Breast cancer related proteins are present in saliva and are modulated secondary to ductal carcinoma in situ of the breast.

Objective: The objective of this study was to determine if protein-by-products secondary to cancer related oncogenes appear in the saliva of breast cancer patients. Methods: Three pooled (n = 10 subjects/pool) stimulated whole saliva specimens from women were analyzed. One pooled specimen was from healthy women, another pooled specimen from women diagnosed with a benign breast tumor and the other one pooled specimen was from women diagnosed with ductal carcinoma in situ (DCIS). Differential expression of proteins was measured by isotopically tagging proteins in the tumor groups and comparing them to the healthy control group. Experimentally, saliva from each of the pooled samples was trypsinized and the peptide digests labeled with the appropriate iTRAQ reagent. Labeled peptides from each of the digests were combined and analyzed by reverse phase (C18) capillary chromatography on an Applied Biosystems QStar LC-MS/MS mass spectrometer equipped with an LC-Packings HPLC. Results: The results of the salivary analyses in this population of patients yielded approximately 130 proteins in the saliva specimens. Forty-nine proteins were differentially expressed between the healthy control pool and the benign and cancer patient groups. Conclusions: The study suggests that saliva is a fluid suffused with solubilized by-products of oncogenic expression and that these proteins may be modulated secondary to DCIS. Additionally, there may be salivary protein profiles that are unique to both DCIS and fibroadenoma tumors.

Keywords: LC-MS/MS; isotope labeling; breast cancer; Saliva

Proteomics was originally defined to represent the analysis of the entire protein component of a cell or tissue ([1]), but now it encompasses the study of expressed proteins, including identification and elucidation of the structure-function relationship under healthy conditions and disease conditions. In combination with genomics, proteomics can provide a holistic understanding of the biology underlying disease processes. Information at the level of the proteome is critical for understanding the function of specific cell types and their role in health and disease ([1], [2]).

Protein expression and function are subject to modulation through transcription as well as through posttranscriptional and translational events. Multiple RNA species can result from one gene through a process of differential splicing. Additionally, there are more than 200 post-translation modifications that proteins could undergo that affect function, protein-protein and nuclide-protein interaction, stability, targeting half-life, and so on ([3], [4], [5]), all contributing to a potentially large number of protein products from one gene. Identifying and understanding these changes are the underlying themes in proteomics ([6], [7], [8]).

Technological advancements have benefited proteomic research to the point where saliva is now being assayed for protein content using the latest available proteomic technology ([9]). The investigators opted to explore saliva as a diagnostic fluid for two reasons:collection of saliva is a non-invasive procedure that can be conducted in any environment requiring no special skills of equipment and the physiology of the oral cavity is such that the flow of secreted fluid is continually flushing and refreshing the fluid content of the mouth. Therefore, the composition of the fluid at any moment temporally reflects the metabolic activity of the secretory elements generating that fluid. There are also significant advantages over the study of plasma. In plasma the concentration of proteins can vary over nine orders of magnitude which severely diminishes the likelihood of detecting those at the lower end of the scale. The second consideration is that blood is composed of peptides, proteins and cells that have half lives ranging from seconds to weeks or even a month or more. As a consequence, the presence of a given substance might not accurately reflect the current state of the organism.

As this is a newly recognized endeavor, there is but a paucity of information regarding the salivary proteome and its constituents in the presence of disease such as carcinoma. Previous studies using immunological techniques have demonstrated that saliva from breast cancer patients exhibited elevated levels of CerbB-2, CA 15-3, EGFR, cathepsin D and p53, suggesting that there is communication between the breast tumor and the salivary gland ([10], [11], [12]). The purpose of this investigation was to determine if protein-by-products secondary to cancer related oncogenes that are over or under expressed appear in the saliva of breast cancer patients.

The investigators protein profiled three pooled, stimulated whole saliva specimens. One specimen consisted of pooled saliva from 10 healthy subjects, another specimen was a pooled saliva specimen from 10 benign tumor patients (fibroadenomas), and the third specimen was from 10 subjects diagnosed with ductal carcinoma in situ (DCIS). Fibroadenomas were selected due to its high prevalence among benign breast tumors. DCIS was selected as this represents the minimal, detectable tumor load ([13]). The cancer cohort, internally, was Tis in lesion size, estrogen, progesterone and Her2/neu receptor status negative as determined by the pathology report. All subjects were matched for age and race and were non-tobacco users. The mean age of the cancer cohort was 57.4 with an age range of 49–62 years. All but two individuals had a family history of cancer. Tumor grade was not available for this study.

All participating subjects were explained their participation rights and signed an IRB consent form. The saliva specimens and related patient data were non-linked and bar coded in order to protect patient confidentiality. This study was performed under the UTHSC IRB approved protocol# HSC-DB-05-0394.

Stimulated whole salivary gland secretion is based on the reflex response occurring during the mastication of a bolus of food. Usually, a standardized bolus (1 gram) of paraffin or a gum base (generously provided by the Wrigley Co., Peoria, Illinois, USA) is given to the subject to chew at a regular rate. The individual, upon sufficient accumulation of saliva in the oral cavity, expectorates periodically into a preweighed disposable plastic cup. This procedure is continued for a period of five minutes. The volume and flow rate is then recorded along with a brief description of the specimen's physical appearance ([14]). The cup with the saliva specimen is reweighed and the flow rate determined gravimetrically. The authors recommend this salivary collection method with the following modifications for consistent protein analyses ([15]). A protease inhibitor from Sigma Co (St. Louis, Missouri, USA) is added along with enough orthovanadate from a 100 mM stock solution to bring its concentration to 1 mM. The treated samples were centrifuged for 10 minutes at top speed in a table top centrifuge. The supernatant was divided into 1 mL aliquots and frozen at −80°C.

Recent advances in mass spectrometry, liquid chromatography, analytical software and bioinformatics have enabled the researchers to analyze complex peptide mixtures with the ability to detect proteins differing in abundance by over 8 orders of magnitude ([16]). One current method is isotopic labeling coupled with liquid chromatography tandem mass spectrometry (IL-LC-MS/MS) to characterize the salivary proteome ([17]). The main approach for discovery is a mass spectroscopy based method that uses isotope coding of complex protein mixtures such as tissue extracts, blood, urine or saliva to identify differentially expressed proteins ([18]). The approach readily identifies changes in the level of expression, thus, permitting the analysis of putative regulatory pathways providing information regarding the pathological disturbances in addition to potential biomarkers of disease. The analysis was performed on a tandem QqTOF QStar XL mass spectrometer (Applied Biosystems, Foster City, California, USA) equipped with an LC Packings (Sunnyvale, California, USA) HPLC for capillary chromatography. The HPLC is coupled to the mass spectrometer by a nanospray ESI head (Protana, Odense, Denmark) for maximal sensitivity ([18]). The advantage of tandem mass spectrometry combined with LC is enhanced sensitivity and the peptide separations afforded by chromatography. Thus, even in complex protein mixtures, MS/MS data can be used to sequence and identify peptides by sequence analysis with a high degree of confidence ([16], [17], [18]).

Isotopic labeling of protein mixtures has proven to be a useful technique for the analysis of relative expression levels of proteins in complex protein mixtures such as plasma, saliva, urine or cell extracts. There are numerous methods that are based on isotopically labeled protein modifying reagents to label or tag proteins to determine relative or absolute concentrations in complex mixtures. The higher resolution offered by the tandem Qq-TOF mass spectrometer is ideally suited to isotopically labeled applications ([17], [19], [20]).

Applied Biosystems recently introduced iTRAQ reagents ([17], [19], [20]), which are amino reactive compounds that are used to label peptides in a total protein digest of a fluid, such as saliva. The real advantage is that the tag remains intact through TOF-MS analysis; however, it is revealed during collision induced dissociation by MSMS analysis. Thus, in the MSMS spectrum for each peptide there is a fingerprint indicating the amount of that peptide from each of the different protein pools. Since virtually all of the peptides in a mixture are labeled by the reaction, numerous proteins in complex mixtures are identified and can be compared for their relative concentrations in each mixture. Thus, even in complex mixtures there is a high degree of confidence in the identification.

Briefly, the saliva samples were thawed and immediately centrifuged to remove insoluble materials. The supernatant was assayed for protein using the Bio-Rad protein assay (Hercules, California, USA), and an aliquot containing 100 μ g of each specimen was precipitated with 6 volumes of −20°C acetone. The precipitate was resuspended and treated according to the manufacturers instructions. Protein digestion and reaction with iTRAQ labels was carried out as previously described and according to the manufacturer's instructions (Applied Biosystems, Foster City, California, USA). Briefly, the acetone precipitable protein was centrifuged in a table top centrifuge at 15,000 × g for 20 minutes. The acetone supernatant was removed and the pellet resuspended in 20L dissolution buffer. The soluble fraction was denatured and disulfides reduced by incubation in the presence of 0.1% SDS and 5 mM TCEP (tris-(2-carboxyethyl)phosphine)) at 60°C for one hour. Cysteine residues were blocked by incubation at room temperature for 10 minutes with MMTS (methyl methane-thiosulfonate). Trypsin was added to the mixture to a protein:trypsin ratio of 10:1. The mixture was incubated overnight at 37°C. The protein digests were labeled by mixing with the appropriate iTRAQ reagent and incubating at room temperature for one hour. On completion of the labeling reaction, the four separate iTRAQ reaction mixtures were combined. Since there are a number of components that can interfere with the LCMSMS analysis, the labeled peptides are partially purified by a combination of strong cation exchange followed by reverse phase chromatography on preparative columns. The combined peptide mixture is diluted 10 fold with loading buffer (10 mM KH2PO4 in 25% acetonitrile at pH 3.0) and applied by syringe to an ICAT Cartridge-Cation Exchange column (Applied Biosystems) column that has been equilibrated with the same buffer. The column is washed with 1 mL loading buffer to remove contaminants. To improve the resolution of peptides during LCMSMS analysis, the peptide mixture is partially purified by elution from the cation exchange column in 3 fractions. Stepwise elution from the column is achieved with sequential 0.5 mL aliquots of 10 mM KH2PO4 at pH 3.0 in 25% acetonitrile containing 116 mM, 233 mM and 350 mM KCl, respectively. The fractions are evaporated by Speed Vac to about 30% of the volume to remove the acetonitrile and then slowly applied to an Opti-Lynx Trap C18 100 μL reverse phase column (Alltech, Deerfield, Illinois, USA) with a syringe. The column was washed with 1 mL of 2% acetonitrile in 0.1% formic acid and eluted in one fraction with 0.3 mL of 30% acetonitrile in 0.1% formic acid. The fractions were dried by lyophilization and resuspended in 10 μL 0.1% formic acid in 20% acetonitrile. Each of the three fractions was analyzed by reverse phase LCMSMS.

The desalted and concentrated peptide mixtures were quantified and identified by nano-LCMS/MS on an API QSTAR XL mass spectrometer (ABS Sciex Instruments Ontario, Canada) operating in positive ion mode. The chromatographic system consists of an UltiMate nano-HPLC and FAMOS autosampler (Dionex LC Packings Sunnyvale, California, USA). Peptides were loaded on a 75 cm × 10 cm, 3 mm fused silica C18 capillary column, followed by mobile phase elution: buffer (A) 0.1% formic acid in 2% acetonitrile/98% Milli-Q water and buffer (B): 0.1% formic acid in 98% acetonitrile/2% Milli-Q water. The peptides were eluted from 2% buffer B to 30% buffer B over 180 minutes at a flow rate 220 nL/min. The LC eluent was directed to a NanoES source for ESI/MS/MS analysis. Using information-dependent acquisition, peptides were selected for collision induced dissociation (CID) by alternating between an MS (1 sec) survey scan and MS/MS (3 sec) scans. The mass spectrometer automatically chooses the top two ions for fragmentation with a 60 s dynamic exclusion time. The IDA collision energy parameters were optimized based upon the charge state and mass value of the precursor ions. Each saliva sample set there are three separate LCMSMS analyses.

The accumulated MSMS spectra are analyzed by ProQuant and ProGroup software packages (Applied Biosystems) using the SwissProt fasta database for protein identification. The ProQuant analysis was carried out with a 75% confidence cutoff with a mass deviation of 0.15 Da for the precursor and 0.1 Da for the fragment ions. The ProGroup reports were generated with a 95% confidence level for protein identification.

The Swiss-Prot database was employed for protein identification while the PathwayStudio bioinformatics software package (Ariadne Genomics, Inc., Rockville, Maryland, USA) was used to determine Venn diagrams were also constructed using the NIH software program (http://ncrr.pnl.gov). Graphic comparisons with log conversions and error bars for protein expression were produced using the ProQuant software (Foster City, California, USA).

Table 1 summarizes the results of the iTRAQ analysis and illustrates protein comparisons between benign vs. healthy, cancer vs. benign and cancer vs. healthy subjects. In total, 130 proteins were identified at a confidence level > 95 and 72 at > 99. Of these, there were 40 proteins that were determined to be expressed significantly different (p < 0.05) in the benign or tumor saliva compared to healthy control. Fig. 1 represents a Venn diagram of the overlapping proteins between the three groups of women.

Table 1 Comparative Protein Counts Between Healthy, Benign, and Cancer Subjects

Graph: Figure 1 Represents a Venn diagram of the overlapping proteins between the three groups of women.

Table 2 represents the up (n = 14) and down (n = 9) regulated proteins for the pooled saliva sample composed of individuals diagnosed with a fibroadenoma. The fold-increase of protein and p values are also presented. As shown in Table 2, 9 of the 29 proteins were significant at the p < 0.001 to p < 0.0001 levels, and 7 proteins had a greater that 50% change in concentration.

Table 2 Benign vs. Healthy

Table 3 is a list of the up (n = 20) and down (n = 12) regulated proteins observed in the Stage 0 cancer saliva compared to controls. There were 15 proteins that showed a 1.5 fold increase in levels in the cancer compared to control subjects. Of these 15 differentially expressed proteins, 12 were significant at the p < 0.001 to p < 0.0001 levels. Additionally, the Table is referenced in the literature for the presence of these proteins in both blood from cancer subjects and cell supernatants from cancer cell lines. Of the 32 proteins that were up or down regulated secondary to carcinoma of the breast, 79% of these proteins were cited in the literature as being involved, molecularly, with the breast cancer ([27], [28], [29], [30], [31], [32], [33], [34], [35], [36]). The protein functions are illustrated in Fig. 2.

Table 3 Cancer vs. Healthy

Graph: Figure 2 Represents protein functions.

A comparison of the differentially expressed proteins is shown in graphical form in Fig. 3. In this Figure, the log of the ratios for benign vs. control and cancer vs. control is plotted for each of the proteins. The error bars are the log of the error factor calculated by the ProQuant software. This comparison illustrates that there are a number of significant differences in the expression levels of several proteins between the cancer and benign saliva samples. A direct comparison of protein expression ratios between the benign and cancer pooled specimens that exhibited overlap or commonality among the proteins is shown in Table 4. Among the comparison of the overlapping proteins, 10 fell at or below a p value of 0.001, and 9 proteins were greater than 50% difference in the cancer compared to benign. These data are replotted in the log form in Figure 4.

Table 4 Cancer vs. Benign

Graph: Figure 3 Differential protein expression in cancer saliva or benign saliva versus normal control.

To the best of our knowledge, this is the first attempt to determine numerous cancer related proteins in saliva. As a consequence, we have only a few references by which to compare our data.

With respect to the overall analyses, the total number of salivary proteins reported in healthy individuals at the 95% confidence levels was 130. In comparing our findings to other studies, which used 2D gel and mass spectrometry, Wilmarth reported 102 proteins ([16]), and Ghafouri100 proteins ([24]). However, Hu reported 309 using both 2D gel and shotgun proteomic techniques ([21]). Despite the wide variety of total protein reported in the literature, all the proteins that are mentioned in this report (Table 4) have been previously identified in the aforementioned literature. Differences in the number of total proteins identified are probably a result of using single individual profiling (non pooled specimen) or in collection and/or sampling techniques ([37]).

Table 3 and Table 4 list the proteins for the healthy pool vs. benign tumor pool and healthy pool vs. cancer tumor pool respectively. As illustrated in Table 4, many of these proteins have been reported as have been either up or down regulated in blood and cancer tissue. There is also an overlap of 13 up regulated proteins and five down regulated proteins between the protein profiles leaving the benign group with five proteins that are unique to fibroadenomas (ENOA, IGHA2, IL-1ra, S10A7, SPLC2) and 11 proteins unique to DCIS (CAH6, K2C4, CYTA, FABP4, IGHGI, TRFL, BPIL1, CYTC, HPT, PROF1, ZA2G). Figure 1 represents a Venn diagram of the overlapping proteins between the three groups of women. For reasons which the authors cannot explain, two-thirds of the total "overlap" proteins were up-regulated. One could speculate that a portion of the up-regulated proteins that exhibited overlap can be associated with pathways which are common to both disorders. This would include the proteins associated with cytoskeleton and cell growth. The investigators targeted the benign tumor in order to increase the specificity of the panel of markers. If there are markers specific to a benign tumor and specific only to the malignancy, then probability of making the correct clinical assessment is further increased.

Table 4 represents a comparison of the benign and cancer proteins that overlapped each group of pooled subjects. In this comparison, only seven proteins remained significantly different in the presence of carcinoma (p < 0.005). This would include the proteins associated with exocytosis, cytoskeleton and immuno-response. As the cell proliferation process is further enhanced in the presence of carcinoma, it stands to reason that these proteins should be significantly up-regulated in the presence of carcinoma. Fig. 3 and Fig. 4 provide further illustration of the protein comparisons.

Graph: Figure 4 Differential protein expression in cancer versus benign saliva.

The authors cannot explain the mechanism by which these proteins are altered in the presence of carcinoma of the breast. The authors, in single analyte reports, have found additional low abundance proteins, such as HER2/neu, Waf-1, pantropic p53, EGFR and cathepsin D, to be altered in the presence in addition to those cited in this manuscript ([12]). We can only hypothesize that since the histophysiology is very similar between the ductal tissues of the breast and those of the salivary glands that there may be extra-cellular communication between the two distant tissues ([38], [39]). This phenomenon has also been observed in nipple aspirates ([40], [41]), which have yielded many of the same protein constituents as observed in Table 4. Cell studies are required to support this hypothesis.

The authors have examined the salivary proteome that is altered in the presence of carcinoma of the breast; however, further study is required to determine the sensitivity and specificity of these proteins with respect to their diagnostic utility. We do not want to over emphasize the findings at this point, but we are encouraged to find that these protein dispositions have also been found to be altered in blood, cancer tissues and nipple aspirates, which provide further support of our findings.

The authors further encourage the exploration of saliva as a diagnostic media for several reasons. The fluid contains numerous proteins and protein fragments, which may have analytical value. The salivary proteome is in its infancy. Additionally, saliva can also be described as a media which provides "real time" results ([42]). The fluid is continually produced and excreted in an open ended circuit, unlike blood which exists in a "closed loop." Blood, a circulating media, may contain proteins that are a day, a week, or a month old as well as proteins which have passed numerous times through many organ systems or have been excreted ([42]). Saliva, with its continuous flow, is not subject to the aforementioned effects. Consequently, saliva and nipple aspirates may be a more useful diagnostic fluid than blood.