The distribution and expression of the Bloom's syndrome gene product in normal and neoplastic human cells

Bloom's syndrome (BS) is an autosomal recessive disorder associated with a predisposition to cancers of all types. Cells from BS sufferers display extreme genomic instability. The BS gene product, BLM, is a 159 kDa DNA helicase enzyme belonging to the RecQ family. Here, we have analysed the distribution of BLM in normal and tumour tissues from humans using a recently characterized, specific monoclonal antibody. BLM was found to be localized to nuclei in normal lymphoid tissues, but was largely absent from other normal tissues analysed with the exception of the proliferating compartment of certain tissues. In contrast, expression of BLM was observed in a variety of tumours of both lymphoid and epithelial origin. A strong correlation was observed between expression of BLM and the proliferative status of cells, as determined by staining for markers of cell proliferation (PCNA and Ki67). We conclude that BLM is a proliferation marker in normal and neoplastic cells in vivo, and, as a consequence, is expressed at a higher level in tumours than in normal quiescent tissues. © 2001 Cancer Research Campaign http://www.bjcancer.com

abnormalities, some features of premature ageing, and a predisposition to certain cancers.
Cells from Bloom's syndrome individuals show a variety of defects in the maintenance of genome stability. In addition to an elevated frequency of SCEs, Bloom's syndrome cells have a mutator phenotype and an elevated rate of genetic recombination events between homologous sequences, including the formation of quadriradial chromosomes (German, 1993(German, , 1995. Although the precise role of the BLM protein (or indeed of any of the RecQ family helicases) in DNA metabolism remains to be elucidated, many of these proteins appear to play roles in genetic recombination, particularly during DNA replication (Chakraverty and Hickson, 1999). Consistent with such a role, the BLM protein has been shown to interact physically and functionally with replication protein A (RPA) (Brosh et al, 2000), and to bind selectively to the Holliday junction recombination intermediate (Karow et al, 2000b). Indeed, BLM has been shown to promote efficient branch migration of Holliday junctions (Karow et al, 2000b), suggesting a possible role in the production of mature DNA recombinants during repair of DNA damage.
Previous studies using cultured cell lines have indicated that the BLM protein is poorly expressed in non-proliferating cells and peaks during S/G2-phases of the cell division cycle (Dutertre et al, 2000). Moreover, BLM protein has been shown to localize to the nucleus and at certain stages of the cell cycle to be present both in nucleoli and in PML nuclear bodies (Ishov et al, 1999;Zhong et al, 1999;Wu et al, 2000aWu et al, , 2000bYankiwski et al, 2000). However little is known about the expression of the BLM protein in cells and tissues in vivo, or whether BLM is generally expressed in human cancers. Using a specific anti-BLM antibody, which we have validated previously using cell lines and recombinant proteins , we have characterized the expression of BLM in normal and neoplastic tissues using immunohistochemistry. We show that BLM is a nuclear protein in human tissues and is expressed at a higher level in proliferating cells of lymphoid origin than in other normal tissues. In contrast, BLM is widely expressed in tumour cells of both lymphoid and epithelial origin.

Antibodies
Generation of the BFL-103 antibody has been described previously . The antigen used to raise this mouse monoclonal antibody was full-length recombinant BLM expressed in Saccharomyces cerevisiae (Karow et al, 1997). Antibodies specific for proliferating cell nuclear antigen (PCNA; PC10) and Ki67 were obtained from Dako (Denmark).

Tissues
A range of normal tissues (tonsil, spleen, lymphoid node, thymus, skin, pancreas, testis, colon, kidney, liver, brain and lung) were obtained from the Cellular Pathology Department at the John Radcliffe Hospital, Oxford, UK. Tumour samples from several organ sites were obtained from the frozen tissue bank in the same University Department. The diagnosis was reviewed and confirmed in each case. Cryostat sections of 8 µm were prepared and mounted on poly-L-lysine coated multi-well microscope slides (C. A. Hendley-Essex Ltd, Loughton, Essex, England.) After overnight drying, slides were processed for immunostaining immediately or stored at -20°C until required.

Immunohistochemistry
Slides were fixed for 10 min in 4% formaldehyde in phosphatebuffered saline. Endogenous peroxidase activity in the samples was blocked for 5 min in Peroxidase Block (Dako K4007) before rinsing in distilled water and washing in PBS. Primary antibodies were then applied to the slides for 90 min at the following concentrations: BFL-103, undiluted tissue culture supernatant from hybridoma cultures; Ki67 (10 µg ml -1 ); PC10 (1 µg ml -1 ). For BFL-103, the secondary antibody used was an Envision Mouse kit (1:200 dilution, Dako K4006). For the Ki67 and PC10 samples, the secondary was a goat anti-mouse HRP conjugate (1:200; Dako PO447). All secondary antibodies were applied for 30 min at room temperature. Slides were counterstained with haematoxylin before mounting in Aquamount (BDH). Substitution of the primary antibody with PBS was used as a negative control

Characterization of a BLM-specific monoclonal antibody
The generation of the BFL-103 anti-BLM mouse monoclonal antibody has been described previously . BFL-103 was raised in Balb-C mice following injection of purified, recombinant BLM protein. BFL-103 recognizes full-length recombinant BLM using Western blotting, as well as a recombinant fragment representing residues 1-447 of BLM, indicating that the BFL-103 epitope lies in the N-terminal domain of the 1417 amino acid BLM protein. BFL-103 was shown to be specific for BLM by demonstrating a lack of staining of GMO8505 cells isolated from an individual with BS and shown to lack expression of the BLM protein .

Normal tissue distribution of the BLM protein
In all tissues examined, where seen, BLM protein expression was confined to cell nuclei. In normal tissues, no positive staining with the BFL-103 antibody was observed (Figure 1 shows kidney as an example) except in areas where cell proliferation is ongoing. There was high expression in thymic cortex, lymphoid follicles of the tonsil, and around the periarteriolar lymphoid sheath of the spleen. Scattered staining for BLM was also evident in cells from the medulla of the thymus, the cortex and paracortex of the tonsil, the red pulp of the spleen, the placenta, and the basal layers of the skin. These are the areas of cell proliferation in these tissues.

Expression of the BLM protein in tumour tissues
Positive staining for BLM was observed in tumours of both lymphoid and epithelial origin (Figure 1; Tables 1 and 2). Generally, 5-50% of cells in each tumour sample stained positive for BLM using BFL-103, but in rare cases more than 60% of cells expressed detectable levels of BLM (Tables 1 and 2). There was a good correlation between the expression of BLM and that of 2 independent, well-established markers of proliferation; PCNA and Ki67 (Table 2). In general, the percentage of cells scoring positive for BLM expression was equal to or a little below the equivalent percentage values for PCNA and Ki67.

DISCUSSION
We have conducted a detailed analysis of the distribution and expression of the BLM protein in normal and neoplastic human tissues using a specific monoclonal antibody. We have shown that BLM is expressed in only a limited range of normal tissues, such as in lymphoid tissue, and in the skin and digestive tract, with a pattern of expression indicative of BLM being a marker of cellular proliferation. Consistent with this finding, BLM was shown to be widely expressed in tumours of both lymphoid and epithelial origin. In the range of tumours that we studied, the percentage of cells expressing BLM was very similar to the equivalent values for two established proliferation markers, PCNA and Ki67. The slightly lower percentage values for BLM probably reflects the fact that the intensity of staining using the BFL-103 anti-BLM antibody was consistently somewhat weaker than that seen when using the PC-10 or Ki67 antibodies. BLM is defective in BS, a rare disorder which is associated with several phenotypic abnormalities. Of greatest relevance to the findings of this study, is the fact that BS individuals are cancer prone and, more particularly, succumb at an early age to cancers of all (or nearly all) types, including leukaemias, lymphomas and epithelial tumours (German, 1993(German, , 1995. There is the possibility, therefore, that the BLM gene will be defective in a proportion of tumours that arise sporadically in the general population. Although we have not addressed this question directly, it might have been expected that we would have detected loss of BLM staining in some of the tumour samples if BLM expression was lost in a high percentage of any of the tumour types studied. However, we did not obtain any evidence of loss of BLM expression in these tumour samples and therefore we would tentatively conclude that loss of BLM expression does not occur at a high frequency in any of the tumour types studied here. This does not exclude the possibility that mechanisms of BLM inactivation other than loss of BLM protein expression will be found in a significant number of sporadic cancers in humans.
BLM is much more widely expressed in human tumours than in normal tissues, particularly those of epithelial origin. This presumably reflects the far greater fraction of proliferating cells that exists  in tumours than in normal tissues of the same origin. As with other proliferation markers of this sort there is the possibility that BLM might be considered a suitable target for the development of antitumour strategies aimed at the elimination of proliferating tumour cells. Moreover, measurement of BLM expression could have some utility in cancer prognosis. Clearly, further work is required to obtain additional evidence that there is any validity to these suggestions.
In summary, we have developed a BLM-specific antibody that is suitable for use in immunohistochemical analysis of human tissues, including primary human tumours. We have shown that the expression of BLM is restricted to the proliferating compartment of normal human tissues, but is widely expressed in tumour tissue of both lymphoid and epithelial origin. BLM is apparently localized exclusively to the nucleus in vivo. With the development of BFL-103, we now have the capability to analyse BLM expression more generally in tumour biopsy material to ascertain whether alterations/abnormalities in BLM expression occur at a measurable frequency in human tumours both before and after antitumour therapy.