Genome-wide association study of loss of heterozygosity and metastasis-free survival in breast cancer patients

Deryusheva I.V.1, Tsyganov M.1, Garbukov E.Y.1, Ibragimova M.K.1, Kzhyshkowska J.G.2, Slonimskaya E.1, Cherdyntseva N.V.3, Litviakov N.V.3

Summary. One of the factors providing the diversity and heterogeneity of malignant tumors, particularly breast cancer, are genetic variations, due to gene polymorphism, and, especially, the phenomenon of loss of heterozygosity (LOH). It has been shown that LOH in some genes could be a good prognostic marker. Aim: To perform genome-wide study on LOH in association with metastasis-free survival in breast cancer. Materials and Methods: The study involved 68 patients with breast cancer. LOH status was detected by microarray analysis, using a high density DNA-chip CytoScanTM HD Array (Affymetrix, USA). The Chromosome Analysis Suite 3.1 (Affymetrix, USA) software was used for result processing. Results: 13,815 genes were examined, in order to detect LOH. The frequency of LOH varied from 0% to 63%. The association analysis identified four genes: EDA2R, PGK1, TAF9B and CYSLTR1 that demonstrated the presence of LOH associated with metastasis-free survival (log-rank test, p < 0.03). Conclusions: The presence of LOH in EDA2R, TAF9B, and CYSLTR1 genes is associated with metastasis-free survival in breast cancer patients, indicating their potential value as prognostic markers.

Submitted: March 16, 2017.
*Correspondence: E-mail:
Abbreviations used: AI — allelic imbalanced; BC — breast cancer; cnLOH — copy-neutral loss of heterozygosity; LOH — loss of hete­rozygosity; NAC — neoadjuvant chemotherapy; SNP — single nucleotide polymorphism.

It is well known that one of the factors determining the individual diversity and heterogeneity of the tumor, including breast cancer, is a normal genetic variability caused by gene polymorphism (single nucleotide polymorphism — SNP). SNP in the tumor tissue can be manifested as the phenomenon of allelic imbalance (AI). Also, the special case — the loss of heterozygosity (LOH) — is often detected. LOH is a loss (structural or functional) of one of the alleles of a heterozygous genotype, resulting in a reduced frequency of heterozygous genotypes compared with genomic DNA. AI is presented by allelic deletions (loss of one copy of the locus), duplication or amplification of one allele [1]. A. Knudson was the first to describe the phenomenon of AI in tumors in 1971. According to his two-hit carcinogenesis model, the inactivation of tumor suppressor genes required two successive mutational events — loss of one allele as a result of LOH and somatic mutation in the other allele. The loss of the allele provides the opportunity of a manifestation of lethal recessive mutation in the remaining allele [2, 3].

In addition, LOH and AI can lead to the activation of oncogenes and inactivation of tumor suppressor genes, which can result in uncontrolled cell growth and metastasis [4, 5]. Currently, allele imbalance and LOH are well explored and shown for many genes in different types of cancer. The biological meaning of LOH in carcinogenesis is suggested to be associated with inactivation of heterozygous loci of pathogenetically significant genes, thus providing tumor progression, including metastasizing [6–12]. From perspective of breast cancer (BC), the most important AI and LOH were shown for cancer-related genes, such as the ERBB2 (HER2) [13], BRCA1 and BRCA2 [14–16]. Recently, a large study on LOH in BC was performed for the following genes: EGFR, TERT, TP53, CASP8, PARP2, GATA3, and BRCA1 [17]. Also, it was demonstrated that the LOH in telomeric sites of chromosomes in BC and ovarian tumor cells appeared to be a good predictor of clinical course of diseases [18]. However, most studies on AI, especially the LOH in BC, were focused on genes, contributing to malignant transformation of normal epithelial breast cells, but not on cancer progression [19].

Thus, the genome-wide LOH study in relation to cancer progression is necessary to understand its contribution to disease outcome and also to find new effective prognostic markers. In this work we performed the genome-wide study on the LOH in association with metastasis-free survival in BC patients.


The study group. A total of 68 patients aged from 28 to 68 years (median age 53 years) were enrolled in the study (Table 1). The diagnosis of BC was verified morphologically. The tumor stages were IIA–IIIB. In accordance with the Consensus Conference on Neoadjuvant Chemotherapy in Carcinoma of the Breast, April 26–28, 2003, Philadelphia, Pennsylvania [20], all patients underwent 2–4 courses of neoadjuvant chemotherapy (NAC) by FAC scheme (5-fluorouracil, doxorubicin, cyclophosphamide), CAX scheme (cyclophosphamide, doxorubicin, capecitabine). Surgery was performed 3–5 weeks after the completion of NAC followed by two cycles of adjuvant chemotherapy with FAC regimen. Radiotherapy and/or hormonal treatment were given if required.

The study was carried out in accordance with Helsinki Declaration of 1964 (amended in 1975 and 1983) and was approved by The Ethical Committee of the Institute of Oncology. Signed informed consent was obtained from all participants.

Table 1. The clinicopathological parameters of BC patients (n = 68)
Trait Value Number of patients, n (%)
Age, years ≤ 45 21 (30.9)
> 45 47 (69.1)
Menstrual status Premenopausal 36 (52.9)
Postmenopausal 32 (47.1)
Histological type Invasive ductal carcinoma 58 (85.3)
Invasive lobular carcinoma 3 (4.4)
Medullary carcinoma 2 (2.9)
Others 5 (7.4)
Tumor size T1 9 (13.2)
T2 52 (76.5)
T3 3 (4.4)
T4 4 (5.9)
Lymph node status N0 27 (39.7)
N1 31 (45.6)
N2 4 (5.9)
N3 6 (8.8)
Estrogen receptor Positive 33 (48.5)
Negative 31 (42.6)
No data 4 (5.9)
Progesterone receptor Positive 35 (51.5)
Negative 29 (42.3)
No data 4 (5.9)
HER2 0/+ 47 (69.1)
++ 10 (14.7)
+++ 6 (8.8)
No data 5 (7.4)
Molecular subtype Luminal B 40 (59.7)
Triple negative 17 (25.4)
HER2-positive 10 (14.9)
Histological form Unicentric 45 (66.2)
Multicentric 23 (33.8)
NAC regimen CAX 28 (41.2)
FAC 40 (58.8)

DNA extraction. Biopsies of tumor tissues were obtained before treatment under ultrasound guidance. The tissues were placed in RNAlater (Ambion, USA), incubated for 24 h at room temperature and stored at –80 °С until DNA and RNA extraction.

DNA was extracted from 68 biopsy specimens of tumor tissues using QIAamp DNA mini Kit (Qiagen, Germany). DNA concentration and purity were assessed using NanoDrop 2000 instrument (Thermo Scientific, USA). The concentration varied between 50 to 150 ng/µl and А260/А280 and А260/А230 ratios were 2.10–2.35 and 2.15–2.40, respectively. The integrity of DNA was assessed using TapeStation instrument (Agilent Technologies, USA); the fragments of the DNA were no longer than 48 kbp, thus suggesting its high integrity.

Microarray analysis. The CytoScanTM HD Array chips (Affymetrix, USA) were used. They contained probes for 2,670,000 markers including 1,900,000 non-polymorphic markers for the analysis of copy number variations and more than 750,000 SNPs. The sample processing, arrays hybridization and scanning were performed according to the manufacturer’s protocols for Affymetrix GeneChip® Scanner 3000 7G. The results were analysed using Chromosome Analysis Suite 3.1 software (Affymetrix, USA).

Statistical analysis. Statistical analysis was performed using STATISTICA 8.0 software package (StatSoft Inc., USA). A significance level of p < 0.05 was considered for statistically significant differences between treatment groups. p-value < 0.05 was considered statistically significant. The Kaplan — Meier curves were used for the analysis of overall and meta­stasis-free survival [21]. Comparison of survival curves of the treatment groups was done using the log-rank test. Two-sided Fisher’s exact test was used for analyzing comparison of frequencies for qualitative data.


First of all the frequency of LOH in genes from the OMIM (Online Mendelian Inheritance in Man) database was studied ( A total of 13,815 genes were evaluated for heterozygosity loss in tumors of BC patients. The frequency of LOH varied from 0% to 63% for certain genes, thus 63% of patients showed LOH in some genes. The LOH frequency for 13,815 genes annotated in the OMIM database is presented in Fig. 1. In accordance with the localization of genes on the chromosomes, they are plotted on the abscissa, and the percentage of patients with the LOH in the certain gene is depicted on the ordinate. Fig. 1 clearly shows that the highest incidence of the LOH events is characteristic for genes on chromosomes 17 and X.

112313 Genome wide association study of loss of heterozygosity and metastasis free survival in breast cancer patients
Fig. 1. A pattern of the LOH presence in the genes, annotated in the OMIM database, in relation to their chromosome localization

The highest LOH frequency (from 30 to 63%) was observed in loci of ZNF267, PAGE3, MAGEH1, RRAGB, KLF8, UBQLN2, SPIN2B, SPIN2A, FAAH2, ZXDB, ZXDA, ARHGEF9, FAM123B, MSN, MIR223, VSIG4, HEPH, EDA2R, AR, SLC16A2, RLIM, KIAA2022, ABCB7, UPRT, ZDHHC15, MAGEE2, and MAGEE1 genes. It is important to test the functional significance of these genes as tumor suppressors. In concordance with the information of the OMIM database, chromosomes 6, 7, 11 and 14 possess large regions, where the LOH was not observed in any of the 68 BC samples examined. In total, no LOH was detected in 873 out of 13,815 genes, annotated in the OMIM database. Well-known tumor suppressor genes, such as TP53, RB1, BRCA1 etc., also showed the low LOH frequency (Table 2).

Table 2. Frequency of LOH in known tumor suppressor genes
Tumor suppressor genes Frequency of LOH, n (%)
MEN1, TNFAIP3 0 (0.0)
CEBPA, JAK2, WRN 1 (0.7)
APC, ATM, CDH11, EXT1, EXT2, NF2, RUNX1, SDHD, SMARCA4 4 (2.7)
BLM, BRCA2, CBFA2T3, PML, RB1 6 (4.1)
FLT3, MAP2K4 7 (4.8)
CDH1, SUZ12 8 (5.4)
CDKN2C, NF1, TP53 9 (6.1)
BRCA1 10 (6.8)
GPC3 11 (7.5)
DDX5 12 (8.2)

The next step of investigation was to analyze the relationship between distant metastasis of breast tumors and the LOH frequency in OMIM genes. Distant metastases occurred in 17 (25%) of the 68 patients within 10 to 77 months from the time of diagnosis. The two-, three- and five-year metastasis-free survival rates were 86.8; 82.4, and 76.5%, respectively.

The patients were divided into two groups: the first group consisted of 17 patients with metastases and the second group consisted of 51 patients without metastases.

Based on a difference between the frequency of the normal status of the gene and the LOH in patients with metastases and without metastases, we have selected genes in which this difference reached the maximum value. In result of our study seven genes were selected, namely, FGD1, GNL3L, TRO, EDA2R, PGK1, TAF9B, and CYSLTR1 for those the difference ranged from 25.5 to 35.3%.

The next step was to evaluate the association between the LOH and metastasizing, using a Fisher’s exact test. The association between the LOH and metastasizing was demonstrated only for 4 genes: EDA2R, PGK1, TAF9B, and CYSLTR1. Using the Kaplan — Meier survival analysis, we showed that the metastasis-free survival rate in patients with the LOH in these genes was significantly lower, compared with those observed in the group of patients, who had a normal status of these genes (Fig. 2).

231231 Genome wide association study of loss of heterozygosity and metastasis free survival in breast cancer patients
Fig. 2. Metastasis-free survival of BC patients with the detected LOH in a set of genes: (a) Ectodysplasin A2 Receptor (EDA2R) gene; (b) phosphoglycerate kinase 1 (PGK1) gene; (c) RNA polymerase II, TATA Box-Binding Protein-Associated Factor (TAF9B) gene; (d) cysteinyl leukotriene receptor 1 (CYSLTR1) gene.
Note: a red line — patients with the LOH in genes examined; a blue line — patients with the normal status of genes examined. p — a significance value, calculated by a log-rank test

It should be noted that all of the genes identified in the study were located on the X chromosome ( EDA2R gene was localized on the long arm of the X chromosome (Xq12), while PGK1, TAF9B, and CYSLTR1 genes were loca­lized in Xq21.1.

Deletions with LOH can provide a decrease in the expression of genes with LOH at the expense of to gene deficiency. Amplification with LOH results in an increase in gene expression. The copy-neutral LOH (cnLOH) may increase or decrease the expression of the gene with cnLOH. cnLOH might confer a phenotypic advantage for tumor cells due to gain of imprinting through duplication of a methylated allele (decrease the expression of the gene with cnLOH), loss of imprinting through duplication of unmethylated allele (increase the expression of the gene) or homozygosity of an initial heterozygous mutation [22]. The expression of genes with LOH it was necessary to explore in order to understand the impact of LOH on gene expression.

In the study Ching et al. [23] was to explicate the underlying chromosomal copy number alterations and LOH implicated in a cohort of Malaysian hospital-based primary breast carcinoma samples (n = 70) using a SNP-array platform. The most prevalent gains (≥ 30%) were detected at the 8q arm, whilst the most ubiquitous losses (≥ 20%) were noted at the 8p and 17p regions. The frequency of deletion with LOH of PGK1, TAF9B, and CYSLTR1 genes were 10%. cnLOH was characterized as the most prevailing LOH event, in which the most frequent distributions (≥ 30%) were revealed at 3p21.31, 5q33.2, 12q24.12, 12q24.12‑q24.13, and 14q23.1. The frequency of cnLOH EDA2R gene (ectodysplasin A2 receptor) was 24%, which along with TNFR1 and Fas (CD95) is one of the receptors of death. Yan et al. [24] showed that EDA2R gene appeared to activate expression of tumor necrosis factor gene. Studies suggest that EDA2R is a potential downstream effector of p53-induced apoptosis in cancer cells [25, 26] and may therefore be a potential tumor suppressor, it is down-regulated in breast and colorectal cancers [27, 28], and mutations and promoter hypermethylation of EDA2R have been identified in colorectal cancer cells [25, 26]. A recent study conducted in 2016 using a microarray CytoScan HD Arraу showed that the presence of cnLOH in EDA2R gene and in several other genes was associated with ameloblastoma cases [29].

The protein encoded by PGK1 gene is a glycolytic enzyme that catalyzes the conversion of 1,3-diphosphoglycerate to 3-phosphoglycerate. Additionally, this protein is secreted by tumor cells where it participates in angiogenesis by functioning to reduce disulfide bonds in the serine protease, plasmin, which consequently leads to the release of the tumor blood vessel inhibitor angiostatin ( CpG-island methylation was evaluated on a 56-gene cancer-specific biomarker microarray in metastatic vs non-metastatic BC in a multi-institutional case series of 123 BC patients. This identified 11 genes (including the PGK1 signaling kinase gene) as the highest differentially methy­lated genes between progressing and non-progressing BC [30]. The PGK1 gene showed differential splicing in TNBC, non-TNBC (luminal A and B) and HER2-positive BC subtypes [31].

The TAF9B (TATA-Box Binding Protein Associated Factor 9b) is involved in transcriptional activation as well as repression of distinct but overlapping sets of genes. Among it related pathways is RNA Polymerase II transcription initiation. Microarray screening identified a TAF9B gene whose expression was significantly changed by combination treatment with inhibitors of LSD1 and HDAC of human BC cells [32]. Collectively, p53, SNHG1, sno-miR-28, and TAF9B form a regulatory loop which affects p53 stability and downstream p53-regulated pathways [33].

CYSLTR1 (Cysteinyl Leukotriene Receptor 1) is involved in mediating bronchoconstriction via activation of a phosphatidylinositol-calcium second messenger system. Upregulation of this gene is associated with asthma and dysregulation may also be implicated in cancer. Patients with breast tumors characterized by high CysLT1R and low CysLT2R expression levels exhibited increased risk of cancer-induced death in univariate analysis for both the total patient group (hazard ratio [HR] = 2.88, 95% confidence interval [CI] = 1.11–7.41) [34].

Thus, the presence of LOH in EDA2R, PGK1, TAF9B, and CYSLTR1 genes were shown to be associated with low metastasis-free survival in patients with BC, indicating their potential value as prognostic markers.


This work was supported by RFBR grant 16-54-76015 ERA and Tomsk State University Competitiveness Improvement Program.


No Comments » Add comments
Leave a comment

ERROR: si-captcha.php plugin says GD image support not detected in PHP!

Contact your web host and ask them why GD image support is not enabled for PHP.

ERROR: si-captcha.php plugin says imagepng function not detected in PHP!

Contact your web host and ask them why imagepng function is not enabled for PHP.