Differential p53 protein expression in breast cancer fine needle aspirates: the potential for in vivo monitoring

Fine needle aspiration (FNA) biopsy is the least invasive method of sampling breast cancer in vivo and provides material for breast cancer diagnosis. FNA has also been used to examine cellular markers to predict and monitor the effects of therapy. The aim of this study was to assess the accuracy of using FNA material compared with resected cancer for Western blotting studies of the p53 pathway, a key to tumour response to radiotherapy and chemotherapy. Paired samples of breast cancer FNAs collected pre-operatively and post-operatively were compared with tissue samples obtained at the time of surgical resection. Western blots were probed for p53 using the antibodies DO12 and DO1, and for levels of downstream proteins p21/WAF1 and p27. The protein extracted by FNA was sufficient for up to 5 Western blot studies. p53 expression and phosphorylation did not differ significantly pre- and post-operatively, indicating that intra-operative manipulation does not affect p53 expression or downstream activation in breast cancer. However, expression of p53, p21 and p27 varied between individual patients suggesting a range of p53 pathway activation in breast cancer. Immunohistochemistry confirmed that the cancer cells accounted for the protein expression detected on Western blots. FNA yields adequate protein for Western blotting studies and could be used as a method to monitor p53 activity in vivo before and during anti-cancer treatment possibly providing early evidence of tumour response to therapy. © 2001 Cancer Research Campaign  http://www.bjcancer.com

The p53 gene is central to the cellular response to genotoxic damage (Kastan et al, 1991) and chemotherapeutic agents (Linke et al, 1996). A coordinated downstream response from p53 involves activation of a number of target genes containing p53binding sites, such as mdm-2 and p21/WAF1 (Bueso-Ramos et al, 1996;Steele et al, 1998;Jones et al, 1999).
Mutation or overexpression of p53 protein has been observed in up to 52% of primary breast cancers (Ziyaie et al, 2000). Unlike most other epithelial malignancies, in breast cancer abnormalities in p53 protein function rather than specific p53 mutations are more common (Andersen and Borresen, 1995;Ziyaie et al, 2000). p53 status is an important determinant of tumour response to antineoplastic agents (Lowe et al, 1994), with specific mutations of p53 associated with poor response to systemic therapy (Elledge et al, 1995;Aas et al, 1996).
Absolute levels of p53 can be identified using the monoclonal antibody DO12, which recognises a cryptic epitope in the core domain of p53 (Vojtesek et al, 1995). DO1, an anti-p53 antibody, recognises amino acids 20-26 at the n terminus of p53 and is commonly used for immunohistochemical staining for p53 (Stephen et al, 1995).
As a consequence of p53 activation there is increased expression of effector genes including mdm2, p21/WAF1 and p27. Whilst the mdm-2 protein controls the biological activity of p53 and targets it for destruction (Haupt et al, 1997), both p21/WAF1 and p27 act as cyclin-dependent kinase (cdk) inhibitors, preventing phosphorylation of Rb and release of E2F thus interfering with S-phase signalling (Sherr and Roberts, 1999). High expression of p21/WAF1 may be related to short relapse-free survival in breast cancer (Barbareschi et al, 1996) although the evidence is conflicting (Elledge and Albred, 1998), and p21/WAF1 mutations are not common in cancers (Shiohara et al, 1997). p27 is rarely mutated in human primary breast carcinomas and breast cancer cell lines (Sgambato et al, 1997); however high levels have been found in many cell lines even during exponential growth phase, despite its known role as an inhibitor of cell cycle progression (Sgambato et al, 1997).
If it were possible to examine the p53 pathway in vivo in patients, and thus predict the individual tumour response to chemotherapy, patients likely to respond to chemotherapy could proceed with subsequent courses of treatment while others were spared further, ineffective, treatment and directed to alternative therapy. Fine needle aspiration (FNA) is used as a routine diagnostic investigation, furthermore FNA samples obtain representative tumour cell populations from a breast cancer (Brotherick et al, 1995) causing minimal tumour disruption and are both acceptable to patients and yield good material for study. FNAs have been found to be a viable alternative to intraoperative conventional surgical biopsy for biochemical analysis of prognostic factors (Kuner et al, 2000). Cancer cells extracted by FNA have been used for cytology (Pelosi et al, 1994), polymerase chain reaction (Gudlaugsdottir et al, 2000), DNA sequencing analysis (Lavarino et al, 1998), fluorescence in situ hybridisation analysis (Rao et al, 1998) and flow cytometry (Makris et al, 1997). In addition, FNA reliably allows immunocytochemical determination of biological variables in primary breast carcinoma (Nizzoli et al, 2000). The functional significance of key genes in a given cancer may be most accurately determined by detecting the protein. The aims of this study were to assess the feasibility of using fine needle aspiration to examine protein expression of genes from the p53 pathway in breast cancer in vivo at multiple time points and to confirm that the cancer tissue (rather than the normal tissues) accounted for the proteins detected by Western blotting.

Fine needle aspiration and human tissue
Fine needle aspirates (FNA) were obtained by a single operator with a 21G needle and 10 ml syringe, passing the needle through the cancer 10 times with application of 1-2 ml suction (Hartley et al, 1988). 500 µl of chilled phosphate-buffered saline (PBS) was then drawn up, the needle contents expelled into a chilled eppendorf on ice then spun at 5000 g (4˚C) for 5 minutes. The supernatant was aspirated and the pellets snap frozen in liquid nitrogen stored at -70˚C.
FNAs were obtained from 16 consenting patients prior to mastectomy or wide local excision for primary untreated breast cancer and immediately post-operatively from the resected specimen adjacent to blocks used for histological assessment. Postoperative samples were collected no later than 15 minutes following excision. Within 15 minutes of excision, 3 mm 2 tumour biopsies were also excised and snap frozen in liquid nitrogen, then stored with the corresponding aspirates at -70˚C. Local Ethical Committee approval was obtained.

Protein extraction and analysis
Protein was extracted using urea lysis buffer (8.7 ml urea (8 M), 1 ml dithiothreitol (1 M), 50 µl Triton (10%), 50 µl NaCl (5 M), 200 µl Hepes (1 M, pH 7.6). Tumour aspirate pellets, Hela cell pellets (as a positive control for p53, p21 and p27) and snap frozen tumour were homogenised in three times the volume of urea lysis buffer. Samples were left on ice for 15 minutes then spun at 13 000 g (0˚C) for 10 minutes. The supernatant was transferred into prechilled eppendorfs, snap frozen and stored at -70˚C. Protein concentration was determined at 595 λ (Bradford, 1976). Bradford assay showed a 3-fold dilution of protein when the barrel of the syringe was washed through with PBS for the second time. To maximise yield, both washes were combined before being spun down.
The protein extracted from the FNAs and tumour tissue was used to perform repeat gels and Western blots of the samples for confirmation of expression profile comparing pre-operative FNA vs. post-operative FNA vs. pathology block in all 16 patients.

SDS/PAGE gel electrophoresis
SDS/PAGE was carried out using both 10% and 15% polyacrylamide resolving gels and standard Laemlli buffer. The 16 tumour lysates were run on a 10% polyacrylamide gel, before being probed with the appropriate antibody. Arn8 Lysates with χ radiation and DRB were used as the positive control (Blaydes et al, 2000).

Antibodies
Anti-p53 antibodies DO-1 (p53 n-terminus specific) and DO-12 (for absolute levels of p53) supplied by Moravian Biotechnology, Czech Republic, have been described previously (Blaydes and Hupp, 1998;Blaydes et al, 2000). p21 WAF1 (Ab1) was supplied by Oncogene Research Products, Calbiochem and p27 (C-19) antibody from Santacruz Biochemistry. Immunoblotting was carried out using standard methodology. Blots were examined and interpreted independently by 3 of the authors (TRH, AMT, HMB); and there was unanimity for each antibody for all samples.

Immunohistochemistry
Immunohistochemistry was performed on 10 neutral-buffered formalin-fixed, paraffin-embedded sections taken from tissue adjacent to that used for Western blotting. Sections were microwaved for 15 minutes in citric acid (0.1 M, pH 6.0) for antigen retrieval and probed with monoclonal antibodies DO1, WAF1 or p27. A horseradish peroxidase (HRP) detection system using diaminobenzidine (DAB) substrate and cells were counterstained with haematoxylin. Coverslips were mounted using organic mountant (DPX). Slides were examined independently by two of the authors (CP, NK) blinded to the Western blotting data. There was unanimity in the scoring. Immunohistochemical data were cross-tabulated with the Western blotting results.

RESULTS
Protein was extracted successfully from all FNA and cancer samples. Fine needle aspiration yielded up to 350 µg of protein for Western blots.
There was no difference in p53 expression between FNA samples taken pre-operatively and post-operatively on Western blotting (Figure 1; tumour samples 13a and 13b). However, an additional (third) band was observed in the homogenised cancer specimen 13c not seen in 13a or 13b (the FNAs) for one cancer. Matched venous lymphocyte protein from the same patient did not demonstrate any p53 expression with DO12 antibody. For DO1, p21/WAF1 and p27 expression was consistent for the FNAs and tumour sample both in terms of level of expression and species detected.
Repeat gels, Western blots and probing demonstrated consistent expression/absence of expression for each sample and every protein species.

Immunohistochemistry
Sections of the 16 tumours stained by immunohistochemistry (IHC) confirmed cancer cells were responsible for the positive bands on Western blotting. In every case IHC-staining intensity correlated with the Western blotting data (Figure 3). This study demonstrates that FNA can be used to assess protein expression in the p53 pathway using Western blotting. The technique thus provides a means to perform serial sampling over time in vivo and could be used to identify and monitor early tumour response (or failure to respond) at the biochemical level before clinical responses are evident during neo-adjuvant chemotherapy.
No difference was observed in p53, p21 or p27 expression (either in quality or quantity) between samples taken pre-operatively and post-operatively. This suggests that manipulation and potential short-term hypoxia during surgery does not affect p53 induction at the protein level, unlike human oesophageal cancer, where there is a marked effect on p53 induction by hypoxia (Hopwood et al, 1997). The additional species seen in the homogenised tumour (but not in the FNAs), is unlikely to be due to blood contamination. It could however be due to additional p53 species in the normal tissue given the greater proportion of cancer cells in an FNA than in a biopsy (Taxin et al, 1997), or due to early p53 protein degradation. In clinical practice, blood staining of FNAs is not uncommon; our data suggest that blood-stained FNA material is still suitable for p53 protein studies, although absolute protein concentration may not reflect that of the cancer cell population alone.
As expected, individual cancers showed different p53 levels within the panel probed with DO12 ( Figure 2A) and with DO1 (data not shown) as we have previously described (Craig et al, 1999). This highlights the individuality of the cancers and their potentially different responses to treatment. DO1 recognises a site at the n-terminus of p53, which may be cleaved in some tumours, thus explaining why some samples were DO1 negative, but DO12 positive on Western blotting. Although DO1 is commonly used in histopathology to detect p53, DO12 is not amenable to IHC techniques. This emphasises the importance of using different antibodies (and techniques) when trying to unravel the p53 pathway in human breast cancer.
To further investigate the individuality of the p53 pathway in each tumour, the expression of p21/WAF1 and p27 (downstream targets of p53) was also examined by Western blotting and immunohistochemistry in the same set of tumours. These data confirm the complexity of the p53 pathway in breast cancer and suggest that there is another pathway involved in p21 activation possibly controlled by IRF-1 (Tanaka et al, 1996), as, unexpectedly, some samples were p53 negative, and p21 positive. Future in vivo studies could also study the mutational status of p53 in the cancer (Lavarino et al, 1998;Gudlaugsdottir et al, 2000), perhaps using the yeast functional assay (Flaman et al, 1995) which could be performed on FNA samples prior to chemotherapy.
The immunohistochemistry demonstrates concordance with detectable DO1, p21 and p27 expression from Western blots and confirmed that it was cancer cells rather than normal tissue expressing these proteins. However, the Western blots remain useful to identify which protein species predominate and could be used to identify and quantify changes between samples.
Based on this small series, FNA of breast cancers provides reproducible data on protein expression of p53, p21 and p27. FNA has the advantage over core biopsy or surgical biopsy, in that it is well tolerated by patients and can be performed on multiple occasions. Thus it could be used in vivo in the same patient to monitor changes in protein expression during treatment and hence predict and monitor the response to chemotherapy or radiotherapy of primary cancers or metastases in vivo.