The aim of this study was to investigate BORIS expression in breast cancer cells and tumours. Although the antibody used in this investigation was previously characterized for their specificity for BORIS (Loukinov et al, 2002
), we performed additional experiments to confirm these observations (Supplementary Figure 1
). In these experiments, the polyclonal chicken anti-BORIS antibody was shown to be specific by the following criteria: (1) intensity of the immunohistochemical staining of the breast tumour tissue was significantly reduced after preincubation of the antibody with the blocking peptide, (2) intensity of the immunofluorescent staining of the cells in culture was significantly reduced after preincubation of the antibody with the blocking peptide, (3) western analysis of the N-terminal domain of BORIS expressed in E. coli
and the protein produced from the full-length BORIS cDNA in the eukaryotic-expressing vector in 293T cells demonstrated the presence of the recombinant products, recognised by the anti-BORIS antibody and (4) preparative immunoprecipitation of the cell lysates with the anti-BORIS antibody and subsequent analysis of the 85-kDa band characteristic for BORIS by mass-spectrometry (MS/MS sequencing), revealed the presence of peptides matching BORIS in this band.
Using this antibody, we found that BORIS was expressed at different levels in all 18 breast cancer cell lines used in the experiments. It is, however, difficult to correlate the levels of BORIS detected by western analysis and immunohistochemical staining of the same cells (e.g. MDA-MB-435 and MDA-MB-468, ). Thus, the apparently denser accumulation of staining in MDA-MB-468 cells as seen by the immunohistochemistry can be due to different cell morphology of MDA-MB-468 (small cytoplasm, more rounded cell shape).
As BORIS was detected in 70.7% of breast tumours, the presence of BORIS in 100% cell lines may indicate that the selective pressure in vivo
on some tumours may be absent in vitro
. The availability of the component(s) in the growth media that may be important for BORIS activation could be a factor; access to such component(s) may be restricted in the microenvironment of some tumours. Given the complex nature of BORIS regulation at the transcriptional level it may be possible that different promoters are involved in the control of BORIS expression in cultured cells and tumours; thus, providing different molecular mechanisms of BORIS activation in vivo
and in vitro
(Renaud et al, 2007
) (see below).
There was no obvious correlation between the amount of BORIS in non-malignant and malignant cells, which suggests that BORIS may not play a key role in the establishment of the malignant phenotype. This is supported by the lack of correlation between BORIS levels in breast tumours and lymph node metastasis, which is used in clinical practice as an indicator of the malignant breast cancer (Tobler and Detmar, 2006
; Eccles et al, 2007
). On the other hand, BORIS was not detected in primary normal breast cells, which suggests that BORIS is likely to be associated with the immortalised and malignant cells.
The immunohistochemical staining revealed that the levels of BORIS protein were significantly higher in all breast tumours compared with normal and PP tissues; 70.7% of breast tumours were BORIS-positive. All normal breast tissues from BR specimens were negative for BORIS. Only 1 tissue out of 15 from the group of the PP tissues was very weakly positive; we explain this by influence of the tumour on normal tissue, which may take place in some cases. Such occurrences are well documented in the literature for mRNA/proteins for both CTA and non-CTA genes (Janz et al, 2002
; Zhao et al, 2004
In our investigation, the levels of the BORIS
mRNA were considerably higher in breast tumour tissues compared with the PP tissues. However, no correlation was observed between the levels of BORIS
mRNA and protein (). Lack of coordinated expression between mRNA and protein is not unusual – poor correlation between expression of an mRNA and the corresponding protein is now generally acknowledged and accepted (Abbott, 1999
). Genes showing no obvious direct link between expression of mRNA and corresponding protein in various cell lines and tumours are described in the literature and include p27Kip
(Ciaparrone et al, 1998
), ERs and PRs (Tong et al, 1999
). The heterogenous nature of breast tumour tissues may also account for this discrepancy as different parts of the tissues are used for extraction of protein and mRNA. The latter may also be the reason for the lack of good correlation between the immunohistochemical staining (IRS for BORIS) and western analysis data ().
We noted the heterogenous pattern of BORIS staining in breast tumour tissues, showing single BORIS-positive cells or clusters of positive cells, with varying staining intensity. This may explain the fact that the IRS for BORIS has not reached the values higher than 9. Such pattern of expression is common to CTA, examples include distribution of MAGE antigens in various tumours (Jungbluth et al, 2000a
Other CTA have also been detected in breast tumours. Although some CTA can be detected in breast tumours of different type, other CTA are observed only in particular breast tumours. Thus, XAGE-1 was observed in invasive lobular as well as ductal carcinoma (Egland et al, 2002
), and NY-BR-1 was detected in various breast carcinomas (Theurillat et al, 2007
). MAGE-A was seen frequently in primary ductal breast carcinomas, but not in lobular carcinomas (Otte et al, 2001
). Almost 81% of the ductal carcinomas have been shown to express PLU-1 (Barrett et al, 2002
Cancer testis antigens are considered to be promising candidates for tumour vaccines because of their immunogenicity and tissue-restricted expression (Scanlan et al, 2004
; Kalejs and Erenpreisa, 2005
). Recent studies using a mouse model revealed that immune response to BORIS can be developed in the organism; and furthermore, such immune response has protective effects against several mouse tumours of different origin (Loukinov et al, 2006
; Ghochikyan et al, 2007
). These data indicate that BORIS may be an attractive candidate for the development of the future cancer vaccine. For such a therapy, it will be critical to determine the presence of BORIS in the tumour given that a proportion of breast tumours (29.3%) are BORIS negative.
The incidence of BORIS expression at the protein level (70.7%) observed in the current study corresponds well to the reported findings of BORIS
mRNA expression in endometrial and uterine tumours (77%) (Risinger et al, 2007
). The fact that not all tumours contain BORIS may reflect different pathways of evolution in different tumours. Another possibility may be the presence in the tissues of the alternative forms of BORIS not detected by the existing antibodies. Indeed, recently 25 full-length and partial splice variants of BORIS messages coding for 17 different proteins have been cloned and submitted to the GenBank (E Pugacheva et al
, manuscript in preparation). Six full-length BORIS splice isoforms (GenBank accession nos. DQ778125, DQ778124, DQ778126, DQ778129, DQ778128 and DQ778127) and two partial BORIS cDNA isoforms (GenBank accession nos. DQ778131 and DQ778130) do not have the exon-1 sequence for the peptide recognised by the affinity-purified chicken anti-BORIS antibody (Loukinov et al, 2002
), which was used in the present study for staining breast tissue sections. Therefore, it is likely that results obtained by the immunostaining with the chicken anti-BORIS antibody may underestimate the actual frequency of presence of BORIS in breast cancers if splice variants undetectable by this antibody are expressed.
In this investigation, we also assessed the clinical significance of BORIS in breast tumours and found that high BORIS levels correlated with high levels of PR and ER. Both oestrogen and progesterone are important hormones in mammary development in humans stimulating cell growth, proliferation and differentiation (Barron et al, 1997
; Humphreys et al, 1997
; Anderson, 2002
). Moreover, both hormones have been demonstrated to promote breast tumorigenesis (Medina, 2005
; Montero Girard et al, 2007
). It is conceivable that BORIS may stimulate production of both PR and ER, which in turn may support tumour progression. Further experiments confirmed that BORIS could activate promoters of both ER
genes in the reporter assays in cells of breast and non-breast origin; thus, pointing at the possible molecular mechanism of direct transcriptional regulation of these genes. Inspection of the promoter regions of the ER
genes tested in the reporter assays indeed revealed potential-binding sites for CTCF/BORIS (Supplementary Figures 4 and 5
). However, further investigations will need to be carried out to confirm direct binding of BORIS to these sites and clarify the role of BORIS in the regulation of ER
The association of BORIS with immortalised and malignant cells indicates that BORIS may be important in the establishment and maintenance of cell proliferation. This may be linked to the original BORIS function, regulation of gametogenesis, which BORIS is likely to coordinate with other CTA family members (Old, 2001
). The appearance of common CTA (including BORIS) during gametogenesis and tumorigenesis prompts the hypothesis that induction of the gametogenetic programme in somatic cells may be associated with tumour development (Old, 2001
; Kalejs and Erenpreisa, 2005
). As BORIS seems to function as an upstream regulator for several CTA (Hong et al, 2005
; Vatolin et al, 2005
), it is tempting to speculate that BORIS could be the ‘master switch' in the process of cell reprogramming, which can guide the epigenetic machinery to a set of target genes eventually leading to their activation. It is important to note that BORIS may also act as an activator of genes responsible for proliferation in particular tissues, such as ER
genes in mammary glands.
An intriguing question is why and how BORIS itself is activated. Recent reports reveal that both genetic and epigenetic mechanisms are likely to be implicated in this process. Thus, DNA methylation, functional p53 and CTCF play an important role in the negative regulation of the promoters of the BORIS
gene (Renaud et al, 2007
). Demethylation of DNA, knockout of CTCF and absence of functional p53 can lead to strong activation of BORIS. Selective utilisation of different promoters (Renaud et al, 2007
) and alternative splicing (Pugacheva et al
, manuscript in preparation) are also likely to contribute to regulation of BORIS expression in different cell types. The research efforts of several laboratories are currently focused on uncovering the details of the molecular mechanisms of activation and regulation of BORIS.