From bioinformatic screening to low-invasive molecular diagnostics of the cervical tumor sensitivity to radiation therapy.

Abstract

e17512 Background: Cervical cancer (CC) is one of the most common cancers in women worldwide (4th in frequency). Traditionally, radical hysterectomy or radiation therapy (RT) is used as the standard treatment for early-stage CC, and locally advanced cancers are treated with RT alone. RT results in complete clinical response only in a part of patients due to formation of malignant cell radioresistance. To date, a significant marker list has been identified to predict the response to RT, but none of them has yet entered clinical practice. Therefore, the aim of this study was bioinformation and laboratory screening of molecular markers for low invasive determination of the CC sensitivity to RT. Methods: The study was performed on 300 CC patients (IB1, IB2, IIA1, tumor < 4 cm) and 30 donors without cancer. The Cancer Genome Atlas database was analyzed to identify potential markers. To obtain data from the Genomic Data Commons Data Portal, we used the TCGABiolinks package of the R language in the Rstudio shell. The GISTIC, MutSig and RAE algorithms were used to identify genome regions whose sizes varied (copy number variation, CNV) significantly in a number of tumor samples. The identified markers (CNV) were validated by RT-PCR in extracellular DNA (cfDNA). Blood was collected before the RT. The remote RT was performed on the Varian TrueBeam linear accelerator in VMAT/IMRT mode (TFD 50Gr). Differences were assessed using the Mann-Whitney test (Bonferroni correction was used). Results: RT result analysis allowed to divide patients into 2 groups - sensitive to RT (n = 170, group 1) and resistant (n = 130, group 2). Bioinformation analysis allowed to isolate a number of genes that altered copying and were associated with sensitivity to RT - ERBB2, BIRC2, TRPC6, YAP1, MIR569, LRRC31, SPRED3, MIR4456, CYP1A, CYP1A, FOXO1, ENOX1, EPSTI1, NEK5, KCTD4, SERP2, MIR621, PTEN, SOD2, MIR3939, ATM, CASP1, CASP4, CASP5, CHEK1 and H2AFX. The CNV of these genes was analyzed in cfDNA. In group 1, a decrease (p < 0.05) in the CNV of H2AFX, ATM, CHEK1, LINC00558, LINC00400 and an increase (p < 0.05) in the CNV of CASP1, CASP4, CASP5, CYP1A1, CYP1A2 and GPX4 were found. In group 2, a decrease (p < 0.05) in the CNV of CASP4, CASP5, CYP1A1, YAP1 and MIR569 and an increase in the CNV (p < 0.05) of H2AFX, ATM, CHEK1, ERBB2 and BIRC2 were found in relation to these indicators in donors without cancer. CNV of H2AFX, ATM, CHEK1, ERBB2, BIRC2, MIR569, LINC00400, CASP4, CASP5 and CYP1A1 genes differed statistically significantly (p < 0.005) in 2 groups of CC patients. Conclusions: The markers identified with a combination of bioinformatics and molecular genetic approaches - the CNV of H2AFX, ATM, CHEK1, ERBB2, BIRC2, MIR569, LINC00400, CASP4, CASP5 and CYP1A1 genes in cfDNA - can become the basis for a low invasive determination of the CC sensitivity to RT.