervical lymph node is uncommon suggesting a relativelystrong immune surveillance. However, in some cases tumorgrowth is faster and it becomes inoperable in a shorter time [2].We earlier reported defective cellular immune responses in pa-tients with OSCC [3–5], particularly the relationship between im-paired DTH response, poor T-cell proliferation and shorterdisease free survival [6]. Abnormalities in immune regulation inOSCC have been reported in several other studies in different eth-nic populations [7–9].

Cytotoxic T lymphocyte Antigen-4 (CTLA-4) or CD125 receptorprotein expressed by activated T-cells, belong to immunoglobulinsuperfamily. It exists in two isoforms, soluble (sCTLA-4) and fulllength (flCTLA-4) protein. It is similar to other T-cell co-stimulatory

n Society for Histocompatibility an

V. Bharti), drbkmohanti@m, satyandas@hotmail.com,

molecule, CD28. Although both binds to B7.1 and B7.2 moleculeson antigen presenting cells, the binding affinity of CTLA-4 to B7is higher than that of the CD28. While CD28 and B7 interactionactivates T-cells, CTLA-4 and B7 interaction down regulates T-cellactivation; hence CTLA4 has been established as an important neg-ative regulator of T-cell functions through multiple mechanisms.CTLA-4 and B7 interaction results in reduced expression of IL-2and IL-2 receptor and causes arrest of T-cells at the G1 phase in cellcycle during T-cell activation [10]. Such impairment of T-cell acti-vation by CTLA4-B7 interaction leads to down-regulation ofimmune surveillance mechanisms and may be related to tumordevelopment. It has been shown that mice deficient in CTLA-4 genedevelop lymphoproliferative disease [11], and CTLA-4 blockade viaantibody leads to enhancement of immune response and rejectionof tumors [12]. Thus CTLA-4 antigen seems to play an importantrole in inflammatory response and susceptibility/resistance to can-cer and autoimmune diseases.

CTLA-4 gene located on chromosome in 2q33 is composed offour exons that encodes functional domain of the molecule sepa-rately [13] and possess several important polymorphic sites [14].Previous studies have indicated significant association betweenSNPs in CTLA-4 gene and various types of cancer such as lung,esophageal, gastric, breast and cervical carcinoma including headand neck cancer [15,16] and autoimmune diseases such as graves’

d Immunogenetics. Published by Elsevier Inc. All rights reserved.

V. Bharti et al. / Human Immunology 74 (2013) 348–352 349

disease, autoimmune hypothyroidism and type-1 diabetes [14].We have earlier reported association between CTLA-4 promoterpolymorphism and type-1 diabetes in North Indians [17]. However,the association between SNPs in CTLA-4 gene and risk of OSCC isnot well understood. In one study on German population CTLA-4-1661 SNP has been reported as predisposing factor for OSCC[18] whereas another in Taiwanese population reported associa-tion between CTLA-4 49 A/A genotype with early onset and poorsurvival of OSCC patients [19]. However, in these studies onlytwo SNPs (�1661A/G and +49A/G) were studied. Since severalSNPs are present in CTLA-4 gene, the contribution of other SNPsto the susceptibility of OSCC remains largely unknown. Secondly,since SNPs in the promoter region (�318C/T, �1661A/G and�1722), CT60A/G in 30UTR region and 49A/G SNP in exon 1 havebeen associated with DNA binding and promoter activities, expres-sion and stability of CTLA-4 mRNA, we assessed these SNPs andhaplotypes in OSCC patients for their possible role in disease sus-ceptibility in North Indian population. To the best of our knowl-edge, no reports are available on CTLA-4 gene polymorphismsand its association with the risk of tobacco related OSCC in NorthIndians.

Table 1Clinopathological features of oral cancer patients.

2. Subjects and methods

2.1. Study population

The study included 130 North Indian OSCC patients of either sexwith histologically proven oral squamous cell carcinoma, attendingoutpatient department of BRA-Institute Rotary Cancer Hospital ofAll India Institute of Medical Sciences (AIIMS), New Delhi, India.Among patients 116/130 (89.3%) were tobacco users. Majority ofthem (107/116) were smokeless tobacco users while none of thenormal subjects had history of tobacco-use. All patients had biopsyproven OSCC of various sites. Clinical staging of the tumor was per-formed as per UICC criteria [20]. One hundred and eighty age, sexand ethnicity matched normal healthy individual served as controlpopulation. The study protocol was approved by ‘Ethics Commit-tee’ of the All India Institute of Medical Sciences, New Delhi, anda written informed consent was obtained from each study subject.

Characteristics Values

Age (yr)Range (Mean ± SD) 22–88 (50.6 ± 12.8)

No. (%)

GenderMale 117 (90.0)Female 13 (10.0)

Habit (Tobacco)User 116 (89.3)Non user 14 (10.7)

SiteTongue 65 (50.0)Buccal mucosa 25 (19.2)Retro molar trigone 10 (7.6)Othersa 30 (23.0)

Tumor sizeT1 + T2 39 (30.0)T3 + T4 91 (70.0)

Lymph node metastasesN0 50 (38.4)N+ 80 (61.6)

Clinical stageEarly stage(I + II) 21 (16.2)Late stage (III + IV) 109 (83.8)

a Floor of mouth (4); Alveolus (5); Lips (3); Gingival Buccal Sulcus (8); Palate (6);Orophrynx (1); and Tonsil (3).

2.2. Genotyping for CTLA-4 �1772C/T, 1661A/G, �318C/T, CT60A/Gand 49A/G SNPs by PCR-RFLP

Ten milliliters of venous blood was aseptically drawn by veni-puncture from each subject and DNA was extracted by sodium per-chlorate method as described elsewhere [17]. The genotyping ofthe CTLA-4 SNPs was performed by PCR–RFLP as described earlier[17]. The primer sequences, restriction endonucleases, annealingtemperatures and product sizes have been summarized in Supple-mentary Table. Briefly, PCR was performed using 0.2 lg of genomicDNA, 10� buffer, 2.5 mM MgCl2, 1 U Taq DNA polymerase,40 pmols of polyacrylamide gel electrophoresis (PAGE)-purifiedoligonucleotide primers (Bio Basic Inc., Ontario, Canada), 200 lMdNTPs and H2O to a final volume of 25 ll. The PCR conditions were:initial denaturation at 94 �C for 5 min, 40 cycles consisting of 94 �Cfor 30 s, 60 �C for 30 s, 72 �C for 30 s and a final extension for 7 minat 72 �C. The reaction volume and conditions for the PCR amplifica-tions were essentially the same for all SNPs, except that the anneal-ing temperature was 58 �C for �1722, �1661 and �318 promoterregion and 60 �C for CT60A/G as well as 49A/G SNPs. The PCR prod-ucts were checked on 1% agarose gel and the positive samples werefurther digested with respective endonucleases. The target frag-ments obtained in each case was resolved on 10% native polyacryl-amide. The DNA samples were randomly selected for confirmation

of the polymorphisms by sequencing and the results were found tobe 100% concordant.

2.3. Statistical analysis

The difference in the frequencies of various allele and genotypebetween normal control and OSCC patients were performed by Chisquare (v2) test using Graph pad software (graphpad.com). The Pvalues obtained were further corrected by multiplying with totalnumber of genotypes and alleles. Corrected P values <0.05 at 95%CI was considered significant. The SNPs were tested for Hardy–Weinberg equilibrium, both in patients and control groups by v2-test. The odd ratio (OR) and confidence interval (CI) was deter-mined using Hutchon.net software (www.hutchon.net/confi-dOR.htm). Multivariate logistic regression analysis was used toestimate the risk of mutant genotype or allele with respect to wildtype. Haplotype analysis of CTLA-4 �1772C/T, 1661A/G, �318C/T,CT60A/G and 49A/G SNPs was performed by PLINK software(pngu.mgh.harvard.edu). On-line software (www.osse.bii.a-star.e-du) was used to calculate the study power and sample size.

3. Results

The clinicopathological features of oral cancer patients havebeen shown in Table 1. The age of patients ranged from 22–88 years(Mean ± SD = 50.6 ± 12.8). Among 130 patients, 117 (90%) weremales and 13 (10%) were females. Majority (89.3%) of them re-ported history of tobacco use for at least last 6 months. The mostcommon site of lesion was tongue (50.0%) followed by buccal mu-cosa (19.2%) and retro molar trigone (7.6%), other sites such as alve-olus, lip, gingivo-buccal sulcus etc were rare. Larger tumor (T3/T4)and cervical lymph node involvement was presented by most pa-tients (70.0% and 61.6% respectively). Clinical staging revealed thatalmost 84% patients had advanced stage (III+IV) tumors.

The frequencies of different genotypes and alleles have beensummarized in Table 2. The distribution of genotypes and alleles

Table 2CTLA-4 SNPs and risk of OSCC development in Northt Indians.

SNPs Genotype and allele Groups (no. (%)) P ORa (95% CI)

Cases (n = 130) Control (n = 180)

�1722 T/C (rs733618) TT 92 (74.8) 131 (72.7) NS ReferenceTC 25 (20.3) 46 (25.5) NS 0.77(0.44–1.35)CC 6 (4.8) 3 (1.7) NS 2.85(0.69–11.68)T 209 (85.0) 308 (85.6) NS ReferenceC 37 (15.0) 52 (14.4) NS 1.05(0.66- 1.65)

�1661 A/G (rs4553808) AA 94 (78.3) 162 (90.0) 0.03 ReferenceAG 26 (21.7) 18 (10.0) 0.03 2.48(1.29–4.78)GG 00 00 – –A 214 (89.1) 342 (95.0) 0.028 ReferenceG 26 (10.8) 18 (5.0) 0.028 2.31(1.23-4.31)

�318 T/C (rs5742909) CC 112 (93.3) 170 (94.4) NS ReferenceCT 8 (6.7) 10 (5.6) NS 1.21(0.47–3.17)TT 00 00 – –C 232 (96.7) 350 (97.2) NS ReferenceT 8 (3.3) 10 (2.8) NS 1.21(0.46–3.10)

49 A/G (rs231775) AA 67 (51.5) 75 (41.7) NS ReferenceAG 63 (48.5) 80 (44.4) NS 0.88(0.55–1.40)GG 00 25 (13.9) 0.0006 _A 197 (75.8) 230 (64.0) 0.0064 ReferenceG 63 (24.2) 130 (36.0) 0.0064 0.57(0.40–0.81)

CT60 A/G (rs3087243) GG 12 (9.2) 34 (19.0) NS ReferenceAG 47(36.2) 79 (44.0) NS 0.59(0.28–1.25)AA 71 (54.6) 67 (37.0) 0.014 3.00(1.43–6.28)G 71 (27.3) 147 (40.8) 0.002 ReferenceA 189 (72.7) 213 (59.2) 0.002‘ 1.80(1.30–2.59)

a Multivariate logistic regression analysis, NS-not significant (P value <0.05 is considered significant); �1722C/T: cases n = 123, �1661A/G: cases n = 120, �318T/C: casesn = 120.

Table 3CTLA-4 haplotype frequencies and risk of OSCC development in North Indians.

Haplotypes (�1772,�1661, �318, 49, CT60

Patients(n = 100)

Normals(n = 100)

OR (95%CI)a

P

TACAA 0.48 0.53 Reference NSTACGG 0.07 0.22 0.35

(0.14–0.90)

0.002

CACGG 0.04 0.13 0.34(0.1–1.11)

0.022

TACAG 0.1 0.05 2.20 (0.7–6.92)

NS

TGCAA 0.08 0.05 1.76(0.54–5.77)

NS

TACGA 0.07 0.01 7.72(0.91–65.13)

0.03

TATAG 0.02 0.01 2.20(0.19–25.13)

NS

CACAG 0.05 0 – 0.023CACAA 0.05 0 – 0.023CACGA 0.04 0 – 0.043

a Logistic regression analysis; NS, not significant.

350 V. Bharti et al. / Human Immunology 74 (2013) 348–352

in control group was in the Hardy–Weinberg equilibrium. In thepromoter region �1722C/T SNP the frequency of CC homozygousgenotype tended to be higher in patients (4.8 vs. 1.7%) when com-pared with control population. Although difference between twogroups was statically insignificant, in logistic regression analysiswith OR 2.85, 95% CI (0.69–11.68), CC homozygous genotype ap-peared to be at a higher risk while TC heterozygous with lower fre-quency in patients compared to control (20.3 vs. 25.5), OR 0.77 and95% CI (0.44–1.35) seems to be protective genotype. The TT homo-zygotes, T and C alleles were equally distribution in both groups.

At �1661A/G SNP site, frequency of AA genotype wassignificantly lower (P < 0.03) and AG genotype was significantlyhigher in patients (21.7 vs. 10.0%) as compared to controls(P < 0.03). Similarly, the prevalence of A allele at this site wassignificantly lower while G allele was significantly higher inpatients as compared to controls. Furthermore, in logistic regres-sion analysis with OR 2.48, 95% CI (1.29–4.78) AG genotype andG allele OR 2.31, 95% CI (1.23–4.31) appeared to increase the riskof OSCC by more than two-fold. However at �318T/C SNP site,different genotypes and alleles were equally distributed in bothgroups.

In case with 49A/G SNP AA and AG genotypes and A allele wasmore frequent in patients whereas GG genotype and G allele wasmore frequent in controls. In fact none of the OSCC patients were49GG homozygote while almost 14% controls were 49GG homozy-gous. Thus in our study 49GG genotype and G allele appeared asprotective (OR 0.57, 95%CI 0.40–0.81) for OSCC. In case withCT60A/G SNP, GG genotype and G allele was less prevalent(P < 0.002) while AA genotype (P < 0.014) and A allele (P < 0.002)were significantly higher in patient as compared to controls. In lo-gistic regression analysis AA genotype and A allele appeared to in-crease the risk [OR 3.0, 95% CI (1.43–6.28)] while G allele appearedto be protective. Interestingly with higher prevalence in patients(11.6 vs. 4.4%) the combined genotypes �1661AG and 49AA withOR 2.83, 95% CI (1.15–6.99) appeared to impart almost 3-fold high-er risk of tobacco-related OSCC (data not shown).

We further constructed 10 haplotypes from these SNPs as sum-marized in Table 3. Among these TACAA were most frequent inboth patients (48%) and control (53%) group while TATAG was rarein both (2 vs. 1% respectively). Three haplotypes CACAG, CACGAand CACAA were present in patient only. In logistic regressionanalysis haplotypes TACGG, OR 0.35 95% CI (0.14–0.90) andCACGG, OR 0.34 95% CI (0.10–1.11) being significantly lower inpatients, appeared as protective haplotypes. The haplotypes TA-CAG, TGCAA, TATAG and TACGA with relatively higher frequenciesin patients appeared as susceptible haplotypes in patients with to-bacco- related OSCC.

V. Bharti et al. / Human Immunology 74 (2013) 348–352 351

4. Discussion

It has been suggested that the evasion of immune surveillanceand tumour development involves both ‘‘immunologic sculping’’of tumor through loss of human leukocyte antigen (HLA) and tu-mor associated antigen expression from tumor cell surface and‘‘immuno editing’’ leading to skewed balance between immuneregulation and activation mechanisms of the immune system[21]. In T cell-mediated immune functions CTLA-4 play an impor-tant role in negative regulation of T-cell activation, thus it preventsauto reactivity; whereas in cancer it may suppress anti tumor im-mune response. CTLA-4 is constitutively expressed on the surfaceof regulatory T-cells (Tregs) whereas induced expression is foundon activated T-cells and monocytes [22]. It has been shown thatproportion of natural Tregs both in tumor microenvironment andin the peripheral blood was reportedly related to survival of the pa-tients [7]. Recently we reported skewed balance of Tregs and Th17cells in patients with tobacco related OSCC [5].

It is still debatable whether genetic variation in CTLA-4 mayinfluence T-cell activation and susceptibility to cancer. Althoughseveral studies indicated a positive association between CTLA-4SNPs and susceptibility to multiple cancer including lung, breast,esophagus, gastric cardia [15], human papilloma virus associatedcervical [23] and oral squamous cell carcinoma [18,19]. Most ofthese studies genotyped 49A/G SNPs in different European andAsian populations. However, the effect of these SNPs in the expres-sion or function of CTLA-4 protein is not well understood. In thepresent study we genotyped 5 functional SNPs, three in the pro-moter region and one each in 30UTR and exon -1 of CTLA-4 genein North Indian population.

Out of three SNPs in the promoter region, �1722CC homozy-gous tended to be higher in patients and appeared to be suscep-tible genotype while with lower frequency in patients, �1722TCappeared to reduce the risk of tobacco related OSCC. In the pa-tients the frequency of �1661AA genotype was significantly low-er but AG genotype was significantly higher as compared tocontrols. Consequently high frequency of �1661G allele and AGgenotype may be related to increased risk while low frequencyof A allele suggest it to be protective in patients with tobaccorelated OSCC. Although �1661A/G SNP is far from the regulatorysequence of the gene, it can modify the potential DNA bindingsites of various transcription factors [23] and thereby influenceexpression of the gene. Similar to our result, higher prevalenceof �1661AG genotype was reported as possible pre-disposingfactor for OSCC in German population [18]. In another study onChinese Han population �1772CC genotype and C allele and�1661GG genotype and G allele were associated with increasedrisk of breast cancer [24].

Although �318T/C SNP lies within 335-bp promoter region ofCTLA-4 gene and presence of T allele at this position has beenshown to increase CTLA-4 expression [23] by 30%, this SNP seemsto have no role in pathogenesis of OSCC in North Indians. Results ofour study are in agreement with that of Erfani et al., who reportedno association between �318 SNP [25], however Wang et al., re-ported the increase of �318T allele among Han population fromChina [26]. Similarly Su et al. reported the association of �318CTgenotype with HPV-16-associated cervical squamous cell carci-noma in Taiwanese [27]. In our study �318T allele with 2.8% prev-alence rate and with 130 patients and 180 controls appeared asminor allele in North Indians, it would have 80% power at 95% con-fidence level to exclude an odd ratio of 1.21 and above. Thus Alarge cohort of 16,162 cases and 21,819 controls would be requiredto detect significantly increased risk associated with �318T allelein North Indian population. The different risk association due tominor alleles in different studies may be due to variation in sample

size. However, it seems that the genotype-phenotype relationshipmay vary in different ethnic population with respect to differenttumors.

CTLA-4 CT60A/G SNP located in 30UTR region is rich in regula-tory elements and can affect the mRNA stability, degradation andnuclear transport [28], however, its role in disease susceptibilityseems to be controversial and inconclusive [24,29,30]. Earlier studyindicated G allele associated with increased risk of breast cancer[26] whereas recent report from the same group indicated no asso-ciation of this SNP with breast cancer risk in Han population fromnortheastern China [24]. In the present study we found a strongassociation between the risk of OSCC and CT60 AA genotype (OR3.0, 95% CI = 1.43–62.8) as compared to GG genotype (P = 0.014)whereas presence of AG genotype at this site appeared to reducethe risk (0R 0.59, 95% CI, 0.28–1.25) of OSCC. We also observed Aallele to be dominant in patients while G allele in control group(P = 0.002). In contrast, in our previous study we did not observeany association between CT60 SNP and type 1 diabetes in the samepopulation [17]. It seems therefore that CT60A/G SNP may be re-lated to higher expression of CTLA-4, leading to impaired activa-tion of T cells and enhanced tumor growth. The ratio of sCTLA-4to flCTLA-4 mRNA splice form in unstimulated CD4 T cells was50% lower in CT60GG genotype while CT60AG heterozygous geno-type reportedly expresses more sCTLA-4 levels [14], which couldbind to more molecules of B7.1 and impair T cell reactivity.

The 49 A/G SNP is most frequently studied polymorphism incancer patients as well as in autoimmune disorder. In our studyGG genotype and G allele at this site, appeared to be protective,which seems to be reasonable, possibly due to two reasons: firstly,it result in amino acid substitution (17Thr to 17Ala) in the leader se-quence of CTLA-4 protein that may affect its expression and sec-ondly, G allele has lower mRNA efficiency and reduced geneexpression [31], thereby impairs the inhibitory function of CTLA-4. Patients with the 49GG genotype reportedly have increasedmRNA and protein expression of interleukin (IL)-2 and primaryT-cell growth factor [32]. Thus carriers of 49GG genotype individ-uals are expected to have a good T cell-mediated antitumor im-mune responses. Our results are further supported by earlierreport on German OSCC patients where 49AA genotype was asso-ciated with early onset of OSCC than AG and GG genotypes and Gallele was suggested to contribute to active immunity [19]. Thelack of G allele has been reported to attenuate the immune re-sponse and contribute to peripheral tolerance [33,34]. Personswith GG genotype reportedly have 30% less CTLA-4 on their T cellsurface than with the AA homozygous [35]. On the contrary 49AAgenotype has significantly lower activation and proliferation ratescompared to 49GG genotype [15]. Similarly 49GG genotype coulddecrease the breast cancer risk in Iranian women [36]. In the pres-ent study the frequency of variant 49AG genotype was slightlyhigher in patients than in the control group which is in agreementwith a recent meta-analysis in which the variants genotype at thislocus significantly increased the risk of non-epithelial and epithe-lial tumors in Caucasians, Asians and Chinese [16]. SimilarlyCTLA-4 49A/G SNP was shown to contribute to genetic susceptibil-ity to infection- related hepatocellular and cervical carcinomas[37].

It may be argued whether combined genotypes increase the riskof cancer which otherwise could not impart significant risk indi-vidually. We observed significantly higher risk among the carrierof �1661AG and 49AA genotypes. Earlier report also suggestedCTLA-4 genotype as a predisposing factor either alone or in combi-nation with other T cell regulator polymorphisms [18].

There are no reports so far on the CTLA-4 �1722C/T, �1661A/G,�318T/C, CT 60A/G and 49A/G SNPs and their risk association inOSCC patients. There are two more reports on breast cancer

352 V. Bharti et al. / Human Immunology 74 (2013) 348–352

patients where these SNPs have been studied so far. In one study[24] CTLA-4 CAAA (�1722C, �1661A, 49A, CT60A) haplotypeshowed significant association with progesterone receptor statusand increased risk of breast cancer while in other [38] TGTA haplo-type (�1722 T, �1661 G,�318 T, 49 A) was associated with protec-tion against cervical cancer where as TGCG haplotype wasassociated with susceptibility. In the present study multivariate lo-gistic regression analysis of the haplotypes indicated increased riskof OSCC in carriers of TACAG, TGCAA, TATAG and TACGA, whilethose with CACGG and TACGG appeared at a significantly lowerrisk.

In conclusion, our results suggest significant risk modifying ef-fects of genotypes, and alleles in CTLA-4 promoter region (�1722C/T and �1661A/G SNP), 30UTR region CT 60A/G and in exon 1 49A/GSNPs and different haplotypes. Results of the present study show-ing association between several polymorphisms in CTLA-4 geneand OSCC further supports the important role of immune/inflam-matory molecules in the susceptibility to tobacco-related OSCC.In addition, these SNPs may be of prognostic significance in NorthIndian populations with tobacco-related OSCC and may be usefulfor screening individuals at a higher-risk.

Acknowledgments

This work was partly supported by the Department of Biotech-nology, under BTIS program. The technical help from VaishaliKapoor, Manasi Mittal and Poonam Gaur is thankfully acknowl-edged. Authors are grateful to patients and normal subjects whoparticipated in this study.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.humimm.2012.12.008.

References

[1] Biennial Report, National Cancer Registry Program. 1988-89, New Delhi, July1992.

[2] Khanna NN, Srivastava PK, Khanna S, Das SN. Intensive chemotherapy forcancer of the oral cavity. Cancer 1983;52:790–3.

[3] Agarwal A, Rani M, Saha GK, Valarmathi TM, Bahadur S, Mohanti BK, et al.Disregulated expression of the Th2 cytokine gene in patients with intraoralsquamous cell carcinoma. Immunol Invest 2003;32:7–30.

[4] Manchanda P, Sharma SC, Das SN. Differential regulation of IL-2 and IL-4 inpatients with tobacco-related oral squamous cell carcinoma. Oral Dis2006;12:455–62.

[5] Gaur P, Qadir GA, Upadhyay S, Singh AK, Shukla NK, Das SN. Skewedimmunological balance between Th17 (CD4+IL17A+) and Treg(CD4+CD25+FOXP3+) cells in human oral squamous cell carcinoma. Cell Oncol2012;35:335–43.

[6] Das SN, Khanna NN, Khanna S. In vivo and in vitro observation of cellularimmune parameters in squamous cell carcinoma of the oral cavity and itscorrelation with tumor load and prognosis. Cancer Invest 1986;4:207–16.

[7] Curiel TJ, Coukos G, Zou L, et al. Specific recruitment of regulatory T cells inovarian carcinoma fosters immune privilege and predicts reduced survival. NatMed 2004;10:942–9.

[8] Chang MC, Chiang CP, Lin CL, Lee JJ, Hahn LJ, Jeng JH. Cell-mediated immunityand head and neck cancer: with special emphasis on betel quid chewing habit.Oral Oncol 2005;41:757–75.

[9] Gasparoto T, Malaspina TS, Benevides L, et al. Patients with oral squamous cellcarcinoma are characterized by increased frequency of suppressive regulatoryT cells in the blood and tumour microenvironment. Cancer ImmunolImmunother 2010;59:819–28.

[10] Appleman LJ, Berezovskaya A, Grass I, Boussiotis VA. CD28 costimulationmediates T cell expansion via IL-2-independent and IL-2-dependent regulationof cell cycle progression. J Immunol 2000;164:144–51.

[11] Waterhouse P, Penninger JM, Timms E, et al. Lymphoproliferative disorderswith early lethality in mice deficient in ctla-4. Science 1995;270:985–8.

[12] Sabel MS, Hess SD, Egilmez Jr MK, Chen FA, Bankert RB. CTLA-4 blockadeaugment human T lymphocyte-mediated suppression of lung tumor xenograftin SCID mice. Cancer Immunol Immunother 2005;54:944–52.

[13] Ling V, Wu PW, Finnerty HF, Sharpe AH, Gray GS, Collins M. Completesequence determination of the mouse and human CTLA4 gene loci: cross-species DNA sequence similarity beyond exon borders. Genomics1999;60:341–55.

[14] Ueda H, Howson JM, Esposito L, et al. Association of the T-cell regulatory geneCTLA4 with susceptibility to autoimmune disease. Nature 2003;423:506–11.

[15] Sun T, Zhou Y, Yang M, et al. Functional genetic variations in cytotoxic T-lymphocyte antigen 4 and susceptibility to multiple types of cancer. CancerRes 2008;68:7025–34.

[16] Zheng J, Yu X, Jiang L, Xiao M, Bai B, Lu J, et al. Association between theCytotoxic T-Lymphocyte Antigen 4 +49G > A polymorphism and cancer risk: ameta-analysis. BMC Cancer 2010;10:522.

[17] Baniasadi V, Narain N, Goswami R, Das SN. Promoter region �318 C/T and�1661 A/G CTLA-4 single nucleotide polymorphisms and type 1 diabetes inNorth Indians. Tissue Antigens 2006;67:383–9.

[18] Kämmerer PW, Toyoshima T, Schöder F, Kämmerer P, Kuhr K, Brieger J, et al.Association of T-cell regulatory gene polymorphisms with oral squamous cellcarcinoma. Oral Oncol 2010;46:543–8.

[19] Wong YK, Chang KW, Cheng CY, Liu CJ. Association of CTLA-4 genepolymorphism with oral squamous cell carcinoma. J Oral Pathol Med2006;35:51–4.

[20] Sobin LH, Wittekind CH, editors. UICC, TNM Classification of MalignantTumors. New York: Wiley-Liss; 2002.

[21] Dunn GP, Old LJ, Schreiber RD. The immunobiology of cancerimmunosurveillance and immunoediting. Immunity 2004;21:137–48.

[22] Read S, Malmstrom V, Powrie F. Cytotoxic T lymphocyte-associated antigen 4plays an essential role in the function of CD25+ CD4+ regulatory cells thatcontrol intestinal inflammation. J Exp Med 2000;192:295–302.

[23] Anjos SM, Tessier MC, Polychronakos C. Association of the cytotoxic Tlymphocyte-associated antigen 4 gene with type 1 diabetes: evidence forindependent effects of two polymorphisms on the same haplotype block. J ClinEndocrinol Metab 2004;89:6257–65.

[24] Li D, Zhang Q, Xu F, Fu Z, Yuan W, Li D, et al. Association of CTLA-4 genepolymorphisms with sporadic breast cancer risk and clinical features in Hanwomen of Northeast China. Mol Cell Biochem 2012;364:283–90.

[25] Erfani N, Razmkhah M, Talei AR, Pezeshki AM, Doroudchi M, Monabati A, et al.Cytotoxic T lymphocyte antigen-4 promoter variants in breast cancer. CancerGenet Cytogen 2006;165:114–20.

[26] Wang L, Li D, Fu Z, Li H, Jiang W, Li D. Association of CTLA-4 genepolymorphisms with sporadic breast cancer in Chinese Han population. BMCCancer 2007;7:173.

[27] Su TH, Chang TY, Lee YJ, et al. CTLA-4 gene and susceptibility to humanpapillomavirus-16-associated cervical squamous cell carcinoma in Taiwanesewomen. Carcinogenesis 2007;28:1237–40.

[28] Hughes TA. Regulation of gene expression by alternative untranslated regions.Trends Genet 2006;22:119–22.

[29] Anjos SM, Shao W, Marchand L, Polychronakos C. Allelic effects on generegulation at the autoimmunity-predisposing CTLA4 locus: a re-evaluation ofthe 30 +6230G > A polymorphism. Genes Immun 2005;6:305–11.

[30] Mayans S, Lackovic K, Nyholm C, Lindgren P, Ruikka K, Eliasson M, et al.Holmberg D (2007) CT60 genotype does not affect CTLA-4 isoform expressiondespite association to T1D and AITD in northern Sweden. BMC Med Genet2007;8:3.

[31] Kouki T, Sawai Y, Gardine CA, Fisfalen ME, Alegre ML, DeGroot LJ. CTLA-4 genepolymorphism at position 49 in exon reduces the inhibitory function of CTLA-4and contribute to the pathogenesis of Graves’ disease. J Immunol2000;165:6606–11.

[32] Jonson CO, Hedman M, Karlsson Faresjo M, Casas R, Ilonen J, Ludvigsson J, et al.The association of CTLA-4 and HLA class II autoimmune risk genotype withregulatory T cell marker expression in 5-year-old children. Clin Exp Immunol2006;145:48–55.

[33] Mapara MY, Sykes M. Tolerence and cancer: mechanism of tumor evasion andstrategies for breaking tolerance. J Clin Oncol 2004;22:1136–51.

[34] Maurer M, Loserth S, Kolb-Maurer A, et al. A polymorphism in the humancytotoxic T- lymphocyte antigen 4 (CTLA4) gene (exon 1 +49) alters T-cellactivation. Immunogenetics 2002;54:1–8.

[35] Anjos S, Nguyen A, Ounissi-Benkalha H, Tessier MC, Polychronakos C. Acommon autoimmunity predisposing signal peptide variant of the cytotoxic Tlymphocyte antigen 4 results in inefficient glycosylation of susceptibilityallele. J Biol Chem 2002;277:46478–86.

[36] Ghaderi A, Yeganeh F, Kalantari T, Talei AR, Pezeshki AM, Doroudchi M, et al.Cytotoxic T lymphocyte antigen-4 gene in breast cancer. Breast Cancer ResTreat 2004;86:1–7.

[37] Hu L, Liu J, Chen X, et al. CTLA-4 gene polymorphism +49 A/G contributes togenetic susceptibility to two infection-related cancers-hepatocellularcarcinoma and cervical cancer. Human Immunol 2010;71:888–91.

[38] Rahimifar S, Erfani N, Sarraf Z, Ghaderi A. CTLA-4 gene variations mayinfluence cervical cancer susceptibility. Gynecol Oncol 2010;19:136–9.