Rare germline large rearrangements in the BRCA1/2 genes and eight candidate genes in 472 patients with breast cancer predisposition.

Hereditary breast cancers account for up to 5-10 % of breast cancers and a majority are related to the BRCA1 and BRCA2 genes. However, many families with breast cancer predisposition do not carry any known mutations for BRCA1 and BRCA2 genes. We explored the incidence of rare large rearrangements in the coding, noncoding and flanking regions of BRCA1/2 and in eight other candidate genes--CHEK2, BARD1, ATM, RAD50, RAD51, BRIP1, RAP80 and PALB2. A dedicated zoom-in CGH-array was applied to screen for rearrangements in 472 unrelated French individuals from breast-ovarian cancer families that were being followed in eight French oncogenetic laboratories. No new rearrangement was found neither in the genomic regions of BRCA1/2 nor in candidate genes, except for the CHEK2 and BARD1 genes. Three heterozygous deletions were detected in the 5' and 3' flanking regions of BRCA1. One large deletion introducing a frameshift was identified in the CHEK2 gene in two families and one heterozygous deletion was detected within an intron of BARD1. The study demonstrates the usefulness of CGH-array in routine genetic analysis and, aside from the CHEK2 rearrangements, indicates there is a very low incidence of large rearrangements in BRCA1/2 and in the other eight candidate genes in families already explored for BRCA1/2 mutations. Finally, next-generation sequencing should bring new information about point mutations in intronic and flanking regions and also medium size rearrangements.

ErbB/HER ligands in human breast cancer, and relationships with their receptors, the bio-pathological features and prognosis.

BACKGROUND: The aim of this study is to provide an expression profile of ErbB/HER ligands in breast cancer. We analysed the relationships with their receptors, the bio-pathological features and prognosis. PATIENTS AND METHODS: Epidermal growth factor (EGF), transforming growth factor-alpha (TGFalpha), amphiregulin (AREG), betacellulin (BTC), heparin-binding EGF-like growth factor (HB-EGF), epiregulin (EREG) and neuregulins1-4 (NRG1-4) were quantified in 363 tumours by real-time reverse transcription-polymerase chain reaction using TaqMan probes. RESULTS: Ligands were detected in 80%-96% of the cases, except NRG3 (42%) and EREG (45.5%). At least one ligand was expressed in 304 cases (cut-off: upper quartile). Almost all combinations of receptor and ligand co-expressions were observed, but TGFalpha is preferentially expressed in tumours co-expressing EGFR/HER3, NRG3 in those co-expressing EGFR/HER4, AREG and EREG in those co-expressing HER2/HER4. EGF and AREG were associated with estradiol receptors, small tumour size, low histoprognostic grading, high HER4 levels. TGFalpha, HB-EGF and NRG2 were negatively related to these parameters. In Cox univariate analyses, EGF was a prognostic factor. CONCLUSION: Our study demonstrates that (i) ErbB/HER ligands, including BTC and EREG, are expressed in most breast cancers; and (ii) TGFalpha, HB-EGF and NRG2 high expressions are related to the biological aggressiveness of the tumours.

Measurement of mRNA of 11 biomarkers by RT-PCR to detect lymph node involvement in cervical cancer.

Lymph node metastases are a major prognostic factor in cervical carcinomas. The aim of this study was to characterize the expression of 11 markers in cervical tumors and negative lymph nodes and to determine which ones could be helpful for improving the specificity of molecular diagnosis of nodal involvement. Using TaqMan RT-PCR, we studied the expression of CK19, MUC1, HER1-HER4, VEGF, VEGF-C, uPA, MMP9, and PRAD1 in uterine cervical tumors and in histologically nonmetastatic lymph nodes of 8 patients diagnosed with locally advanced cervical cancer. We observed that CK19, MUC1, HER1-HER3, uPA, and VEGF had a significantly higher expression in cervical tumors than in the negative nodes, whereas VEGF-C expression level was higher in the negative nodes than in the tumors. PRAD1 harbored similar expression levels in the tumors and in the negative nodes. Interestingly, 1 of the 4 patients who presented a clinical recurrence, showed elevated HER1, HER2, uPA, and VEGF in the histologically negative nodes. Our results suggest that CK19, MUC1, HER1-3, uPA, and VEGF are biomarkers that have a higher expression in tumoral cervical tissues compared with the negative lymph nodes and could be useful to diagnose nodal involvement in uterine cervical carcinoma. Our results should encourage us in continue to investigate a greater number of patients, including patients with histologically involved nodes.

Real-time reverse-transcription PCR to quantify a panel of 19 genes in breast cancer: relationships with sentinel lymph node invasion.

At the Centre Oscar Lambret, the anticancer centre of the North of France, sentinel lymph node (SLN) procedures are routinely performed for localized (T0-T1, N0, M0) breast carcinoma without any previous treatment, in order to prevent the deleterious effects of axillary lymph node dissection. The present study was undertaken to assess if the expression in the tumor of a panel of 19 genes would allow to predict histological SLN involvement. We looked at cytokeratin 19 (CK19), mucin-1 (MUC1), mammaglobin (MGB1), cyclin D1 (CCND1), the four members of the HER/ErbB growth factor receptor family (EGFR, HER2-4), insulin-like growth factor-1 receptor (IGF-1R), estradiol receptors (ERalpha, ERbeta), progesterone receptor (PR), vascular endothelial growth factors (VEGF, VEGF-C), urokinase-like plasminogen activator (uPA), matrix metalloproteinases 2 and 9 (MMP2, MMP9), ets-related transcription factor ERM, and E-cadherin (CDH1). Their expression was quantified by real-time RT-PCR in 134 breast cancer samples and the relationships with SLN metastases were analyzed. A slight increase (35-40%) in CK19 and HER3 expression was observed in the tumors of patients with SLN metastases compared to those of patients without metastases, even if neither CK19 expression nor HER3 expression allowed to distinguish patients with micrometastases from patients with macrometastases. We conclude that the tumoral expression of biological parameters involved in cell proliferation or playing a critical role in the metastatic process, including tumor invasion and angiogenesis, is not strongly associated with SLN metastases.

Prognostic significance of insulin-like growth factor 1 receptors in human breast cancer.

Overall survival (OS) and relapse free survival (RFS) were studied in 297 patients according to the presence of insulin-like growth factor 1 receptors (IGF1-R). All the patients were surgically treated for locoregional disease in the same institution from January 1986. The median duration of follow-up was 40 months. RFS was better in patients with IGF1-R in their tumors as assessed by actuarial survival (P = 0.014) as well as Cox analysis (P = 0.016). OS was better in IGF1-R positive tumors studied by actuarial (P = 0.007) as well as Cox analysis (P = 0.010). By Cox analysis the other prognostic factors on RFS were estrogen receptor (P = 0.002), progesterone receptor (P = 0.002), axillary node metastases (P = 0.032), histoprognostic grading (GHP) according to the standard of Scarff and Bloom (P = 0.004), and tumor diameter (P = 0.019). The other prognostic factors on OS (Cox analysis) were estrogen receptor (P = 0.001), axillary node metastases (P = 0.010), GHP (P = 0.009), progesterone receptor (P = 0.012), and tumor diameter (P = 0.007). When combining IGF1-R, GHP, and axillary node metastases, it appeared that IGF1-R, GHP, and axillary node metastases had independent prognostic significance. In this prospective study IGF1-R had a prognostic significance on RFS as well as on OS studied by actuarial as well as Cox analysis.

Prognostic significance of prolactin receptors in human breast cancer.

Overall survival and relapse free survival (RFS) were studied in 547 patients according to the presence of prolactin receptors (PRL-R), either free or total (after 3 M MgCl2 desaturation). All these patients were surgically treated for locoregional disease in the same institution between 1978 and 1984. In actuarial survival studies, RFS was higher in total PRL-R positive patients in the whole population (P less than 0.02). When the population was divided into two groups, according to either the presence or the absence of node metastasis or the presence or absence of estradiol receptor, the higher RFS was restricted to node positive (P less than 0.001) and to estradiol receptor positive patients (P less than 0.01). The Cox analysis on RFS showed that free PRL-R alone was a significant prognostic factor in estradiol receptor positive patients; total PRL-R alone was never significant; when considered together with steroid receptors, free as well as total PRL-Rs were significant prognostic factors in some subgroups of patients.

Insulin-like growth factor 1 receptors in human breast cancer and their relation to estradiol and progesterone receptors.

Insulin-like growth factor 1 (IGF1) binding sites were characterized in breast cancer. We demonstrate the presence of one high affinity binding site. Chemical cross-linking of 125I-IGF1 to breast cancer membranes in reducing condition and sodium dodecyl sulfate-polyacrylamide gel electrophoresis revealed one band with an apparent molecular weight of 130,000. The specificity of the binding was studied. IGF2 was a good competitor whereas insulin competed with a potency lower than 1/100 that of IGF1. This IGF1 binding corresponded to the previously described type 1 IGF receptor (IGF1-R). IGF1-R was determined in 76 human breast cancer biopsies. Ninety-three % of the tumors were positive. The specific binding range was 0-16.4%; the geometric IGF1-R mean level was 3.9%. There was a relation (chi 2 test) between IGF1-R and progesterone receptor positivity rates (P = 0.002). The IGF1-R concentrations were correlated (Spearman test) with those of estradiol receptor (P = 0.0018) and progesterone receptor (P = 0.0011). A positive linear correlation existed between IGF1-R and estradiol receptor (P = 0.006) and between IGF1-R and progesterone receptor (P = 0.003). Our demonstration of the presence of IGF1-R in human breast cancer biopsies suggests that IGF1, acting either via the endocrine, paracrine, or autocrine pathways, could stimulate tumor growth.

Nuclear estradiol-binding sites in human breast cancer.

The binding of estradiol to nuclear fractions extracted from human breast carcinomatous tissue was demonstrated. The material, which was extracted with KCl, had an approximate molecular weight of 37,000 and bound estradiol with both high and low affinity (Kd congruent to 1 nM, type A receptors; Kd congruent to 30 nM, type B receptors) as calculated according to the method of Scatchard. Competition studies indicated that both components were specific for estradiol, and among the 134 tumors studied the receptors were found to be linked in almost all cases. Thirty-six % of the tumors were nuclear receptor positive. Cytoplasmic estradiol and progesterone receptors were also measured. Among the cytoplasmic tumors positive for cytoplasmic and progesterone receptors, 37% were devoid of both types of nuclear receptors; this may explain the failure of endocrine therapy in some cases. The determination of nuclear binding sites in human breast tumors appeared to be an interesting criterion for the assessment of estradiol-dependent cell growth.

Multiplex reverse transcription polymerase chain reaction assessment of sialyltransferase expression in human breast cancer.

Increased sialylation, especially involving the Sialyl-Lewisa and Sialyl-Lewisx determinants, has been reported in breast cancer. A multiplex reverse transcription-PCR method was used here to determine the expression of five sialyltransferases (ST3Gal III, ST6Gal I, ST3Gal IV, ST3Gal I, and ST3Gal II) in 49 patients surgically treated for locoregional breast cancer. We assessed the relationship between these expressions and clinical, pathological, and biological features. The most expressed sialyltransferase was ST3Gal 1II, which is involved in Sialyl-Lewisa synthesis. ST3Gal III expression was positively correlated to ST6Gal I and ST3Gal IV expressions, to tumor size, and to the number of involved axillary nodes. Patients with high ST3Gal III expression had a shorter overall survival. High ST6Gal I expression was associated with histoprognostic grade III. ST6Gal I expression was negatively correlated to expression of progesterone receptor. In conclusion, high ST3Gal III and ST6Gal I expressions in human breast tumors are associated with poor prognosis markers.

Expression of nerve growth factor receptors and their prognostic value in human breast cancer.

Nerve growth factor (NGF) has been shown recently to be mitogenic for human breast cancer cells. In the present study, we have assayed the expression of NGF receptors (NGFRs: TrkA and p75) mRNAs in 363 human primary breast cancers, using real-time quantitative reverse transcription-PCR. NGFRs were found in all of the tumor biopsies. TrkA and p75 were positively correlated and were respectively associated with the histoprognostic grading and the tumor type. NGFRs were both related to progesterone receptors. In univariate analyses, TrkA (>upper quartile) was associated with longer overall survival. Histoprognostic grading, tumor size, node involvement, and steroid receptors were also prognostic factors. In Cox multivariate analyses, TrkA was not a prognostic parameter. This study demonstrates the expression of NGFRs in breast cancer and points out that patients with high levels of TrkA have a more favorable overall survival prognosis.

Proteomic analysis reveals that 14-3-3sigma is down-regulated in human breast cancer cells.

The class of molecular chaperones known as 14-3-3 is involved in the control of cellular growth by virtue of its apparent regulation of various signaling pathways, including the Raf/mitogen-activated protein kinase pathway. In breast cancer cells, the sigma form of 14-3-3 has been shown to interact with cyclin-dependent kinases and to control the rate of entry into mitosis. To test for a direct role for 14-3-3 in breast epithelial cell neoplasia, we have quantitated 14-3-3 protein levels using a proteomic approach based on two-dimensional electrophoresis and matrix-assisted laser desorption/ionization mass spectrometry (MALDI-TOF). We show here that 14-3-3sigma protein is strongly down-regulated in the prototypic breast cancer cell lines MCF-7 and MDA-MB-231 and in primary breast carcinomas as compared with normal breast epithelial cells. In contrast, levels of the alpha, beta, delta, or zeta isoforms of 14-3-3 were the same in both normal and transformed cells. The data support the idea that 14-3-3sigma is involved in the neoplastic transition of breast epithelial cells by virtue of its role as a tumor suppressor; as such, it may constitute a robust marker with clinical efficacy for this pathology.

Truncation at conserved terminal regions of BRCA1 protein is associated with highly proliferating hereditary breast cancers.

The existence of two subgroups of BRCA1-associated breast cancer (BC) families has been recently posited: the first with highly proliferating tumors, and the second composed of cases with a low proliferation rate. Our aim was to test whether the proliferation rate of BRCA1-associated breast cancers was affected by the site of the germ line mutation in the BRCA1 gene. We analyzed the distribution of the mitotic index, a histoprognostic grade component shown to segregate in families, matching for germ line mutation location in a series of 28 breast cancers from 20 kindreds. We observed a prevalence of highly proliferating tumors when the mutation occurs in the two terminal conserved domains of the BRCA1 protein, ie., in the amino and carboxyl termini (P = 0.0024). Our data provide evidence for a genotype-phenotype correlation and along with their strong conservation during evolution argue for the importance of these two regions in the control of mammary cell growth.

Mutations at BRCA1: the medullary breast carcinoma revisited.

BRCA1-associated breast cancers (BRCA1-BCs) frequently harbor a high histoprognostic grade, p53 alterations, and estrogen receptor negativity. Although these parameters predict a poor outlook, the overall survival in BRCA1-BCs is equivalent to or even better than that in sporadic cases. These features are reminiscent of what is observed for breast carcinoma of the medullary type, a high-grade tumor with a particular favorable course. To explore a possible relationship between this phenotype and BRCA1 mutations, we first compared 32 BRCA1-BCs and 200 consecutive cases of breast cancer without familial history for the prevalence of typical medullary breast carcinoma (TMC) using the criteria given by Ridolfi et al. [R. Ridolfi et al, Cancer (Phila.), 40: 1365-1385, 1977]. Second, we searched for BRCA1 mutations in a set of 18 cases of TMC, using denaturing gradient gel electrophoresis and Cleavase fragment length polymorphism scanning. Six of 32 (19%) BRCA1-BCs were of the TMC type, compared to 0 of 200 controls (P < 0.0001). Among the 18 TMCs, 2 BRCA1 nonsense mutations were found. This corresponds to almost 7 times the contribution of BRCA1 mutations in the general population. Two additional missense mutations were identified. Together, these results suggest that, although TMC and BRCA1-BCs are not strictly coincidental, an important connection between the two populations does exist.

Analysis of p53 antibodies in patients with various cancers define B-cell epitopes of human p53: distribution on primary structure and exposure on protein surface.

p53 antibodies have been found in sera of patients with breast and lung carcinomas and in children with B-lymphomas. We report here the presence of p53 antibodies in sera of patients with 11 different types of cancer. The frequency of seropositives for p53 varied among the different types of cancer, but a correlation with the frequency of p53 gene alteration was established. Using a powerful peptide enzyme-linked immunosorbent assay, we demonstrated that the immune response of patients with p53 antibodies was restricted to a small subset of peptides localized in the amino and carboxy termini of p53, whatever the type of cancer. Given the similarities of the patterns of immune responses in patients with p53 antibodies and animals hyperimmunized with human p53, we propose that the p53 humoral response is the result of a self-immunization process which is itself the consequence of p53 protein accumulation in tumor cells.

Detection of BCR-ABL transcripts in chronic myeloid leukemia (CML) using a 'real time' quantitative RT-PCR assay.

Quantitative competitive RT-PCR techniques have been developed to detect BCR-ABL fusion transcripts in CML but they are hardly reproducible. In this work, we have developed BCR-ABL quantification by real time RT-PCR using the ABI PRISM 7700 (Perkin Elmer), a new technique which allows simple and rapid quantification of a target sequence during the extension phase of PCR amplifications. A fluorogenic probe labeled with both a reporter dye at the 5' end and a quencher-dye at the 3' end hybridizes to the target sequence on the third exon of the ABL gene. The exonuclease activity of the Taq DNA polymerase cleaves the probe and releases the reporter dye, resulting in an increase in the fluorescence signal. The absolute copy number of the target sequence (BCR-ABL) or a control gene (ABL) in an unknown sample can then be calculated using a calibration curve prepared from a set of BCR-ABL RNA standards, and results are expressed as a BCR-ABL/ABL ratio. In our hands, the sensitivity of a serial dilution of total RNA from a positive cell line (K562) in a negative cell line (HL60) was 10(-4). Fifteen CML patients in cytogenetic CR, including 11 allografted patients, two autografted patients and two patients treated by IFN were studied sequentially by this new real time quantitative RT-PCR technique in parallel with conventional qualitative two round nested RT-PCR. The two autografted patients showed high BCR-ABL/ABL ratio in all samples. The two patients treated by IFN showed a progressive decrease in the ratio. In the 11 allografted patients, four were sequentially studied 2 years or more after allo-BMT, and all ratios were below 10(-4). The four patients remained in clinical and cytogenetic CR. The seven other allografted patients were studied immediately after the procedure. Three of them showed a progressive decrease in the BCR-ABL/ABL ratio which reached 10(-4) 7 months after allo-BMT. The three patients remained in hematologic and cytogenetic CR. The remaining four allografted patients had progressive increase of BCR-ABL ratio; three developed cytogenetic relapse 9, 11, 28 months after allo-BMT, and the last patient remained in cytogenetic CR in the bone marrow but developed granulocytic sarcoma. Results of real-time quantitative RT-PCR were in agreement with those of qualitative two round nested PCR. However, evolution changes in the results of real-time quantitative RT-PCR often preceded those of the conventional technique: a decrease of the BCR-ABL/ABL ratio preceded progression from first round to second round positivity and then negativity with the classical technique; conversely, an increase in the ratio preceded evolution with the classical technique. Thus, real-time quantitative RT-PCR may show better correlation with clinical and cytogenetic evolution than conventional qualitative techniques and may help in making early therapeutic decisions in CML, especially after molecular relapse.

Prognostic value of the type I growth factor receptors in a large series of human primary breast cancers quantified with a real-time reverse transcription-polymerase chain reaction assay.

We measured the expression of the type I growth factor receptor gene family [epidermal growth factor receptor (EGFR), c-erbB-2, c-erbB-3 and c-erbB-4] in a series of 365 unselected primary breast cancers. The expression was quantified with a real-time one-step reverse transcriptase-PCR (RT-PCR) assay, based upon the 5' nuclease activity of the Taq polymerase and using an Abi Prism 7700 Sequence Detector System (Perkin-Elmer, Courtaboeuf, France). c-erbB-3 and c-erbB-4 were positively correlated to each other (Spearman test) and negatively correlated to EGFR. EGFR and c-erbB-2 were inversely correlated to the presence of estradiol receptors (ER) and progesterone receptors (PgR), and positively correlated to the histoprognostic grading (HPG). Conversely, c-erbB-3 and c-erbB-4 were positively correlated to the presence of ER and PgR, and inversely correlated to the grading HPG. EGFR was inversely related (chi2 test) to the presence of ER and PgR, and positively associated with HPG. In contrast, both c-erbB-3 and c-erbB-4 were inversely related to HPG, and positively associated with the presence of ER and PgR. The expression level of EGFR and c-erbB-2 was significantly higher in ER- and PgR-negative tumors compared with ER- and PgR-positive tumors (Student's t test), and in tumors with higher grade compared with tumors with lower grade. The expression level of c-erbB-3 and c-erbB-4 was significantly higher in ER- and PgR-positive tumors compared with ER- and PgR-negative tumors and in tumors with lower grade compared with tumors with higher grade. In overall survival studies, Cox univariate analyses showed prognostic values of EGFR [> or = median; P = 0.026; risk ratio (RR), 1.6], c-erbB-3 (> or = median; P = 0.0093; RR, 0.58), c-erbB-4 (> or = median; P = 0.0024; RR, 0.52), HPG, node involvement, tumor diameter, ER, and PgR. In Cox multivariate analyses, tumor diameter, ER, and PgR had a prognostic value. In relapse-free survival studies, univariate analyses demonstrated prognostic values of tumor diameter, node involvement, and c-erbB-4 (P = 0.015; RR, 0.65). These three parameters maintained their prognostic value in multivariate analyses (c-erbB-4, P = 0.035; RR, 0.67). This study confirms that EGFR expression and c-erbB-2 expression are markers of tumor aggressiveness in breast cancer. Conversely, we demonstrate that c-erbB-3 and c-erbB-4 elevated expressions are associated with a better prognosis.

Prognostic value of p53 and urokinase-type plasminogen activator in node-negative human breast cancers.

We measured the levels of p53 and urokinase-type plasminogen activator (uPA) in 634 tumor tissues from 634 different node-negative primary breast cancer patients who underwent locoregional surgery in the Center Oscar Lambret between July 1989 and September 1994. p53 and uPA were assayed using commercially available kits in cytosols prepared for estradiol receptor (ER) and progesterone receptor (PgR) assays. The optimum clinical thresholds were chosen for prognostic studies: 4 ng/ml for p53 and 0.5 ng/ml for uPA. p53 was elevated in 13.7% of the tumors, and uPA was elevated in 27.5% of the tumors; they were negatively related (chi 2 test) to ER and PgR and positively related to histoprognostic grading (HPG) and tumor diameter. uPA was negatively correlated to ER and PgR, and p53 and uPA were positively correlated to each other (P = 0.0001; Spearman test). In the prognostic studies, the 316 patients who did not receive adjuvant chemotherapy were included to avoid treatment interference; this number corresponds to all of the patients operated on between 1989 and 1992. The mean duration of follow-up of living patients was 4 years. In overall survival studies, Cox univariate analyses demonstrated a prognostic value of p53 (P = 0.011; risk ratio, 1.59), uPA (P = 0.038; risk ratio, 2.32), PgR, HPG, and tumor diameter. In Cox multivariate analyses, only HPG had a statistically significant prognostic value. In relapse-free survival studies, univariate analyses demonstrated prognostic values of uPA (P = 0.0011) and of age, and both parameters retained their prognostic value in multivariate analyses (uPA: P = 0.0004). This study demonstrates not only that p53 and uPA have prognostic value but also that these two parameters are linked to other classical clinical, histological, or biological prognostic parameters, as well as to each other. Moreover, because uPA is of prognostic value in multivariate relapse-free survival studies, uPA is an important prognostic factor in node-negative breast cancer patients.

Prognostic value of circulating soluble E-selectin concentrations in node-negative breast cancer patients.

Several studies have suggested that endothelial cells participate in tumor development. Soluble E-selectin (sE-selectin) is specifically released by activated endothelial cells, and its serum concentration can be considered a marker of endothelial activation. In this study, we assessed the prognostic value of sE-selectin concentrations in node-negative breast cancer patients. Serum sE-selectin concentrations were measured by an ELISA method prior to surgery in 456 node-negative breast cancer patients. We analyzed also tumor size (TS), histoprognostic grading, and steroid hormone receptor status. The mean sE-selectin concentration was 24.9 +/- 15.0 ng/ml. The sE-selectin concentrations were mildly correlated with the TS but not with the other factors. For prognostic analyses, the median follow-up duration was 7.5 years. The cutoff sE-selectin concentration used was 40 ng/ml. In overall survival studies, univariate analyses demonstrated a prognostic value of sE-selectin, TS, and histoprognostic grading, and multivariate analyses demonstrated a prognostic value of sE-selectin and TS. For disease-free survival, univariate and multivariate analyses demonstrated a prognostic value of sE-selectin and TS. sE-selectin concentration is an easily measurable and strong prognostic factor in node-negative breast cancer patients. These results provide further evidence for the role of adhesion molecules expression by endothelial cells in tumor progression.

The relationship between concentrations of circulating soluble E-selectin and clinical, pathological, and biological features in patients with breast cancer.

Increasing evidence suggests that E-selectin contributes to tumor growth and metastasis. E-selectin may increase tumoral angiogenesis and the adhesion of tumoral cells to endothelial cells at distant sites. The aim of this study was to assess the relationship between concentrations of circulating soluble E-selectin (sE-selectin) and clinical, pathological, and biological features in patients with breast cancer (BC). Concentrations of sE-selectin were analyzed by an ELISA method in sera from 113 patients with metastatic BC, 30 patients with primary inflammatory BC, 105 patients with primary noninflammatory BC, and 42 healthy controls. These concentrations were analyzed in terms of the clinical and pathological features of the tumors as well as in terms of the concentrations of serum inflammatory parameters (erythrocyte sedimentation rate, C reactive protein, interleukin 1 beta, and tumor necrosis factor alpha), the response to chemotherapy or hormone therapy, and the survival duration. Tumoral angiogenesis was also assessed in 68 patients with primary noninflammatory BC who had had primary surgery. The mean concentration of sE-selectin in the metastatic BC group was significantly higher than the mean concentration found in the healthy control group (33.5 versus 21.8 ng/ml; P < 0.01). In metastatic BC, the mean concentration of sE-selectin was significantly higher in patients with liver metastasis than in patients without liver metastasis (55.3 versus 26.0 ng/ml; P < 10(-5). The univariate analysis showed that high concentrations of sE-selectin were associated with reduced overall survival (P < 0.05), but this probably reflected the association between high concentrations of sE-selectin and liver metastasis. In patients with primary noninflammatory BC, a negative correlation was found between sE-selectin concentrations and the tumoral microvessel count (r = -0.47; P = 10(-4). In patients with primary inflammatory or noninflammatory BC, no correlation was found between concentrations of sE-selectin and tumor size, lymph node involvement, response to chemotherapy or hormone therapy, and survival. No correlation was found between the concentrations of sE-selectin and serum inflammatory parameters in any of the patient groups. This study suggests that in patients with metastatic BC, levels of sE-selectin are higher in the presence of liver metastasis. In patients with primary BC, high concentrations seem to be associated with reduced tumoral angiogenesis. Although several studies have previously demonstrated that the expression of cell surface E-selectin enhances the metastatic process, the shedding of sE-selectin in circulation may be considered a mechanism of inhibition of tumor progression.

Basic fibroblast growth factor receptors and their prognostic value in human breast cancer.

We performed a saturation binding study with 125I-labeled FGF (fibroblast growth factor)-2 in a nonselected series of 250 human primary breast cancers. Two hundred twenty-five breast cancer biopsies possessed bFGFR (basic FGF receptor). The median dissociation constant was 0.35 nM (range, 0.014-1.9), and the median concentration was 1126 fmol/mg protein (range, 49-7328). FGFR-1 was localized, using a specific monoclonal antibody, in cancerous cells and in epithelial cells in normal breast or in benign tumors. In all of the tissues studied, light stromal cell staining was also observed. Thus, the localization of FGFR-1 in carcinoma cells supports the hypothesis that an important part of FGF-2 binding reflects binding to FGFR-1. bFGFR concentrations were positively correlated to estrogen receptor and progesterone receptor levels. Cox univariate analyses showed that the bFGFR (> or = upper quartile) was associated to longer relapse-free survival [P = 0.004; RR (risk ratio), 0.46] and overall survival (P = 0.001; RR, 0.35); age, estrogen receptor levels, progesterone receptor levels, node involvement, tumor diameter, and histoprognostic grading were prognostic, also. In Cox multivariate analyses, only the bFGFR, age, node involvement, and histoprognostic grading were prognostic factors; the bFGFR was associated with longer relapse-free survival (P = 0.03; RR, 0.4) and overall survival (P = 0.009; RR, 0.3). The present study confirms that FGF could be an important regulator of human breast cancer growth and that patients with a high level of bFGFR had a better prognosis.