Differential Expression of Peroxisomal Proteins in Distinct Types of Parotid Gland Tumors

Salivary gland cancers are rare but aggressive tumors that have poor prognosis and lack effective cure. Of those, parotid tumors constitute the majority. Functioning as metabolic machinery contributing to cellular redox balance, peroxisomes have emerged as crucial players in tumorigenesis. Studies on murine and human cells have examined the role of peroxisomes in carcinogenesis with conflicting results. These studies either examined the consequences of altered peroxisomal proliferators or compared their expression in healthy and neoplastic tissues. None, however, examined such differences exclusively in human parotid tissue or extended comparison to peroxisomal proteins and their associated gene expressions. Therefore, we examined differences in peroxisomal dynamics in parotid tumors of different morphologies. Using immunofluorescence and quantitative PCR, we compared the expression levels of key peroxisomal enzymes and proliferators in healthy and neoplastic parotid tissue samples. Three parotid tumor subtypes were examined: pleomorphic adenoma, mucoepidermoid carcinoma and acinic cell carcinoma. We observed higher expression of peroxisomal matrix proteins in neoplastic samples with exceptional down regulation of certain enzymes; however, the degree of expression varied between tumor subtypes. Our findings confirm previous experimental results on other organ tissues and suggest peroxisomes as possible therapeutic targets or markers in all or certain subtypes of parotid neoplasms.


Introduction
Salivary gland cancers are rare neoplasms accounting for 3-6% of all head and neck cancers with an annual incidence of 16/1,000,000 population, a 5-year survival rate of 95% if confined to gland and a 44% survival rate in the case of distant metastasis [1][2][3]. In this case, 80% of those tumors arise in the parotid gland, 10-15% arising in the submandibular gland, and the remainder arising in sublingual and minor salivary glands [4]. Owing to the mixed composition of epithelial and non-epithelial tissue in affected salivary glands, up to 33 histological subtypes of parotid tumors are possible, leading to the variable clinical behavior of neoplasms [5]. The most frequent benign salivary gland tumor is pleomorphic adenoma (PMA). Common histopathological types are mucoepidermoid carcinoma (MEC) and acinic cell carcinoma (ACC) [6]. We focused our experiments on deciphering peroxisome characteristics in the above three prevalent parotid gland histopathological tumor types.
Peroxisomes are subcellular, single membrane-bound and morphologically heterogeneous eukaryotic structures characterized by a homogenous, granular matrix rich in concentrated and diverse proteomes [7,8]. They contribute an important part to cell metabolism and are responsible for divergent biochemical anabolic and catabolic processes including ether phospholipid biosynthesis as well as the αand β-oxidation of fatty acids, which is essential for the integrity of cell membranes [9,10]. Peroxisomes also catalyze a multitude of key processes, such as scavenging bioactive hydrogen peroxide, redox homeostasis, lipidic intra-and inter-cell signaling and cellular transportation [11,12]. Furthermore, peroxisomes regulate early embryonic development as well as cellular differentiation and survival and were recently shown to modulate the innate immune system [13][14][15]. Despite this, many of physiological roles of peroxisomes remain enigmatic [16].
The efficient functioning and co-ordination of peroxisomes relies on peroxisome machinery. This entails the regulation of peroxisomal proliferation, maintenance, pexophagy (degradation) and compartmentalization of matrix proteins by an array of specialized proteins [17]. These constitute at least 130 different peroxisomal proteins, and their abundance is specific to cell type, metabolic demand and tissue microenvironment [18]. These peroxisomal proteins include: oxidases for the β-oxidation of fatty acids, such as D-amino acid oxidase, acyl-coA oxidase and acyl-coA synthetase; catalases for hydrogen peroxide scavenging; HMG-CoA reductase for cholesterol synthesis; ATP Binding Cassette subfamily D (ABCD), responsible for transporting matrix proteins; 14 peroxins, such as PEX19, PEX3 and PEX16; and peroxisome-proliferator-activated receptors (PPARs) with their three identified isotopes-PPARα, PPARβ/δ and PPARγ-which are involved in peroxisome proliferation, membrane transport and macropexophagy via regulation of peroxisomal genes [19].
Peroxisomes are central to pathways that are paramount for healthy living cells; they are critical for normal human health and development. It is not surprising, therefore, that disorders of peroxisomes have proven to contribute to the pathophysiology of human neurodegenerative diseases that are attributed fundamentally to altered anti-oxidative stress, such as diabetes, obesity, aging and age-related disorders [20,21]. With the advent of highthroughput next-generation genetic sequencing and advances in molecular approaches, it has been demonstrated that cancer has aberrant metabolism compared to healthy tissue [22]. Subsequently, deregulation of peroxisome dynamics has been hypothesized for its pro-tumorigenic role [23,24]. The earliest plausible peroxisomal role in carcinogenesis was examined in human breast, liver and colonic carcinoma specimens in the 1990s. In these studies, a downregulation of peroxisomal enzymes was found [25][26][27]. No significant difference was noted between numbers of peroxisomes in normal and tumoral breast epithelia; however, catalase-positive organelles and enzyme activity was found to be less numerous and lower in neoplastic colonocytes [28]. Recent data has indicated that enzymes involved in peroxisomal lipid processing are elevated in many tumor types including prostate cancer [29], colorectal carcinomas [30], as well as breast, ovarian and bladder cancer [31]. Intriguingly, it was observed that the activity of specific peroxisomal enzymes correlate with the histopathological grade of the tumor, suggesting peroxisomes as one means of tumor grading [32]. On the other hand, depletion of PEX2 was lethal to hepatic cell carcinoma in a xenograft mouse model via inhibiting the mTOR pathway, a common signaling pathway involved in proliferation and survival of tumor tissue, thus suggesting a context or tissue-specific response to peroxisomal deregulation [33]. However, there is still a significant gap in our knowledge of how peroxisomes influence pro-or anti-tumor metabolism. To the best of our knowledge, there are no studies that either compare peroxisomal dynamics in normal human parotid salivary gland specimens against their neoplastic counterparts or the differences across its common clinical histopathological subtypes (PMA, MEC and ACC). Therefore, we attempt to use immunofluorescence (IF) tissue staining and quantitative polymerase chain reaction (qPCR) techniques to ascertain if significant differences of peroxisomal proteins can be observed at a morphological and or molecular biological level between healthy and neoplastic cells of the human parotid salivary gland.

Pathology of Parotid Tumor Entities Using HE Staining
Using conventional HE staining, resected normal parotid gland tissue specimens appeared as parenchymal (secretory) tissue divided into lobules by the stromal fibrous connective tissue septa in which large excretory interlobular ducts were seen under lower magnification ( Figure 1A,J). Under high magnification, 5-7 round to oval clusters of darkly stained uniform-sized basophilic glandular cells with round basal nuclei and indistinct borders appeared, surrounded by few flat-shaped myoepithelial cells characteristic of normal serous acinar cells. These acinar cells were admixed with few adipocytes and surrounding cuboidal epithelium-lined intra-lobular striated ducts ( Figure 1E,F). In PMA, the normal architecture, as outlined, was partly replaced by pseudo-capsular pleomorphic sheets of duct-like structures, clumps and/or interlacing strands of polygonal or stellate-shaped myo-epithelial cells, but some normal acinar structures are still visible ( Figure 1B,G,K). MEC appeared as non-capsulated solid sheets of epidermoid polygonal cell nests with areas of necrosis and clear cell changes in the form of glycogen accumulation ( Figure 1C,H,L). ACC showed biphasic neoplastic tissue composed of a mixture of cord-like or trabecular glandular cells and basaloid myoepithelial cells with dark angulated nuclei and limited cytoplasm ( Figure 1D,I,M).

Myoepithelial and Acinar Cell Markers in Parotid Tumor Entities
IF staining showed the periductal cytoplasmic staining pattern of αSMA heavy chains, a marker of myoepithelial cells, around striated ducts in healthy tissue (Figure 2A,E) compared to a higher intensity staining around residual acinar and ductal structures in PMA, correlating with the predominant secretory nature of those tumors ( Figure 2B,F). MEC specimens showed classic negative staining for αSMA ( Figure 2C,G), while ACC showed scant, patchy positive staining for αSMA, correlating with islands of activated myofibroblasts ( Figure 2D,H). The parotid specific protein (PSP) antibody showed a specific positive staining of the secretory granules at ultrastructural level and is highly abundant in the human parotid gland [34]. Therefore, we used it as a parotid specific marker protein to identify the correct tissue before characterizing the peroxisomal compartment. Indeed, the labeling of PSP showed a positive staining of the secretory granules ( Figure 2I,M) in the control tissue, suggesting the isolated tissue was a typical parotid gland. Higher magnification of the acini showed many big and spherical secretory granules ( Figure 2M). However, in tumor entities PSP labeling was significantly less abundant compared to the control tissue ( Figure 2I-P). Surprisingly, analysis of mRNA expression levels for PSP did not corroborate with the morphological findings. The Psp mRNA showed a significantly higher expression in the PMA in comparison to the human parotid gland ( Figure 2Q). To ascertain the PMA tumor tissue that was used for experiments, we performed a gene expression profile of possible markers, which were used to differentiate between the different entities. In PMA, compared to healthy tissue, quantitative-PCR showed classic differential mRNA over-expression of carbonic anhydrase 6 (Ca6) and anoctamin 1 (Ano1), markers of differentiated serous acinar cells; P63, a myoepithelial marker; and Sox10, a non-specific marker of salivary glandular tissue. Striking over-expression of pleomorphic adenoma gene 1 (Plag1), a proto-oncogene, was detected in PMA compared to healthy tissue. In contrast, the mRNAs encoding for S100 and Gfap were significantly downregulated in PMA tumors in comparison to control tissue ( Figure 2S,T).

Peroxisomal Biogenesis Proteins
As shown by IF analysis, control parotid tissue showed the typical punctate staining pattern of peroxisomal PEX14p, an established marker protein used to visualize the abundance of peroxisomes, with high fluorescence corresponding to peroxisome membranes. PEX14p intense IF staining was detected in the striated ducts of the control tissue only. In tumorous tissues, PEX14p IF staining was less abundant suggesting less peroxisome biogenesis ( Figure 3A-H). Further, qRT-PCR analysis of mRNA encoding for peroxisomal biogenesis protein 6 (Pex6) showed significantly lower expression in PMA in comparison to healthy tissue. In contrast, analysis for mRNAs encoding for peroxisomal biogenesis proteins (Pex7, Pex10, Pex12), and proteins for peroxisomal proliferation (Pex11α, Pex11β) were significantly upregulated in PMA tumor compared to control tissue ( Figure 3J,K).

Peroxisomal β-Oxidation Enzymes of Pathway 1 and 2
IF showed higher expression of the distinct ATP binding cassette subfamily D3 (ABCD3), required for the transport of lipid into the peroxisomes. ABCD3 was highly expressed in PMA tumor tissue compared to healthy tissue. In contrast, in ACC and MEC, ABCD3 seems to be expressed to a lesser extent in tumor than in healthy tissue ( Figure 4A-H). In qRT-PCR, the mRNA expression levels of Abcd1 and Abcd3 were higher in PMA tumor tissue compared to those in healthy tissue ( Figure 4I). Interestingly, in addition to the ABCD family, all other enzymes involved in β-oxidation, namely Acox1, Acox2, Acox3, Mfp1, Mfp2 and Acaa1, showed higher mRNA expression in PMA type of parotid tissue than healthy tissues in qRT-PCR ( Figure 4J,K).

Catalase, Superoxide Dismutase and Extra-Peroxisomal Antioxidative Enzymes
Catalase exhibits a punctuate pattern in healthy tissue and is particularly abundant in striated ducts ( Figure 6A,E). Since the normal tissue architecture of parotid glands is lost during neoplastic transformation ( Figure 6B-D,F-H), we were only able to observe an apparently higher expression of catalase in tumor tissue without reference to a specific glandular structure. Moreover, catalase was mis-localized into cytoplasm in all tumor entities in comparison to the healthy tissue ( Figure 6B-H). In qRT-PCR, catalase showed a significant upregulation in tumor tissue in comparison to control tissue ( Figure 6I). It is worth noting that mRNA expression levels of other antioxidative enzymes, namely peroxiredoxin 1 (Prdx1), glutathione peroxidase (Gpx), but not superoxide dismutase 1 (Sod1), were similarly higher in the PMA type of tumor tissue than in healthy parotid tissue ( Figure 6J-L). Specifically, superoxide dismutase 2 (SOD2) showed a scattered pattern in tumor tissue under IF ( Figure 7B-D,F-H) as opposed to its typical mitochondrial localization in healthy tissue ( Figure 7A,E). Moreover, thioredoxin reductase 1 (TRX1), an antioxidant protein enzyme ubiquitous in mammals and essential for life, showed typical IF staining in healthy tissue but an incremental decrease in intensity from PMA, to MEC and ACC accordingly. The ACC type of tumor tissue showed a complete absence of TRX1 IF staining ( Figure 7I-P). In qRT-PCR, mRNA expression levels of Sod2, heme oxygenase 1 (Ho-1) and glutathione reductase (Gr) were higher in the PMA type of parotid tumor tissue ( Figure 7Q-T) in comparison to healthy tissue. Surprisingly, mRNA expression analysis for Trx1 showed a significant upregulation in PMA in comparison to healthy tissue, which is not in agreement with IF staining. Moreover, TRX2 was slightly upregulated in PMA; however, this was not statistically significant ( Figure 7T).

Peroxisome Proliferator-Activated Receptors (PPARs)
PPARs regulate peroxisomal proteins by binding to promoters of peroxisomal genes. Therefore, all three PPAR family members (PPARα, -β and -γ) were characterized in parotid tumors. Interestingly, PPARα was downregulated in all tumor entities in comparison to healthy parotid tissue ( Figure 8A-H). However, a clear upregulation of PPARβ was detected in PMA and in MEC while a downregulation was detected in ACC parotid tumor entities in comparison to healthy tissue ( Figure 8I-P). The antibody against PPARγ did not yield any reaction. In contrast, qRT-PCR results demonstrated a significant upregulation of Pparß and -γ in neoplastic tissues compared to healthy tissue ( Figure 8R, S). Further, a downregulation of Pparα was noted in PMA in comparison to healthy tissue; however, this was not statistically significant ( Figure 8Q).

Discussion
Rapidly proliferating tumor cells have their cellular metabolism reprogrammed by the direct and indirect consequences of underlying oncogenic mutations. Tumor cells with rewired metabolic pathways confer the selective advantage of enhanced influx of necessary and unconventional nutrients that are required to sustain deregulated proliferation towards creation or expansion of its biomass [11,23]. One of the many metabolic demands of a tumor cell is meeting its need for the massive amount of cell membrane production required to progress. As peroxisomes play a cornerstone role in lipid metabolism, we started our experiments by examining the gene expression of proteins that regulate peroxisomal biosynthesis, the peroxins (PEX), which are peroxisomal biogenesis factors essential in regulating peroxisomal assembly and function. To date, 34 PEX genes encoding peroxins have been identified [35,36]. Of these, at least 13 PEX genes are required for proper peroxisomal biogenesis [37]. The universal lower expression of PEX14p, a docking protein required for translocation of matrix proteins into the peroxisomes, in our tumor samples suggests an impaired peroxisomal biogenesis parotid neoplasia. Only PEX11 and PEX12 were differentially upregulated in parotid tumor tissue compared healthy parotid tissue. Downregulation of some peroxisomal biogenesis proteins, such as PEX6, alongside upregulation of others (PEX 7, 10, 11α, 11β, 12) in our PMA samples may suggest tumorigenic addiction, and thus could be targeted therapeutically. However, while the implications of upregulation of certain PEX genes compared to others in tumor parotid cells are unclear, this could indicate the presence of a deregulated peroxisomal biogenesis process. Mutations in PEX genes typically result in defective peroxisomes due to the mis-localization of peroxisomal matrix proteins to the cytosol [38]. Inherited bi-allelic mutations in PEX genes cause peroxisomal biogenesis disorders (PBD). PBDs often have severe neurological manifestations attributed to defective neuronal myelin sheath, which results from peroxisomal defects in fatty acid metabolism [39] Peroxisomal enzymes oxidize both linear and branched forms of very long chain unsaturated fatty acids to generate fatty acyl-CoA and acetyl-CoA that substrates enter in plasma membrane biosynthesis [40]. The transport of these substrates across peroxisomal membranes is dependent on the ATP-binding cassette (ABC) transporter subfamily D (ABCD1, ABCD2, ABCD3) [41]. Functions of these subfamily members overlap; however, ABCD1 and ABCD2 preferentially transport hydrophobic saturated very long chain fatty acids, while ABCD3 (PMP70) distinctly imports hydrophilic branched chain fatty acyl-CoAs across the peroxisomal membrane [42,43]. Another family of enzymes is the FADdependent acyl-coA oxidases and multifunctional proteins (MFP) involved in subsequent βoxidation of the imported acyl-coA substrates [44][45][46]. Indeed, we found a characteristically higher expression of ABCD3 in parotid tumor cells compared to their healthy counterpart, suggesting an enhanced lipid trafficking across peroxisomes. This is consistent with previous findings that dysregulation of ABC protein expression takes place in both solid and diffuse tumors [47,48].
Interestingly, the lower overexpression of ABCD3 in our malignant MEC and ACC tissue compared to PMA, a tumor of benign nature, may suggest a metabolic shift to alternative pathways during malignant transformation of benign parotid tumors. Thus, in parotid gland tumors, downregulation of ABCD3 transporter may confer an association with progression to a malignant phenotype such as to MEC and ACC; however, further studies remain to confirm or negate our hypothesis [49]. Moreover, the aberrant overexpression of the ACOX and MFP families of enzymes found exclusively in PMA parotid tumor cells are in concordance with a previous finding in certain subtypes of breast cancer cells [50]. However one should keep in mind that not all protein expression is attributed to correlative changes in mRNA levels [51].
In the same context, our parotid tumor cells showed elevated levels of lipogenic enzymes (i.e., plasmalogen and cholesterol synthesis enzymes). This is consistent with other studies suggesting that high levels of ether phospholipids in some tumor types implies that elevated peroxisomal lipid synthesis is associated with tumor progression [52].
It cannot be ignored that peroxisomal enzymes of β-oxidation yield byproducts such as hydrogen peroxide (H 2 O 2 ), reactive oxygen and nitrogen species (ROS/RNS) [53]. Since these byproducts are genotoxic and, if chronic, can result in tumorigenesis, then parotid cells possess protective mechanisms through which they can regulate resident radicles. To do so, peroxisomes harbor six enzymes with antioxidant function, namely peroxiredoxin 1 (PRDX1), glutathione peroxidase (GPX), superoxide dismutase 1 (SOD1), glutathione S transferase (GTS1), GTS-kappa and catalase. This also includes other cytoplasmic enzymes such as thioredoxin reductase 1 (TRX1) and heme oxygenase 1 (HO-1). It is thus conceivable that higher expression levels of the aforementioned antioxidant enzymes in parotid tumor tissue are either adaptation or selection to chronic redox imbalance, such that secondary high levels of antioxidant enzymes promote survival of tumor cells or precipitate apoptosis when targeted by a pro-oxidant chemotherapeutic [24,54,55]. Such findings further support the potential role of peroxisomes in cancer through their β-oxidation of fatty acids and offer perspective therapeutic targets.
Furthermore, an active area of research is examining the role of peroxisome proliferatoractivated receptors (PPARs) in cancer. PPARs are ligand-activated transcription factors that belong to the nuclear hormone-receptor family and are activated by fatty acid and lipid ligands [56]. Accumulating evidence shows that long-term activation of PPARα induces hepatocellular carcinoma in rodent, but not human, liver, suggesting that disorders in lipid metabolisms can be implicated in carcinogenesis. Conversely, PPARγ was over expressed and localized to the cytoplasm in salivary gland duct cancer suggesting it as a potential therapeutic target [57]. In fact, parotid tumor cells showed overexpression of all PPAR family members, which was inconsistent with their immunohistochemistry (IHC) staining. We thus believe that even though PPARβ and PPARγ were shown previously to have anti-proliferative, pro-apoptotic effects, their implication with regards to cancer is still unclear and warrants further investigation [56,58].

Materials and Methods
All subjects gave their informed consent for inclusion before they participated in the study. This study was conducted in accordance with the 1964 Declaration of Helsinki, its latest amendments and in concordance with our institutional and national research committee standards. The Ethics Committee of Justus Liebig University Giessen approved the study protocol (AZ 95/15, 25 June 2015).

Tissue Harvesting and Preparation
After obtaining informed consent from all participants, fresh human tissue samples were obtained following elective surgical resection of five parotid gland tumors from five patients. A healthy parotid tissue sample from each patient was included among the excised tumor mass. All Parotid tissue specimens were then immediately cryopreserved in liquid nitrogen at −80 • C for storage and transferred to our laboratory for subsequent tissue processing. One portion of our parotid specimens was processed for routine staining by hematoxylin and eosin (HE), immunofluorescence (IF), as well as Alcian blue (AB) and periodic acid Schiff (PAS) staining, while the other portion was designated for RNA extractions for detection and quantification of gene expression level.

Paraffin Embedding and Antigen Retrieval
Parotid tissue specimens, containing both healthy and tumor tissue, were transferred to 4% PFA in PBS at a pH of 7.4 and kept at 4 • C overnight. On the next day, Parotid tissue specimens were then embedded in paraffin (Paraplast Plus, St. Louis, MO, USA) using a Leica TP1020 embedding machine (3× 70%, 80%, 90%, 100% alcohol-90 min each; 2× xylene-90 min each; 2× paraffin-120 min each). Paraffin-embedded parotid specimens were sectioned and cut into 2, 3 and 5 µm sections with a rotation microtome (Leica SM 2000 R, Leica Instruments Nussloch, Germany) and applied on SuperFrost Plus (+) slides (Shandon, Frankfurt, Germany). Paraffin embedding, tissue sectioning and subsequent IF staining (discussed below) were carried out according to our previous publication [59,60]. Paraffin-embedded parotid sections were then immersed in xylene solvent and incubated over 3 min for deparaffinization, followed by rehydration in a series of ethanol (2× 99%, 96%, 80%, 70%, 50%, 2× aqua dest.-3 min each). Subsequently 0.01% trypsin was applied for 10 min at 37 • C to improve antigen retrieval and accessibility of epitopes. After washing in PBS, parotid sections were then put in citrate buffer (pH 6) for 3× 5 min in a microwave (600 W) followed by gradual cooling to room temperature (RT). A 4% bovine serum albumin in PBST was then used as a blocking solution for two hours.

Hematoxylin and Eosin (HE) Staining
To identify the general morphology, we used a regular hematoxylin and eosin (HE) staining. Alcian blue and periodic acid Schiff (PAS) staining were performed to detect parotid secretory protein (PSP), a human salivary protein that is expressed in acinar and ductal cells of the parotid gland, submandibular gland and in gingival epithelial cells. PSP staining allows us to distinguish parotid glandular tissue from surrounding non-glandular stromal cells within our tissue sections [34].

Immunofluorescence (IF) Staining
After establishing the general morphology of parotid tissue sections, we aimed at distinguishing our parotid tumor samples using known parotid tumor markers. We then aimed at investigating the distribution, numerical abundance and protein composition of the peroxisomal compartment in parotid gland tumor entities in comparison to the healthy tissue. For this purpose, we used a wide array of primary antibodies against peroxisomal, mitochondrial and cytoskeleton proteins ( Table 1). The validity and specificity of our chosen antibodies against peroxisomal membrane and matrix proteins have been previously confirmed in the parotid gland [34] and other human tissue organs [59][60][61][62]. Following an overnight incubation of those antibodies with our antigen-retrieved parotid sections, sections were washed three times to remove unbound primary antibodies. Next, commercially available fluorescently labeled secondary antibodies were used: anti-rabbit IgG Alexa Fluor 488 or anti-mouse Texas Red (Table 1) were subsequently added for visualization as previously described. Parotid sections without primary antibodies were incubated in parallel and served as our negative control. Finally, we counterstained the nuclear structure of our parotid cells using TOTO-3 iodide and DAPI diluted in PBS, in which blue-fluorescence stands out in vivid contrast to green, yellow or red fluorescent probes of other cellular structures in our parotid cells. Afterwards, all sections were covered with Mowiol 4.88 and N-propyl gallate in a ratio of 3:1. A Leica TCS SP5 confocal laser scanning microscope with a 63× objective and "Airy 1" setting was used for examinations and image acquisition of all our histopathological parotid gland sections.

RNA Extraction and Isolation
Using TissueLyser LT (Qiagen, Hilden, Germany), stored parotid tissue sections were homogenized to comparatively analyze the peroxisomal compartment in both healthy and neoplastic cells. An RNeasy Mini Kit (Qiagen, Hilden, Germany) was then used for purpose of RNA isolation and extraction. We adopted the manufacturer's instructions previously proven to yield a high-quality RNA extract [34]. Finally, we used Agilent 2100 Bioanalyzer system and the Agilent RNA 6000 Nano Kit (Agilent Technologies, Santa Clara, CA, USA) to verify the quality and concentration of our RNA extract.

Quantitative Reverse Transcriptase Polymerase Chain Reaction (qPCR)
In order to obtain a gene-encoding DNA copy, the quantitative reverse transcriptase polymerase chain reaction (qRT-PCR) was adopted. We used the high-capacity RNA-to-cDNA Kit (Applied Biosystems, Weiterstadt, Germany) and the C1000 Thermal Cycler PCR system (BioRad, Dreieich, Germany). First, primers were designed using the PrimerQuest Tool (http://eu.idtdna.com/Primerquest/Home/Index, accessed date: 3 June 2018), and manufactured by Eurofins MWG Operon (Table 2). We started with a primer concentration of 5 pmol/µL, and qRT-PCR was conducted using the SYBR Select Master Mix Kit (Life Technologies, Carlsbad, California, United States of America) and the StepOnePlus Real-Time PCR System (Life Technologies, Carlsbad, California, United States of America). The following operational standards were used: denaturation at 94 • C for 4 min; 45 cycles of denaturation at 95 • C for 15 s; annealing at 60 • C for 60 s, and extension at 7 • C for 1 min.
As a final step, RT-values were divided by housekeeper HPRT values in order to normalize for different mRNAs. The RNA of three different human healthy and tumor parotid tissues were used in duplicates.

Statistical Methods and Tools
To calculate the statistical significances, we used the unpaired t-test with the Graphpad Prism software, version 6.01. Fold changes were analyzed by an approximation method taking the ∆∆Ct values into consideration.

Conclusions
We demonstrated the presence of differential expression of peroxisomal proteins between healthy and tumor tissue of the human parotid salivary gland. Upregulation of biosynthesis alongside downregulation of antioxidant key enzymes suggests that peroxisomes in parotid gland tumors have a pro-tumorigenic role. Such a role appears variable between subtypes of parotid tumors. Our findings supplement current research in the field and add perspectives on the effectiveness of peroxisomes as potential molecular targets and or predictive biological markers.  Institutional Review Board Statement: All subjects gave their informed consent for inclusion before they participated in the study. This study was conducted in accordance with the 1964 Declaration of Helsinki, its latest amendments and in concordance with our institutional and national research committee standards. The Ethics Committee of Justus-Liebig-University Giessen approved the study protocol (AZ 95/15, 25 June 2015).

Informed Consent Statement:
Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The data presented in this study are contained within the article. Further supporting material is available upon request.