From Proteomic Analysis to Potential Therapeutic Targets: Functional Profile of Two Lung Cancer Cell Lines, A549 and SW900, Widely Studied in Pre-Clinical Research

Abstract

Lung cancer is a serious health problem and the leading cause of cancer death worldwide. The standard use of cell lines as in vitro pre-clinical models to study the molecular mechanisms that drive tumorigenesis and access drug sensitivity/effectiveness is of undisputable importance. Label-free mass spectrometry and bioinformatics were employed to study the proteomic profiles of two representative lung cancer cell lines and to unravel the specific biological processes. Adenocarcinoma A549 cells were enriched in proteins related to cellular respiration, ubiquitination, apoptosis and response to drug/hypoxia/oxidative stress. In turn, squamous carcinoma SW900 cells were enriched in proteins related to translation, apoptosis, response to inorganic/organic substances and cytoskeleton organization. Several proteins with differential expression were related to cancer transformation, tumor resistance, proliferation, migration, invasion and metastasis. Combined analysis of proteome and interactome data highlighted key proteins and suggested that adenocarcinoma might be more prone to PI3K/Akt/mTOR and topoisomerase IIα inhibitors, and squamous carcinoma to Ck2 inhibitors. Moreover, ILF3 overexpression in adenocarcinoma, and PCNA and NEDD8 in squamous carcinoma shows them as promising candidates for therapeutic purposes. This study highlights the functional proteomic differences of two main subtypes of lung cancer models and hints several targeted therapies that might assist in this type of cancer.

Fig 1. Global proteome analysis of the lung cancer cell lines.

Venn diagram highlighting the distribution of the identified proteins per cell line in numbers and in percentage, evidencing the overlapped and unique proteins (Venny 2.0.2).

Fig 2. GO analysis of the specific proteins of adenocarcinoma and squamous carcinoma cell lines: cell components.

Enriched GO terms were retrieved using DAVID database. For the A549 cell line 127 proteins out of 239 (53%) and for the SW900 cell line 174 proteins out of 293 (59%) were classified in the GO terms. The enrichment was performed considering a p-value of 0.05 and a minimum number of 3 genes per term. Parts of the figure were adapted from Servier Medical Art templates available at /www.servier.co.uk/content/servier-medical-art. Servier Medical Art is licensed under a Creative Commons Attribution 3.0 Unported License (http://creative-commons.org/licenses/by/3.0/).

Fig 3. GO analysis of the specific proteins of adenocarcinoma and squamous carcinoma cell lines: biological process and molecular function.

Enriched GO terms were retrieved using DAVID database. Biological process: for the A549 cell line 141 proteins out of 239 (59%) and for the SW900 cell line 152 proteins out of 293 (52%) were classified in the GO terms and clustered. Molecular function: for the A549 cell line 114 proteins out of 239 (48%) and for the SW900 cell line 136 proteins out of 293 (46%) were classified in the GO terms clustered. The clustering was performed considering a p-value of 0.05 and a minimum number of 2 terms per cluster.

Fig 4. Expression levels of several altered proteins in adenocarcinoma and squamous carcinoma cell lines.

Four independent extracts of A549 and SW900 cell lines were prepared and analyzed using the Western blot technique for expression comparison purposes. (A) Western blot images of the four replicates in both cell lines. Actin was used as loading control. (B) Protein band densitometries were obtained, values were normalized using the internal actin control and finally averaged. In the graph (Mean ± SE, n = 4), * p < 0.05, ** p < 0.01 and *** p < 0.001, indicate significant changes between the analyzed cell lines following one-way ANOVA.

Fig 5. Integration of proteome and interactome data of the adenocarcinoma cell line.

SteinerNet webserver was used to reveal hidden components in A549 network by integrating the proteome (MS, fold-regulation) and the interactome data (HIPPIE, interaction scores). From the original network, 175 terminal nodes were excluded (23.7%) and 563 terminal nodes included (76.3%). Circular nodes denotes proteins obtained from MS, whereas diamond nodes are proteins obtained from HIPPIE database. Node and letter size are related to the betweeness centrality (high betweeness centrality represent important nodes in the network, also called bottlenecks) of the proteins and was calculated using the Cytoscape NetworkAnalyzer tool. Edge width shows the interaction score confidence. Node color is depicted as following (A549 vs SW900): green, proteins upregulated (fold-regulation > 2); red, proteins downregulated (fold-regulation > -2); yellow, unaltered proteins; violet, A549-specific proteins.

Fig 6. Integration of proteome and interactome data of the squamous carcinoma cell line.

SteinerNet webserver was used to reveal hidden components in SW900 network by integrating the proteome (MS, fold-regulation) and the interactome data (HIPPIE, interaction scores). From the original network, 172 terminal nodes were excluded (21.8%) and 618 terminal nodes included (78.2%). Circular nodes denotes proteins obtained from MS, whereas diamond nodes are proteins obtained from HIPPIE database. Node and letter size are related to the betweeness centrality (high betweeness centrality represent important nodes in the network, also called bottlenecks) of the proteins and was calculated using the Cytoscape NetworkAnalyzer tool. Edge width shows the interaction score confidence. Node color is depicted as following (SW900 vs A549): green, proteins upregulated (fold-regulation > 2); red, proteins downregulated (fold-regulation > -2); yellow, unaltered proteins; violet, SW900-specific proteins.

Fig 7. Functional network of altered proteins present in both cell lines.

ClueGO plugin of Cytoscape was used to generate a functional network (biological process). Node size is related to the degree (high degree represent important nodes in the network, also known as hubs) of the proteins and was calculated using the Cytoscape NetworkAnalyzer tool. Proteins node color is depicted as following (A549 vs SW900): green, proteins upregulated (fold-regulation > 2); red, proteins downregulated (fold-regulation > -2). Biological process node color is represented on the right side of the image. On the right bottom side of the image are shown the genes that does not fit these biological processes and that do not have any interaction with the proteins that are altered.