Proteomic Analysis of Whole Human Saliva Detects Enhanced Expression of Interleukin-1 Receptor Antagonist, Thioredoxin and Lipocalin-1 in Cigarette Smokers Compared to Non-Smokers

A gel-based proteomics approach was used to screen for proteins of differential abundance between the saliva of smokers and those who had never smoked. Subjecting precipitated proteins from whole human saliva of healthy non-smokers to two-dimensional electrophoresis (2-DE) generated typical profiles comprising more than 50 proteins. While 35 of the proteins were previously established by other researchers, an additional 22 proteins were detected in the 2-DE saliva protein profiles generated in the present study. When the 2-DE profiles were compared to those obtained from subjects considered to be heavy cigarette smokers, three saliva proteins, including interleukin-1 receptor antagonist, thioredoxin and lipocalin-1, showed significant enhanced expression. The distribution patterns of lipocalin-1 isoforms were also different between cigarette smokers and non-smokers. The three saliva proteins have good potential to be used as biomarkers for the adverse effects of smoking and the risk for inflammatory and chronic diseases that are associated with it.


Introduction
Cigarette smoking is the most preventable cause of addiction, sickness and mortality in the world. Death attributed to cigarette smoking is estimated to rise from 5.4 million in 2005 to 6.4 million by 2015 [1]. Chronic cigarette smoking is the single most important risk factor for lung and oral cancers, cardiovascular diseases, chronic obstructive pulmonary disease (COPD) and other tobacco related oral diseases, including periodontitis [2][3][4][5][6][7]. Cigarette smoke contains more than 60 carcinogens and around 4,000 chemicals, including bacteria-derived endotoxins, which are toxic to cells [8][9][10]. The risk of developing tobacco smoking-related diseases increases with the total exposure time to the cigarette smoke, which generally includes the number of cigarettes a person smokes each day and the number of years a person has been smoking [11].
The oral cavity is the first organ in the human body to be exposed to the cigarette smoke. The tobacco smoke alters normal homeostasis of the oral cavity, including the saliva's antioxidant and other protective systems. This may lead to oral inflammatory diseases and oral cancers [12][13][14][15]. Early tumorigenic activities have been detected in normal oral mucosa of heavy smokers who have no overt precancerous or cancerous lesions [16]. The mucosal changes in smokers may also arise from the drying effects of the mucosa, high intraoral temperatures, intraoral pH changes, local alteration of membrane barriers and immune responses, or altered resistance to bacteria, fungal and viral infections.
Smoking-related cell damage may leave molecular footprints in the saliva, offering the potential for non-invasive early diagnosis of tobacco-related oral diseases.
Human saliva contains a large number of proteins and peptides that are easily accessible and may serve as a potential source of biomarkers to monitor changes that occur under pathological conditions. The value of saliva as a biological fluid for the detection of diagnostic and prognostic biomarkers has become increasingly well established [17][18][19][20][21][22][23][24]. Collection of human saliva is a simple, non-invasive and cost-effective approach for screening large populations. It is easy to handle and may be repeated without inflicting much discomfort to the subjects [17,18].
Proteomic analysis is an important investigative tool used to systematically explore cellular proteins that are responsive to adverse environmental challenges. Several proteomic approaches, including those involving separation of proteins by two-dimensional electrophoresis (2-DE), have been applied in the investigation of biomarker candidates in the human saliva [25][26][27][28][29]. Recently, saliva has been shown to harbor potential informative biomarkers for oral cancer [30][31][32], head and neck cancer [33,34], and breast cancer [35]. While effects of the cigarette smoke on proteins expressed in the bronchoalveolar lavage [36][37][38], nasal lavage fluid [39], urine [40], lung tissue [41], bronchial airway epithelium and pooled exhaled breath condensate samples [42] have been analyzed, little information is available regarding the effects of smoking on the whole saliva proteome.
To the best of our knowledge, there had been no reported studies that specifically compared the expression of proteins in the saliva of smokers and non-smokers. In this study, 2-DE-based proteomics was used to screen for saliva proteins of differential abundance between smokers and subjects who had never smoked. The aberrantly expressed proteins, when correlated to those similarly altered in the saliva of patients with tobacco-related diseases including oral cancer, may potentially be used as biomarkers to indicate risks for the various diseases.

Collection of Whole Saliva
Unstimulated whole saliva samples were collected from 24 healthy Malay male volunteers aged between 35 and 55 years (12 smokers and 12 non-smokers), with no history of diabetes, autoimmune diseases or exposure to radiation and chemotherapy. Characteristics of the participants who were considered heavy smokers in this study are shown in Table 1. Saliva samples were collected with the volunteers' consent and approval granted by the Ethical committee (Institutional Review Board) of the Faculty of Dentistry, University of Malaya. Each subject answered a questionnaire concerning personal data, smoking and alcohol drinking habits, health or dental problems, oral hygiene habits, previous dental examinations, use of prescriptions, quantity and length of cigarette smoking. Unstimulated whole saliva was collected in the morning between 9 and 11 am to minimize the circadian effect, and subjects refrained from eating, drinking, smoking or performing any oral hygiene for at least 2 h prior to the collection. The difference between the mean flow rates of non-smokers (0.39 ± 0.04 mL/min) and smokers (0.42 ± 0.04 mL/min) was not statistically significant. Protease inhibitor cocktail was added to the saliva immediately after collection as previously described [43]. To remove debris and cells, the saliva was centrifuged at 14,000 g for 20 min at 4 °C and the proteins were precipitated in 10% TCA/acetone/20 mM DTT. Saliva proteins were quantified using the Bradford protein assay kit (Bio-Rad, Hercules, USA) according to the manufacturer's instructions.

Silver Staining
The 2-DE gels were developed by silver staining as described by Heukeshoven and Dernick [44]. For mass spectrometry analyses, gels were stained with compatible silver staining with slight modifications according to Yan et al. [45].

Image Analysis
The LabScan image scanner was used to capture and store images of 2-DE gels. The GE ImageMaster TM 2D Platinum Software version 7 was used to evaluate the protein profiles and perform protein analyses. To detect proteins that were differentially secreted in the saliva, the percentage volume contribution (% vol) of a protein spot, which refers to the spot volume of a protein expressed as a percentage of the total spot volume of all detected saliva proteins, was calculated. Data expressed this way are independent of variations attributed to protein loading and staining. The 2-DE profiles and relative spot intensities obtained were reproducible when performed in triplicate.

In Gel Trypsin Digestion and Mass Spectrometry
Highly resolved protein spots were initially identified by visual comparison with previously published protein maps obtained from the human whole saliva [22][23][24][25][26]. The protein spots (1-2 mm diameter) were excised from silver-stained gels with pipette tips and kept hydrated in clean microfuge tubes containing Milli-Q water, prior to the in-gel digestion. Trypsin digestion and precise identification by mass spectrometry, using the MALDI-TOF/TOF instrument (Applied Biosystem 4800 Proteomic Analyzer), were performed as previously described [46].

Database Searches
Spectra were processed and analyzed by the Global Protein Server Workstation (Applied Biosystems), which uses the internal MASCOT (Matrix Science, London, UK) software for search of the peptide mass fingerprints and MS/MS data. Searches were performed against the Swiss-Prot database (Last update: October 23, 2008, containing 261513 sequences). Database search parameters were set as follows: The enzyme trypsin was used; up to one missed cleavage was allowed; variable modification included were carbamidomethylation of cysteine and oxidation of methionine; the mass tolerance for MS precursor ion and MS/MS fragment ion were 100 ppm and 0.2 Da, respectively; and only monoisotopic masses were included in the search.

Statistical Analysis
All values are presented as mean ± S.E.M (standard error of the mean). The Student's t-test was used to analyze the significance of difference between non-smokers and smokers. The false discovery rate control was performed using the method of Benjamini and Hochberg [47]. Figure 1 shows a typical 2-DE profile of saliva proteins separated between pH 4 and 7 in healthy non-smokers. This range of pH was chosen as our earlier 2-DE results performed at a pH range of 3 to 10 showed that most of the saliva proteins were located in the acidic region between pH 4 to 7. More than 120 spots were detected in the whole saliva samples using the 2-DE that was performed under the conditions of our study. Identities of 108 spots belonging to 57 different proteins were established by MS and database search ( Table 2). Some of these proteins, including polymeric immunoglobulin receptor (spots 3-9), carbonic anhydrase VI (spots 27-32), prolactin inducible proteins (spots 81-86), zinc-alpha-2-glycoprotein (spots 43 and 44), short palate, lung and nasal epithelium carcinoma-associated protein 1 (spots 58-61), cystatin S (spots 90 and 91) and lipocalin-1 (spots 87-89) were resolved in several isoforms and thus separated into distinct spots in the 2-DE gels. A total of 108 protein spots (circled and numbered) were identified by mass spectrometry and database search (please refer to Table 2). Acid side of 2-DE gel is to the left and relative molecular mass declines from the top.    Among the total of 57 saliva proteins, 35 had been previously identified using 2-DE [32][33][34][35][36][37][38][39], whereas 16, including plastin-2, actin-related protein-3, C3 complement precursor, macrophage capping protein, F actin capping protein, annexin A3, protein kinase C inhibitor protein-1, rho-GDP-dissociation inhibitor 1, rho-GDP-dissociation inhibitor 2, actin-related protein 2/3 complex subunit 5, alpha-1-acid glycoprotein 1, chloride intracellular channel protein 1, protein disulfide-isomerase, leukotriene A-4 hydrolase, IgGFc-binding protein and long palate, lung and nasal epithelium carcinoma-associated protein 1, were previously detected using liquid-based proteomics [21,23]. The other six saliva proteins, eosinophil lysophospholipase, beta-microseminoprotein, coactosin-like protein, nucleoside diphosphate kinase A, calreticulin and synaptic vesicle membrane protein VAT-1, are reported for the first time by this study.

Results
When 2-DE was performed on whole saliva samples of heavy smokers, the profiles obtained were similar to those from non-smokers. All 57 different proteins that were expressed in the saliva of the non-smokers were also detected in the saliva of the heavy smokers although the rates of presence of 16 proteins in the 2-DE profiles of the cigarette smokers were different from those of the non-smokers. When the 2-DE protein profiles obtained from the non-and heavy smokers were subjected to densitometry analysis, initially a significantly enhanced expression of seven proteins including polymeric immunoglobulin receptor, complement C3, α1-antitrypsin, calgranulin B, interleukin-1 receptor antagonist, thioredoxin and lipocalin-1, was detected between the two subject groups. However, only three of the proteins, i.e., interleukin-1 receptor antagonist (+3 fold), thioredoxin (+2.5 fold) and lipocalin-1 (+4.4 fold) were found to be truly significant when the p-values were corrected for false significant results using the method of Benjamini and Hochberg [47] (Table 3). Figure 2 demonstrates examples of 2-DE spot clusters of proteins whose levels were altered in the saliva obtained from the heavy smokers as compared to those of the non-smokers.  0.758 (a) volume of a protein expressed as a percentage of the total spot volume of all proteins; (b) rate of presence of the protein spots in the 12 2-DE profiles that were analyzed; (c) p-values of less than 0.0068 (p < 0.0068) were considered statistically significant when the false discovery rate procedure of Benjamini and Hochberg [47] was performed to the data set.  Table 2.
When the different isoforms of polymeric immunoglobulin receptor (spots 3-9), carbonic anhydrase VI (spots 27-32), prolactin inducible proteins (spots 81-86), zinc-alpha-2-glycoprotein (spots 43 and 44), short palate, lung and nasal epithelium carcinoma-associated protein 1 (spots 58-61) and cystatin S (spots 90 and 91) were similarly analyzed by densitometry, their volume distribution patterns were found to be consistent between the saliva of non-smokers and smokers. In contrast, the 2-DE volume distribution pattern for isoforms of lipocalin-1 in the saliva of non-smokers was different from that detected in the saliva of the heavy smokers ( Figure 3). Among the seven isoforms analyzed, the isoform f was almost exclusive to the saliva of the smokers (Table 4).  Table 4.  Figure 3; (b) volume of a protein expressed as a percentage of the total spot volume of all proteins; (c) rate of presence of the protein spots in the 12 2-DE profiles that were analyzed; (d) fold change is the ratio of %vol of smokers to non-smokers.

Discussion
Human whole saliva contains fluid from the salivary glands, gingival crevicular fluid, bronchiol and nasal secretions, desquamated epithelial cells, oral tissues, and very often, the components of blood, bacteria and viruses [48][49][50]. Therefore, whole saliva-in contrast to serum-is a hostile environment with proteins subjected to the effects of many host-and bacteria-derived enzymes. Some saliva proteins are synthesized in the salivary glands and subsequently subjected to intracellular processing including glycosylation, phosphorylation and proteolysis. Once the secretions enter the non-sterile oral environment, additional and continuous protein modifications by host-and bacteria-derived enzymes occur. This results in the possible generation of many modified proteins in whole saliva [51].
The 2-DE profiles of proteins in whole saliva from healthy non-smokers that were generated in the present study showed strong resemblance to those that were previously reported [22][23][24][25][26]. Almost 90% of the protein spots that were highly resolved were eventually identified. The remaining spots were unidentifiable as the proteins generated low intensity spectra probably due to their low amounts, resistance to proteolytic cleavage, low recovery of digested peptides, and/or low efficiency in peptide ionization. Nevertheless, it is also possible that some of the unidentified proteins were of bacterial origin since the mouth is likely to harbor a lot of microorganisms.
The establishment of highly resolved 2-DE protein profiles enabled investigations on protein changes associated with cigarette smoking. Densitometry analyses on the 2-DE protein profiles obtained from the non-and heavy smokers showed differential abundance of interleukin-1 receptor antagonist, thioredoxin and lipocalin-1 between the saliva samples of the two subject groups ( Table 3). The three proteins have good potential to be used as non-specific complementary biomarkers for the adverse effects of smoking although this requires further evaluation and correlative studies. Some of the proteins may be used as risk indicators for inflammatory and chronic diseases that are associated with smoking as they have been shown to be of increased levels in the saliva of the patients. In the case of lipocalin-1, the isoform distribution pattern detected was also found to differ between smokers and non-smokers. This suggests that the carbohydrate moieties of lipocalin-1 of the heavy cigarette smokers were different from those of the non-smokers and that they may be differently glycosylated or modified. However, this remains to be further established.
Despite being distinctly categorized according to their primary biological roles [62], the three saliva proteins that were altered in abundance reflect the body's overall response to the damaging effects of heavy smoking. The high levels of IL-1 receptor antagonist in the saliva of the heavy smokers detected in this study reflect an anti-inflammatory response in the oral cavities of the smokers. Increased generation of the proteins in smokers may be induced by the proinflammatory cytokines that were promoted by oxidative stress [63][64][65]. An imbalance between IL-1 receptor antagonist and IL-1 has been hypothesized to play a role in the pathogenesis of various inflammatory diseases [65].
Lipocalin-1 and thioredoxin are proteins most likely involved in the response to stress in relation to tissue damage. The high levels of lipocalin-1 and thioredoxin in the cigarette smokers' saliva may reflect their function as an oxidative stress-induced scavenger against toxic and pro-inflammatory lipids [66][67][68]. Lipocalin-1 had been suggested to be a cysteine proteinase inhibitor [67] and may have a role in the control of inflammatory processes in oral tissues. Thioredoxin, on the other hand, was shown to modulate remodeling factors in response to the cigarette smoke [68]. Increased secretion of thioredoxin had been previously demonstrated in the saliva of patients with oral cancer [32].

Conclusion
Comparative proteomics analysis of human saliva samples from subjects who were considered heavy cigarette smokers and those who did not smoke detected altered abundance of interleukin-1 receptor antagonist, thioredoxin and lipocalin-1, as well as a change in the isoform distribution patterns of lipocalin-1. These proteins may be used as early biomarkers to indicate risks of tobacco-related diseases.