Significance of Fluorescent Spectroscopy in Screening Oral Potentially Malignant Disorders and Oral Cancer by Characterization of Salivary DNA Using Ethidium Bromide—A Comparative Study

Introduction

Oral cancer is the sixth most common cancer worldwide in which 95% are of squamous cell carcinoma.1 Incidence of oral cancer varies globally. High incidence is reported in South Asian countries such as India and Sri Lanka. In India, oral cancer remains the major cause of mortality in both the males and females.2 Mostly, oral squamous cell carcinoma (OSCC) is developed from oral potentially malignant disorder (OPMDs) such as leukoplakia (LPA) oral submucous fibrosis (OSMF) and erythroplakia, palatal lesions in reverse smokers, oral lichen planus, and oral lichenoid reactions.3 The survival and quality of life of these high-risk people depend upon the detection of OPMD and OSCC at an earlier stage. Early diagnosis of OPMD and OSCC is the corner stone to provide prompt treatment and better prognosis.4 The standard method to identify oral cancer is conventional oral examination followed by scalpel biopsy to confirm the diagnosis. Histopathological test is the gold standard method to confirm oral cancer that is an invasive procedure and requires adequate training.5

Light-based imaging of the tissues and biofluids detects slightest changes such as (i) cell microanatomy, e.g., nuclear/cytoplasmic ratio; (ii) redox status; (iii) expression of specific biomarkers; (iv) tissue architecture and composition; (v) chemical changes; and (vi) vascularity/angiogenesis and perfusion. These characters are appropriate to detect minimal (early) changes, for assessing the margins of lesions and possibly the presence of subclinical changes beyond the clinical margins, for repeated noninvasive monitoring of existing lesions, and for rapidly examining at-risk populations.6 Among the various optical diagnostic methods, fluorescence is one of the most sensitive tools in exploring the subtle changes in tissue conditions by changes in its energy, wavelength, polarization, and direction.7

Many researchers have explored the diagnostic potential of fluorescence spectroscopy and imaging in the characterization of tissue and biofluids.8910 The altered metabolic function and pathological conditions induce significant changes in both biofluids and tissue samples of patients that were noticed from some of the intrinsic fluorophores such as the aromatic amino acids—phenylalanine, tryptophan, and tyrosine, nicotinamide adenine dinucleotide (NADH), flavin adenine dinucleotide (FAD), pyridoxine derivatives, structural proteins like collagen, elastin and their cross-links, and endogenous porphyrins present in biofluids and tissues.81112

DNA biomarkers specific to OSCC have been identified in saliva that can be used for diagnosis and prognosis.1314 Initially, isolation and detection of DNA present in saliva were done using electrophoretic technique and it was a precise method, but long and cumbersome technique. DNA present in the biofluids cannot be detected with fluorescence spectroscopy because of its low quantum yield but by introducing a fluorescent dye like ethidium bromide, DNA can intercalate with a dye and its fluorescent characteristics can be studied that can be used to discriminate disease. Jiang et al reported increased amount of mitochondrial DNA in head and neck cancer patients than normal patients based on quantitative polymerase chain reaction test.15 From the literature, there was not much studies done to characterize DNA based on fluorescence spectroscopy to discriminate oral cancer and oral potentially malignant disorders. Yuvaraj et al studied about the fluorescence spectroscopic characterization of salivary DNA in oral cancer and normal patients and observed a significant spectral change between the groups.16

So, this current study focused to discriminate OSCC and oral potentially malignant disorders rapidly at the earliest with fluorescence spectroscopic characterization of salivary DNA using a fluorescent dye ethidium bromide.

Materials and Methods

Results

Among the 40 patients, the mean age for group A, B1, B2, and C were 48, 41, 36, and 32 years, respectively. Group A included 8 males and 2 females, group B1 included 10 males, group B2 included 8 males and 2 females, and group C included 2 males and 8 females.

Excitation Emission Spectrum

The fluorescence of salivary DNA was obtained by adding the ethidium bromide dye that binds with DNA and showed a fluorescence emission. The saliva was excited at 480 nm and the fluorescence emission spectrum was recorded from 500 to 750 nm in 1 mm increments and the shape of the spectrum appeared similar in all groups with peak intensity at around 605 nm and the average fluorescence intensity was measured (Figs. 1 and 2). The fluorescent intensity of normal group is lesser than the other three groups and the group B1 showed a maximum intensity (Table 1).

Table 1

Descriptive statistics (average fluorescence emission intensity in cycles per second of salivary DNA)

Variables	Groups	Mean	Standard deviation
Salivary DNA	Group A	3771135.50	616089.94
	Group B1	5369227.45	1203353.92
	Group B2	2472118.89	1210762.48
	Group C	1524847.56	874305.8

Close

Fig. 1: Average fluorescence emission intensity of salivary DNA at 480 nm excitation. Group B1 and A showed a maximum intensity than group C.

Export to PPT

Close

Fig. 2: Normalized fluorescence intensity of salivary DNA.

Export to PPT

One-way ANOVA test was done to compare the DNA fluorescence spectrum among the groups and showed a significant difference with p-value of 0.00 as shown in Table 2. Post-hoc Bonferroni multiple comparisons showed that the group A showed a significant difference with group C and group B1 and B2 with p-values of 0.00, 0.07, and 0.04 respectively. No statistically significant difference was observed between the group C and group B2 (Table 3).

Table 2

One-way ANOVA test to compare the average salivary DNA fluorescence intensity among the groups

Group A Mean + SD	Group B1 Mean + SD	Group B2 Mean + SD	Group C Mean + SD	p-Value
3771135.5 ±	5369227.45 ±   1203353.92	2472118.8 ±   1210762.4	1524847.5 ±   874305.8	0.000

Abbreviations: ANOVA, analysis of variance; SD, standard deviation.

Table 3

Post-hoc Bonferroni test for multiple comparisons between the groups for salivary DNA

Dependent variable	Group	Group	Mean difference	p-Value
Salivary DNA	Control	Leukoplakia	−3844379.89	0.000
		OSMF	−947271.33	0.255
		OSCC	−2246287.94	0.000
	Leukoplakia	Control	3844379.89	0.000
		OSMF	2897108.55	0.000
		OSCC	1598091.95	0.007
	OSMF	Control	947271.33	0.255
		Leukoplakia	−2897108.55	0.000
		OSCC	−1299016.6	0.040
	OSCC	Control	2246287.94	0.000
		Leukoplakia	−1598091.95	0.007
		OSMF	1299016.6	0.040

Abbreviations: OSCC, oral squamous cell carcinoma; OSMF, oral submucous fibrosis.

Fig. 3 shows the scatter plot of discriminant score for group C and group A. From this analysis, it showed that the original cases and cross-validated cases were classified correctly with 90% of predictive accuracy. The classification is given in Table 4.

Table 4

PCA–LDA for salivary DNA fluorescence group C and group A

		VAR00001	Predicted group membership		Total
			1.00	2.00
Originala	Count	Normal	9	1	10
		OSCC	2	8	10
	%	Normal	90.0	10.0	100.0
		OSCC	20.0	80.0	100.0
Cross-validatedbc	Count	Normal	9	1	10
		OSCC	2	8	10
	%	Normal	90.0	10.0	100.0
		OSCC	20.0	80.0	100.0

Abbreviations: OSCC, oral squamous cell carcinoma; PCA, principal component analysis; LDA, linear discrimant analysis.

85.0% of original grouped cases correctly classified.

Cross-validation is done only for those cases in the analysis. In cross-validation, each case is classified by the functions derived from all cases other than that case.

85.0% of cross-validated grouped cases correctly classified.

Close

Fig. 3: Scatter plot showing the discriminant function of salivary DNA fluorescence between group C and group A.

Export to PPT

The discriminant function of group C and group B1 is shown in Fig. 4 and the classification is given in Table 5. This analysis showed that 95% of predictive accuracy as the original cases and cross-validated cases were correctly classified.

Table 5

PCA–LDA for salivary DNA fluorescence between group C and group B1

		VAR00001	Predicted group membership		Total
			1.00	2.00
Originala	Count	Normal	10	0	10
		LPA	1	9	10
	%	Normal	100.0	0.0	100.0
		LPA	10.0	90.0	100.0
Cross-validatedbc	Count	Normal	10	0	10
		LPA	1	9	10
	%	Normal	100.0	0.0	100.0
		LPA	10.0	90.0	100.0

Abbreviations: LPA, leukoplakia; PCA, principal component analysis; LDA, linear discriminant analysis.

95.0% of original grouped cases correctly classified.

Cross-validation is done only for those cases in the analysis. In cross-validation, each case is classified by the functions derived from all cases other than that case.

95.0% of cross-validated grouped cases correctly classified.

Close

Fig. 4: Scatter plot showing the discriminant function of salivary DNA fluorescence between group C and group B1.

Export to PPT

Fig. 5 shows the scatter plot of discriminant score for group C and group B2 and the classification is shown in Table 6. From this analysis, it is revealed that original and cross-validated cases were correctly classified with predictive accuracy of 65%.

Table 6

PCA–LDA for salivary DNA fluorescence between group C and group B2

				VAR00001	Predicted group membership		Total
1.00	2.00
Originala	Count	Normal	10	0	10
		OSMF	0	10	10
	%	Normal	100.0	0.0	100.0
		OSMF	0.0	100.0	100.0
Cross-validatedbc	Count	Normal	10	0	10
		OSMF	0	10	10
	%	Normal	100.0	0.0	100.0
		OSMF	0.0	100.0	100.0

Abbreviations: OSMF, oral submucous fibrosis; PCA, principal component analysis; LDA, linear discriminant analysis.

100.0% of original grouped cases correctly classified.

Cross-validation is done only for those cases in the analysis. In cross-validation, each case is classified by the functions derived from all cases other than that case.

100.0% of cross-validated grouped cases correctly classified.

Close

Fig. 5: Scatter plot showing the discriminant function of salivary DNA fluorescence between group C and group B2.

Export to PPT

Discussion

In spite of many advancements in the treatment for oral cancer, 5-year survival rate remains the same without any significant variation.21 Currently, many research projects are mainly done to detect OSCC and OPMD at earlier stage based on optical diagnosis with salivary biomarkers. These salivary biomarkers play a noninvasive role in the diagnosis and inspection of OSCC and OPMDs making it a most sensitive and specific screening method in early detection, staging, and prognosis.22 Molecular markers for the diagnosis of OSCC include initial changes in the cellular DNA, which result in altered mRNA transcripts, leading to altered protein contents.23 Literature search reveals many in vitro and in vivo studies done on fluorescence spectroscopy and imaging based on the fluorophores like NADH, FAD, proteins, and protoporphyrin. Udhyakumar et al11 studied about the normal and malignant tissues using native fluorescence and and time resolved fluorescence spectroscopy. They observed a statistically significant difference in the excitation spectra at 280nm between the normal and malignant tissues because of the fluorophore tryptophan. Vedeswari et al24 studied autofluorescence intensity of pre- and post-treated patients with OSMF and reported variations in the intensity of fluorescence at 385 and 440 nm intensity region corresponding to NADH. The average fluorescence spectrum of the post-treated OSMF mucosa had a lesser intensity around 385 nm and a higher intensity around 440 nm than that of the pretreated OSMF mucosa, thereby mimicking the normal oral mucosa. Raja Pappu et al12 studied about the in vivo native fluorescence spectroscopic characterization of oral tissue and saliva in normal patients and oral carcinoma patients and observed that diagnosis of cancer with the help of saliva has better sensitivity than tissue.12

DNA-based diagnosis is highly sensitive and accurate but isolation of DNA and analysis are tedious. Certain procedures like DNA microarrays and DNA amplification with polymerase chain reaction are highly laborious and expensive and time consuming. The intrinsic fluorescence of DNA is extremely weak because of its short decay times. Florescent bases can be used as a probe for DNA–protein interaction that can increase the quantum yield which in turn reveals the characteristics of the DNA.25 Hoechst 33258 fluorochrome, ethidium bromide, and PicoGreen dsDNA quantitation reagent are some of the probes that can be used to study DNA fluorescence. Ethidium bromide dye is widely used as a probe in spectrometric analysis that strongly binds with nucleic acids and enhances its fluorescence which is also reported in many literatures.2627 Yuvaraj et al in his study observed that the fluorescence emission of ethidium bromide and DNA complex is better than the emission of ethidium bromide alone.16 Based on these concepts, we structured this study to discriminate the OSCC and OPMDs with DNA fluorescence that is rapid, simple, and noninvasive and it can even identify a small fragment of DNA. Gallagher and Desjardins studied the quantification of DNA and RNA and mentioned that minimum DNA concentration of 1000 ng/mL is required for ethidium bromide assay.26

In this study, the salivary samples are excited at 480 nm and emission spectrum is recorded from 500 to 750 nm and the maximum peak obtained in a range from 600 to 610nm. When compared with group C, other groups showed higher intensity in an ascending order of group C (1.6 × 10⁶ cps)> group B2 (2.5 × 10⁶ cps) > group A (3.7 × 10⁶ cps) > group B1(4.9 × 10⁶ cps) as shown in Fig. 1. The maximum intensity is observed in LPA and OSCC because of the continuous cell proliferation that probably results in increased availability of DNA. The fluorescence intensity in OSMF is less than the OSCC but more than the normal, probably because of less availability of DNA. Quantification of DNA should have been done to validate these results. The results of our study are similar to the study done by Yuvaraj et al where they observed the excitation emission spectra of OSCC were higher than the normal.13 Here in this study, we also included the oral potentially malignant disorders that also showed a higher fluorescence emission than normal.

The test results were analyzed statistically that showed a significant difference between group C and group A (p = 0.00) and also between group C and group B1 (p = 0.00). The discriminant analysis also showed a predictive accuracy of 95% for oral LPA, 90% for OSCC, and 65% predictive accuracy for oral submucous fibrosis.

Though we achieved a statistically significant results, the sample size was small. Further studies are required with larger sample size. Quantification of the DNA was not done that is necessary to provide accurate amount of DNA required for fluorescence spectroscopy studies. This pilot study may elucidate the potential of screening significance of fluorescence spectroscopy of salivary DNA using ethidium bromide.

Conclusion

In conclusion, the present study revealed that salivary DNA characterization with the help of fluorescent spectroscopy can be used to screen oral cancer and oral potentially malignant disorders at the earliest. It is a rapid, noninvasive method, accurate and used for mass screening. This is the first of its kind to assess the fluorescence spectroscopic characteristics of salivary DNA with ethidium bromide dye in OPMDs and OSCC. Statistically significant difference was observed in OSCC and OPMDs from group C. However, further multicentric studies are required with more study samples, various age groups, different genders, different stages of OSMF, LPA, and OSCC, and various sites to validate our research findings.