Females with paired occurrence of cancers in the UADT and genital region have a higher frequency of either Glutathione S-transferase M1/T1 null genotype

Upper Aero digestive Tract (UADT) is the commonest site for the development of second cancer in females after primary cervical cancer. Glutathione S-transferase (GSTM1 and / or T1) null genotype modulates the risk of developing UADT cancer (primary as well as second cancer). The aim of this study was to evaluate the difference in GST null genotype frequencies in females with paired cancers in the UADT and genital region as compared to females with paired cancers in the UADT and non-genital region. Forty-nine females with a cancer in the UADT and another cancer (at all sites-genital and non-genital) were identified from a database of patients with multiple primary neoplasms and were analyzed for the GSTM1 and T1 genotype in addition to known factors such as age, tobacco habits, alcohol habits and family history of cancer. Frequencies of GSTM1 null, GSTT1 null, and either GSTM1/T1 null were higher in females with paired occurrence of cancer in the UADT and genital site (54%, 33% and 75% respectively) in comparison to females with paired occurrence of cancer in the UADT and non-genital sites (22%, 6% and 24% respectively). The significantly higher inherited frequency of either GSTM1/T1 null genotype in females with a paired occurrence of cancers in UADT and genital region (p = 0.01), suggests that these females are more susceptible to damage by carcinogens as compared to females who have UADT cancers in association with cancers at non-genital sites.


Introduction
Precise risk estimation for the development of second can-cers in the Head & Neck, Lung, and Esophagus, also known as the Upper Aero-Digestive Tract (UADT), is difficult due to the multifactorial origin and gene-environment interaction [1].
The Glutathione S transferase (GST) family of enzymes coded for by at least five distinct loci-alpha, mu, theta, pi and gamma detoxify tobacco carcinogens [2] and play an important role in the gene-environment interaction associated risk for cancer. The results of a recently published meta-and pooled-analysis of 4635 cases and 5770 controls, support the widely studied association of inherited GST genotype with the risk of developing squamous cell carcinomas (SCC) in the UADT. In addition the results support the notion that the risk increases when genotypes at multiple GST loci are considered [3]. Patients with second (synchronous and metachronous) cancers in the UADT also have a higher frequency of GST null genotype (particularly GSTM1 and T1) as compared to those with single cancers [4][5][6][7].
Similar to the association seen in UADT cancers, GST null genotype has been associated with the risk of developing Human Papilloma Virus (HPV)-positive primary cervical cancer in females [8,9] and also with other cancers such as primary breast cancer [10] and primary thyroid cancer [11].
Females with a primary HPV related cancer (particularly uterine cervix) develop second cancers in the UADT with an increased frequency [12][13][14]. HPV transmission has been suggested to account in part for the paired occurrence of cancers at these anatomically distinct sites of UADT and cervix. Exposure to tobacco carcinogens has been suggested as synergistic cofactor [15,16]. The mechanism by which HPV infection, necessary for the development of cervical cancer [17], causes UADT cancers is unclear [18]. Recently it has been shown that interaction between viral infection (especially HPV) and xenobiotic enzymes such as GSTM1 could modulate cancer risk, however the evidence comes from in vitro studies alone [19].
There are no studies found in the literature evaluating the association of GST null genotypes and the risk of developing second cancers in the UADT in females, particularly after primary cancers at genital sites. We identified females with a paired cancer in the UADT in association with another cancer (at all sites-genital and non-genital) and analyzed GSTM1 and T1 genotype in addition to known factors such as age, tobacco habits, alcohol habits and family history of cancer.

Patients
From a registry of 150 patients with Multiple Primary Neoplasms (MPN) at Tata Memorial Hospital, Mumbai, during the period of 1996 -2004, all the females, who met the following criteria were selected: 1) synchronous or metachronous occurrence of second primary cancer; 2) Histological or cytological confirmation of both primaries; 3) One of the two cancers was a SCC in the UADT. Modified Hong's criteria [20] was used to differentiate the second primary cancer from metastatic or direct spread from the first cancer.
Details about age, site of first cancer, site of second cancer, histology, grade, history of cancer in the first degree relatives, tobacco habits (smoking or smokeless), alcohol intake (regular, occasional or never) and radiotherapy for the first cancer were noted both from patient interviews and medical records. Detailed and exact quantification of exposure to tobacco and alcohol was not available in all cases. Neither were details about reproductive and sexual history available.

DNA extraction
Having obtained an informed consent, 5-10 ml of peripheral blood sample was collected in sterile EDTA tubes. The DNA was extracted from the leukocyte pellet obtained from the blood by sucrose lysis, purified by phenol: chloroform and suspended in Tris-EDTA buffer and stored at -20°c. gene consisted of an initial melting step at 94°C for 9 min followed by 30 cycles at 94°C for 1 min, 55°C for 1 min, and 72°C for 1 min, with a final step at 72°C for 10 min for elongation. The PCR profile for amplification of GSTT1 gene consisted of an initial melting step at 95°C for 9 min followed by 30 cycles at 94°C for 1 min, 66°C for 1 min, and 72°C for 1 min, with a final step at 72°C for 10 min for elongation.

Gel electrophoresis
The PCR products were separated either on 2% agarose gels or on 12% polyacrylamide gels and were seen by ethidium bromide staining or silver staining respectively. In the PCR reaction presence of a 170 bp (IFNβ-1) band alone and absence of a higher 273 bp (GSTM1) or 459 bp (GSTT1) band indicated null genotype of the sample. In the PCR assay when both Interferon and GST fragments were absent then it indicated that either the PCR was not successful or the DNA was degraded.

Statistical analysis
Fishers exact test was used to assess the statistical significance for the difference between proportions of various factors (number of tobacco chewers, GSTM1 null, GSTT1 null, either GSTM1/T1 null).

Overall data of all females
Forty-nine females with a paired SCC in the UADT in association with another cancer (at genital and non-genital related sites) were identified. The median time to development of the paired cancer (or second cancer) was 4 years. The median age at diagnosis of first cancer in all females was 51 years and that of second cancer was 60 years. The commonest sites of first cancer were oral cavity (n = 17), genital region (n = 15), and esophagus including post-cricoid region (n = 12). Information on family history of cancer was available in 42 women and 6 (14.3%) had a first-degree relative affected with cancer. Of the 45 women where details of habits were available, 25 (56%) were tobacco chewers whereas only 1 (2%) smoked. Of the 29 patients in whom GSTM1 and T1 genotyping was done, the frequencies of GSTM1 null, GSTT1 null and either GSTM1/T1 null were 35%, 17% and 45% respectively.

Groupwise analysis
Results were further analyzed by dividing the females into two Groups (Table 1): Group 1 (n = 19), females who had a paired SCC in the UADT along with a cancer in the genital region and Group 2 (n = 30), females who had a paired SCC in the UADT along with a cancer at other sites (non-genital region). In the majority of the cases (80% in Group 1 and 86% in Group 2), UADT cancer occurred second in chronological order. The median time for the development of second cancer was 5-years (range 0-20 years) for females in Group 1 in comparison to 3-years (range 0-24 years) for females in Group 2. There were no significant differences in the median age at the diagnosis of the first cancer, family history of cancer or tobacco habits between the two groups. Uterine cervix was the commonest genital site of cancer in females in Group 1 (16/ 19, 84%), followed by vagina in one female, vaginal vault in one and vulva in one. The sites of second UADT cancer in females in Group 1 were: oral cavity (n = 7), oropharynx (n = 2), hypo pharynx (n = 1), trachea and Lung (n = 4), esophagus including post-cricoid (n = 4), thyroid (n = 1). In the females in Group 2, UADT was the commonest non-genital site of cancer (22/30, 73%), followed by breast (n = 6), duodenum (n = 1), and chronic lymphatic leukemia (n = 1). Histology of the cancer in the genital region in females in Group 1 was confirmed as SCC in the majority (18/19). In the remaining female the histology could not be confirmed as the records were missing, however it was confirmed that her cervical cancer was treated with radiotherapy, a treatment that is never given without histological confirmation at our hospital. Null genotype frequencies for GSTT1, GSTM1 and either T1 or M1 were more frequent in Group 1 as compared to Group 2. The proportion of females who were null for either GSTM1/T1 were found to be significantly higher in the Group 1 than in Group 2.

Discussion
This is the first study evaluating the association of GST genotype with paired occurrence of an UADT cancer and genital site cancer. Of the 49 women with paired occurrence of an UADT SCC with a second cancer at any site, 19 (39%) were in the HPV related genital site. Majority (84%) of the genital cancers were in the Uterine Cervix and in more than half (68%) the genital cancer occurred before or within 6 months of development of the UADT cancer. Association between malignancies of the UADT and genital region (especially uterine cervix) has been observed by various investigators [13,15]. Hemminki et al [14] studied the pattern of second cancers in females and found a consistent increase in second HPV related cancer when the first cancer was at an HPV related site (anogenital, skin, head and neck, oesophageal and rectal cancers). An excess of oral cancers (including lip, mouth, tongue, larynx and pharynx), as found in this study, followed cervical cancers. Other investigators have also found similar results in the past [12]. Using incidence data from Surveillance, Epidemiology & End Results (SEER), Spitz et al [15] found an elevated Standardized Incidence Ratio (SIR) for cervical cancer after an initial buccal cavity cancer and larynx cancer.
HPV transmission has been suggested to account in part for the paired occurrence of cancers at these anatomically distinct sites of UADT and cervix and smoking has been suggested as synergistic cofactor [15]. However, smokeless tobacco is a predominant form of tobacco consumed by females in India [21] and a recent study from India has shown a positive correlation of smokeless tobacco habits with cervical cancer risk [22]. Additionally, results from another study from India found more than 80% of oral cancers in females attributable to smokeless tobacco habit as compared to the negligible influence of smoking and drinking on oral cancer in females [23]. In concordance with these figures, 44% of females in Group 1 in our study had smokeless tobacco habits and none of the patients, except for one, smoked.
The role of GST polymorphisms in the development of second cancers in the UADT has been a focus of few studies in various ethnic groups [4][5][6][7]. In all of these studies the GST null genotype (GSTM1 and or GSTT1) has been seen with increased frequency in patients with second cancers in the UADT as compared to those with single cancers in the UADT. The results of the recently published meta and pooled analysis support the notion of a greater risk of head and neck cancer when genotypes at multiple GST loci are considered [3]. The 35% GSTM1 null genotype frequency (in entire population of females with second cancers in the UADT) in this study is very similar to the GSTM1 null genotype frequency of 36% found in our previous study in male patients with second cancers in the UADT from similar ethnic background [7]. However, the null genotype frequency for GSTT1 in males was 41% vs. 17% in females from this study. Differences in the site of cancer and the form of tobacco use, which predisposes to cancer development at a particular site, between males and females might possibly explain the gender discordance in GSTT1 null genotype frequencies. Nonetheless the 41% and 17% null genotype frequency for GSTT1 in males and females respectively are higher than 8-12% GSTT1 null genotype frequency found in healthy controls from similar ethnicity suggesting a true association [24,25].
GSTM1 and GSTT1 null genotype frequencies (54% and 33%) in the females who had second cancer in the UADT in association with a cancer in the genital region (Group 1) in the present study are higher than a) the respective frequencies (22% & 6%) in females with second cancer in the UADT in association with a cancer at non-genital region (Group 2), b) the respective frequencies (49% & 18%) found in subjects with single UADT cancers in two studies from a similar ethnic background [24,25], c) the respective frequencies (26% & 19%) found in males with single UADT cancers from our previous study [7], and d) the respective frequencies (33% & 8% and 24% & 12%) found in Indian healthy controls [24,25]. This suggests that the females who developed second cancers in the UADT in association with cancer at genital sites (Group 1 in this study) are possibly more susceptible to damage induced by carcinogens.
The association of GST polymorphism and the risk of single primary cervical SCC have been studied in female patients from various ethnicities. GSTM1 null genotype has been found to incur an increased risk for primary cervical cancer in a mixed Caucasian, Hispanic and African-American patient population as compared to controls [9]. Like wise, in a study on 181 Korean cervical carcinoma patients, Kim et al [8], showed a positive correlation between GST genotypes and cervical carcinoma risk as compared to controls. In this study by Kim et al,GSTT1 null genotype was associated with an increased risk for cervical cancer. When the patients were stratified according to age, either GSTM1/T1 null genotype was significantly over represented in patients with cervical carcinoma who were = 40 years old. However, in a study on 190 Caucasian women, Chen et al [26], determined no association of GSTM1 null genotype with invasive cervical cancer risk. Other investigators have found results similar to that found in the study by Chen et al [26][27][28]. Differences in GST genotype frequencies in various control populations could account for the differences seen in various studies. For instance, the prevalence of GSTM1 null genotype in Caucasian, American and Korean population is similar (as high as 50%) whereas the prevalence of GSTT1 null genotypes varies from 15-31% in the Caucasians to 50-58% in the Koreans and Chinese.
In view of the known differences of the population frequency of null genotypes and its associated risks for cervical cancer across various ethnic groups, it is important to compare our results with those from studies conducted on subjects from a similar ethnic background. GSTM1 and GSTT1 null genotype frequencies (54% and 33% respectively) in the females who had at least one of the cancers in the genital region (Group 1) in the present study were similar to the respective frequencies (57% and 20% respectively) found in females with single (unpaired) cervical cancer cases from India [29] suggesting that the association between GST genotype and cervical cancer risk in Group 1 females in our study is more likely to be a true association.
HPV is undoubtedly the main causative agent in cervical SCC's and the prevalence of HPV, mainly HPV16/18 of more than 95%, has been seen in tumours from cervical cancer patients from India [30]. Evidence linking HPV to oral carcinogenesis is also growing and one of the highest prevalence rates of 74% of HPV 16 and 18 in oral tumors from smokeless tobacco habitués have been reported from southern India [31]. As HPV infections are transient, the absence of HPV DNA does not rule out previous exposure [32]. Interestingly, though most of the patients with cervical cancer are positive for HPV [17], not all patients who have HPV infection will go on to develop cancer [33].
Interactions between HPV and GST genotype may be a reason for this. The evidence for this is two-way. Chen et al [19] evaluating the reason for the lack of association between GSTM1 null genotypes and cervical cancer risk in their epidemiological study demonstrated that viral infection could have potential effect on the activity of xenobiotic metabolizing enzymes, particularly GSTM1. The results of their study raised a question of how chronic viral infections could affect cellular defenses against carcinogens in general. Evidence from other studies suggests that inherited susceptibility in the form of GST genotype may modulate the risk of developing HPV related cancer as evidenced by GSTM1 null genotype which, in addition to HPV infection and smoking, has been found to increase the risk of developing cervical cancer [9]. Moreover, in the patients with cervix cancers having a positive association with GST null genotype, as studied by Kim et al [8], HPV 16 or 18 infections were confirmed in all.

Conclusion
The findings of our study suggest an association between the GST null genotype and the occurrence of paired cancers in the UADT and HPV related genital sites in females. Though this is the first study to report such an association, further confirmation from larger studies on patients from different ethnic backgrounds is required.