Putative salivary protein biomarkers for the diagnosis of oral lichen planus: a case-control study

Oral lichen planus (OLP) is a chronic immune-mediated disease that affects the oral mucosa. The prevalence rate of OLP is reported as 0.5–2.2% [1, 2]. A recent systematic review indicated that the malignant transformation rate of OLP-affected tissue is 1.09% [3]. There are several types of OLP, the most common of which are the reticular, atrophic, and ulcerative/erosive forms. Often, the clinical appearance of OLP mimics other types of vesiculobullous diseases; therefore, biopsy of the lesion is required for definitive diagnosis [4]. Histopathological features of OLP include hyperkeratosis, acanthosis, liquefactive degeneration of the basal cell layer, and prominent lymphocyte infiltration at the epithelium-connective tissue interface [1]. Use of direct immunofluorescent (DIF) staining also aids in the diagnosis of OLP. Recently, it was revealed that the most common manifestations found on DIF analysis of OLP biopsies are 1) shaggy expression of fibrinogen in the basement membrane zone; 2) deposition of Immunoglobulin M as colloid bodies; 3) deposition of C3 in granular and linear patterns. Thus, the presence of these proteins on DIF analysis is utilized for diagnosis of OLP [5].

Human saliva is a readily available fluid that contains locally- and systemically-produced compounds and proteins that may serve as markers for diagnosis of several diseases. Following the development of highly sensitive assays with simple and non-invasive saliva collection techniques, saliva-based biomarkers have found increasing usage in the diagnosis of disease [6]. Al-Tarawnej and colleagues performed a systematic review to define and summarize disease-related salivary biomarkers that could be identified by mass spectrometry proteomics [7]. These diseases included Sjögren’s syndrome, squamous cell carcinoma, diabetes, breast cancer, dental caries, periodontitis, systemic sclerosis, bleeding oral cavity, graft versus host disease and OLP. A total of 180 biomarkers were identified; 87 of these were upregulated, 63 were downregulated, and 30 varied in a disease-dependent manner [7].

Some proteins have been previously reported as salivary biomarkers for OLP. Notably, C-reactive protein, Immunoglobulin A (IgA), matrix metalloproteinase 8, CTX I, CD14 and toll-like receptor-2 demonstrate upregulation in the saliva of OLP patients [8‐11]. Thus far, very few studies have used a proteomic approach to identify potential salivary protein biomarkers for OLP [12‐14]. In one of these, Mizukawa et al. used reverse-phase high-performance liquid chromatography and mass spectrometry (MS); they reported a significant increase in defensin-1 (HNP-1) levels in the saliva of patients with OLP, compared with HNP-1 levels in healthy subjects [12]. In another study, Yang et al. evaluated potential biomarkers in whole saliva from OLP patients [13]. They identified 14 proteins with at least a two-fold difference in intensity between OLP patients and healthy controls. Importantly, urinary prokallikrein was elevated, while palate, lung, and nasal epithelium carcinoma-associated protein (PLUNC) was reduced in saliva samples from OLP patients. Lastly, Chaiyarit et al. used matrix assisted laser desorption/ionization time-of-flight/time-of-flight (MALDI-TOF/TOF) MS to analyze the patterns of mass signals from low-molecular-weight proteins that were present in saliva samples from patients with oral cancer, OLP, and chronic periodontitis [14]. Mass signals at both 12,964.55 and 13,279.08 Da were significantly elevated in the saliva of OLP patients, compared with the other groups. However, the corresponding proteins have not been identified.

Though numerous candidate proteins have been discovered, these must be validated in other patient populations. The identification of potential salivary protein biomarkers is an essential first step in the development of effective screening techniques and early diagnosis of OLP, which is expected to improve treatment outcomes and quality of life for affected patients. The objective of this study was to use proteomics approaches to identify potential salivary protein biomarkers for screening and diagnosis of OLP.

In the present study, we used a 2D gel-based proteomic approach to determine the protein profile of WUS from OLP patients (Fig. 1). We reviewed and reported functions of the proteins that were differentially expressed between the saliva of OLP patients and healthy control subjects (Table 5). We then selected three proteins that appear to have strong potential to serve as biomarkers for OLP and validated the salivary levels of these proteins through ELISA and immunoblotting analyses (Tables 2, 3, 4 and Fig. 2). We found that fibrinogen fragment D and complement component C3c exhibited increased expression in the saliva of OLP patients (Table 6 and Fig. 2). Moreover, we found that cystatin SA exhibited decreased expression in the saliva of OLP patients, compared with healthy control subjects (Table 6 and Fig. 2).

Using a 2D gel-based proteomic approach, we found 31 protein spots with either a ≥ 1.5-fold difference in abundance or an exclusive presence in one of the study groups (OLP patients or healthy controls); these spots were excised for further identification by LC-MS/MS. We ultimately identified 20 unique proteins, of which 13 exhibited increased expression in OLP patients, compared with healthy control subjects (Tables 2 and 3). In contrast, the remaining seven proteins exhibited decreased expression in OLP patients (Tables 2 and 4).

Some of the proteins that were identified in this study have been previously reported to exhibit increased expression in patients with oral or skin LP. For example, C3 and fibrinogen have been used for the diagnosis of OLP on immunofluorescent analysis [5]. Haptoglobin, an acute phase reactant protein, and NGAL, a small cationic antimicrobial peptide of epithelial origin and marker for dysregulated keratinocyte differentiation, both demonstrated increased expression in keratinocytes from the skin of patients with LP [21, 22]. The Immunoglobulin J (IgJ) chain is required for dimerization of IgA molecules, allowing them to be secreted as secretory-IgA (sIgA) [23]. Several studies have described increased sIgA expression in the saliva of OLP patients; this is consistent with our detection of increased IgJ expression in the saliva of our cohort of OLP patients [8, 24].

Another protein upregulated in OLP patients, 14–3-3 zeta/delta, has multiple roles in signal transduction, regulation of apoptosis, and oral carcinogenesis [25, 26]. Immunohistochemical analysis has demonstrated overexpression of this isoform of 14–3-3 in premalignant oral lesions and samples of oral squamous cell carcinoma [27, 28]. Using proteomics integrated with bioinformatics to compare metastatic and non-metastatic cancer cell lines, multiple interaction partners of 14–3-3 zeta have been identified, suggesting its involvement in several intracellular protein networks, including evasion of apoptosis, cytoskeletal remodeling, protein synthesis, protein folding/degradation, and nuclear import and export of proteins in oral cancer cells [29]. Since OLP is a potentially malignant disorder, the presence of this protein in the saliva of OLP patients may contribute insights on how OLP may transform into oral cancer.

Several protein biomarkers have been previously reported for the diagnosis of OLP by proteomic analysis. These include salivary CD14, urinary prokallikrein, PLUNC, and defensin-1 [9, 12‐14]. In addition, two unidentified mass signals, with molecular weight values of approximately 12.9 and 13.2 kDa, are significantly increased in the saliva of OLP patients [14]. Importantly, different techniques for the preparation of salivary samples are likely to influence the detection of various proteins in the saliva. Yang et al. detected an increase in urinary prokallikrein levels, and a decrease in PLUNC levels in saliva samples from OLP patients, using 2-DE followed by MALDI-TOF MS [13]; however, these two proteins were not identified as putative biomarkers in our study. Although the techniques used in our study were similar to those used by Yang et al., the discrepancy in these results may be due to differences in preparation of the raw saliva samples. In the Yang method, the saliva was desalted, lyophilized, and dissolved in urea sample solution, then precleaned using the PlusOne 2-D Clean-up kit; however, we did not utilize this purification method in our study [13].

Nevertheless, our study detected an elevated level of fibrinogen fragment D and complement component C3c in the saliva of OLP patients, compared with healthy control subjects. To our knowledge, this work is the first to detect changes in salivary fibrinogen D fragment and complement component C3c in OLP patients.

Fibrinogen is a soluble 340-kDa glycoprotein with three specific domains, including two outer D domains (67 kDa) and a central E domain (33 kDa) [30]. The specific function of fibrinogen in OLP and other vesiculobullous diseases is unknown. However, recent reports have described the effects of fibrinogen on vascular cells, including increased vascular tone, endothelial disorganization, and increased endothelial permeability to albumin [30]. In vitro, fibrinogen has been shown to trigger cytokine and chemokine secretion in several types of cells, and to mediate attachment of monocytes/leukocytes to the endothelium [30, 31]. In addition, the in vivo release of proinflammatory and chemotactic factors, such as monocyte chemoattractant protein-1 and interleukin 6, has been linked to binding of fibrinogen to the vascular wall [32]. Although fibrinogen is primarily synthesized by hepatocytes, the oral epithelium may synthesize and secrete fibrinogen locally in response to proinflammatory mediators, in a manner similar to that of the lung alveolar epithelium [33, 34].

The complement system is a key part of innate immune system, which aids antibodies and phagocytic cells in eradicating pathogens from an organism. Complement component C3 is a 185-kDa glycoprotein that belongs to the α2-macroglobulin family; it is the most abundant complement protein in serum [35]. Although most complement factors exhibit short half-lives, complement C3c is the major stable conversion product of complement C3, and has therefore been used as a surrogate for complement C3 in studies of complement function [36].

Fibrinogen expression at the basement membrane, and the specific pattern of C3 deposition, are both hallmarks in the diagnosis of OLP using DIF [5]. Our findings are consistent with these prior studies, suggesting that both fibrinogen and C3c may be useful salivary markers for OLP. Notably, these findings must be confirmed in a larger group of OLP patients.

Our study is the first to demonstrate decreased expression of salivary cystatin SA in OLP patients. In a previous study, Hu et al. [37] reported a reduction in the expression of salivary cystatin SA in patients with primary Sjögren’s syndrome, compared with healthy control subjects. Moreover, others have reported reduced expression of salivary cystatin SA in the saliva of periodontitis patients, compared with healthy control subjects [38, 39]. Interestingly, Ibrahim et al. [40] recently reported an increase in the expression of cathepsin L in skin specimens from patients with cutaneous LP, compared with controls. Further, this study suggested that cathepsin L could be an essential protease that contributes to the pathogenesis of LP [40]. In another study, Baron et al. showed that salivary cystatin SA can inhibit human cathepsin L and may be involved in proteolytic events in vivo [41]. Therefore, it is possible that proteolytic events may contribute to the pathogenesis of OLP, and that reduced inhibition of these proteolytic events may contribute to the destruction of basal cells in OLP. However, additional studies are required to confirm this hypothesis.

Although we investigated proteins that appear most likely to play a role in the pathogenesis of OLP, some questions still remain. According to a systematic review of MS-based proteomic studies that endeavored to define salivary biomarkers related to specific diseases, some listed biomarkers were not specific for the diseases studied—variations existed in the expression of the same protein in the same disease across multiple studies [7]. This suggests that, instead of using a single protein for diagnosis and monitoring of a specific disease, a combination of biomarkers should be utilized [7]. Based on an assessment of all salivary biomarkers, there are three categories: (1) biomarkers that are specific for a disease, (2) biomarkers that are not specific for a disease, but may indicate an abnormal condition, and (3) biomarkers that do not reliably indicate abnormality but represent diversities or variations between control and diseased samples, sample treatment protocols, or mass spectrometry platforms [7]. Our results suggest that although these three proteins (complement C3c, fibrinogen fragment D and cystatin SA) are differentially expressed in the saliva of OLP patients, they may not be specific for OLP; however, the expression of this group of proteins may be used as a multifactorial biomarker for the diagnosis of OLP and subsequent monitoring of the progression of disease. We are currently gathering more OLP patients and evaluating the sensitivity and specificity of these three biomarkers for the diagnosis of OLP. Further studies of the expression of these proteins in OLP patients, both before and after treatment, are ongoing; this may yield some insight on the ability of these proteins to serve as biomarkers of OLP.

There are some limitations to our study. First, the method that we used, 2D-PAGE followed by LC-MS/MS, might not provide the sensitivity of other recent techniques, such as MALDI-TOF or surface-enhanced laser desorption/ionization (SELDI)-TOF, both followed by LC/MS/MS [7, 42]. It has been shown that platforms based on 2D electrophoresis are affected by poor reproducibility; to avoid bias, it is often necessary to run multiple replicates of the same sample [42]. In our study, we first performed 2D-PAGE in five pairs of saliva samples from patients and controls; we found non-redundant protein spots that were differentially expressed in each pair of gels. Hence, we decided to pool saliva samples and again perform 2D-PAGE in triplicate then we chose the best pair of gels for LC-MS/MS. Moreover, it has been suggested that the best method for robust biomarker identification is analysis of multiple samples using different proteomic methodologies [7, 43]. We are currently performing shotgun proteomic analyses on saliva samples of OLP patients and comparing these with saliva from healthy control participants; we expect confirmation of whether these proteins can reliably serve as potential markers for OLP.

Although our sample size of OLP patients is small, we show novel findings that the combination of these three salivary proteins (complement component C3c, fibrinogen fragment D, and cystatin SA) may serve as salivary biomarkers for diagnosis of OLP. Further studies, with a larger group of OLP patients and a before/after cohort study, are underway.