Is the proteome of bronchoalveolar lavage extracellular vesicles a marker of advanced lung cancer?

cancers Article Is the Proteome of Bronchoalveolar Lavage Extracellular Vesicles a Marker of Advanced Lung Cancer? Ana Sofia Carvalho 1,* , Maria Carolina Strano Moraes 2, Chan Hyun Na 3, Ivo Fierro-Monti 1, Andreia Henriques 1, Sara Zahedi 1, Cristian Bodo 2, Erin M Tranfield 4, Ana Laura Sousa 4, Ana Farinho 5, Luís Vaz Rodrigues 6, Paula Pinto 7, Cristina Bárbara 8, Leonor Mota 7, Tiago Tavares de Abreu 7, Júlio Semedo 7, Susana Seixas 9, Prashant Kumar 10,11 , Bruno Costa-Silva 2, Akhilesh Pandey 10,11,12 and Rune Matthiesen 1,* 1 Computational and Experimental Biology Group, Chronic Diseases Research Centre, NOVA Medical School, Faculdade de Ciencias Medicas, Universidade NOVA de Lisboa, Campo dos Martires da Patria, 130, 1169-056 Lisboa, Portugal; [email protected] (I.F.-M.); [email protected] (A.H.); [email protected] (S.Z.) 2 Systems Oncology Group, Champalimaud Research, Champalimaud Centre for the Unknown, Av. Brasilia, Doca de Pedroucos, 1400-038 Lisbon, Portugal; [email protected] (M.C.S.M.); [email protected] (C.B.); [email protected] (B.C.-S.) 3Department of Neurology, Institute for Cell Engineering, Johns Hopkins University School of Medicine, Baltimore, MD 21205, USA; [email protected] 4Electron Microscopy Facility, Instituto Gulbenkian de Ciência—Rua da Quinta Grande, 6, 2780-156 Oeiras, Portugal; [email protected] (E.M.T.); [email protected] (A.L.S.) 5iNOVA4Health—Advancing Precision Medicine, CEDOC—Chronic Diseases Research Centre, NOVA Medical School/Faculdade de Ciências Médicas, Universidade NOVA de Lisboa, Campo dos Martires da Patria, 130, 1169-056 Lisboa, Portugal; [email protected] 6 Department of Pneumology, Unidade Local de Sa ú de da Guarda (USLGuarda), 6300-659 Guarda, Portugal; [email protected] 7Unidade de Técnicas Invasivas Pneumológicas, Pneumologia II, Hospital Pulido Valente, Centro Hospitalar Lisboa Norte, 1649-028 Lisbon, Portugal; [email protected] (P.P.); [email protected] (L.M.); tavaresdeabr[email protected] (T.T.d.A.); [email protected] (J.S.) 8Instituto de Saúde Ambiental (ISAMB), Faculdade de Medicina, Universidade de Lisboa, Centro Hospitalar Universitário Lisboa Norte, 1649-028 Lisbon, Portugal; [email protected] 9Instituto de Investigação e Inovação em Saúde (I3S), Universidade do Porto, 4200-135 Porto, Portugal; [email protected] 10 Institute of Bioinformatics, Discoverer building, ITPL, Bangalore 560066, India; [email protected] (P.K.); pandey[email protected] (A.P.) 11 Manipal Academy of Higher Education (MAHE), Manipal 576104, India 12 Department of Laboratory Medicine and Pathology, Mayo Clinic, Rochester, MN 55905, USA *Correspondence: [email protected] (A.S.C.); r[email protected] (R.M.) Received: 15 October 2020; Accepted: 19 November 2020; Published: 20 November 2020   Simple Summary: Bronchoalveolar lavage is routinely collected during bronchoscopy for cytology analysis in the diagnostic of lung cancer. Due to low sensitivity of this method, early-stage cancers are undetected, lowering the treatment success. In this study, we analyzed extracellular vesicles isolated from bronchoalveolar lavage of lung cancer suspects by mass spectrometry-based proteomics. The protein composition of bronchoalveolar lavage extracellular vesicles of late-stage cancer showed a higher proteome complexity associated with mortality within the two year follow-up period. We identified a potential therapeutic target DNMT3B complex which was significantly expressed in bronchoalveolar lavage extracellular vesicles as well as in tumor tissue. Bronchoalveolar lavage Cancers 2020,12, 3450; doi:10.3390/cancers12113450 www.mdpi.com/journal/cancers Cancers 2020,12, 3450 2 of 18 extracellular vesicles proteome analysis of immune markers indicates the presence of markers of innate immune and fibroblast cells. Abstract: Acellular bronchoalveolar lavage (BAL) proteomics can partially separate lung cancer from non-lung cancer patients based on principal component analysis and multivariate analysis. Furthermore, the variance in the proteomics data sets is correlated mainly with lung cancer status and, to a lesser extent, smoking status and gender. Despite these advances BAL small and large extracellular vehicles (EVs) proteomes reveal aberrant protein expression in paracrine signaling mechanisms in cancer initiation and progression. We consequently present a case-control study of 24 bronchoalveolar lavage extracellular vesicle samples which were analyzed by state-of-the-art liquid chromatography-mass spectrometry (LC-MS). We obtained evidence that BAL EVs proteome complexity correlated with lung cancer stage 4 and mortality within two years ´ follow-up (pvalue =0.006). The potential therapeutic target DNMT3B complex is significantly up-regulated in tumor tissue and BAL EVs. The computational analysis of the immune and fibroblast cell markers in EVs suggests that patients who deceased within the follow-up period display higher marker expression indicative of innate immune and fibroblast cells (four out of five cases). This study provides insights into the proteome content of BAL EVs and their correlation to clinical outcomes. Keywords: extracellular vesicles; lung cancer; proteomics; immuno oncology; bronchoalveolar lavage 1. Introduction At present, lung cancer persisted as the most prevalent oncological disease with an estimated incidence of more than 2 million new cases worldwide, and a mortality rate of 84% with a total of 1.8 million deaths yearly [ 1 , 2 ]. The late detection of the disease is the main cause of such a dismal outcome as demonstrated by the five-year overall survival rates for non–small cell lung cancer (NSCLC): 14–49% for stages I to IIIA, and <5% for stage IIIB/IV [3]. The characterization of lung tumor tissue and the surrounding microenvironment at the molecular level contributed to the increased knowledge on the physiology of the disease as well as the design of novel therapies. Nevertheless, the collection of tissue biopsies of peripheral adenocarcinomas in distal airways by minimally invasive techniques is limited, and during the past years, several studies have targeted different types of biological specimens, which range from tumor tissues to different types of liquid biopsies [ 4 ]. Bronchoalveolar lavage (BAL) liquid biopsies are obtained from minimally invasive procedures, for example, fiber optic bronchoscopy. Bronchoscopy is typically performed upon suspicion of lung cancer, for example, derived from previously obtained imaging of the lungs. Due to the closer anatomical proximity of BAL compared to saliva, more accurate biomarker signatures are expected from BAL than from biofluids such as blood, saliva, or sputum. Blood and saliva, on the other hand, might serve as potential markers for early screening. In the past, several molecular technologies were applied to profile the molecular content of BAL such as miRNA [ 5 ], mRNA [ 6 ], DNA [ 7 ], DNA methylation [ 8 ], metabolites [ 9 ], microbiota [ 10 ], and proteomics. Over the years, proteomics technologies have evolved and consequently, BAL proteome profiling has been attempted multiple times. The first BAL proteomics profiling applied MALDI TOF-MS [ 11 ]. Later, 2D-PAGE and MALDI-TOF were combined [ 12 ]. Nowadays, liquid chromatography-mass spectrometry (LC-MS) approaches are typically applied for BAL proteome profiling [ 13 – 15 ]. Abundant proteins like albumin partially hamper the protein coverage obtainable by direct LC-MS analysis of acellular BAL samples. Therefore, Sim et al. [ 16 ] established a novel methodology based on combining antibody-based depletion of high abundant BALF proteins, high pH peptide fractionation, and label-free quantitation on a high-resolution Orbitrap Fusion Lumos instrument. However, antibody depletion of abundant Cancers 2020,12, 3450 3 of 18 proteins risks the removal of clinically relevant target proteins. Consequently, it is imperative to explore additional sample preparation methodologies. The enrichment of small (sEVs) and large (lEVs) EVs represents an alternative methodology for the depletion of abundant proteins in biomarker studies based on liquid biopsies [ 17 ]. Currently, no study has targeted BAL sEVs and lEVs from suspected lung cancer patients. We, therefore, performed an exploratory case-control study to evaluate the protein complexity from proteomics profiling of BAL EV fractions. The sEVs proteins constituted the most complex proteome when compared to acellular BAL, vesicle-depleted BAL, and lEVs. We identified and quantified 7484 protein isoforms from 3158 encoding genes from sEVs. The lEVs and sEVs were enriched by sequential centrifugation followed by ultracentrifugation and floatation on sucrose gradient cushion, which captures all the subpopulations of sEVs while depleting from non-vesicular protein aggregates and complexes. Besides, the potential of BAL sEV proteome as a source of biomarkers was explored. The data obtained suggested that BAL sEV proteome complexity correlates with cancer stage IV and death (Wilcoxon rank-sum test pvalue =0.006). The proteomes were compared, by system biology approaches, to previously obtained proteomes from lung tumor tissues and acellular BAL proteome [ 13 , 18 ]. sEVs and tumor tissue displayed a common significant regulation of several enriched functional categories with potential therapeutic value. Finally, sEV protein markers for innate immune and fibroblast cells correlated with poor prognosis (mortality within two years follow-up in four out of five cases). 2. Results 2.1. Outline of Study As a proof of concept, we compared different BAL fractions by high-throughput mass spectrometry analysis such as acellular BAL, vesicle depleted BAL, lEVs and sEVs from control, and lung cancer patients. The proteome content was compared by LC-MS (Figure S1). We observed that sEVs contained the larger proteome complexity with the most significant potential for biomarker discovery (Section 2.2 and Figure S2). Therefore, sEVs were isolated from 24 acellular BAL samples and analyzed by mass spectrometry and further characterized by transmission electron microscopy (TEM), western blot, and nanoparticle tracking analysis. The quantitative MS data were compared with iBAQ values obtained from reanalyzed data of previous studies on acellular BAL (PXD004700) [ 13 ] and patient-matched tumor versus normal tissue (PXD000853) [18]. 2.2. Proteome Content of lEVs, sEVs, Acellular BAL, and EV Depleted BAL (DB) BAL fluids (acellular BAL, aBAL) from two different cancer status samples (NO and YES) were fractionated by sequential centrifugation and ultracentrifugation into three fractions: (1) depleted BAL (DB), (2) lEVs and (3) sEVs (Figure S1). To address which BAL fraction holds the greatest promise in terms of significant cancer status discrimination based on protein identification and quantitation, each fraction was analyzed by LC-MS in parallel with the respective aBAL fluid. Figure 1compares the protein identification in cancer versus control for (a) sEVs, (b) BAL, (c) lEVs, and (d) DB. The sEVs fraction resulted in the highest number of protein identifications and the highest number of proteins unique to cancer (Figure 1a). KEGG functional enrichment of all the identifications from each of the fractions was compared to address the question of which fraction has the greatest potential for separating cancer status (Figure 1e). Figure S2 displays the overlap between the identified proteins in each fraction type obtained from LC-MS analysis. The highest numbers of quantified proteins were obtained in BAL sEVs from YES and NO cancer status samples. On average the number of identified proteins from sEVs was almost double compared to aBAL, while in BAL vesicle-depleted and lEVs fractions the number of identified proteins was lower, with the lEVs proteome being the simplest proteome (Figure S2). KEGG functional enrichment analysis of all identified proteins revealed functional diversity between the BAL fractions analyzed (Figure 1e). The samples clustered according to BAL fractions based on the functional KEGG annotation, strongly suggesting that each fraction contains distinct protein functional groups. The identified proteins overall Cancers 2020,12, 3450 4 of 18 exhibited significant enrichment in pathways related to infection, inflammation, and complement and coagulation pathways. sEVs uniquely displayed significant enrichment in proteasome and ribosome factors while no enrichment in glycolysis and glycogenesis factors was observed in contrast to the other BAL fractions (Figure 1). Proteasome and ribosome factors are known as abundant factors in cancer due to increased protein turnover. Although lEV fractions resulted in the lowest amount of identifications, it presented unique functional enrichment related to, e.g., chemokine and synapse signaling in cancer. In turn, the two vesicles containing fractions displayed the maximum separation in terms of sample clustering. None of the fractions fully encapsulated all enriched functional categories. Proteomics analysis of sEV fractions showed the highest number of identified proteins and the highest diversity of functional enriched KEGG pathways. In this view, we have analyzed the proteome of sEVs for biomarkers discovery in a case-control study of 12 controls and 12 lung cancer cases. Cancers 2020, 12, x FOR PEER REVIEW 4 of 20 suggesting that each fraction contains distinct protein functional groups. The identified proteins overall exhibited significant enrichment in pathways related to infection, inflammation, and complement and coagulation pathways. sEVs uniquely displayed significant enrichment in proteasome and ribosome factors while no enrichment in glycolysis and glycogenesis factors was observed in contrast to the other BAL fractions (Figure 1). Proteasome and ribosome factors are known as abundant factors in cancer due to increased protein turnover. Although lEV fractions resulted in the lowest amount of identifications, it presented unique functional enrichment related to, e.g., chemokine and synapse signaling in cancer. In turn, the two vesicles containing fractions displayed the maximum separation in terms of sample clustering. None of the fractions fully encapsulated all enriched functional categories. Proteomics analysis of sEV fractions showed the highest number of identified proteins and the highest diversity of functional enriched KEGG pathways. In this view, we have analyzed the proteome of sEVs for biomarkers discovery in a casecontrol study of 12 controls and 12 lung cancer cases. Figure 1. Proteome overview in bronchoalveolar lavage (BAL) fractions. Venn diagrams comparing cancer (YES) versus control (NO) in: (a) sEVs, (b) BAL, (c) lEVs, and (d) DB; (e) Heat-map of KEGG Figure 1. Proteome overview in bronchoalveolar lavage (BAL) fractions. Venn diagrams comparing cancer (YES) versus control (NO) in: ( a ) sEVs, ( b ) BAL, ( c ) lEVs, and ( d ) DB; ( e ) Heat-map of KEGG pathway enrichment analysis. Functional enrichment analysis was performed using R. p-value of each KEGG pathway was <0.05. The shades of the colors reflected the –log10 (p-value) of the enrichment analysis of proteins identified in different subcellular BAL fractions for NO (white) and YES (black) cancer status samples. Large extracellular vesicles (lEVs), BAL vesicle-depleted (DB), acellular BAL (BAL), small extracellular vesicles (sEVs). Cancers 2020,12, 3450 5 of 18 2.3. Baseline Characteristic of Clinical Samples for the Enrichment of sEVs To explore the potential of sEVs proteome ability to classify clinical samples, we selected 12 controls (non-lung cancer cases) and 12 lung cancer samples in a case-control design (Table 1) from our previous analyzed cohort of 91 patients suspected of lung cancer [ 13 ]. The lung cancer status reflects the diagnosis after two years of follow-up. Approximately 42% of the cancer patients died within the two year follow-up period. The selection of samples for sEV characterization was optimized to prevent statistical association of age, gender, smoking status, and experimental batch effects with lung cancer status. As expected, the lung cancer status displayed an association to two-year survival and cancer staging. sEVs from the 24 clinical samples were isolated following the protocol outlined in (Figure S1). Table 1. Baseline characteristic of the clinical samples. NO (N=12) YES (N=12) Total (N=24) pValue Age 0.355 <55 4 (33.3%) 1 (8.3%) 5 (12.5%) >55 1 (8.3%) 1 (8.3%) 2 (8.3%) NA 7 (58.3%) 10 (83.3%) 17 (70.8%) Smoking History 0.370 Current smoker 4 (33.3%) 3 (25.0%) 7 (29.2%) Former smoker 2 (16.7%) 6 (50.0%) 8 (33.3%) Nonsmoker 2 (16.7%) 1 (8.3%) 3 (12.5%) Unknown 4 (33.3%) 2 (16.7%) 6 (25.0%) Gender 0.100 NA 0 (0.0%) 1 (8.3%) 1 (4.2%) F 8 (66.7%) 3 (25.0%) 11 (45.8%) M 4 (33.3%) 8 (66.7%) 12 (50.0%) Batch 0.390 1 2 (16.7%) 5 (41.7%) 7 (29.2%) 2 9 (75.0%) 6 (50.0%) 15 (62.5%) 3 1 (8.3%) 1 (8.3%) 2 (8.3%) Stage <0.001 2 0 (0.0%) 1 (8.3%) 1 (4.2%) 3 0 (0.0%) 1 (8.3%) 1 (4.2%) 4 0 (0.0%) 5 (41.7%) 5 (20.8%) NA 1 (8.3%) 5 (41.7%) 6 (25.0%) No 11 (91.7%) 0 (0.0%) 11 (45.8%) Status 0.012 Alive 12 (100.0%) 7 (58.3%) 19 (79.2%) Dead 0 (0.0%) 5 (41.7%) 5 (20.8%) 2.4. Nanoparticle Tracking Analysis of Isolated EVs The enriched sEVs were characterized by nanoparticle tracking analysis (Figure 2and Figure S3). Figure 2displays representative particle size distribution for one control (a) and one lung cancer sample (b). All the normalized particle distributions are plotted in Figure S3 and color-coded with clinical status. The distributions display high similarity across samples. Based on a t-test on the normalized counts for each particle size bin of 0.5 nm, no statistical differences between the distributions for control versus cancer were observed (Figure S3). All distributions displayed a maximum peak around 100 nm, which represents the expected size mode of sEVs. The precise position of the maximum peak differed slightly between samples. The different samples displayed slightly different peak tops at larger particle sizes likely representing different types of non-vesicle particles such as protein complexes, aggregates, and eventually virus particles. A comparison of overall EV protein content Cancers 2020,12, 3450 6 of 18 and overall total particle counts from control versus cancer samples showed no significant differences (Figure 2c,d). Cancers 2020, 12, x FOR PEER REVIEW 6 of 20 clinical status. The distributions display high similarity across samples. Based on a t-test on the normalized counts for each particle size bin of 0.5 nm, no statistical differences between the distributions for control versus cancer were observed (Figure S3). All distributions displayed a maximum peak around 100 nm, which represents the expected size mode of sEVs. The precise position of the maximum peak differed slightly between samples. The different samples displayed slightly different peak tops at larger particle sizes likely representing different types of non-vesicle particles such as protein complexes, aggregates, and eventually virus particles. A comparison of overall EV protein content and overall total particle counts from control versus cancer samples showed no significant differences (Figure 2c,d). Figure 2. Nanoparticle tracking analysis of sEV samples: (a) particle counts versus particle size in nm for a control sample; (b) and a lung cancer sample; (c) protein concentration in sEVs versus cancer status; (d) particle concentration versus cancer status. 2.5. Transmission Electron Microscopy and Overall Proteome of EVs In accordance with nanoparticle tracking analysis, TEM analysis of sEVs revealed vesicles of similar sizes around 100 nm (Figure 3a). Western blot analysis of the exosome marker CD63 comparing BAL to sEVs fraction showed a fortyfold enrichment (Figure S4). All identified proteins from controls and lung cancer in sEV were compared to proteins deposited in the exosome database ExoCarta, showing an overlap >70% for the two different cancer status samples (Figure 3b). However, such comparisons exclude protein quantitation information and consequently are not suited for Figure 2. Nanoparticle tracking analysis of sEV samples: ( a ) particle counts versus particle size in nm for a control sample; ( b ) and a lung cancer sample; ( c ) protein concentration in sEVs versus cancer status; (d) particle concentration versus cancer status. 2.5. Transmission Electron Microscopy and Overall Proteome of EVs In accordance with nanoparticle tracking analysis, TEM analysis of sEVs revealed vesicles of similar sizes around 100 nm (Figure 3a). Western blot analysis of the exosome marker CD63 comparing BAL to sEVs fraction showed a fortyfold enrichment (Figure S4). All identified proteins from controls and lung cancer in sEV were compared to proteins deposited in the exosome database ExoCarta, showing an overlap >70% for the two different cancer status samples (Figure 3b). However, such comparisons exclude protein quantitation information and consequently are not suited for assessing the exosome content in the sEV fractions. We, therefore, developed a bioinformatics methodology to quantitatively estimate exosome content in sEV fractions (Figure 3c). The heatmap displays average iBAQ values ranked from 0 to 1 for frequently reported exosome markers in the literature and the 10 most frequently reported proteins in ExoCarta. The quantitative values from cell line sEVs were obtained by re-analyzing previously published MS data and serve as reference samples. The reference data for sEVs from cell lines were obtained by two different EV enrichment methodologies: ( 1) PEG -based precipitation [ 19 ] and (2) differential ultra-centrifugation [ 20 ]. The two methods resulted in the identification and quantification of similar exosome markers. We compared the quantitative proteomics data on sEVs Cancers 2020,12, 3450 7 of 18 from human clinical samples to the reference data on sEVs from cell lines. This constitutes a prudent MS-based assessment of exosome content. The clinical BAL sEVs isolated in this study exhibited similar expression levels of exosome markers when compared to the previous data from sEVs isolated from cell lines (Figure 3c) [ 19 , 20 ]. Additionally, the level of non EV proteins is considerably lower in BAL sEVs compared to cell lines sEVs enriched using the two methods while PEG isolated EVs from cell lines displayed the highest level of non EV protein markers. Cancers 2020, 12, x FOR PEER REVIEW 7 of 20 assessing the exosome content in the sEV fractions. We, therefore, developed a bioinformatics methodology to quantitatively estimate exosome content in sEV fractions (Figure 3c). The heatmap displays average iBAQ values ranked from 0 to 1 for frequently reported exosome markers in the literature and the 10 most frequently reported proteins in ExoCarta. The quantitative values from cell line sEVs were obtained by re-analyzing previously published MS data and serve as reference samples. The reference data for sEVs from cell lines were obtained by two different EV enrichment methodologies: (1) PEG-based precipitation [19] and (2) differential ultra-centrifugation [20]. The two methods resulted in the identification and quantification of similar exosome markers. We compared the quantitative proteomics data on sEVs from human clinical samples to the reference data on sEVs from cell lines. This constitutes a prudent MS-based assessment of exosome content. The clinical BAL sEVs isolated in this study exhibited similar expression levels of exosome markers when compared to the previous data from sEVs isolated from cell lines (Figure 3c) [19,20]. Additionally, the level of non EV proteins is considerably lower in BAL sEVs compared to cell lines sEVs enriched using the two methods while PEG isolated EVs from cell lines displayed the highest level of non EV protein markers. Figure 3. Extracellular vehicle (EV) quality assessment: (a) representative transmission electron microscopy images of enriched sEVs; (b) Venn diagram indicating overall protein identification overlap with ExoCarta database; (c) MS-based quantitative comparison between enriched BAL sEVs Figure 3. Extracellular vehicle (EV) quality assessment: ( a ) representative transmission electron microscopy images of enriched sEVs; ( b ) Venn diagram indicating overall protein identification overlap with ExoCarta database; ( c ) MS-based quantitative comparison between enriched BAL sEVs and cell line isolated EVs of frequently reported exosome protein markers and 10 most abundant exosome marker from ExoCarta. Red labeled proteins indicate non EV proteins. 2.6. Dysregulated Proteins in BAL Exosomes Principal component analysis of all the quantitative values separated NO versus YES cancer samples based on three different principal components (Figure S5). The statistical analysis of iBAQ expression values by the R package limma applying a correction for gender and smoking resulted in 166 genes significantly regulated (p-value <0.05) and one protein after correction of multiple testing (adjusted p-value <0.05) between cancer positive and negative status (Table S1 and Figure 4a,b). Cancers 2020,12, 3450 8 of 18 Previously, we reported 133 significant regulated proteins after correction of multiple testing [ 13 ], in which 91 acellular BAL samples were analyzed. We speculate that this difference primarily results from the difference in the number of cases studied (24 versus 91) rather than reflecting BAL sEVs potential as a biomarker source. In other words, due to cancer heterogeneity, a higher number of samples are required to define proteins significantly dysregulated after correction of multiple testing. To further explore protein heterogeneity of sEVs samples, uniquely identified proteins in each sample were extracted. Figure 4c depicts the number of unique proteins, finding 607 unique proteins (Table S2) in total for sEVs from lung cancer versus 176 for control. Counting the number of uniquely identified proteins per sample revealed that stage IV cancer and mortality displayed a significant tendency to contain more unique proteins identified compared to other cases (p-value =0.006, Figure 4c). The p-value was calculated based on Wilcoxon rank-sum test without removing any outliers. This trend of increased protein complexity in cancer was confirmed in lung cancer tissue (Figure S6b) but not for acellular BAL proteome (Figure S6a). Unique cancer sEVs proteins displayed KEGG functional regulation in pathways associated with cancer (Figure 4d). 1 Figure 4. Dysregulated lung cancer proteins in BAL EVs: ( a ) Volcano plot summarizing the main dysregulated proteins where –log10 pvalues were corrected for gender and smoking; ( b ) top regulated protein after correction for multiple testing; ( c ) number of unique proteins per sample (red bars indicate lung cancer and black bars represent controls); ( d ) KEGG functional enrichment of cancer unique proteins. The numbers in the bars represent the number of cancer unique proteins in the functional category. The dashed line indicates the p-value threshold (<0.05). Cancers 2020,12, 3450 9 of 18 2.7. Complete Functional Regulation Complete functional regulation analysis tests for significant regulation of entities (genes, proteins) within a functional group together with significant enrichment in terms of the identification of a given functional annotated group [ 21 ]. The methodology combines functional regulation and enrichment analysis into a single visual summary. Figure 5compares the results from the complete functional regulation analysis of acellular BAL, sEVs, lung tumor tissue, and mouse xenotransplant lung tumor tissue. For each functional group, the number of proteins identified in each sample type is displayed as well as the total number of proteins in the respective category. A statistical test is performed by assessing the significance of the enrichment and the significance of the overall regulation of the category. Complete functional regulation analysis revealed higher similarity between sEVs and tumor tissue when compared to acellular BAL (Figure 5). The large functional groups up-regulated in both tissue and sEVs include NADP binding and ERK pathway. ECM receptor interaction constitutes the largest down-regulated functional category common to tissue and sEVs (Figure 5). Several functional categories revealed reverse regulation when comparing sEVs with tumor tissue, for example, triglyceride metabolic processes. DNMT3B complex showed up-regulation in lung cancer for sEVs and tumor tissue. Aberrant DNA methylation caused by the DNMT3B complex in lung cancer is well established and proposed as a possible therapeutic target. Cancers 2020, 12, x FOR PEER REVIEW 10 of 20 tumor tissue, for example, triglyceride metabolic processes. DNMT3B complex showed up-regulation in lung cancer for sEVs and tumor tissue. Aberrant DNA methylation caused by the DNMT3B complex in lung cancer is well established and proposed as a possible therapeutic target. Figure 5. Complete functional regulation and enrichment analysis of proteins identified in acellular BAL (BAL), small EVs, lung tumor tissue (TissueTN), and xenotransplant lung tumor tissue (TissueXN). The color code in the heatmap reflects the regulation level (black unregulated, green significant up-regulated in tumor, and red significantly down-regulated in tumor). The text in the cells indicates the number of proteins identified in each functional category, the number of proteins annotated in the given functional category and the significance level of regulation, and significance level of enrichment (* p <0.05, ** p < 0.01, and *** p < 0.001). 2.8. Quantitative Evaluation of Immune Cell Lineage Markers BAL contains cells from various sources like tumor cells, epithelial cells, immune cells, and fibroblasts [22,23]. Based on the R package MCPcounter [24], we estimated the average abundance scores of markers for eight different immune and two stromal cell populations (Figure 6). MCPcounter identified six protein markers in BAL, 15 in sEVs, and 17 in tissue samples. In BAL sEVs, B lineage markers resulted in the highest average abundance scores. BAL sEVs from lung cancer tend to display higher abundance scores for monocytic lineage and fibroblasts than controls (Figure 6). This trend was not confirmed for acellular BAL proteome (Figure S7a). Markers of eight immune and two stromal cell populations could separate samples into normal, tumor, and xenotransplant tumor tissues (Figure S7b). Figure 5. Complete functional regulation and enrichment analysis of proteins identified in acellular BAL (BAL), small EVs, lung tumor tissue (TissueTN), and xenotransplant lung tumor tissue (TissueXN). The color code in the heatmap reflects the regulation level (black unregulated, green significant up-regulated in tumor, and red significantly down-regulated in tumor). The text in the cells indicates the number of proteins identified in each functional category, the number of proteins annotated in the given functional category and the significance level of regulation, and significance level of enrichment (* p<0.05, ** p<0.01, and *** p<0.001). Cancers 2020,12, 3450 16 of 18 References 1. Fidler, M.M.; Bray, F.; Soerjomataram, I. The global cancer burden and human development: A review. Scand. J. Public Health 2018,46, 27–36. [CrossRef] 2. Barta, J.A.; Powell, C.A.; Wisnivesky, J.P. Global Epidemiology of Lung Cancer. Ann. Glob. Health 2019 ,85, 8. [CrossRef] [PubMed] 3. Ko, E.C.; Raben, D.; Formenti, S.C. The Integration of Radiotherapy with Immunotherapy for the Treatment of Non-Small Cell Lung Cancer. Clin. Cancer Res. Off. J. Am. Assoc. Cancer Res. 2018 ,24, 5792–5806. [CrossRef] [PubMed] 4. Matthiesen, R. MS-Based Biomarker Discovery in Bronchoalveolar Lavage Fluid for Lung Cancer. Proteom. Clin. Appl. 2020,14, e1900077. [CrossRef] 5. Lee, H.; Groot, M.; Pinilla-Vera, M.; Fredenburgh, L.E.; Jin, Y. Identification of miRNA-rich vesicles in bronchoalveolar lavage fluid: Insights into the function and heterogeneity of extracellular vesicles. J. Control. Release Off. J. Control. Release Soc. 2019,294, 43–52. [CrossRef] 6. Venosa, A.; Katzen, J.; Tomer, Y.; Kopp, M.; Jamil, S.; Russo, S.J.; Mulugeta, S.; Beers, M.F. Epithelial Expression of an Interstitial Lung Disease-Associated Mutation in Surfactant Protein-C Modulates Recruitment and Activation of Key Myeloid Cell Populations in Mice. J. Immunol. 2019 ,202, 2760–2771. [CrossRef] [PubMed] 7. Lu, R.; Zhao, X.; Li, J.; Niu, P.; Yang, B.; Wu, H.; Wang, W.; Song, H.; Huang, B.; Zhu, N.; et al. Genomic characterisation and epidemiology of 2019 novel coronavirus: Implications for virus origins and receptor binding. Lancet 2020,395, 565–574. [CrossRef] 8. Zhang, C.; Yu, W.; Wang, L.; Zhao, M.; Guo, Q.; Lv, S.; Hu, X.; Lou, J. DNA Methylation Analysis of the SHOX2 and RASSF1A Panel in Bronchoalveolar Lavage Fluid for Lung Cancer Diagnosis. J. Cancer 2017 ,8, 3585–3591. [CrossRef] 9. Walmsley, S.; Cruickshank-Quinn, C.; Quinn, K.; Zhang, X.; Petrache, I.; Bowler, R.P.; Reisdorph, R.; Reisdorph, N. A prototypic small molecule database for bronchoalveolar lavage-based metabolomics. Sci. Data 2018,5, 180060. [CrossRef] 10. Gomes, S.; Cavadas, B.; Ferreira, J.C.; Marques, P.I.; Monteiro, C.; Sucena, M.; Sousa, C.; Vaz Rodrigues, L.; Teixeira, G.; Pinto, P.; et al. Profiling of lung microbiota discloses differences in adenocarcinoma and squamous cell carcinoma. Sci. Rep. 2019,9, 12838. [CrossRef] 11. Oumeraci, T.; Schmidt, B.; Wolf, T.; Zapatka, M.; Pich, A.; Brors, B.; Eils, R.; Fleischhacker, M.; Schlegelberger, B.; von Neuhoff, N. Bronchoalveolar lavage fluid of lung cancer patients: Mapping the uncharted waters using proteomics technology. Lung Cancer 2011,72, 136–138. [CrossRef] [PubMed] 12. Pastor, M.D.; Nogal, A.; Molina-Pinelo, S.; Melendez, R.; Salinas, A.; Gonzalez De la Pena, M.; Martin-Juan, J.; Corral, J.; Garcia-Carbonero, R.; Carnero, A.; et al. Identification of proteomic signatures associated with lung cancer and COPD. J. Proteom. 2013,89, 227–237. [CrossRef] [PubMed] 13. Carvalho, A.S.; Cuco, C.M.; Lavareda, C.; Miguel, F.; Ventura, M.; Almeida, S.; Pinto, P.; de Abreu, T.T.; Rodrigues, L.V.; Seixas, S.; et al. Bronchoalveolar Lavage Proteomics in Patients with Suspected Lung Cancer. Sci. Rep. 2017,7, 42190. [CrossRef] [PubMed] 14. Almatroodi, S.A.; McDonald, C.F.; Collins, A.L.; Darby, I.A.; Pouniotis, D.S. Quantitative proteomics of bronchoalveolar lavage fluid in lung adenocarcinoma. Cancer Genom. Proteom. 2015,12, 39–48. 15. Ortea, I.; Rodriguez-Ariza, A.; Chicano-Galvez, E.; Arenas Vacas, M.S.; Jurado Gamez, B. Discovery of potential protein biomarkers of lung adenocarcinoma in bronchoalveolar lavage fluid by SWATH MS data-independent acquisition and targeted data extraction. J. Proteom. 2016,138, 106–114. [CrossRef] 16. Sim, S.Y.; Choi, Y.R.; Lee, J.H.; Lim, J.M.; Lee, S.E.; Kim, K.P.; Kim, J.Y.; Lee, S.H.; Kim, M.S. In-Depth Proteomic Analysis of Human Bronchoalveolar Lavage Fluid toward the Biomarker Discovery for Lung Cancers. Proteom. Clin. Appl. 2019,13, e1900028. [CrossRef] 17. Carvalho, A.S.; Baeta, H.; Silva, B.C.; Moraes, M.C.S.; Bodo, C.; Beck, H.C.; Rodriguez, M.S.; Saraswat, M.; Pandey, A.; Matthiesen, R. Extra-cellular vesicles carry proteome of cancer hallmarks. Front. Biosci. 2020 ,25, 398–436. 18. Li, L.; Wei, Y.; To, C.; Zhu, C.Q.; Tong, J.; Pham, N.A.; Taylor, P.; Ignatchenko, V.; Ignatchenko, A.; Zhang, W.; et al. Integrated omic analysis of lung cancer reveals metabolism proteome signatures with prognostic impact. Nat. Commun. 2014,5, 5469. [CrossRef] Cancers 2020,12, 3450 17 of 18 19. Hurwitz, S.N.; Rider, M.A.; Bundy, J.L.; Liu, X.; Singh, R.K.; Meckes, D.G., Jr. Proteomic profiling of NCI-60 extracellular vesicles uncovers common protein cargo and cancer type-specific biomarkers. Oncotarget 2016 , 7, 86999–87015. [CrossRef] 20. Hoshino, A.; Kim, H.S.; Bojmar, L.; Gyan, K.E.; Cioffi, M.; Hernandez, J.; Zambirinis, C.P.; Rodrigues, G.; Molina, H.; Heissel, S.; et al. Extracellular Vesicle and Particle Biomarkers Define Multiple Human Cancers. Cell 2020,182, 1044–1061. [CrossRef] 21. Carvalho, A.S.; Molina, H.; Matthiesen, R. New insights into functional regulation in MS-based drug profiling. Sci. Rep. 2016,6, 18826. [CrossRef] [PubMed] 22. Pollock, K.; Albares, L.; Wendt, C.; Hubel, A. Isolation of fibroblasts and epithelial cells in bronchoalveolar lavage (BAL). Exp. Lung Res. 2013,39, 146–154. [CrossRef] 23. Liao, M.; Liu, Y.; Yuan, J.; Wen, Y.; Xu, G.; Zhao, J.; Cheng, L.; Li, J.; Wang, X.; Wang, F.; et al. Single-cell landscape of bronchoalveolar immune cells in patients with COVID-19. Nat. Med. 2020 ,26, 842–844. [CrossRef] [PubMed] 24. Becht, E.; Giraldo, N.A.; Lacroix, L.; Buttard, B.; Elarouci, N.; Petitprez, F.; Selves, J.; Laurent-Puig, P.; Sautes-Fridman, C.; Fridman, W.H.; et al. Estimating the population abundance of tissue-infiltrating immune and stromal cell populations using gene expression. Genome Biol. 2016,17, 218. [CrossRef] [PubMed] 25. Webber, J.; Clayton, A. How pure are your vesicles? J. Extracell. Vesicles 2013,2. [CrossRef] [PubMed] 26. Purcell, A.W.; Ramarathinam, S.H.; Ternette, N. Mass spectrometry-based identification of MHC-bound peptides for immunopeptidomics. Nat. Protoc. 2019,14, 1687–1707. [CrossRef] 27. Bassani-Sternberg, M.; Barnea, E.; Beer, I.; Avivi, I.; Katz, T.; Admon, A. Soluble plasma HLA peptidome as a potential source for cancer biomarkers. Proc. Natl. Acad. Sci. USA 2010,107, 18769–18776. [CrossRef] 28. Cvjetkovic, A.; Jang, S.C.; Konecna, B.; Hoog, J.L.; Sihlbom, C.; Lasser, C.; Lotvall, J. Detailed Analysis of Protein Topology of Extracellular Vesicles-Evidence of Unconventional Membrane Protein Orientation. Sci. Rep. 2016,6, 36338. [CrossRef] 29. Roberts, P.J.; Der, C.J. Targeting the Raf-MEK-ERK mitogen-activated protein kinase cascade for the treatment of cancer. Oncogene 2007,26, 3291–3310. [CrossRef] 30. Chiu, L.Y.; Hsin, I.L.; Yang, T.Y.; Sung, W.W.; Chi, J.Y.; Chang, J.T.; Ko, J.L.; Sheu, G.T. The ERK-ZEB1 pathway mediates epithelial-mesenchymal transition in pemetrexed resistant lung cancer cells with suppression by vinca alkaloids. Oncogene 2017,36, 242–253. [CrossRef] 31. Yang, Y.F.; Chang, Y.C.; Jan, Y.H.; Yang, C.J.; Huang, M.S.; Hsiao, M. Squalene synthase promotes the invasion of lung cancer cells via the osteopontin/ERK pathway. Oncogenesis 2020,9, 78. [CrossRef] [PubMed] 32. Jones, G.G.; Del Rio, I.B.; Sari, S.; Sekerim, A.; Young, L.C.; Hartig, N.; Areso Zubiaur, I.; El-Bahrawy, M.A.; Hynds, R.E.; Lei, W.; et al. SHOC2 phosphatase-dependent RAF dimerization mediates resistance to MEK inhibition in RAS-mutant cancers. Nat. Commun. 2019,10, 2532. [CrossRef] 33. Maacha, S.; Bhat, A.A.; Jimenez, L.; Raza, A.; Haris, M.; Uddin, S.; Grivel, J.C. Extracellular vesicles-mediated intercellular communication: Roles in the tumor microenvironment and anti-cancer drug resistance. Mol. Cancer 2019,18, 55. [CrossRef] [PubMed] 34. Aung, T.; Chapuy, B.; Vogel, D.; Wenzel, D.; Oppermann, M.; Lahmann, M.; Weinhage, T.; Menck, K.; Hupfeld, T.; Koch, R.; et al. Exosomal evasion of humoral immunotherapy in aggressive B-cell lymphoma modulated by ATP-binding cassette transporter A3. Proc. Natl. Acad. Sci. USA 2011,108, 15336–15341. [CrossRef] [PubMed] 35. Ostler, K.R.; Davis, E.M.; Payne, S.L.; Gosalia, B.B.; Exposito-Cespedes, J.; Le Beau, M.M.; Godley, L.A. Cancer cells express aberrant DNMT3B transcripts encoding truncated proteins. Oncogene 2007 ,26, 5553–5563. [CrossRef] [PubMed] 36. Cheng, Y.; He, C.; Wang, M.; Ma, X.; Mo, F.; Yang, S.; Han, J.; Wei, X. Targeting epigenetic regulators for cancer therapy: Mechanisms and advances in clinical trials. Signal Transduct. Target. Ther. 2019,4, 62. [CrossRef] 37. Sandhu, R.; Rivenbark, A.G.; Coleman, W.B. Enhancement of chemotherapeutic efficacy in hypermethylator breast cancer cells through targeted and pharmacologic inhibition of DNMT3b. Breast Cancer Res. Treat. 2012 , 131, 385–399. [CrossRef] 38. Wu, C.; Guo, E.; Ming, J.; Sun, W.; Nie, X.; Sun, L.; Peng, S.; Luo, M.; Liu, D.; Zhang, L.; et al. Radiation-Induced DNMT3B Promotes Radioresistance in Nasopharyngeal Carcinoma through Methylation of p53 and p21. Mol. Ther. Oncolytics 2020,17, 306–319. [CrossRef] 39. Yao, P.; Sun, H.; Xu, C.; Chen, T.; Zou, B.; Jiang, P.; Du, W. Evidence for a direct cross-talk between malic enzyme and the pentose phosphate pathway via structural interactions. J. Biol. Chem. 2017 ,292, 17113–17120. [CrossRef] Cancers 2020,12, 3450 18 of 18 40. Desgrosellier, J.S.; Cheresh, D.A. Integrins in cancer: Biological implications and therapeutic opportunities. Nat. Rev. Cancer 2010,10, 9–22. [CrossRef] 41. Hamidi, H.; Pietila, M.; Ivaska, J. The complexity of integrins in cancer and new scopes for therapeutic targeting. Br. J. Cancer 2016,115, 1017–1023. [CrossRef] [PubMed] 42. Domagala-Kulawik, J.; Guzman, J.; Costabel, U. Immune cells in bronchoalveolar lavage in peripheral lung cancer—Analysis of 140 cases. Respir. Int. Rev. Thorac. Dis. 2003,70, 43–48. [CrossRef] [PubMed] 43. Bruno, T.C.; Ebner, P.J.; Moore, B.L.; Squalls, O.G.; Waugh, K.A.; Eruslanov, E.B.; Singhal, S.; Mitchell, J.D.; Franklin, W.A.; Merrick, D.T.; et al. Antigen-Presenting Intratumoral B Cells Affect CD4(+) TIL Phenotypes in Non-Small Cell Lung Cancer Patients. Cancer Immunol. Res. 2017,5, 898–907. [CrossRef] [PubMed] 44. Carmi, Y.; Spitzer, M.H.; Linde, I.L.; Burt, B.M.; Prestwood, T.R.; Perlman, N.; Davidson, M.G.; Kenkel, J.A.; Segal, E.; Pusapati, G.V.; et al. Allogeneic IgG combined with dendritic cell stimuli induce antitumour T-cell immunity. Nature 2015,521, 99–104. [CrossRef] [PubMed] 45. MacMahon, H.; Austin, J.H.; Gamsu, G.; Herold, C.J.; Jett, J.R.; Naidich, D.P.; Patz, E.F., Jr.; Swensen, S.J. Guidelines for management of small pulmonary nodules detected on CT scans: A statement from the Fleischner Society. Radiology 2005,237, 395–400. [CrossRef] [PubMed] 46. Wisniewski, J.R.; Zougman, A.; Nagaraj, N.; Mann, M. Universal sample preparation method for proteome analysis. Nat. Methods 2009,6, 359–362. [CrossRef] 47. Carvalho, A.S.; Ribeiro, H.; Voabil, P.; Penque, D.; Jensen, O.N.; Molina, H.; Matthiesen, R. Global mass spectrometry and transcriptomics array based drug profiling provides novel insight into glucosamine induced endoplasmic reticulum stress. Mol. Cell. Proteom. MCP 2014,13, 3294–3307. [CrossRef] 48. Cox, J.; Mann, M. MaxQuant enables high peptide identification rates, individualized p.p.b.-range mass accuracies and proteome-wide protein quantification. Nat. Biotechnol. 2008,26, 1367–1372. [CrossRef] 49. Matthiesen, R.; Prieto, G.; Amorim, A.; Aloria, K.; Fullaondo, A.; Carvalho, A.S.; Arizmendi, J.M. SIR: Deterministic protein inference from peptides assigned to MS data. J. Proteom. 2012 ,75, 4176–4183. [CrossRef] 50. Hackenberg, M.; Matthiesen, R. Annotation-Modules: A tool for finding significant combinations of multisource annotations for gene lists. Bioinformatics 2008,24, 1386–1393. [CrossRef] 51. Gu, Z.; Eils, R.; Schlesner, M. Complex heatmaps reveal patterns and correlations in multidimensional genomic data. Bioinformatics 2016,32, 2847–2849. [CrossRef] [PubMed] 52. Kanehisa, M.; Furumichi, M.; Tanabe, M.; Sato, Y.; Morishima, K. KEGG: New perspectives on genomes, pathways, diseases and drugs. Nucleic Acids Res. 2017,45, D353–D361. [CrossRef] [PubMed] 53. Gene Ontology Consortium. The Gene Ontology Resource: 20 years and still GOing strong. Nucleic Acids Res . 2019,47, D330–D338. [CrossRef] [PubMed] 54. Goel, R.; Harsha, H.C.; Pandey, A.; Prasad, T.S. Human Protein Reference Database and Human Proteinpedia as resources for phosphoproteome analysis. Mol. Biosyst. 2012,8, 453–463. [CrossRef] [PubMed] 55. Ruepp, A.; Brauner, B.; Dunger-Kaltenbach, I.; Frishman, G.; Montrone, C.; Stransky, M.; Waegele, B.; Schmidt, T.; Doudieu, O.N.; Stumpflen, V.; et al. CORUM: The comprehensive resource of mammalian protein complexes. Nucleic Acids Res. 2008,36, D646–D650. [CrossRef] 56. Liberzon, A.; Birger, C.; Thorvaldsdottir, H.; Ghandi, M.; Mesirov, J.P.; Tamayo, P. The Molecular Signatures Database (MSigDB) hallmark gene set collection. Cell Syst. 2015,1, 417–425. [CrossRef] 57. Perez-Riverol, Y.; Csordas, A.; Bai, J.; Bernal-Llinares, M.; Hewapathirana, S.; Kundu, D.J.; Inuganti, A.; Griss, J.; Mayer, G.; Eisenacher, M.; et al. The PRIDE database and related tools and resources in 2019: Improving support for quantification data. Nucleic Acids Res. 2018,47, D442–D450. [CrossRef] Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).