ASSOCIATION OF THE COMBINATION OF STEMNESS GENE AMPLIFICATIONS AND COPY NUMBER ABERRATIONS OF WNT-SIGNALING GENES IN BREAST TUMORS WITH METASTASIS

We studied the association between the presence of 2 or more stemness gene amplifications as well as copy number aberrations (CNAs) of WNT signaling genes in residual breast tumor and metastasis. WNT pathway genes associated with metastasis were identified. Material and Methods . The study included 30 patients with breast cancer, who had 2 or more stemness gene amplifications in the residual tumor after neoadjuvant chemotherapy. Fifteen of the thirty patients developed hematogenous metastases; they constituted a group with metastases, the remaining 15 patients entered the second group without metastases. The tumor dNA was examined using a CytoScanHd Array microarray (Affymetrix, USA). Results. By subtracting amplification and deletion frequencies in 852 cytobands between groups with metastases and without metastases, 21 cytobands were identified with the largest difference in deletion and amplification frequencies. They contain 19/150 of WNT genes (12 activators: SKP1 , WNT8A , MAPK9 , CCND3 , FZD9 , WNT8B , CCND1 , PLCB2 , PRKCB , FZD2 , WNT3 , WNT9B and 7 negative regulators: GSK3B , APC , CSNK2B , SFRP5 , BTRC , TCF7L2 , CSNK2A2 ). A point system was developed: when amplifying WNT-signaling activators or deletion of negative regulators, one point was added to the total score, and vice versa when deleting WNT-signaling activators or amplification of negative regulators, one point was taken from the total amount. it was shown that 93% (14/15) of patients with metastases had a total score higher than 0, while 93% (14/15) of patients without metastases had a total score of zero or less than zero. The differences between the groups were statistically significant according to the two-sided Fisher test with a high level of confidence probability (p=0.000003) and the log-rank test (p=0.00004) when assessing non-metastatic survival by the Kaplan-Mayer method. Conclusion Nineteen WNT signaling genes were identified. Copy number aberrations of these genes in combination with stemness gene amplifications in residual tumors were associated with metastasis. A new highly effective prognostic factor for breast cancer was identified. The clinicopathological parameters of breast cancer patients and stemness gene amplifications in residual


Introduction
Previously, we showed the association of amplifications of chromosome regions of stemness genes localization (3q (26.33), 5p (15.33; 13.1), 6p (24.3; 22.3; 21.33; 21.32), 7q (11.23; 21.13; 31.2; 32.1), 8q (11.21; 24), 9p (21.2), 9q (34.3; 21.13; 31.2, 22.33), 10p (15.2; 13; 12.2; 11.22), 10q22. 1, 12p (13.31) 13q (34; 32.3; 22.1; 13.3; 12.2), 16p (11.2; 13.3), 18q (21.1; 21.2) 19p (13.3; 13.2; 13.12) with hematogenous metastasis. It was found that in cases when tumor cell clones with 2 or more amplifications of the above chromosomal regions remained in a residual tumor after neoadjuvant chemotherapy, 50 % of the patients developed metastases. When all amplifications were eliminated or only one amplification remained, 100 % of patients did not develop metastases [1]. The results of this study showed that the presence of 2 or more amplifications at regions of stemness gene localization in residual tumors was a necessary, but not sufficient condition for metastasis development. There are some other genetic changes that, along with amplification of stemness genes, are necessary and sufficient for development of metastases after presurgery chemotherapy. Milanovic et al. (2018) reported that replicative aging of breast tumor cells occurred after treatment with doxorubicin and tamoxifen. After cancellation of doxorubicin and tamoxifen, the activity of the WNT signaling pathway significantly increased in tumor cells and the number of stem tumor cells increased LABORATORY AND ExPERIMENTAL STUDIES sharply. Tumor growth became even more active than it was before chemo-and hormone therapy [2]. It remains unknown why in some cases WNT signaling pathways are activated, while in other cases they are not activated and some tumors do not progress.
Our working hypothesis is that activation of WNT signaling pathway in tumor cells is due to the presence of copy number aberration (CNA) of WNT signaling pathway genes. We believe that a tumor emerges from replicative aging caused by NAC and metastasizes if there are 2 or more amplifications of stemness genes (a necessary condition) and CNA genes of WNT signaling pathway. If there are only amplifications of stemness genes or CNA genes of WNT signaling pathway, a tumor does not metastasize.
In addition to emerging from replicative aging, WNT can directly induce stemness genes, thereby providing a significant increase in the number of tumor stem cells and the likelihood of metastasis. Some authors have shown that MYC, which is one of the stemness genes and is part of the Yamanaki cocktail, is directly connected to WNT signaling pathway and when WNT is attached to Frizzled receptors, expression of markers of epithelial-mesenchymal transition and stemness of SNAI2, MYC and others is activated. Canonical WNT signaling caused by LRP6 Cyclin Y/CDK14 priming helps stabilize β-catenin-independent proteins, block polyphosphorylation and polyubiquitination of target proteins, including c-Myc [3,4]. Activation of β-catenin in WNT signaling pathway also stimulates hyperexpression of MYC [5].
In accordance with the working hypothesis, we identified WNT pathway genes, whose CNA together with the amplification of stemness genes, were associated with metastasis.

Material and Methods
Between 2009 and 2014, 30 patients with IIbIIIa (T2-3N0-1M0) luminal B breast cancer were treated at Cancer Research Institute of Tomsk NRMC (Tomsk, Russia). The age of the patients ranged from 35 to 62 years (average age 52.4 ± 0.5) All procedures followed were in accordance with the Helsinki Declaration (1964, amended in 1975 and 1983). This study was approved by the institutional review board, and all patients signed an informed consent for voluntary participation (tabl. 1). All patients received 6 to 8 cycles of systemic neoadjuvant chemotherapy with the AC (Adriamycin, and Cyclophosphamide) or AT/ACT (Adriamycin and Taxotere or Adriamycin, Cyclophosphamide and Taxotere), or Taxotere. Clinical and imaging responses were categorized according to WHO and International Union Against Cancer criteria [7]. A partial response (PR) was determined as a tumor reduction >50 % and stable disease (SD) as a tumor reduction <50 % or a tumor size increase of <25 % and progressive disease (PD) as a tumor size increase of >25 %. Physical examination was performed before NAC and before surgery to determine the clinical response. Surgery (radical resection, sectoral resection or mastectomy) was performed within three to four weeks after the last administration of chemotherapy in responsive patients. After surgery hormonal therapy was given.
DNA was extracted from fresh samples of post-NAC residual tumor tissues using the QIAamp DNA mini Kit (Qiagen, Germany #51304).
Microarray analysis. To study CNAs of breast tumor, microarray analysis was performed using high density microarray platform Affymetrix (USA) CytoScan TM HD Array, (http://www.affymetrix.com/ esearch/search. jsp?pd= prod520004&N=4294967292). Procedures of sample preparation, hybridization and scanning were performed in accordance with the manufacturer's protocol using the system Affymetrix GeneChip® Scanner 3000 7G (Affymetrix, USA). The Chromosome Analysis Suite 4.0 software (Affymetrix, USA), which was specifically devised for analyzing microarray results from the CytoScan TM HD Array, was used. Unbalanced chromosomal aberrations (deletions and amplifications, or Loss and Gain) were detected in all chromosomal regions.
Bioinformatics Methods. For each of 852 cytobands of patient groups with 2 or more amplifications of stemness genes with metastases and without metastases, the frequency of amplifications and deletions was determined and a histogram was constructed for each group. Afterwards, these histograms were combined. Thus, cytobands were calculated with the greatest difference in amplification and deletion frequencies.
Statistical analysis. A two-sided p-value was calculated using Fisher's exact test http://vassarstats. net/odds2x2.html. Metastasis-free survival was calculated using the Kaplan-Meier method, and differences among patient groups were evaluated using the log-rank test.
We divided all identified WNT-signaling genes into three large groups. The first group included activators of WNT-signaling (lilac); it was composed of receptor genes of the WNT signal pathway, receptor ligands, secondary messengers and transcription factors that play a key role in the work of the WNT-pathway. Also, this group included two cyclins D3 and D1, which trigger the cell cycle via WNT-signaling and are the endpoint of the WNT signaling pathway. In total, this group included 12 genes: SKP1, WNT8A, MAPK9,CCND3,FZD9,WNT8B,CCND1,PLCB2,PRKCB,FZD2,WNT3,WNT9B. Activation of WNT signaling should be substantially facilitated by amplification of chromosomal regions of localization of these genes and substantially hindered by deletions. The second group of genes included genes whose products, according to OMIM and GeneCards, negatively regulate the WNT signaling pathway (green). This group included 7 genes: GSK3B, APC, CSNK2B, SFRP5,BTRC,TCF7L2,CSNK2A2. Activation of WNT signaling should be significantly facilitated by deletions of chromosomal regions of these genes localization and significantly hindered by their amplifications. The third group was composed of positive regulators of canonical and noncanonical WNT signaling pathways (PPP2CA, TCF7, PPARD, PPP2R5D, CUL1, CHP1, FRAT1, FRAT2). It included some transcription factors, secondary messengers, kinases. They can exert a noticeable but not as critical effect as activators on the WNT-signaling pathway, while the activity of the products of these genes can either be suppressed by other factors, or they belong to the noncanonical pathway, or they regulate the manifestations of WNT signaling little associated with proliferation, migration, adhesion and stemming, i.e. those mechanisms that are necessary for metastasis. CNA regions of localization of these genes were excluded from further analysis due to the high pleiotropy of mutual influences and low significance for metastasis mechanisms. CHP1 belong to the noncanonical WNT-signaling pathway.

LABORATORY AND ExPERIMENTAL STUDIES
Рис. 1. Частота опухолевых CNA (амплификации и делеции) у пациенток с метастазами и без них Fig. 1 Tumor CNA (amplifications and deletions) frequencies in patients with and without metastases After such preliminary bioinformatics selection of WNT-signaling genes, the relationship with metastasis was assessed only for the first (lilac) and second (green) gene groups (tabl. 2). In accordance with the working hypothesis, metastasis should be facilitated by amplification of gene loci of the first group and deletion of loci of genes of the second group, while inhibition of metastasis should be done by deletion of gene loci of the first group and amplification of loci of genes of the second group. In accordance with this formulation, a point system was developed. One point was added to the total score when amplifying WNT-signaling activators or deletion of negative regulators, and vice versa, when deleting WNT-signaling activators or amplification of negative regulators, one point was taken from the total amount. The distribution of amplifications and deletions at all loci of the genes of the first and second groups and the total score for all 30 patients studied are presented in Fig. 2. As seen in Fig. 2, 93 % (14/15) of patients with metastases have a total score greater than 0, while 93 % (14/15) of patients without metastases have a total score of zero or less than zero. The differences between the groups are statistically significant according to the two-sided Fisher test with a high level of confidence probability (p=0.000003) and the log-rank test (p=0.00004) when assessing metastasis-free survival by the Kaplan-Mayer method ( fig. 3).
Thus, our data indicate that CNA genes of WNTsignaling are associated with metastasis and the prognostic value of CNA genes of WNT-signaling in residual tumors is highly predictive, which confirms our working hypothesis that CNA genes of WNT signaling are involved in the implementation of the metastasis process. Amplifications of stemness genes give tumor cells the ability to do stem transition [1], amplifications of activators SKP1, WNT8A,MAPK9,CCND3,FZD9,WNT8B,CCND1,PLCB2,PRKCB,FZD2,WNT3,WNT9B and / or deletions of negative regulators of WNT signaling genes: GSK3B, APC, CSNK2B, SFRP5,BTRC,TCF7L2,CSNK2A2, contribute to the release of tumor cells from replicative aging and dormant state, and activation of stem transition.

Discussion
The association of the combination of stemness gene amplifications and WNT-signaling genes in residual tumors with hematogenous metastasis was studied. Based on the microarray study and using bioinformatics methods, the most important WNT-signaling genes were identified for 12 WNT signaling activator genes and 7 negative regulator genes, amplification and deletion (respectively) of which were associated with metastasis (according to the Fisher's two-sided criterion, p=0.000003 and the log-rank test, p=0.00004, when evaluated by the Kaplan-Mayer method).
According to modern recommendations, adjuvant chemotherapy (ACH) is given to patients, who have not previously received 6-8 cycles of NACH. Current methods for predicting the risk of relapse (in particular, the Oncotype DX platform), which show the need for ACT are often inefficient, and even if they are used, the majority of patients (more than 40-50 % of patients) receive postoperative chemotherapy [8], while according to our data, only 25 % of patients need adjuvant chemotherapy. These are patients with 2 or more amplifications of stemness genes and a positive total CNA score of WNT signaling genes. Our new prognostic factor is currently one of the most highly effective for breast cancer. Some of the WNT signaling genes that we have identified are considered in the modern literature in terms of metastasis mechanisms. Gene FZD9 gene

LABORATORY AND ExPERIMENTAL STUDIES
is activated after exposure of breast tumor cells to chemotherapy and is involved in the induction of the stem phenotype in these cells [9]. Amplification of gene CCND1gene is associated with a poor outcome in breast cancer [10]. Knockdown of CCND1 inhibits mammosphere formation of breast cancer cells [11]. MiR-4779 inhibits CCND3 in colon cancer cells, causing cell cycle arrest and apoptosis [12]. Inhibition of SKP1 leads to mitotic arrest of lung cancer cells [13]. Inhibitors of other activators of WNT signaling genes are being developed: PPP2CA [14], WNT3 [15] and others. For FZD2, on the contrary, it was shown that its hyperexpression was accompanied by suppression of metastasis of salivary adenoid cystic carcinomas [16]. Previous studies have shown that during EMT and metastasis, Wnt5a/b ligand and/or its cognate receptor Fzd2 are generally overexpressed in cell lines derived from late-stage mesenchymal-type cancers, such as melanoma and cancers of the, lung, colon, liver, and the gastric tract [17][18][19]. The WNT8A gene is associated with an unfavorable prognosis in patients with breast cancer treated with NAC [20]. As for negative regulators, there is much less information. The SNP of BTRC rs61873997 G>A gene was associated with the prognosis for skin cancer, and the presence of the wild allele had an unfavorable effect with HR=0.61 (0.46-0.80), p=0.00031) [21]. Suppression of negative WNT regulators does not always lead to inhibition of metastasis. In some cases, WNT activation inhibits metastasis. Thus, it was shown that pharmacological inhibition of GSK3B activity stabilized the TRAF6 protein, promoted CTNNB1 degradation, and effectively suppressed EMT and CRC metastasis [22]. On the other hand, tamoxifen treatment of breast cancer cells increases the phosphorylation of GSK3B and, therefore, its activity [23]. Low expression of SFRP5 in human pancreatic ductal adenocarcinoma (PDAC) was a poor prognostic factor for human PDAC [24]. Epigenetic inactivation of the SFRP5 gene in human breast cancer is associated with unfavorable prognosis [25]. The APC gene is a known tumor suppressor gene and in breast cancer its deletion is associated with resistance to doxorubicin and cisplatin [26,27].
The indirect effect of negative regulators should also not be neglected. It is well known that TP53 and CDKN2A genes are considered the main barrier to efficient conversion of normal cells into induced pluripotent stem cells. They inhibit spontaneous transformation of differentiated cells into pluripotent ones [28]. Some negative regulators are able to activate TP53 [3]. Apparently, deletions of these genes lead to a decrease in TP53 activity and, as a result, thresholds for transition of differentiated tumor cells to tumor stem cells are reduced. This may be one of the explanations of the effect observed by Milanovic [2] which consists in a sharp increase in the number of tumor stem cells upon activation of WNT signaling after cancellation of chemo and hormonal therapy.

Conclusion
The results of the study confirm the working hypothesis that CNAs of WNT signaling genes are involved in the metastatic process. Nineteen WNT signaling genes were identified. Copy number aberrations of these genes in combination with stemness gene amplifications in residual tumors were found to be associated with metastasis. A new highly effective prognostic factor for breast cancer was developed.