High expression of MCM10 is predictive of poor outcomes in lung adenocarcinoma

Accession of data

Three unique expression microarrays were used for the present analysis: GSE32863 (Selamat et al., 2012), GSE31210 (Okayama et al., 2012) and GSE75037 (Girard et al., 2016). All datasets were downloaded from the Gene Expression Omnibus (GEO). GSE32863 and GSE75037 datasets were produced using Illumina HumanWG-6/Ref-8 v3.0 expression beadchip platform while GSE31210 used Affymetrix Human Genome U133 Plus 2.0 Array. The pertinent details of each dataset are presented in Table 1.

Identifying DEG and HUB genes

Bioinformatic analysis was performed based on our previous protocols (Dong, Zhu & Zhang, 2020). Briefly, Differentially Expressed Genes (DEGs) were identified from the three accessed datasets using GEO2R with a cutoff criteria set to a P-value of <0.05 and log fold change >1. Values were averaged In the event probe sets did not contain exact gene symbols or two or more probe sets contained the same gene symbols. An interaction network between identified DEGS was constructed using The Search Tool for Retrieval of Interacting Genes (STRING) database (Szklarczyk et al., 2015). Cytoscape version 3.4.0 software was then used to visualize the output. Molecular Complex Detection (Bader & Hogue, 2003) (MCODE) was used to arrange the topology, cluster the connected genes, and search through the resulting network. Cutoffs for inclusion were MCODE scores >5, degree cutoff of 2, node score cutoff of 0.2, node density cutoff of 0.1, k-score of 2, and max depth of 100. HUB genes were initially selected if their degree was >10. Hub genes were initially selected if their degree of connectivity was >25. We then performed a functional analysis HUB genes using Cytoscape’s Biological Networks Gene Ontology (GO) (Maere, Heymans & Kuiper, 2005) (BiNGO) and validated these findings using the Database for Annotation, Integrated Discovery, and Visualization (Da Huang, Sherman & Lempicki, 2009) (DAVID).

Kyoto Encyclopedia of Genes and Genomes and GO enrichment analyses of DEGs

The functional annotation tools version 6.7 of the DAVID (http://david.ncifcrf.gov) were used to extract biological information about our DEGs. GO was also used to annotate genes and further analyze their biological functions. The DAVID online database was used to analyze the function and biological process of the screened DEGs. P < 0.05 was considered to indicate statistical significance.

Identification of MCM10 as a crucial HUB gene

Hub genes were further stratified by evaluating degree of mutation using cBioPortal’s Oncomine (Rhodes et al., 2004) plug in to search The Cancer Genome Atlas provisional sample set of 586 samples. Expression at various stages of cancer was evaluated using UCSC Xena cancer browser. MCM10’s relevance in cancer and in lung adenocarcinoma was further evaluated by again using Oncomine and UCSC XenaBrowser to identify transcript level mutations and expression levels. Finally, the correlation between MCM10 expression and overall survival was evaluated using the Kaplan–Meier plotter (Nagy et al., 2018).

In vitro analysis of MCM10

Bioinformatics results were confirmed using in vitro experiments on A549 lung adeno carcinoma cells (Shanghai Cell Bank, Shanghai, China). The expression level of A549 was confirmed on Human Protein Atlas (http://www.proteinatlas.org/). A549 cells were cultured in RPMI 1640 medium supplemented with 1× penicillin/streptomycin (Gibco, Waltham, MA, USA) and 10% Fetal Bovine Serum (Gibco). RNA was extracted using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) as per manufacturer’s recommendations. RNA was converted to cDNA using PrimeScript RT reagent kit (Takara Bio, Dalian, China) and reverse transcription was performed as per manufacturer’s recommendations. Transcription levels were evaluated using reverse transcription quantitative real time polymerase chain reaction (RT-qPCR) RR901A Premix Taq^™ (TaKaRa Taq^™ Version 2.0 plus dye). The MCM10 shRNA sequence was as follows: 3′-CCGGGACGGCGACGGTGAATCTTATCTCGAGATAAGATTCACCGTCGCCGTCTTTTT-5′ (Jikai, Shanghai, China). The PCR primers for MCM10 or GAPDH were as follows: MCM10 sense, 5′ GAAGAAGGTTACGCCACAGAG′ and reverse, 5′ TTTACAGGTTCCCAGGTCAAG3′; GAPDH sense, 5′-CAA TGACCCCTTCATTGACC-3′ and reverse, 5′- TGGAAGATGGTGATGGGATT-3′. PCR cycling conditions were 50 °C for 2 min, followed by 95 °C for 10 min and then 40 cycles for 95 °C for 15 s and 60 °C for 1 min. The 2^−ΔΔCt method was used to calculate the relative expression level by normalizing to GAPDH levels. Proliferation and colony formation assays were performed using the CCK8 cell counting kit (Dojindo Molecular Technologies, Rockville, MD, USA) on cultures initially seeded at 30,000 cells/mL. Results are displayed as mean (n = 3) ± SEM. Differences were considered statistically significant for P < 0.05.

Validation in clinical samples

A total of 72 paired lung adenocarcinoma tissues and adjacent normal tissues were collected from patients with a median age of 60 undergoing surgery at our hospital. Collections took place between June 2018 and June 2019. MCM10 expression was detected using qPCR. The research was approved by the Ethics Committee of The First hospital of China Medical University (2018-88). Written informed consent was obtained from all of the patients.

Identifying lung adenocarcinoma HUB genes

Genome arrays provide invaluable data for bioinformatics-based analysis of DEGs. To fully identify and explore DEGs in lung adenocarcinoma samples we evaluated three separate arrays: GSE32863, GSE31210 and GSE75037. Characteristics of each are listed in Table 1. We chose these datasets because they were composed of matching samples of tumorous and non-tumorous tissues, allowing us to control for patient to patient differences. In all, we found 41 unique DEGs correlated with late stage lung adenocarcinomas in all three microarray samples (Fig. 2A). These 41 co-DEGs made the basis of our further analysis.

Gene IDs were entered into STRING database to predict protein-protein interactions and the resulting interactions were mapped using Cytoscape. We found ten of the genes had no connections, while five of the associated genes, CD69, CD83, PSAT1, MTHFD2 and FOLR1 were not highly connected (Fig. 2B, Red). The 26 additional genes showed high interconnectivity (Fig. 2B, yellow). These HUB genes have been found to be essential in cancer pathogenesis and progression. Further analysis via cBioPortal’s network application demonstrates high level of pathway interactivity. Interestingly, while many of the genes are highly connected, four of the HUB genes upregulated during lung adenocarcinoma (MCM10, MCM4, GINS2 and CD45) are predicted to interact nearly independently through direct binding mechanisms. We utilized the BiNGO application within Cytoscape to explore the HUB GO (Fig. 2D). While the HUB genes were highly connected in ontology, the most over represented were S phase of mitotic cell cycle, DNA replication initiation, and double strand break repair, insinuating a disruption in the cell cycle mechanism.

MCM10 is a key HUB gene in lung adenocarcinoma

We chose to further evaluate the function of the HUB genes which are over expressed in Lung Adenocarcinoma. Using DAVID, we found that the genes that are mostly over expressed are involved in cell cycle regulation, progesterone-mediate oocyte maturation, and negative regulation of apoptotic processes (Fig. 3A). To narrow down the list of potential key genes, we chose the top 15 most connected HUB genes. Of these, all except for MAD2L1 have a precedence of mutation in the Lung Adenocarcinoma TCGA Pan Cancer atlas (Fig. 3B). Furthermore, RNAseq analysis shows that each of these 15 are more highly expressed in both primary and recurrent tumors than in solid tumors (Fig. 3C). Based on the Cytoscape BiNGO analysis, it was evident that pertinent HUB genes should be connected to S phase of the mitotic cell cycle and DNA replication mechanisms (Fig. 2D). Based on an analysis of PathwayCommons, it was determined only four of the most connected HUB genes were involved in both pathways. Perhaps not surprisingly, many of these markers have been reported and extensively studied for their role in lung adenocarcinoma. However, of these four markers, MCM10 was found to be mutated the most (1.3% of samples), and therefore it was chosen as the most prominent target for the rest of the analysis. Through this, we have demonstrated that MCM10 may be a viable marker for lung adenocarcinoma.

MCM10 expression is upregulated in lung adenocarcinoma

The results thus far have led to the hypothesis that MCM10 is a novel marker for lung adenocarcinoma. To further validate this theory, we further explored MCM10’s clinical value. This was accomplished by screening MCM10’s role in several organotypic cancers. We found that MCM10 is highly expressed in a variety of cancers in addition to lung cancer, including colorectal cancer and breast cancer (Fig. 4A). The incidences of cancer typically resulted in single copy alterations (i.e., gain). Interestingly, there were only a few mutations within the data set, suggesting that while MCM10 may be correlated with cancer pathogenesis, it may not be the leading cause (Fig. 4B). However, the mutations that do occur in MCM10 during lung adenocarcinoma arise from a missense mutation. We further evaluated our hypothesis by evaluating MCM10 transcript expression using the UCSC Xena Browser (Figs. 4C and 4D). Nearly all expressed transcripts are upregulated in lung adenocarcinoma samples, with the most prominently expressed transcript (MCM10-202) demonstrating a 2.5-fold increase in expression.

The results thus far support the theory that MCM10 expression is a novel marker for lung adenocarcinoma. We corroborate this using Oncomine, which shows that there is an average of 2.9-fold increase in MCM10 in cancer samples compared to normal tissues (Fig. 5A). The normalized level of transcript expression is relatively similar across stages 1–3 (~1.3–1.5-fold increase), however there is a higher level in later stage (overall 2-fold increase, Fig. 5B). Most convincingly, individuals who have low MCM10 expression have over a 2-fold increase in overall survival rate (Fig. 5C). Furthermore, individuals with higher levels of MCM10 expression have a 1.6-fold increase in mortality after successful treatment (Fig. 5D). Together, this suggests MCM10 is a clinically relevant prognostic biomarker, with increased expression resulting in poor outcomes.

MCM10 is crucial for proliferation in an in vitro lung adenocarcinoma model

While these results suggest that MCM10 expression is important in lung adenocarcinoma pathogenesis, they do not fully implicate MCM10 in rampant tumor growth. MCM10 RNA expression is high in a variety of cancer cell lines, with A549 cells demonstrating a much higher level of normalized expression (NX) than the pulmonary epithelial cell line HBEC3 (15.3 vs. 8.3, Fig. 6A). To regulate the MCM10 expression levels, we exposed A549 cell lines to corresponding siRNAs (Fig. 6B). Decreasing the expression resulted in a net loss of 50% of the proliferation capacity (Fig. 6C). Similarly, exposing the A549 cell lines to MCM10 siRNA results in a 50% decrease in colony formation (Fig. 6D). These results demonstrate that MCM10 regulation may be a relevant strategy for therapeutic intervention, and is worth further examination.

Validation in human samples

Thus far, we have determined that MCM10 is a potential diagnostic marker for lung adenocarcinoma and demonstrated upregulation is necessary for excessive tumor growth. We chose to further validate these results using clinical samples taken from patients with lung ACA as well as matching adjacent normal tissue. Results from qPCR align with our bioinformatic results, suggesting lung adenocarcinoma samples have a much higher level of MCM10 than did normal tissue (Fig. 7A). While there was no significant difference due to lymphatic metastasis and distant metastasis (Figs. 7C and 7D), there were significant differences in tumor size (Fig. 7B). These results suggest MCM10 is a valuable marker which can be used to assist in diagnosing lung adenocarcinoma.