Identification of Potentially Novel Molecular Targets of Endometrial Cancer Using a Non-Biased Proteomic Approach

Abstract

The present study was aimed at identifying novel proteins in endometrial cancer (EC), employing proteomic analysis of tissues obtained after surgery. A differential MS-based proteomic analysis was conducted from whole tissues dissected from biopsies from post-menopausal women, histologically confirmed as endometrial cancer (two endometrioid and two serous; n = 4) or normal atrophic endometrium (n = 4), providing 888 differentially expressed proteins with 246 of these previously documented elsewhere as expressed in EC and 372 proteins not previously demonstrated to be expressed in EC but associated with other types of cancer. Additionally, 33 proteins not recorded previously in PubMed as being expressed in any forms of cancer were also identified, with only 26 of these proteins having a publication associated with their expression patterns or putative functions. The putative functions of the 26 proteins (GRN, APP, HEXA, CST3, CAD, QARS, SIAE, WARS, MYH8, CLTB, GOLIM4, SCARB2, BOD1L1, C14orf142, C9orf142, CCDC13, CNPY4, FAM169A, HN1L, PIGT, PLCL1, PMFBP1, SARS2, SCPEP1, SLC25A24 and ZC3H4) in other tissues point towards and provide a basis for further investigation of these previously unrecognised novel EC proteins. The developmental biology, disease, extracellular matrix, homeostatic, immune, metabolic (both RNA and protein), programmed cell death, signal transduction, molecular transport, transcriptional networks and as yet uncharacterised pathways indicate that these proteins are potentially involved in endometrial carcinogenesis and thus may be important in EC diagnosis, prognostication and treatment and thus are worthy of further investigation.

Keywords: biomarker; endometrial cancer; prognosis; protein; proteomics; therapy.

Figure 1

Scaffold Proteome analyses for correct spectral identification. The upper panel shows the number of peptide spectra assigned correctly (pink bars) compared to number of peptide spectra assigned incorrectly (green bars) together with their relative distributions (red and blue lines). The lower panel shows the log-transformed data for the number of spectra plotted against the negative log of the expected value of spectra based on the discriminant score shown in Panel A. Only 1 spectral line is a higher level than expected and that associates with serum albumin in both the control and cancer samples. The software can be accessed from www.proteomesoftware.com, last accessed 6 September 2023.

Figure 2

Scatterplot for the relative quantities of proteins identified in 4 control (atrophic) and 4 endometrial cancer (cancer) samples. The upper panel (A) shows the raw data, whilst the middle panel (B) shows the log-transformed data once the proteins outside the 3 standard deviations had been omitted. The protein highlighted with a yellow circle is serum albumin and was the most abundant protein in the atrophic samples (the data point for the cancer sample is outside the 3 standard deviations range). The spread of data is consistent with a high probability of protein identification. The lower panel (C) shows a histogram of positive peptide sequences identified in the biological samples. The data indicate the number of positive peptide sequences identified in the individual biological samples (control/atrophic: pink bars) and (endometrial cancer/cancer: blue bars) for the protein with the most abundant peptide spectra—serum albumin (highlighted in panel A).

Figure 3

Venn diagram and bar charts for proteins identified using X! Tandem and Mascot. Venn diagrams (A) are used to show that a total of 1851 proteins were originally identified, of which 888 were common to both control (atrophic) and endometrial cancer (cancer) samples. Only 300 proteins were found solely in the control tissues, and 663 were only found in the cancer tissues. After close examination of the peptide identities, 4 proteins were duplicated in the cancer tissues and so were eliminated from further analyses (second Venn diagram). Gene ontology (GO) pathway analysis of the 663 proteins found only in endometrial cancer matched the genes identified to 21 biological processes (B) and to 8 molecular functions (C). The % of genes in each category is indicated.

Figure 4

Functional analysis of cancer-specific protein pathways and interactions. The diagram depicts functional densities for the proteins identified as being exclusive to endometrial cancer. The data output is direct from the Reactome application. Note the enrichment of proteins in signal transduction, cell cycle, programmed cell death, metabolism, disease processes and reproduction. Also note that two distinct networks are prevalent: one on the left-hand side of the figure and focused on signal transduction and another on the right-hand side focused on the cell cycle.

Figure 5

Expression of proteins unique to endometrial cancer. Panel (A) shows an expression heatmap for the 26 proteins identified to be uniquely expressed in EC. The fold-expression is related to the mean number of peptide spectra of the atrophic samples. When no peptide signal was identified in the atrophic samples, the expression in the serous samples was relative to the mean of all peptide spectra in the 4 atrophic controls. Panel (B) shows a list of proteins found only in the endometrioid (Type 1 EC) samples, serous (Type 2 EC) samples or both types of EC.