Integrity and quantity of salivary cell‐free DNA as a potential molecular biomarker in oral cancer: A preliminary study

Abstract Background Differences in cell‐free DNA (cfDNA) fragments have been described as a valuable tool to distinguish cancer patients from healthy individuals. We aim to investigate the concentration and integrity of cfDNA fragments in saliva from oral squamous cell carcinoma (OSCC) patients and healthy individuals in order to explore their value as diagnostic biomarkers. Methods Saliva samples were collected from a total of 34 subjects (19 OSCC patients and 15 healthy controls). The total concentration of salivary cfDNA (scfDNA) was determined using a fluorometry method and quantitative real‐time polymerase chain reaction (qPCR). To evaluate the scfDNA quantity and integrity, qPCR targeting Arthobacter luteus (ALU) sequences at three amplicons of different lengths (60, 115, and 247 bp, respectively) was carried out. ScfDNA integrity indexes (ALU115/ALU60 and ALU247/ALU60) were calculated as the ratio between the absolute concentration of the longer amplicons 115 bp and 247 bp and the total scfDNA amount (amplicon 60 bp). Results The total scfDNA concentration (ALU60) was higher in OSCC than in healthy donors, but this trend was not statistically significant. The medians of scfDNA integrity indexes, ALU115/ALU60 and ALU247/ALU60, were significantly higher in OSCC, showing area under the curve values of 0.8211 and 0.7018, respectively. Conclusion Our preliminary results suggest that scfDNA integrity indexes (ALU115/ALU60 and ALU247/ALU60) have potential as noninvasive diagnostic biomarkers for OSCC.

deaths worldwide in 2020. 1 There is an urgent need to develop noninvasive biomolecular approaches for improving early detection and therapeutic success rates. 2 Liquid biopsies have emerged as a potential noninvasive source of tumor-related biomarkers. 3 The analysis of cell-free DNA (cfDNA) has aroused great interest as a noninvasive biomarker. 4 The cfDNA comprises both nuclear and/or mitochondrial double-stranded DNA released from the cells into the physiological body fluids, such as blood, saliva, urine, cerebrospinal fluid, or pleural fluid. 5 The structural composition of cfDNA has been related to its origins including apoptosis, necrosis, or active release. Apoptotic cells secreted small and uniform DNA fragments, with a length around 167 bp as a result of caspaseactivated DNase (CAD) cleavage at internucleosomal chromatin regions. Conversely, necrotic cells release high-molecular-weight DNA due to random and incomplete genomic DNA fragmentation by DNases. 6 The determination of cfDNA size distribution in plasma from cancer patients showed the presence of short and long DNA fragments derived from both apoptotic and necrotic cells, 7 while plasma cfDNA from healthy individuals is only derived from apoptotic cells. 8 Several studies have reported the presence of elevated levels of longer cfDNA fragments in serum, plasma, or urine from different types of tumors. Therefore, the DNA integrity index calculated as the ratio of the quantity of long to short DNA fragments has been described as a promising tumor biomarker. 9,10 Salivary cfDNA (scfDNA) has emerged as a potential biomarker in cancer 11,12 ; however, there is little information about its profile.
Since saliva is the biological fluid of the oral cavity, scfDNA fragmentation could reflect a different origin under physiological and pathological conditions, representing a noninvasive biomarker for tumor detection. Our proposed hypothesis is that scfDNA fragments from OSCC patients are longer than those of healthy individuals due to tumor necrosis. Therefore, this study aimed to investigate whether the total scfDNA quantity and DNA integrity could be used for discriminating OSCC patients from healthy individuals.

| Saliva collection and DNA extraction
All the participants were asked to refrain from eating, drinking, smoking, and using oral hygiene products for at least 1 h prior to sample collection. Unstimulated saliva samples were collected using

| Bacterial scfDNA quantification assay
To determine the percentage of bacterial scfDNA, we carried out a qPCR-based assay using microbial DNA qPCR Assay kit (Qiagen, Ref. No. 3520510), which is designed to detect bacterial 16S rRNA gene.
The reaction mixture for each direct qPCR contained 12.5 μL microbial qPCR mastermix, 1 μL microbial DNA qPCR Assay, and 11.5 μL of scfDNA in a total reaction volume of 25 μL. In addition, for the qPCR, a positive PCR control, a negative control, and a microbial DNA positive control were included following the manufacturer's protocol. A standard curve was performed with serial dilutions (20 ng/μL to 0.02 ng/μL) of ZymoBIOMICS™ microbial community DNA standard (Zymo, D6305). Thermal cycling was conducted on the QuantStudio™ 3 Real-Time PCR system (Life Technologies, Foster City, CA, USA) using the following parameters: initial denaturation at 95 C for 10 min followed by 40 cycles of amplification at 95 C for 15 s and 60 C for 2 min.

| qPCR of ALU fragments
The quantity of the cfDNA in saliva was evaluated by qPCR targeting

| scfDNA integrity index
Each integrity index was calculated as the ratio between the absolute concentration of the longer amplicons (amplicon 115 and 247 bp) and the total scfDNA amount (amplicon 60 bp). Then, scfDNA integrity was considered as ALU115/ALU60 and ALU247/ALU60. In addition, we evaluated the fraction of scfDNA fragments with lengths varying from 60 to 115 bp and from 115 to 247 bp by subtracting the absolute concentration of the longer fragment to that of the shorter one.
Also, longer scfDNA fragments (≥247 bp) were calculated. These results were normalized on the total scfDNA amount and expressed as percentages.

| Statistical analysis
The sample size was calculated using G*Power version 3.   Table 1. According to TNM stage, 12 patients were at Stages I and II and 7 at Stages III and IV. Moreover, the histological grade was well/moderate in 18 OSCC patients.
Although scfDNA concentration was increased in cancer patients, no

| scfDNA integrity indexes by qPCR
The median salivary DNA integrity ALU115/ALU60 was significantly Analyzing the amount of scfDNA fragments by qPCR (Figure 2), we observed that 62% of the total scfDNA in healthy controls comprised fragments with lengths ranging from 60 to 115 bp. However, this percentage was significantly lower in OSCC patients (p = 0.0016).
Regarding the percentage of fragments comprised between 115 and 247 bp, no significant differences were observed between OSCC patients and healthy individuals, representing around 35% of total scfDNA fragments (p = 0.8642). On the contrary, the percentage of longer fragments (≥247 bp) with respect to total scfDNA quantity was significantly higher in OSCC patients (35%) than in healthy controls (25%) (p = 0.048).
The scfDNA integrity index ALU247/ALU60 also yielded 83.33% sensitivity and 73.33% specificity at the cut-off level of >0.2661 ng/ml.

| DISCUSSION
Liquid biopsies based on tumor biomarkers represent an opportunity for improving early cancer detection. Among the various liquid biopsy samples, saliva has attracted great interest in identifying tumorrelated biomarkers since its collection is easy, noninvasive, and inexpensive. 14 Like in blood, in addition to the cellular fraction, saliva comprises an extracellular fraction that reflects local and systemic biological information. 15,16 As we previously noted, scfDNA could be released by apoptosis, necrosis, or active secretion from malignant and nonmalignant cells. 12 Recently, evidence has demonstrated the feasibility to detect tumor-derived circulating DNA in saliva. 11 Nevertheless, the potential clinical role of scfDNA concentration and integrity for OSCC diagnosis was not already established.
To the best of our knowledge, this is the first study that investigates the potential application of cfDNA concentration and cfDNA integrity using saliva for discriminating OSCC patients from healthy controls. The cfDNA concentration has been shown to act as a diagnostic biomarker due to its ability to differentiate cancer patients from controls subjects. 17,18 Opposite to other tumor types, few studies analyzed the quantity of cfDNA in oral cancer. [19][20][21] In this preliminary study, we have observed using fluorometry and ALU60-qPCR that scfDNA levels were higher in OSCC patients compared to healthy controls, but this increase was not statistically significant, probably because of the limited cohort analyzed. Similarly, Ding et al. analyzed the concentration of scfDNA in lung cancer by LINE1-qPCR, but no significant differences were observed among the groups. 22 19 Based on our findings, the total concentration of scfDNA may have limited diagnostic value for oral cancer detection and not provides information about the mechanism of releasing cfDNA into saliva.
In addition to the total scfDNA concentration, the scfDNA fragment size may be used to differentiate cancer patients from healthy controls. In the last few years, several studies have focused on the analysis of cfDNA integrity as a promising tumor biomarker. 9,23,24 Since apoptosis is the main source of uniform and small fragments of circulating cfDNA in healthy individuals, the presence of longer cfDNA fragments has been referred to as a marker of necrotic tumor cell death. Therefore, the ratio of longer to shorter fragments, known as cfDNA integrity, reflects the relative amount of nonapoptotic cell death. 9 ALU-targeting qPCR represents the most common method for assessing DNA integrity, specifically ALU115 and ALU247 are the sequences analyzed to determine DNA originating from apoptotic and necrotic cell death, respectively. 9 represents a potential diagnosis tool that should be further explored in this type of tumor. Interestingly, our study revealed for the first time also the presence of scfDNA of high molecular weight in healthy individuals indicating that nontumoral scfDNA is innately a mixture of longer and shorter DNA fragments. We hypothesized that the global profile of scfDNA could be conditioned by the content of salivary exosomes and the mechanisms of epithelial cell death modulated by the oral microbiota. 29,30 However, our study is not exempt from shortcomings. Our findings should be validated in longitudinal studies with larger sample sizes and including a control group of oral potentially malignant disorders for a better understanding of the potential clinical impact of scfDNA, not only for diagnosis but also for prognosis and monitoring the response to therapy in OSCC. Also, future studies should investigate the correlation between cfDNA integrity in saliva and plasma to determine the effectiveness of each biofluid in early-and late-stage cancer. To opti-