IntroductionOral cancer is one of the most common types of head and neck malignancy, with oral squamous cell carcinoma (OSCC) accounting for nearly 90% of all oral cancer cases. OSCC typically arises from squamous epithelial cells lining the inner oral mucosa and is known for its aggressive nature. The five-year survival rate varies significantly, ranging from 80% to 90% when diagnosed at an early stage, but drops to approximately 30% in advanced stages. Although the etiology of OSCC is multifactorial, key risk factors include tobacco use (smoking or chewing), excessive alcohol consumption, and betel nut chewing, with a higher prevalence observed in males1,2. The global disease burden of OSCC varies geographically, with significantly higher incidence rates reported in Asian countries3. A recent national-level study identified India as a major contributor to the global burden, accounting for nearly one-third of all oral cancer cases worldwide4. Clinically, OSCC presents with symptoms such as non-healing ulcers and oral pain in the initial phase, progressing to enlarged oral masses, dysphagia, and odynophagia in more advanced stages. Currently, the gold standard for OSCC diagnosis involves clinical oral examination followed by histopathological evaluation of the tissue obtained through biopsy5. However, tissue biopsy methods are invasive, tumour molecular heterogeneity may not be captured, and timely screening and monitoring of the therapeutic response may be challenging.Advances in oral cancer diagnostic research have increasingly focused on the development of minimally invasive or non-invasive molecular biomarker-based liquid biopsies to improve diagnostic accuracy, offer comprehensive insights into tumour biology, and enable real-time monitoring5,6,7. Various biological fluids such as serum, saliva, urine, cerebrospinal fluid, and prostatic fluid are being explored for tumour-derived molecule profiling, among which serum and saliva are the most extensively studied and clinically reliable for developing diagnostic platforms for oral cancer8,9. Serum contains a diverse pool of tumour-derived molecules released into the circulation, including genomic, proteomic, transcriptomic, epigenetic markers, and circulating tumour cells (CTCs), offering a systemic snapshot of tumour dynamics for various cancers including OSCC. In contrast, saliva, owning to its direct contact with the tumour site in OSCC, is enriched with locally derived biomarkers such as RNA, proteins, enzymes, and extracellular vesicles like exosomes10,11. Recent studies have explored both serum and saliva for OSCC detection, with growing interest in saliva owing to its non-invasive collection and tumour proximity. For instance, several studies have identified serum biomarkers such as chemerin, miR-31-5p, and IL612,13,14, while others have demonstrated the clinical relevance of DNA, RNA, and protein analysis in saliva15,16,17. However, most of these studies were conducted on small cohorts, and there is a lack of population-specific research in India, despite its high OSCC incidence.Exosomes are small extracellular vesicles ranging from 30 to 200 nm in diameter, known to play key roles in cancer pathophysiology through their cargo, and are stably enclosed within a lipid bilayer, which includes mRNAs, miRNAs, and proteins18. In recent years, there has been a growing interest in developing exosome-based platforms for cancer diagnostics and therapeutics owning to their demonstrated efficiency19,20,21,22. Consequently, few studies have focused on identifying exosomal biomarkers in serum and saliva for OSCC due to their high diagnostic sensitivity and specificity. Notably, salivary exosomal proteins such as AMER3, LOXL2, and AL9A1 shown potential as OSCC biomarkers, while serum exosomal miRNAs like miR-155 and miR-21 may serve as valuable diagnostic and prognostic indicators11,23. However, the diagnostic potential of exosomes in OSCC remains in its infancy, and to the best of our knowledge, no studies have explored exosomal mRNAs in this context.In this study, we aimed to evaluate nine mRNAs including MMP9, IL1, IL6, IL8, TNF-α, Chemerin, OAZ1, SAT, and S100P in both serum and salivary exosomes from histopathologically confirmed OSCC patients, to gain a comprehensive understanding of the tumour microenvironment (TME). These genes were selected based on their established roles in key hallmarks of OSCC, including proinflammation, tumour cell proliferation, and invasion. The proinflammatory cytokines IL1, IL6, IL8, and TNF-α are known to promote a tumour-supportive microenvironment, with elevated levels reported in OSCC patients24,25. OAZ1, SAT, and S100P are involved in regulating cancer cell growth and survival, whereas MMP9 and Chemerin play critical roles in promoting tumour invasion and metastasis, with their expression shown to correlate with histological grade26,27,28,29. Thus, the altered enrichment of these molecules in exosomes reflects the tumour status and aggressiveness, and analysing them through liquid biopsy presents a promising approach for diagnosing and monitoring of OSCC progression. Our findings demonstrated that salivary exosomal biomarkers possess greater diagnostic potential than those derived from serum, showing significant correlation with tumour grade and lymph node metastasis. Based on this pilot analysis, we propose a salivary exosomal biomarker panel, particularly TNF-α in combination with OAZ1, as a promising molecular diagnostic tool for OSCC, with potential relevance for the Indian population.Material and methodsSample size estimationThis preliminary study aimed to explore exosome gene panels derived from serum and saliva as potential biomarkers for OSCC, establishing a proof of concept by analysing expression patterns and correlation trends before conducting large-scale studies. The sample size was estimated based on the standard formula for sensitivity analysis, assuming a 95% confidence interval, expected sensitivity of 98%, specificity of 92.1%, and disease prevalence of 20%. An acceptable margin of error (precision) of 10% (d = 0.10) was considered. Based on this, at least 38 samples were required for the study to achieve statistically meaningful data.Patient recruitment and sample selectionA total of 40 histopathologically confirmed OSCC patients were recruited from the Apollo Cancer Institute, Apollo Hospitals, Jubilee Hills, Hyderabad, between May 2023 and September 2024. Exclusion criteria included patients with recurrent disease, those unwilling to provide informed consent, individuals with a history of other malignancies, and patients with sexually transmitted diseases (STDs). Age-matched healthy controls (40 individuals) were selected following an oral cavity examination by a specialist to confirm the absence of any oral pathological lesions that might influence saliva composition. All participants, including OSCC patients and healthy controls, voluntarily agreed to participate in the study and provided informed consent. Demographic parameters, including age and gender, were recorded during sample collection, whereas clinical data such as tumor grade, site, and lymph node involvement were retrieved from medical records following histopathological evaluation. The study protocol was reviewed and approved by the Institutional Ethics Committee at Apollo Hospitals (IEC No. AHJ-C-S-006/02–25). Participants were informed about the study objectives and the option to withdraw at any time without consequences. All participants provided informed consent, and all serum and saliva samples were anonymized with code labels, ensuring that personal identifiers were not included in the data. All experiments in this study were performed in accordance with the relevant guidelines and regulations.Collection of serum and saliva samplesPaired blood and saliva samples were collected from all OSCC patients and healthy controls (HC) enrolled in the study. For each individual, 5 mL of peripheral blood was drawn into blood collection tubes (BD biosciences, USA). After allowing the samples to rest at room temperature for 1 h, they were centrifuged at 2000 rpm for 10 min at 4 °C to separate the serum. The resulting supernatant was carefully transferred into 1.5 mL microcentrifuge tubes and subjected to a second centrifugation at 3000 rpm for 10 min at 4 °C to remove any residual cellular debris. The purified serum was then aliquoted and stored at -80 °C until further use for exosome isolation. In parallel, approximately 1 mL of unstimulated saliva was collected using the Norgen Biotek Saliva Collection and Preservation Kit (Norgenbiotek, Canada). Participants were instructed to refrain from eating or drinking for at least one hour prior to saliva collection to ensure sample consistency. Following collection, saliva samples were centrifuged at 1000 rpm for 10 min at 4 °C to eliminate any debris. The clarified saliva was then stored at -80°C until exosome isolation. All samples were processed within 3 h of collection to preserve their integrity. Notably, blood and saliva were collected from the same individual on the same day to ensure sample comparability. Informed consent was obtained from all participants prior to sample collection, and individuals who were unwilling to provide consent were excluded from the study without prejudice.Isolation of exosomes from serum and salivaExosomes were isolated from both serum and saliva samples using commercial Total Exosome Isolation Kits (Thermo Fisher Scientific, USA), following the manufacturer’s protocols. Frozen serum and saliva samples were thawed on ice and centrifuged at 2000 × g for 30 min at 25 °C to remove any residual cellular debris. The clarified supernatants were then transferred to fresh 1.5 mL microcentrifuge tubes for exosome isolation. For serum exosome isolation, 500 µL of the sample was mixed with 100 µL of the Total Exosome Isolation Reagent for serum (0.2 volume ratio). The mixture was gently vortexed and incubated at 4 °C for 30 min. For saliva, 250 µL of the sample was first diluted with 250 µL of phosphate-buffered saline (PBS), followed by the addition of 250 µL of the Total Exosome Isolation Reagent for other body fluids (0.5 volume ratio). The mixture was incubated at 4 °C for 60 min. Following incubation, both serum and saliva mixtures were centrifuged at 10,000 × g for 10 min at room temperature to pellet the exosomes. The resulting exosome pellets were resuspended in 1X PBS and stored at -80 °C until further analysis.Identification of exosomesExosomes isolated from serum and saliva were comprehensively characterized using biochemical, biophysical, and molecular methods to evaluate their biochemical composition, size, surface charge, morphology, and specific protein markers, in accordance with the MISEV2024 guidelines, ensuring their structural integrity and quality30.Biochemical characterizationBiochemical characterization included quantification of total protein and lipid content. The protein concentration of the isolated particles was measured using the Bicinchoninic Acid (BCA) assay, while the sulfo-phospho-vanillin (SPV) assay was employed to estimate the lipid content, following established protocols. From these measurements, the protein-to-lipid ratio was calculated, providing insight into the purity of isolated vesicles22.Biophysical characterizationNanoparticle Tracking Analysis (NTA) was performed using the ZetaView® x30 system (Particle Metrix, Germany) to determine the size distribution and zeta potential of the particles. This helped confirm that the isolated vesicles were within the expected exosomal size range and exhibited surface charges consistent with stable colloidal particles. Transmission Electron Microscopy (TEM) was used to examine the morphology of the particles. For this, 50 µL of the exosome suspension was loaded onto carbon-coated copper grids and air-dried. The grids were then visualized under the TEM, where the vesicles appeared as round shaped structures, confirming successful exosome isolation31.Molecular characterizationWestern blotting analysis was performed to analyse the protein expression of exosome specific markers in the isolated particles. In brief, equal amounts of total protein (20 µg) from serum- and saliva-derived exosomes from both OSCC patients and HC were used. The samples were lysed using Mammalian Protein Extraction Reagent (M-PER), denatured by adding Laemmli buffer, and incubated at 95 °C for 10 min. Proteins were resolved on a 10% SDS-PAGE gel and transferred onto nitrocellulose membranes. After blocking the membranes with 5% BSA to prevent nonspecific binding, they were incubated overnight at 4 °C with primary antibodies against human CD81, CD9, TSG101, and HSP70 (Abcam, UK). Following TBST washes, membranes were incubated with HRP-conjugated secondary antibodies for 1 h at room temperature. Signal detection was performed using a Bio-Rad ECL kit, and the bands were visualized using the ChemiDoc Imaging System32. Exosomal mRNA extraction and expression profilingTotal RNA was extracted from isolated exosomes using TRIzol™ Reagent (Invitrogen, USA) in accordance with the manufacturer’s protocol. The concentration and purity of the extracted RNA were assessed using a biospectrophotometer, with 1 µL of each sample measured. Subsequently, 500 ng of total RNA was reverse-transcribed into complementary DNA (cDNA) utilizing the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, USA), following the supplier’s instructions. For gene expression analysis, nine target genes including IL1-β, IL6, IL8, TNF-α, OAZ1, SAT, S100P, MMP9, and Chemerin, were selected. Specific primers for these genes were procured from Bioserve Biotechnologies (India) PVT. Ltd and their sequences detailed in Supplementary Table 1.To enhance detection sensitivity, 100 ng of cDNA from each sample underwent pre-amplification through 15 PCR cycles using High-Fidelity PCR Master Mix (DXBIDT Laboratories, India) in a thermal cycler (Bio-Rad T100 thermal cycler). The cycling conditions as follows: initial denaturation at 98 °C for 30 s, denaturation at 98 °C for 10 s, annealing at 58–62 °C (gene-specific) for 30 s (Supplementary Table 1), extension at 72 °C for 60 s, and final extension at 72 °C for 10 min. The resulting pre-amplified products were diluted into 1:4 ratio with nuclease-free water. Subsequently, 1 µL of the diluted product was subjected to two-step quantitative real-time PCR (qRT-PCR) over 30 cycles using gene-specific primers and High-Throughput qPCR Master Mix (DXBIDT Laboratories, India) on a Bio-Rad Opus 96 Real-Time PCR System. Detailed PCR conditions are provided in Supplementary Table 1. Gene expression levels were quantified using the comparative Ct (ΔΔCt) method, with B2M serving as the internal control gene33,34.Statistical analysesStatistical analyses were conducted using IBM SPSS Statistics (Version 24, Chicago, USA) and GraphPad Prism (Version 8, Boston, USA). Descriptive statistics, including mean, meadian, and standard deviation (SD), were calculated for continuous variables to summarize the data. Chi-squire test was used to study the association between gender and groups. To assess differences between HC and OSCC groups in saliva and serum samples, the non-parametric Mann-Whitney U test was employed, given its suitability for comparing two independent groups without assuming normal distribution, as the data homogeneity has significant deviation. Accordingly, the fold change was represented as the Median with 95% confidence interval (CI). For evaluating gene expression differences across multiple groups, one-way analysis of variance (ANOVA) was utilized, followed by the Least Significant Difference (LSD) post hoc test to identify specific group differences. In instances where assumptions of normality and homogeneity of variances were not met, the Kruskal–Walli’s test, a non-parametric alternative to ANOVA, was applied. To determine the diagnostic performance of identified gene markers for OSCC, receiver operating characteristic (ROC) curve analyses were performed, calculating the area under the curve (AUC) along with 95% confidence intervals to assess sensitivity and specificity. The diagnostic effectiveness was assessed by Youden index. Furthermore, a combined panel of top-performing individual genes from each category was proposed by computing the cumulative AUC, aiming to enhance diagnostic accuracy using stepwise discriminant function analysis. Throughout the study, a p-value of less than 0.05 was considered statistically significant.ResultsDemographic and clinical profiling of the selected cohortThis study included 80 individuals, comprising 40 patients diagnosed with OSCC and 40 HC. The mean age of OSCC patients was 47.60 ± 11.02 years, while the age-matched control group had a mean age of 42.8 ± 8.3 years. In terms of gender distribution, the OSCC group predominantly consisted of males (92.5%, n = 37), with only 7.5% females (n = 3), reflecting the known higher incidence of OSCC among males. The healthy control group also had a male predominance (72.5%, n = 29), though with a slightly higher proportion of females (27.5%, n = 11) compared to the OSCC group. Tumor localization in the OSCC cohort revealed that the most common sites were the left side of the tongue, right buccal mucosa, and left border of the tongue (n = 3 each). These were followed by the left buccal mucosa, right border of the tongue, left lower alveolus, and middle third mandible (n = 2 each), while the remaining 23 patients presented with lesions at various other intraoral sites, reflecting the heterogeneity of tumor localization in OSCC. Histopathological grading showed that 40% of tumors were classified as grade I (n = 16), another 40% as grade II (n = 16), and 20% as grade III (n = 8), indicating that the majority of patients had moderately to well-differentiated tumors. TNM staging further revealed that 21 patients (52.5%) had evidence of nodal involvement or metastasis, while 19 patients (47.5%) showed no lymph node involvement (Table 1). These demographic and clinical characteristics of the selected cohort highlight the diversity in tumor location, histological grade, and metastatic status in OSCC, providing a foundation for the diagnostic investigations.Table 1 Demographic and clinical characteristics of oral cancer patients and HC included in the study.Full size tableExosome characterization revealed successful isolation and integrity of serum and saliva-derived exosomesExosomes isolated from serum and saliva of OSCC patients and HC were comprehensively characterized using biochemical, biophysical, and molecular approaches to confirm their integrity and purity (Fig. 1). Initial biochemical assessment of the isolated particles from both serum and saliva serum showed the higher levels of total protein in OSCC patients, when compared to HC (Fig. 1A and B). However, lipid concentration of the serum derived particles was not shown significance between HC and OSCC, while saliva derived particles had a comparable difference with heightened levels in OSCC patients (Fig. 1B). Additionally, the protein-to-lipid ratio of the vesicles was estimated to be 1–2, supporting the bilayered membrane structure characteristic of exosomes (Fig. 1C). NTA analysis demonstrated that the isolated particles are in the average size of 150 to 200 in diameter with a negative zeta potential, consistent with the size and surface charge typically associated with exosomes (Fig. 1D and E). TEM analysis further determined their spherical morphology and most vesicles measured below 100 nm (Fig. 1F), indicating the confirmation of exosomes. Further, western blot analysis of exosomal lysates revealed the presence of classical exosome-specific protein markers CD81, CD9, HSP70, and TSG101 in both serum- and saliva-derived samples (Fig. 1G), validating the exosomal origin of the isolated particles. These findings confirm the successful isolation and structural integrity of exosomes from both biofluids, supporting their utility for subsequent molecular and diagnostic applications.Fig. 1Characterization of exosomes isolated from serum and saliva of HC and OSCC patients. (A) Protein concentrations of serum and saliva derived exosomes quantified using BCA assays from HC and OSCC patients revealed distinct biomolecular profiles. (B) Analysis of lipid levels in serum and saliva derived exosomes using Sulfo-phospho-vanillin assay (SPVA) (C) The protein-to-lipid ratio was approximately 2 for serum exosomes and 1 for salivary exosomes, consistent with the characteristic lipid bilayer structure of exosomes and indicating good sample quality. (D) NTA analysis showed an average particle size of 150–200 nm, representing the overall size distribution of isolated particles. (E) Zeta potential measurements confirmed a net negative surface charge of the particles, supporting their colloidal stability. (F) TEM analysis revealed the round shaped morphology with