Comparative evaluation of genotoxicity by micronucleus assay in the buccal mucosa over comet assay in peripheral blood in oral precancer and cancer patients

Early detection and quantification of DNA damage in oral premalignancy or malignancy may help in management of the disease and improve survival rates. The comet assay has been successfully utilised to detect DNA damage in oral premalignant or malignancy. However, due to the invasive nature of collecting blood, it may be painful for many unwilling patients. This study compares the micronucleus (MN) assay in oral buccal mucosa cells with the comet assay in peripheral blood cells in a subset of oral habit-induced precancer and cancer patients. For this, MN assay of exfoliated epithelial cells was compared with comet assay of peripheral blood leucocytes among 260 participants, including those with oral lichen planus (OLP; n = 52), leukoplakia (LPK; n = 51), oral submucous fibrosis (OSF; n = 51), oral squamous cell carcinoma (OSCC; n = 54) and normal volunteers ( n = 52). Among the precancer groups, LPK patients showed significantly higher levels of DNA damage as reflected by both comet tail length ( P < 0.0001) and micronuclei (MNi) frequency ( P = 0.0009). The DNA damage pattern in precancer and cancer patients was OLP < OSF < LPK < OSCC, and with respective oral habits, it was multiple habits > cigarette + khaini > cigarette smokers > areca + khaini > areca. There was no significant difference in the comet length and MNi frequency between males and females who had oral chewing habits. An overall significant correlation was observed between MNi frequency and comet tail length with r = 0.844 and P < 0.0001. Thus, the extent of DNA damage evaluation by the comet assay in peripheral blood cells is perfectly reflected by the MN assay on oral exfoliated epithelial cells, and MNi frequency can be used with the same effectiveness and greater efficiency in early detection of oral premalignant conditions.


Introduction
Worldwide oral cancer ranks eighth position in cancer incidence (1). Oral squamous cell carcinoma (OSCC) constitutes ~96% of all oral malignancies. OSCC is an imperative cause of morbidity and mortality that differ in rate of occurrence and in geological location with estimations of >500 000 new cases per year (2). The aetiology of oral cancer is multifactorial. The established factors that cause oral cancer are mostly smoking, chewing betel quid/ tobacco and alcohol intake separately or synergistically (3,4). The commercially available smokeless forms such as khaini (5) or pan (betel quid), gutkha and pan masala consisting of a mixture of lime, areca nut, slaked lime, catechu and other condiments are highly associated with oral cancer (6). Of the malignancies of the oral cavity, 30-80% arise from premalignant lesions such as oral lichen planus (OLP), leukoplakia (LPK) and oral submucous fibrosis (OSF), the range of malignancy transformation rate being 0-5.3%, 0.13-17.5% and 2.3-7.6%, respectively (7,8). Several studies have indicated that the above deleterious habits may induce oxidative stress/reactive oxygen species generation that may lead to DNA damage in form of DNA breaks, double-strand (ds) or single-strand (ss) breaks (9)(10)(11)(12)(13). DNA breaks or misrepaired DNA breaks may lead to a lagging acentric chromosome or chromatid fragments during anaphase from which micronuclei can originate (14). Hence, DNA damage due to oral habits in precancer and malignancy patients may provide a wide array of useful diagnostic and prognostic information along with easy and reliable techniques for their early detection.
The alkaline comet assay (ACA) or single-cell gel electrophoresis is based on the principle that damaged DNA (ds or ss) moves faster than undamaged DNA in an agarose gel. A cell with DNA damage appears in the form of comet, whereas an undamaged cell appears as a halo (15,16). ACA is rapid, versatile and inexpensive, and a small number of cells is required. It is applicable to all eukaryotic cells and sensitive to a smaller amount of DNA damage due to ds, ss and alkali-labile sites. The tail length is directly proportional to DNA damage. The MN assay has emerged as a method for assessing chromosome damage as it facilitates measurement of both chromosome loss and chromosome breakage reliably (17). Micronuclei are characterised by the presence of both a main nucleus and one or more smaller nuclear structures, round/oval in shape, diameter ranges between 1/3 and 1/16 of the main nucleus. There are possibilities of multiple micronuclei depending upon the extent of DNA damage (18). The MN assay is suitable method for DNA damage analysis at differential stages of cancer progression in the response of exposure/accumulation of carcinogen/mutagens as micronuclei start to form much earlier than is observed.
In the last few decades, the comet and MN assays have been widely used tools for evaluation of DNA damage and genomic instability in various forms of cancer (19,20); this has been shown primarily in lymphocytes but has not yet been validated in buccal mucosa cells. Hence, this study mainly aims firstly to evaluate the differential DNA damage induced by habitual usage of processed areca nut products and smoking/smokeless tobacco in differential oral precancer conditions, and secondly, in patients where peripheral collection of blood is not possible, whether the comet assay in peripheral blood lymphocytes can be replaced by the MN assay in oral buccal mucosa cells for monitoring cancer progression.

A. Katarkar et al.
OLP (n = 52), LPK (n = 51), OSF (n = 51) and OSCC (n = 54), together with healthy normal volunteers (controls; n = 52) were recruited in order to study the extent of DNA damage using the comet and MN assays. Detailed oral examination was carried out by an oral pathologist. The healthy volunteers without (or <6 months) oral habit with normal oral epithelium were included. The clinically selected and histologically diagnosed patients with OLP, LPK, OSF and OSCC having habitual use of areca nut, khaini or cigarettes in any form were included in the study. Patients suffering from any infectious or contagious disease or with any intractable medical or radiological abnormality, any other white patch-like candidiasis, hypertrophic OLP and lichenoid-like lesions, scleroderma, previous history of surgery, radiotherapy or chemotherapy, or any vitamin or dietary supplement use were excluded from the study. The patients were educated about the study and after obtaining consent a detailed questionnaire of various demographic parameters was recorded. The questionnaire primarily included personal information (age and sex) and asked about risk factors for oral cancer development; habit of chewing pan masala, gutkha, betel quid, areca nut, khaini, cigarette smoking, alcohol ingestion (frequent or occasional), oral hygiene and use of mouthwash. Exposure/duration to deleterious oral habits, number per day of areca nut, khaini and cigarettes used in all habitual users and multiple habits were recorded. The study was duly approved by the human ethics review committee of the institute.

Sample collection
For the comet assay, a peripheral blood sample (2-3 ml) was collected from the antecubital vein, labelled and stored at 4°C with EDTA and protected from light until processed in the laboratory. For the MN assay, individuals were asked to rinse their mouths thoroughly for 2 min with tap water. The exfoliated buccal cells for micronuclei (MNi) analysis were collected from one or both cheeks in the control group, and in the area with lesion in the cases groups, depending on the region where the lesion was located, including cheek, soft and hard palate, dorsal, ventral and lateral surfaces of the oral cavity. The sites of buccal cell collection for smear preparation depended upon the location of the lesion, precancer condition and oral habit. In OLP, sites with white striations, white papules, white plaques, erythema, erosions or blisters affecting predominantly the buccal mucosa were selected. In LPK sites with white or white-and-red lesions that may be either irregular or uniformly flat and thin, nodular or exophytic, wrinkled or corrugated surface with a consistent texture throughout were selected. In OSF sites with tough, leathery texture of mucosa, blanching of mucosa (persistent, white, marble-like appearance which may be localised, diffuse or reticular), quid-induced lesions (fine, white, wavy, parallel lines that do not overlap or criss-cross, are not elevated and radiate from a central erythematous area) were selected. In OSCC site with red or white, painless, non-healing, indurated ulcers were selected along with an unaffected site. The site of the smear was located and wiped with cotton moistened in normal saline to remove slough/surface coatings. In hyperkeratotic lesions such as LPK, surface keratin layer may be scraped off with a curette. Exfoliated oral mucosal cells were spread over glass slides using a spatula. The prepared smears were fixed in the cytology fixative (alcohol and ether in 1:1 ratio) for 30-35 min.

Comet assay
Slides were dipped in 1% normal-melting-point agarose (NMPA) and allowed to dry at 37°C. Once NMPA is solidified, 80 μl of peripheral blood diluted with 80 µl of 1× phosphate-buffered saline was added in a micro-centrifuge tube containing an equal volume (160 µl) of low-melting-point agarose (LMPA, 37°C). A cover slip was placed carefully over the slide so that a uniform layer over the NMPA coat was formed avoiding trapping of air bubbles. The slides were kept over an ice-pack to solidify the gel for 10-15 min. The cover slip was carefully removed and 100 μl of LMPA was added over the gel mixture layer and a fresh cover slip was placed (avoiding trapping of air bubbles), this was also kept over an ice-pack to solidify the gel for 10-15 min. The cover slip was finally removed and the slide was dipped into lysis buffer solution (2.5 M NaCl, 0.1 M Na 2 EDTA and 10 mM Trizma base, 1% Triton X and 10% dimethyl sulphoxide) and refrigerated overnight. After lysis at 4°C, DNA was allowed to unwind under alkaline conditions. The slides were allowed to stay in the cold (4°C) alkaline buffer (300 mM NaOH and 1 mM EDTA, pH > 13) for 30 min in order to unwind DNA strands and expose the alkali labile sites (alkali unwinding). Electrophoresis was performed for 20 min under alkaline conditions in refrigerator (4°C) at 280 mA and 24 V (~0.74 V/cm). The slides were gently picked up from the alkaline electrophoresis buffer and were placed on a staining tray. The slides were carefully flooded with neutralising buffer (0.4 M Tris, pH 7.5) for 5 min for three times. Following electrophoresis, the slides were neutralised using neutralising buffer (0.4 M Tris, pH 7.5) followed by staining with a fluorescent dye (ethidium bromide 2 mg/ml, Sigma Aldrich). The DNA comet was visualised using a fluorescent microscope (Leica DM 3000, Leica Microsystems). For each sample, two slides were prepared, and the head and tail lengths were measured collectively in 100 cells per sample (50 cells per slide were measured) using Leica QWinPlus digital image processing and analysis software (16,21). The DNA damage was evaluated by the length of the comet tail for each cell.

Micronucleus assay
Fluorescent stain 4ʹ,6ʹ-diamidino-2-phenylindole (DAPI) was used for MNi analysis. DAPI has the highest specificity for DNA and therefore reduces false positives that can arise from hyalokeratin granules (keratin bodies), which may be misinterpreted as micronuclei using a non-specific DNA-stain-like Giemsa stain (22). Also, it gives high fluorescence yield, which is suitable to detect small signals from micronuclei. It also yields the best fluorescence signal for image analysis without cytoplasmic background. Briefly, the slides smeared with buccal cells were washed with 1× PBS for 5 min for three times. The stock solution of DAPI (Invitrogen) was diluted (1 μg/ml) with distilled water and stored at 4°C. Fixed slides were stained for 20 min, rinsed thrice in 1× PBS, air-dried and mounted in the same buffer (22,23). Observations were carried out in a dark room using a fluorescence microscope (Leica DM 3000, Leica Microsystems) equipped with a band-pass filter of 450-490 nm (excitation range: blue). Stained slides were read at ×40 magnification. The criterion for scoring of human buccal micronuclei was followed as per briefly explained by Bolognesi et al. (24). MNi identification and structures to be considered were as per the following characteristics: (i) micronuclei are 1/3-1/16 diameter of the main nucleus; (ii) the same plane of focus as the main nucleus; (iii) the same colour, texture and refraction as the main nucleus; (iv) smooth oval or round shape and (v) the nuclear boundary of the micronucleus (MN) should be clearly distinguishable from that of the main nucleus. The buccal cytome model includes biomarkers of (i) DNA damage (micronuclei and nuclear buds); (ii) cell death (condensed chromatin, karyorrhexis, pyknotic and karyolytic cells); (iii) cytokinetic defects or arrest (binucleate cells) and (iv) proliferative activity (basal cell frequency) (24). In this study, DNA damage in buccal cells in terms of micronuclei in differentiated cells (which include both transitional and terminally differentiated cells) with intact nuclei and cytoplasm, non-fragmented, non-clumped and non-overlaid were scored and included in the analysis. The detailed criteria for identification of both transitional and terminally differentiated cells were followed as given by Bolognesi et al. (24). Basal cells for MNi analysis were impractical due to low frequency of this type of cell and thus excluded from the study. Nuclear buds, the mechanism behind the formation of which is not known, are thought to arise from the elimination of amplified DNA or from DNA repair. They contain a nucleus with sharp constriction forming a bud, which attaches to main nucleus and has a similar staining intensity to the main nucleus with a diameter a quarter to half that of the nuclear diameter. They were excluded from this analysis because they were observed with very low frequency. A total of 1000 cells per individual were evaluated for the presence of micronuclei. In order to avoid bias, two observers were used to score micronuclei. The measures of DNA damage were evaluated as MNi frequency and expressed as percentages.

Statistical analysis
GraphPad Prism software 5.0 was used for statistical analysis. The data were checked for normality by D'Agostino & Pearson omnibus normality test for age and gender distribution using chi-square test. Non-parametric Kruskal-Wallis test followed by Dunn's post-test was used for comparison of data in studied groups and different oral habits. The one-way analysis of variance (ANOVA) was performed to estimate DNA damage associated with studied groups and differential oral habit. Spearman correlation coefficients were used to calculate the correlation between different oral diagnostic groups. Spearman correlation coefficient was calculated separately to estimate the correlation between MNi frequency and comet tail length (micrometre). Inter-observer variability and inter-individual variability were calculated by using GraphPad Prism. P-values of <0.05 were regarded as statistically significant.

Demographic characteristics of the study participants
A representative clinical presentation of control, potentially malignant and cancer patients is described in Figure 1 and their demographic characteristics are summarised in Table I Figure 2A). The mean % tail lengths in different oral diagnostic conditions were allowed for non-parametric Kruskal-Wallis test followed by Dunn's post-test ( Figure 2C). Though comet mean tail length was significantly higher (<0.0001) in LPK (16.66 ± 2.23 µm) and OSF (14.13 ± 1.76 µm) compared with controls, it was not as high as in OSCC cases, the most nearest was LPK (Table II, Figure 2C). The OLP cases presented characteristics more like controls (8.154 ± 2.99 µm). The presence of deleterious oral habits worsened the case in OSCC patients (28.45 ± 5.43 µm) ( Figure 2C). No significant differences were observed in % tail length compared with male and female in all precancer and cancer groups (supplementary Table 1, available at Mutagenesis Online, Figure 2E). When the precancer and cancer group were stratified according to the oral habits, the OLP cigarette smokers showed a significant increase in mean % tail length (P < 0.0001) compared with control and in OSF areca with khaini had significant mean % tail length compared with only areca chewers (P = 0.0033; Figure 3). LPK and OSCC had similar patterns of DNA damage and multiple habits in all oral lesion groups has highest DNA damage (Figure 3). The order of genotoxicity from precancer to cancer was observed as OLP < OSF < LPK < OSCC by the comet assay. An overall 6.02% inter-observer variability (% CV I-Ob ) was observed in the comet assay. The mean % CV I-Ob with habits, without habit and in different type of oral habits was 6.18%, 6.42% and 8.15%, respectively (Table II). An overall 21.42% inter-individual variability (% CV I-Id ) was observed. The mean %CV I-Id with habits, without habit and in different type of oral habits was 18.11%, 18.54% and 35.40%, respectively (Table II).  Figure 2D). Similar results to those obtained with the comet assay were observed. LPK shows significantly higher mean MNi frequency compared with OLP and OSMF (P < 0.0001; Figure 2D). Likewise, no significant differences were also observed in MNi frequency between males and females in all precancer and cancer groups (supplementary Table 1, available at Mutagenesis Online; Figure 2F), suggesting that DNA damage was independent of any gender ( Figure 2E and F). After stratification of precancer and cancer group according to differential oral habits, results were similar to those obtained with the comet assay. In OLP, a significantly higher MNi frequency (P < 0.0001) was observed among the patients who smoked only cigarettes (5.85 ± 3.8) and those who had multiple habits (6.05 ± 0.7). In LPK, MNi frequency was significantly (P = 0.0296) correlated among those who smoked only cigarettes (6.10 ± 3.66) and those with multiple habits (9.12 ± 3.50). In OSF, areca chewing was significantly correlated with multiple habits (7.71 ± 1.82) along with areca (P = 0.0322; Figure 3). In OSCC, the mean MNi frequency observed is almost equal in all oral habit groups. We observed the multiple micronuclei (>2) in certain patient samples and their

Genotoxicity measured by the MN assay in exfoliated oral buccal cells
MNi frequency was calculated (supplementary Table 2, available at Mutagenesis Online). In LPK, 52.94% of patients showed multiple micronuclei and significantly higher MNi frequency was recorded (12.11 ± 2.49) than overall. Moreover, in this study, we did not found any significant difference in MNi frequency between unaffected (13.71 ± 3.56) and affected sites (P = 0.898; supplementary Table 3, available at Mutagenesis Online). The order of genotoxicity from precancer to cancer was the same as observed with the comet assay. The overall 6.82% CV I-Ob was observed in MNi analysis. The mean CV I-Ob with habits, without habit and in different type of oral habits was 6.79%, 9.35% and 14.71%, respectively (Table II). The overall 40.51% CV I-Id was observed in MNi analysis. The mean CV I-Id with habits, without habit and in different type of oral habits was 37.32%, 26.23% and 56.57%, respectively (Table II).

Genotoxicity association with exposure of oral habits
The combined description of genotoxicity and exposure of oral habits (in pack years) in different precancer and cancer groups is shown in Table I Figure 4A). Likely, the same extent of DNA damage was observed in terms of MNi frequency (Table II, Figure 4A). The comet and MN assays showed overall DNA damage patterns, irrespective of oral lesion, were in the order multiple habits > cigarette + khaini > cigarette smokers > areca + khaini > areca ( Figure 4A).

Correlation analysis
The Spearman correlation coefficient was calculated in between comet tail length and MNi frequency to ascertain the extent of damage associated either with different oral premalignant and malignant conditions (Table III)  When stratified according to different oral habits, areca nut or smoking in association with smokeless tobacco (khaini) emerged as the most genotoxic habit with highly correlated increased tail length and MNi formation (r = 0.91 and 0.9, respectively) compared with any other habit either singly or in combination ( Figure 5). Moreover, correlation analysis between overall MNi frequency vs. % tail length irrespective of oral lesion and oral habit was highly significant (r 2 = 0.844, P < 0.0001) suggesting that the MN assay (locally) and the comet assay (systemic) having similar accuracy and precision for detecting DNA damage ( Figure 4B).

Discussion
Cancer progression is a multistep process; genetic alterations induced by DNA damage represent the genetic, epigenetic and phenotypic changes, which are the hallmarks of carcinogenesis. The complex patterns of genetic alteration include premalignant stages to fully developed cancer. It is possible to take advantage of the recent discoveries in specific biomarker development for early detection of cancer progression. These biomonitoring techniques should be cost-effective, easy, rapid, non-invasive, accurate and sensitive. A range of tests can be used for early detection of DNA damage, among which the comet assay is widely practiced (25,26). It has been successfully applied to diagnose DNA damage in screening several cancers like breast cancer (27), lung cancer (28), head and neck cancer (29), bladder cancer (30), thyroid cancer (31), ovarian cancer (32) and prostate cancer (33). However, collection of peripheral blood is invasive and can be unsuitable for anaemic patients. The extent   of DNA damage in oral premalignancy and malignancy has been successfully ascertained by the comet assay (21,34). Buccal cells form the first barrier to potential carcinogens that can be metabolised to generate potentially reactive products present in pan masala, gutkha, areca nut, betel quid and cigarettes. Nearly 90% of oral cancers are epithelial in origin (18). Hence, oral epithelial cells represent the preferred target site for early genotoxic effects induced by carcinogens. It has been observed that there is a higher comet tail length in buccal epithelium than in leukocytes (35). The comet assay has certain drawbacks when performed in buccal cells. Though lower pH (9.1) was suitable for buccal cells to evaluate DNA damage, it was highly associated with background levels of DNA strand breaks. At this pH, buccal cells are highly resistant to lysis. Even trypsin and proteinase K aided practice caused a high background DNA damage and a homogenous comet head (even fluorescence intensity throughout the nucleoid) (35). Higher pH (>12) aided cell lysis but induced DNA and cellular disintegration, which collectively increased baseline DNA damage and very few cells could be seen in the agarose gel (35)(36)(37). The MN assay was used extensively to monitor DNA damage in buccal cells owing to its non-invasive sampling, short turnover time, low DNA repair capacity of the cells and direct correlation with cancer progression, whereas in lymphocytes sampling is invasive, has long turnover time, efficient DNA repair capacity and is indirectly correlated with cancer progression (38,39). In our earlier studies, difficulties were found in collecting peripheral blood samples both from anaemic or trypanophobic patients. The MN assay from the exfoliated buccal cells taken directly from the lesion has certain advantages over the comet assay. Firstly, its collection process is non-invasive, patient friendly and faster. Secondly, the extent of local damage surrounding the lesion (both the affected and the unaffected side) can be ascertained, which helps to determine the spread. And thirdly, the comet assay cannot be performed using exfoliated buccal cells due to its lesser number of collected cells. MNi staining has been evaluated by varieties of stains in exfoliated buccal cells, among them DNA-specific stains (Feulgen, acridine orange, DAPI and propidium iodide) and non-specific stains (Giemsa, May-Grünwald-Giemsa, Papanicolaou [PAP] and haematoxylin and eosin [H&E]) (40). The results of our study show increases in MNi frequency in smoker compared with control are in contrast with Nerseseyan et al. (22) according to which no significant difference in MNi counts were observed between smokers and non-smokers by DAPI staining. The results of Grover et al. (41) were though in accordance with our results for increase in MNi count in premalignant conditions, but micronuclei counted by PAP and H&E stains were higher, and so low by Feulgen stain compared with DAPI stain in our case. The results of Jyoti et al. (42) stained with acridine orange for OSF are in agreement with our results but show higher MNi counts. Bloching et al. (43) showed there was a 2.2-fold increase in the MNi frequency in LPK stains with Giemsa: in contrast, we observed a 3.6-fold increase in overall MNi frequency in LPK. Haldar et al. (44) showed a gradual increase in MNi frequency from precancer to cancer compared with normal stained with Giemsa, in agreement with our results and with no significant deviation in MNi frequency/count. In contrast, Fareed et al. (45) (MNi stain with Giemsa) and Sellappa et al. (46) (MNi stain with May-Grünwald Giemsa) in exfoliated buccal cells of areca chewers found increased MNi frequency, but the MNi count was low and high, respectively, compared with our data. Sanchez-Siles et al. (23) observed increased MNi frequency stained with DAPI in OLP compared with normal, and this result and the MNi count are in agreement with our data. Moreover, several studies used PAP stain were positively correlated, in agreement, and with basal MNi frequency/counts not widely different from our results (47)(48)(49)(50). These inter-study differences may be due to the exposure to the genotoxic material.
In India, the rate of malignant transformation of LPK is predominant in males particularly in association with chewing tobacco and smoking (51,52). Cigarette smoking is a major risk factor associated with oral LPK (53) and increased MNi formation in exfoliated buccal mucosal cells due to smoking has studied by different groups and has agreement with our results (54,55). A previous report from our lab showed DNA damage measured by the comet assay was greater in LPK associated with smoking (21). As 25% female and 75% male were incident with LPK, among 77% patient has cigarette smoking habit with exposure of 332.3 ± 65.79 pack years observed in our study. Interestingly, a high quantity of multiple micronuclei in single buccal cells was frequently observed in LPK. These observations are in agreement with Mahimkar et al. (56), which suggests increased DNA damage in the buccal mucosa of LPK and may be strongly correlated with high incidence with which LPK progresses to malignancy. Collectively in LPK, areca or cigarette smoking and areca + khaini or cigarettes + khaini show relatively similar patterns of DNA damage with the comet assay and with the MN assay. Moreover, MNi frequency was significantly higher with cigarettes + khaini compared with areca, which demonstrates the impact of cigarette smoking in combination with other oral habits.
OSF is a chronic, debilitating disease characterised by progressive fibrosis of the lamina propria and deeper connective tissues, manifest into rigidity and inability to open the mouth (8). OSF is associated with regular use of areca nut chewing, betel nut, gutkha and pan masala (57). Areca nut and arecoline, which is the major contributing factor in the pathogenesis of OSF, have been shown to be cytotoxic and genotoxic to oral epithelial cells via DNA damage by oxidative reactive oxygen species generation (9)(10)(11) and N-nitrosamine formation (5,58,59). In OSF, MNi frequency and % tail length show similar patterns of DNA damage with respect to corresponding oral habits. Recently, several studies have shown increases in MNi in buccal mucosa cells of OSF associated with areca nut and tobacco chewers. Smoking and khaini results in more MNi formation in addition with areca nut and has agreement with our present results (42,47,60).
In India the overall prevalence of OLP, which is a chronic inflammatory condition of unknown origin affecting the lining of mouth with lacy white patches, the malignant transformation rate (7) was 1.5%, relative risk was highest to 13.7% among those who smoked and chewed tobacco. A marked predominance of females was found irrespective of oral habits (61). In this study, 58% female and ~72% subjects with no oral habits were presented with OLP. Increased MNi frequency was observed compared with controls in cigarette smokers than in areca chewers, which may be due to oxidative stress (62). Similar results were observed with the comet assay, but the MN assay was the reliable method for genotoxicity detection in OLP.
The major source of variability in the comet assay arises from the key steps of the protocol, e.g. slide preparation, homogeneity of agarose layer, cell lysis step, electrophoresis condition and scoring (63). Variability in the MN assay in buccal cells arises from differences in scoring by observers, inter-individual variability and type of stain used (64). In this study, inter-observer variability was acceptable in both the comet assay (4.3-8.4%) and the MN assay (4.5-8.8%). CV I-Id difference in the comet assay does not vary considerably from premalignant to malignant condition in the subjects with oral habit (17.79-19.07%), but the CV I-Id difference in the MN assay was more among premalignant to malignant conditions in the subjects with oral habit (43.26-19.48%). A moderately high CV I-Id (35.4%) was shown by the comet assay, whereas a much higher CV I-Id (56.57%) was shown by MN assay when observed in types of habit group. This moderate variability for the comet assay in lymphocytes may be due to confounding factors responsible for precancer or cancer reflected in blood due to mutagenic exposure that leads to DNA damage. Subsequent accumulation of DNA damage and the DNA repair capacity of lymphocyte varies with age and sex, which indirectly correlated with disease progression. High CV I-Id in the MN assay may be due to sensitivity and susceptibility to mutagens to which the buccal cells are directly exposed.
We observed no significant changes in DNA damage in relation to gender; this suggests that DNA damage is independent of gender and reliant on type and duration of exposure to carcinogens and oral habits. The negative correlation of OLP with LPK and OSF may be due to larger difference in the genomic damage accumulated over the time with exposure from oral habits. The significant positive correlation of LPK with OSCC suggests that LPK has the severe DNA damage, which can be directly correlated with its high possibilities to malignancy transformation. The positive correlation of LPK with OSF indicates the quantum of genotoxic insult required to cross the barrier, which leads to neoplasmic transformation.
The major objective of this study was to evaluate DNA damage in buccal cells as a non-invasive sampling, which is of prime importance in early detection of oral cancer. As identification of such abnormalities as cellular proliferation, cell death, DNA damage in the pool of buccal cells is very difficult and only a few studies are documented in the literature, the validation of such a study is quite difficult and may be erroneous. MN identification is very simple, easy and the protocol for MN detection is well standardised. Evaluation of DNA damage extent can be done more accurately by the MN assay by the clinician. However, a more detailed 'cytome-based assay' of oral precancer and cancerous condition will be required for generating more information regarding the biomarkers of cellular proliferation and damage.

Supplementary data
Supplementary Tables 1-4 is available at Mutagenesis Online.

Funding
Council of Scientific and Industrial Research (Government of India).