Introduction

Head and neck squamous cell carcinoma (HNSCC) treatment is multimodal, including surgery, radiotherapy (RT), and chemotherapy (CT) [1,2,3]. Despite the notable benefit of RT, this treatment modality can be associated with several acute and chronic side effects that affect non-target tissues present within the radiation field [4]. In this context, oral mucositis (OM) is a highly prevalent acute toxicity that affects up to 80 to 100% of patients undergoing RT for oral and oropharyngeal cancer. According to the literature, chemoradiotherapy (CRT) protocols can increase the frequency and clinical severity of OM [4, 5]. Besides the effect of combining CT with RT, other factors that have been correlated as risk factors for RT-induced OM are poor oral hygiene, young age, female gender, poor renal function, poor nutritional status, total cumulative radiation dose, smoking, and decreased salivary secretion [6].

OM has a complex pathophysiology characterized by persistent oral mucosal ulcers associated with severe pain and reduced oral functions, such as swallowing, speech, taste, and chewing. These consequent morbidities can, in severe cases, lead to the interruption of cancer treatment and, consequently, cause tumor progression and negatively impact on survival rates [1,2,3, 7]. These comorbidities may also increase the cost of healthcare due to increased hospitalization, expensive drugs such as opioids or growth factors (e.g., keratinocyte growth factor (KGF)), and the use of probes for nasogastric feeding [8,9,10,11].

Improved molecular characterization of oncologic treatment–induced OM will not only aid in an improved understanding of the disease process but also enable development of robust prognostic biomarkers to improve effective interventions. A study by Jehmlich et al. [12] outlined the salivary proteomic profile of HNSCC cancer patients and noted the expression of 48 proteins related to an increased risk for the development of OM. Considering the above-mentioned information and the need for better understanding of the proteins associated with OM development, the aim of the present study was to characterize two proteins, alpha-1-antitrypsin (A1AT) and macrophage migration inhibitory factor (MIF), and correlate them with the risk of developing severe OM post-oncologic treatment.

Materials and methods

Study design

This was a single-center cohort study designed to evaluate the salivary proteomic profile of HNSCC patients submitted to RT or CRT protocols at the Instituto do Câncer do Estado de São Paulo (ICESP, Brazil) from January 2011 to February 2018. This study was approved by the Ethics Committee of the School of Medicine, University of São Paulo, Sao Paulo, Brazil (Protocol # 2.647.153). Written informed consent was obtained from all participants and the study was conducted in accordance with the Declaration of Helsinki.

Inclusion criteria

This study included oral and oropharynx cancer patients who underwent RT or CRT protocols (with or without previous surgery) using a 6MV linear accelerator and 3-dimensional conformal or intensity-modulated radiation therapy technique (IMRT) (Synergy Platform, Elekta AB, Stockholm, Sweden). The target radiation volumes encompassed the primary site and areas of regional lymph nodes at risk and received cumulative doses that ranged from 60 to 70 Gy. We applied the recommendations for treatment planning and constraints for organs at risk, as previously reported [13,14,15].

All included patients completed the institutional dental conditioning protocol prior to beginning RT. Besides routine oral care, all patients were submitted to the standard-of-care daily photobiomodulation (PBM) protocol for prevention of OM as per our institutional (Dental Oncology Service, ICESP, Brazil) protocol [11]. Finally, all patients needed to have complete demographic and clinicopathological data available on electronical medical charts, including gender, age, tumor location, clinical cancer stage (according to the American Joint Committee on Cancer Staging System, 7th edition [16]), cancer treatment modalities, and information regarding weekly dental follow-up during RT and OM outcomes.

Exclusion criteria

Patients who missed one or more RT or PBM sessions were considered to have received incomplete treatment and were excluded from the study.

Oral mucositis assessment

A trained dental surgeon conducted OM grading using the National Cancer Institute Common Terminology Criteria for Adverse Events (NCI CTCAE, Version 4.0, 2010 [17]) on the last day of each week of treatment (day 5, 10, 15, 20, 25, 30, and 35) as a part of standard of care. The highest OM grade developed throughout RT was used for comparison with the salivary proteomic profiles. Patient self-reported OM pain was recorded using the visual analogue scale (VAS) with scores ranging from 0 to 10, where 0 represents no pain and 10 is the highest level of pain. Assessments were recorded at the end of each week of RT and mean VAS values throughout RT were noted.

Saliva collection and preparation

Saliva samples were obtained voluntarily and with signed consent from HNSCC patients. Whole saliva collection was performed immediately prior to the beginning of RT. Individuals first rinsed their mouths with 5 mL of drinking water and then harvested the saliva into a sterile recipient. Saliva samples were aliquoted in 2-mL tubes and immediately frozen at − 80 °C for long-term storage.

Whole saliva protein preparation

The preparation of saliva samples was performed as previously described [18, 19]. Briefly, saliva was first centrifuged for 5 min at 1500g at 4 °C to remove intact cells and debris. A volume of 100 μL of whole saliva was then used in the protein extraction procedure by homogenizing with 100 μL of urea buffer (100 mM Tris-HCl pH 7.5, 8 M urea, and 2 M thiourea) containing the complete Mini Protease Inhibitor Cocktail (Roche, Auckland, New Zealand), 5 mM EDTA, 1 mM PMSF, and 1 mM DTT. Samples were sonicated for 10 min and centrifuged at 10,000g for 5 min. Total protein was quantified using a Bradford assay kit (Bio-Rad, São Paulo, Brazil).

Mass spectrometry and data analysis

The proteotypic peptides and their respective transitions of the proteins alpha-1-antitrypsin (A1AT) and macrophage migration inhibitory factor (MIF) were selected based on the criteria described [20, 21]: MIF_HUMAN (ASVPDGFLSELTQQLAQATGK, m/z 721.04, +3 and its respective transitions [y6] m/z 575.31+; [y4] m/z 376.21+; [y2] m/z 204.13+); A1AT_HUMAN (LQHLENELTHDIITK, m/z 451.74, +4; and its respective transitions [y2] m/z 248.16+; [y8] m/z 470.77++; [y7] m/z 414.23++ and EINDYVEK, m/z 505.24,+2 and its respective transitions [y6] m/z 767.35+; [y5] m/z653.31+; [y4] m/z 538.28+).

Three proteotypic peptides were purchased as crude heavy-isotope-labeled peptide standards (Thermo Fisher Scientific). The stable isotope-labeled peptides (SIL) were synthesized with heavy isotopes on lysine (+8 Da), localized at the C-terminal of the peptide (Thermo Fisher Scientific). Three transitions were monitored for the light and heavy counterparts of each peptide. Eight or nine peptides with their respective transitions of the internal retention time standard (Pierce™ Peptide Retention Time Calibration Mixture, Thermo Fisher Scientific) were monitored as a control for retention time shifts in liquid chromatography.

In order to avoid bias in the analyses, samples were randomized using the software R (v3.4.0) and separated in blocks with each sample analyzed in triplicate for SRM analysis. All samples were assessed within each block in different orders to further reduce systematic bias.

Samples were analyzed on a Xevo TQ-XS triple quadrupole mass spectrometer (Waters, Milford, MA, USA) equipped with an electrospray ion source (Ion Key, Waters, Milford, MA, USA) with MassLynx software (version 4.2), as described by Carnielli et al. [19]. The data analysis was manually performed in Skyline.

Statistical analysis

Demographic data, clinicopathological features, and clinical outcome results were grouped into a spreadsheet for descriptive statistical analyses based on mean, median, standard deviation, and proportion values. For the SRM analysis, the comparison of the levels of the monitored peptides between the patients who developed grades 1 or 2 OM (M1) and those who developed grades 3 or 4 OM (M2) was performed using a Wilcoxon Mann-Whitney test (not log-transformed data) with the significance level set at α = 0.05.

Results

Patient characteristics

A total of 27 patients were included. The major clinicopathological features, including age, gender, tumor site, clinical staging, treatments, and RT modality, are summarized in Table 1. The mean age was 57 years (range 20–81 years) and 24 (88.9%) were males. The most frequent primary tumor site was the lateral border of the tongue (12 cases, 44.5%) followed by oropharynx (10 cases, 37%), and the floor of the mouth (2 cases, 7.4%). Most patients (n = 21, 77.8%) were diagnosed with an advanced stage of disease (III/IV), and none of the 27 (100%) patients had their surgery at the time of saliva collection (presented with visible oral cancer lesions). All included patients received complete and uninterrupted RT and 19 (70.4%) received concomitant CRT protocols. Regarding the RT technique, 20 (74.1%) of patients were treated by 3-dimensional conformal RT; 7 (25.9%) with IMRT, with an overall mean dose of 67.7 Gy (ranging from 66 to 68.3 Gy).

Table 1 Clinicopathological features of 27 patients included in the study
Full size table

Oral mucositis assessment

The majority of the patients (n = 24, 88.9%) developed OM during the treatment period, even when performing PBM protocol. During the first 2 weeks of treatment (days 5 to 10), OM grades varied from 0 to 2. Grade 3 lesions started to develop from the third week of treatment (day 15), while grade 4 lesions developed from the 5th week (day 25) of treatment (Fig. 1). Three (11.1%) patients did not develop OM throughout the treatment. From the patients that developed OM, a total of 14 (51.9%) patients developed minimal OM (grade 1 (3/11.1%) or 2 (11/40.8%)) and 10 (37%) patients developed severe OM (grade 3 (9/33.3%) or 4 (1/3.7%)) as the highest grade throughout the treatment. The mean highest reported VAS throughout the treatment was 3.22 (range 0 to 9) and the highest level of OM-related pain was observed on the last week of radiotherapy with a mean VAS of 1.6. Data regarding OM grades and mean VAS per week of radiotherapy are summarized in Table 2.

Fig. 1
figure 1

Representative clinical images of subjects in the present study who developed oral mucositis lesions. a Subject presenting with grade 1 OM demonstrating erythema on the right buccal mucosa. b Subject presenting with grade 2 OM demonstrating small, punctuate ulcers covered with a pseudomembrane. c Subject presenting with grade 3 OM demonstrating confluent ulcers with pseudomembrane affecting the lateral border, dorsum, and ventral surface of the tongue. d Subject presenting with grade 4 OM demonstrating deep ulcers and bleeding, note the crust formations on the upper and lower lip

Full size image
Table 2 Oral mucositis grades and mean VAS scores per radiotherapy weeks
Full size table

Salivary protein assessment

Targeted proteomic analysis was performed for the samples of the 27 patients. Figure 2 demonstrates, individually, the light/heavy intensity ratio (not log transformed) of two proteins, MIF and A1AT, between M1 and M2 saliva samples. We demonstrated that alpha-1-antitrypsin and macrophage migration inhibitory factor showed a statistically significant correlation between patients who developed severe grades 3 or 4 OM (p < 0.02 and 0.04, respectively) as compared with patients who developed grades 1 or 2 OM. The increase in A1AT and MIF abundance was 2.97- and 21.4-fold, respectively, in grades 3 or 4 subjects compared with patients who developed low grades 1 or 2 or no OM.

Fig. 2
figure 2

The graph demonstrates individually the L/H (light/heavy) intensity ratio (not log transformed) of two proteins, MIF and A1AT, between M1 (oral mucositis grades 1 or 2) and M2 (oral mucositis grades 3 or 4) saliva samples. *p value < 0.05, Wilcoxon Mann-Whitney test

Full size image

Discussion

This study evaluated the proteomic profile of whole saliva in HNSCC patients treated with RT or CRT, associated or not with previous surgery for tumor resection, and correlated this with OM severity. We found two proteins A1AT and MIF that could be possible prognostic biomarkers for the development of severe oncologic treatment–induced OM. The relevance of this finding is driven by the great interest in the development of personalized treatments based on the molecular profile of diseases that could lead to increased personalized treatment efficacy [22].

The patient-related findings in the present study are consistent with those reported in the literature, with the highest prevalence of HNSCC noted among middle-aged men diagnosed with advanced squamous cell carcinomas at the lateral border of the tongue and oropharynx, with most of these lesions considered clinically unresectable [23, 24]. In terms of the frequency and severity of OM, the present study also observed conventionally reported incidences, with over 80% of patients presenting some degree of lesions for patients submitted to RT or CRT protocols and treated with prophylactic PBM therapy. Over 40% of the patients in this study developed grade 2 OM as the highest grade throughout the treatment which is similar to reports in the literature of approximately 50% of grade 2 OM incidence by the end of RT with PBM treatments [25, 26]. The use of PBM therapy prophylactically is now considered standard of care and has likely provided additional benefit in this study to mitigate OM severity [11, 27].

The A1AT protein, also known as serpin 1, is a major liver-derived circulating protein that functions as a natural inhibitor of various serine proteases. It serves as a key component of the acute-phase response with roles in modulating the local and systemic inflammatory responses [28]. Jehmlich et al. [12] were the first to correlate the A1AT protein with the presence of OM, based on proteome analysis of whole saliva samples [12]. We also observed a significant modulation of this marker in the present study and this appeared to correlate with increased susceptibility to develop severe OM post-oncologic treatment. However, the precise role of A1AT in the OM pathogenesis remains unclear and further studies are warranted [12].

The MIF protein is a T cell–derived factor and its primary role is as a pro-inflammatory protein [29, 30]. MIF acts by activating pathways such as mitogen-activated protein (MAP) kinase and phosphoinositide-3-kinase and enhancing the expression of pro-inflammatory genes [31]. Additionally, MIF can both act on macrophages and be secreted by them in response to stimulation by cytokines such as tumor necrosis factor-alpha (TNF-α) [29, 32]. Interestingly, MIF is also responsible for the production and release of TNF-α, a pro-inflammatory cytokine that, in turn, further upregulates production of MIF in an autocrine manner and thereby promoting a pro-inflammatory response [29]. MIF also regulates the innate immune response and has been shown to be important in various autoimmune diseases. It has been noted to be involved in cell proliferation, cell survival, migration, and metastasis in several types of cancers, particularly oral squamous cell carcinomas [30, 31, 33].

The role of MIF as a pro-inflammatory marker appears to be compatible with the results of the present study, in that we found a higher abundance of MIF in samples from subject who developed severe grades 3 or 4 OM. This is the first clinical demonstration, to our knowledge, of this correlation in human subjects that has been previously observed in animal studies [34, 35]. Other studies have reported reduced cytokine levels of TNF-α and MIF following effective OM treatment, likely reflecting a reduction in the underlying inflammatory pathophysiology [34, 35].

As previously mentioned, OM presents a complex pathophysiology which involves several signaling pathways [12, 34]. One of the known signaling pathways implicated in OM is the nuclear factor kappa beta (NFκB) signal transduction pathway [36]. This pathway, when activated, enhances the expression of the pro-inflammatory cytokine TNF-α which, in turn, drives tissue damage leading to the development of OM lesions. In addition, both TNF-α and NFκB signaling have been shown to modulate MIF expression, potentially corroborating the results in this study for an important role for MIF in OM pathogenesis [34, 35].

Strengths and limitations

The major strengths of the current study were the rigorous clinical inclusion and exclusion criteria that enabled a relatively homogeneous cohort of subjects. Nonetheless, limitations of the present work include a single institution study design, availability of only pre-RT/CRT (and not intermediate and post-treatment) saliva samples, and variations in tumor burden during saliva sampling. These limitations justify the need for a larger and independent cohort and long-term prospective studies that are being pursued currently.

Conclusions

This study indicates the correlation of two protein levels with the development of severe OM and noted increased MIF and A1AT levels correlated with more severe grades of OM following oncologic treatment. Besides providing potential insights into their role in OM pathogenesis, these potential salivary biomarkers could serve as valuable prognostic aids to enhance and precisely calibrate effective OM interventions such as PBM treatments, patient-controlled analgesia, or personalized medicaments.