Abstract P4-15-07: MicroRNA signature of breast cancer brain metastasis for novel targeted therapies

Rationale:

MicroRNAs (miRNAs) are recently described novel small non-coding endogenous RNAs which critically regulate cancer progression, invasion and metastasis. Altered expression of miRNAs has been proposed to control the invasive capability of cancer cells. Modulating altered miRNAs is a novel approach for targeted therapy for cancer. Breast cancer is the most common cancer among women in the US. Breast cancer-related mortality and morbidity are primarily due to metastatic disease especially brain metastasis which also impacts patients’ quality of life. Molecular mechanisms of brain metastasis of breast cancer are largely unknown.

Study Objectives:

Identifying molecular miRNA signatures in breast cancers which metastasize to the brain can assist in designing novel targeted therapies to prevent and eliminate brain metastasis.

Design:

RNA was extracted from formalin fixed paraffin embedded tumor tissues from a cohort of breast cancer patients with brain metastasis, and age, stage and follow-up matched breast cancer cases without brain metastasis. miRNA expression profiling was done using EQIXON microarray. Quantitative real-time PCR (qRT-PCR) was used to validate abnormal expression of miRNAs. Data was statistically analyzed using Kruskal-Wallis test to determine the clinical significance of the findings.

Results:

Over 2000 unique miRNA sequences were profiled in each case. The upregulated miRNAs included miR-3927-5p, miR-216b and miR-10b and the down regulated miRNAs were miR-4280, miR-3174 and miR-1244 and miR20b. Validation of the deregulated miRNAs by q RT-PCR showed statistically significant difference in the expression levels in the breast cancers with brain metastasis compared to breast cancers without brain metastasis (Kruskal-Wallis test; p < 0.001).

Conclusions:

A unique set of miRNAs were deregulated in breast cancers with brain metastasis. This panel of altered miRNAs have clinical relevance since they can be silenced by antagomirs (chemically modified anti-miRNA oligonucleotides) to treat and prevent brain metastasis. These molecular microRNA signature-based novel targeted prophylactic and therapeutic strategies could improve patients’ quality of life and overall survival in the era of personalized and precision medicine.

Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P4-15-07.

P53-independent induction of p21(waf1) pathway is preserved during tumor progression

The p21(WAF1) gene encodes a cyclin-dependent kinase inhibitor and plays an important role in controlling the cell cycle. Its expression can be induced through wild-type p53-dependent or -independent pathways. Since the p53-dependent pathway is disrupted in more than 50% of human tumors, we wondered whether the p53-independent pathway is also altered during tumor progression. In the present study, we have determined p21(WAF1) induction by mitogenic stimuli, which is known to be a p53-independent process. p21(WAF1) is induced by mitogenic stimuli in all cell lines tested regardless of the status of p53, i.e. wild-type, wild-type inactivated by SV40T or mutant, and the transformation of cells from immortal to tumorigenic and to metastatic. These results indicate that the p53-independent induction of p21(WAF1) pathway is preserved during tumor progression.

Dual blockade of epidermal growth factor receptor (EGFR) and cyclooxygenase 2 (COX 2) may be dependent upon the EGFR mutational status in non-small cell lung cancer (NSCLC) cell lines

7170

Background: Recent data has shown that the benefit of EGFR TKIs (tyrosine kinase inhibitors) in NSCLC is in patients with tumors that have EGFR mutations and /or gene amplification. Even among these patients, the median survival with EGFR TKIs is only 22 months. Pre-clinical data has shown that dual blockade of EGFR and COX 2 pathways may be beneficial. A recent clinical trial conducted at our center suggested that this combination primarily impacted NSCLC tumors that have the EGFR biomarkers. We therefore hypothesized that the EGFR biomarker status determines the outcome of dual blockade of these pathways in NSCLC. Methods: Three different cell lines with varying biomarker status and sensitivities to EGFR TKIs were used- H3255 (L858R; gene amplified), H1650 (del E746-A750; gene amplified), H1781 (wild type). Cells were treated with erlotinib, EGFR TKI- 10nM-100nM or celecoxib, COX 2 inhibitor- 5 μM and with the combination of two drugs. Cell survival was determined by a standard MTT assay and apoptosis was measured by ELISA method. Western blot analysis was conducted to assess COX 2, EGFR and pAkt levels. Results: Celecoxib by itself had no effects on any of the cell lines. Erlotinib showed a concentration dependent growth inhibition of both H3255 (IC50–41.72nM; surviving fraction at 50nM was 52%) and H1650 (IC50 > 100nM; Surviving fraction at 50nM was 70%) but had no effect on H1781. Celecoxib added to erlotinib significantly enhanced the growth inhibition of H3255 (p = 0.001) and H1650 (p = 0.014) as well as apoptosis (H3255- p = 0.016; H1650- p = 0.011) at all concentrations of erlotinib but had no effect in H1781 cells. In western Blot analysis the combination significantly reduced levels of COX-2, EGFR and pAkt compared to baseline and either agent alone in H3255 cells. Conclusions: The addition of celecoxib to erlotinib has a differential effect in NSCLC cell lines based on their EGFR biomarker status. This beneficial effect of celecoxib addition may be through improved inhibition of each pathway.

[Table: see text]

Expression of inflammatory modulator COX-2 in pancreatic ductal adenocarcinoma and its relationship to pathologic and clinical parameters

Despite the exceedingly poor prognosis of pancreatic cancer, it is often histologically well to moderately differentiated. The apparent resistance to conventional therapeutic modalities is poorly understood and may be related to the molecules involved in its progression or its propensity for perineurial invasion. Cyclooxygenase-2 (COX-2) is an inducible enzyme homologous to COX-1 that is responsible for production of prostaglandins at sites of inflammation. It is activated by a variety of growth factors and tumor promoters, and it has been implicated in cancer progression. It may also have a role in the resistance to therapy. Anti-COX-2 agents have been documented to have antitumor activity, and some are now being tested in the therapy for various cancers, including those of the pancreas. Experience regarding the rate of COX-2 expression in pancreatic cancer and its relationship to the clinical and biologic parameters is very limited. In this study, immunohistochemical stains for COX-2 have been performed on 120 cases of pancreatic ductal adenocarcinoma. The stains were scored according to the percentage (0: no staining, 1: < 10%, 2: 10-50%, and 3: >50% of the cells staining) and intensity (0 for no staining, 1 for mild staining, and 2 for dark staining) of staining. Based on the combined score for each case, they were divided into low expressors (percentage and intensity < or =1) and high expressors (percentage or intensity >1). In addition to global scoring for each case, the glandular and solid (poorly differentiated) components, when present, were scored separately. The global scores were correlated with clinical and biologic parameters. Seventy-four percent of the cases exhibited expression of COX-2 and 53% were high expressors. No significant association was observed when comparing the global COX-2 expression to survival, tumor size, stage, and vascular invasion. Increased perineural invasion was found to be significantly associated with COX-2 expression (p < 0.05). Increased expression was also more common in the glandular component as compared with the solid component of the tumors (68% versus 35%, p < 0.05). Of the 34 patients who received radiotherapy, 9 were low expressor (median survival 19.5 months) and 25 were high expressors (median survival 14 months). The difference in survival was not statistically significant.

Colloid (mucinous noncystic) carcinoma of the pancreas

In the past, colloid (mucinous noncystic) carcinoma (CC) of the pancreas had been included under the category of ordinary ductal adenocarcinoma, a tumor with a dismal prognosis, or was frequently misdiagnosed as mucinous cystadenocarcinoma. The clinicopathologic features of CC have not yet been well characterized, because most cases on record have been parts of studies on either mucinous cystic neoplasms (MCN) or intraductal papillary mucinous neoplasms (IPMN), with which colloid carcinomas are frequently associated. To determine the clinicopathologic characteristics of CC, 17 pancreatic tumors composed predominantly (>80%) of CC (defined as nodular extracellular mucin lakes with scanty malignant epithelial cells) and in which the invasive carcinoma measured larger than 1 cm were studied. Ten of these were originally classified as mucinous ductal adenocarcinoma and four as mucinous cystadenocarcinoma. The mean age of the patients was 61 years; 9 were men and 8 were women. The mean size of the CC was 5.3 cm (range, 1.2-16 cm). In more than half of the patients, CC represented the invasive component of an IPMN (in nine cases) or MCN (in one case). The tumors were composed of well-defined pools of mucin with sparse malignant cells in various patterns of distribution. Signet-ring cells floating in the mucin (but not as individual cells infiltrating stroma, a characteristic finding of signet-ring cell adenocarcinomas) were commonly identified and were prominent in five cases. Perineurial invasion was noted in six cases and regional lymph node metastases in eight. Mutation in codon 12 of the k-ras gene was detected in only 4 of 12 cases studied and p53 mutation in 2 of 9. Immunohistochemical and histochemical mucin stains suggested luminalization of the basal aspects of the cells. Five-year survival was 57%. At an overall mean follow up of 57 months, 10 patients were alive with no evidence of disease (median, 79 mos), including four with lymph node metastasis, three others with perineurial invasion, and another with vascular invasion. Four patients died of disease (18, 18, 25, and 26 mos), and three died of thromboembolism (with persistent disease) at 2, 5, 10 months. All seven patients who died with or of tumor had undergone incisional biopsy of the tumor either before the operation or intraoperatively, whereas none of the patients who were alive had incisional biopsy. When compared with 82 cases of resectable ordinary ductal adenocarcinoma on whom follow-up and staging information was complete, it was found that the patients with CC present with larger tumors (p = 0.03) but lower stage (p = 0.01). The prognosis of CC is significantly better: 2-year and 5-year survival are 70% versus 28% and 57% versus 12%, respectively (p = 0.001). In conclusion, pancreatic CC may occur with or without an identifiable IPMN and MCN component, and should be distinguished from mucinous cystadenocarcinoma, ordinary ductal adenocarcinoma, and signet-ring cell adenocarcinoma. CC of the pancreas is associated with a significantly better prognosis than ordinary ductal adenocarcinoma. In addition to its distinctive morphologic and clinical characteristics, CC of the pancreas also appears to have a low incidence of mutation in codon 12 of the k-ras gene. In cases with a clinical suspicion of colloid carcinoma, the possibility that an incisional biopsy may contribute to thromboembolic complications or even dissemination of the tumor may need to be considered. The luminalization of the basal aspects of the tumor cells may be the cause of stromal mucin accumulation that characterizes colloid carcinoma and may act as a containing factor.

Foamy gland pattern of pancreatic ductal adenocarcinoma: a deceptively benign-appearing variant

Pathologic diagnosis of pancreatic adenocarcinoma is frequently a challenge, particularly in small biopsies, frozen sections, and in metastatic foci. Here we report a deceptively benign-appearing and morphologically distinctive pattern of ductal adenocarcinoma with prominent microvesicular cytoplasm, giving the cells a foamy appearance similar to that described in the prostate (Am J Surg Pathol 1996;20:419). This variant, which we refer to as foamy gland pattern (FGP), was frequently misdiagnosed in frozen sections or biopsies and its pathologic stage underestimated in surgical specimens. Histologically, the diagnostic features were: (1) white and crisply foamy, "microvesicular" cytoplasm; (2) often basally located and compressed, hyperchromatic nuclei reminiscent of endocervical glands (and so-called "adenoma malignum") or gastric foveolar glands; (3) irregular nuclear contours forming wrinkled (raisinoid) nuclei in some areas; and (4) a distinctive chromophilic condensation of the cytoplasmic material in the luminal aspect of the cells forming a brush border-like zone (BLZ). Histochemically, this BLZ was positive for mucicarmine, alcian blue, and high iron diamine, but not PAS. The remainder of the cytoplasm was negative for all these stains. In contrast, benign mucinous ducts, which constitute the major differential diagnosis, had more homogeneous acidophilic cytoplasm, lacked BLZ, and showed cytoplasmic staining with PAS. Immunohistochemically, the tumor cells were diffusely and strongly positive for CEA and cytokeratin 8 whereas B72.3 staining was focal and weak. MUC1 staining was largely confined to the BLZ. MUC2 was negative. P53 staining was detected in 16 of the 20 cases studied and was strong and diffuse in five. K-ras mutation was detected in 6 of 8 cases studied. The clinical findings in the 20 patients in this study (4 pure and 16 mixed with usual ductal carcinoma) did not appear to differ significantly from those of ordinary ductal adenocarcinoma of the pancreas. Eleven patients were men and nine were women; the mean age was 62 years and the mean tumor size was 4.4 cm. Follow-up information was available in 17 patients of whom 7 were alive at an average follow up of 23 months (range, 7-104 mos), and 10 were dead of disease at a median follow up of 15 months (range, 4-42 mos). The median survival of the four patients with pure FGP was 18 months. The median survival did not appear to be significantly longer than that of the patients with resectable ordinary ductal adenocarcinoma in the authors' experience (109 patients, median survival of 12 mos, p = 0.48). In conclusion, foamy gland pattern of invasive pancreatic ductal carcinoma is morphologically distinctive and is prone to misdiagnosis as a benign process. The pathologic stage is often underestimated as a result of the lack of its recognition and misinterpretation as mucinous ducts. Careful attention to its microscopic features is adequate for accurate diagnosis. Histochemical and immunohistochemical stains are useful in confirming the diagnosis of malignancy in challenging cases.

Mutations in the arginine-rich protein gene (ARP) in pancreatic cancer

The ARP gene encodes a highly conserved arginine-rich protein from chromosomal band 3p21.1. At the cytogenetic level this region is frequently deleted in a variety of different solid tumors, although not in pancreatic cancer. We have reported the presence of a specific mutation (ATG50-->AGG) or deletion of codon 50 of the ARP gene in different tumor types (Shridhar et al., 1996, 1996a). In the present study, we have observed mutations involving codon 50 in 11 of 37 pancreatic tumors. The frequency of codon 50 mutation is roughly the same in pancreatic tumors as in the other types of tumors previously examined. In addition, we have detected mutations at codon 51 in multiple PCR subclones in two other pancreatic tumors. Mutations in the ARP gene are thus commonly observed in pancreatic cancer, as well as many other cancers.

Frequent breakpoints in the 3p14.2 fragile site, FRA3B, in pancreatic tumors

FRA3B, at chromosomal band 3p14.2, is the most active common fragile site in the human genome. Homozygous deletions in the region of FRA3B have been observed in many types of solid tumors. Recently, the FHIT gene was reported to span FRA3B and was shown to be homozygously deleted in several gastric and colonic tumor cell lines. Several microsatellite markers that precisely define the 1.0-Mb region surrounding FRA3B and FHIT have been utilized, along with other 3p microsatellites, to analyze the loss of 3p sequences in 25 primary pancreatic adenocarcinomas. The high density of microsatellite markers in the 3pl4.2 region enabled us to both identify losses within and flanking FRA3B in pancreatic cancer and define the breakpoints. We observed loss of het erozygosity of 3pl4.2 markers in 16 of 25 pancreatic tumors and loss of heterozygosity of 3p markers outside of 3pl4.2 in only 2 of 25 tumors of this type. There appears to be a dramatic clustering of chromosomal breakpoints at 3pl4.2 in and immediately distal to FRA3B in pancreatic cancer. We detected no homozygous deletions in this region.

Clinicopathologic analysis of bcl-2 immunostaining in breast carcinoma

Tissue sections of 81 breast carcinomas and 19 benign breast tissues were immunostained with a monoclonal antibody to the bcl-2 gene product, a cytoplasmic protein that regulates apoptosis. The degree of immunoreactivity was then compared with clinicopathologic parameters and to immunostaining for mutated p53 gene product. Immunoreactivity for bcl-2 was present consistently in lymphocyte populations and in residual benign lobules. Apocrine metaplasia (n = 6) and lactating breast (n = 1) exhibited minimal bcl-2 expression, whereas duct hyperplasia (n = 10) showed staining of cells primarily at the periphery of the involved structure and adenosis (n = 7) displayed staining in a majority of cells. Neoplastic epithelial bcl-2 immunoreactivity was negative or minimally positive (staining in 1-5% of cells) in 42% of cases, heterogeneous (staining in 6-30% of cells) in 27% of cases, and diffuse (> 30% of cells) in 31% of cases. Immunostaining for bcl-2 correlated with the presence of estrogen receptor (bcl-2 negative, 16% estrogen receptor positive versus bcl-2 positive, 88% estrogen receptor-positive; P < 0.001), with differentiation (bcl-2 negative, 62% poorly differentiated versus bcl-2 positive, 8% poorly differentiated; P < 0.001) and with better disease-free survival (bcl-2 negative, 82% recurrence versus bcl-2 positive, 28% recurrence; P = 0.0001; 52-mo mean follow-up). Immunostaining for p53 in greater than 5% of tumor cells was observed in 39% of cases and was more frequent in bcl-2-negative tumors (18/35, 51%) as opposed to bcl-2-positive tumors (14/46, 30%); P = NS. Disease recurrence correlated with p53 staining, which was observed in 51% of tumors that relapsed versus only 22% of tumors that did not recur. We conclude that bcl-2 is expressed in benign breast tissues that retain proliferative capacity and partial differentiation. Moreover, in neoplastic breast tissue, it is better correlated with a differentiated, "hormonally responsive," prognostically favorable phenotype than with disabled p53 gene function.

Human papillomavirus DNA in malignant and hyperplastic prostate tissue of black and white males

This study hypothesized that human papillomavirus (HPV) infection is associated with increased prostate cancer risk, and that the 40% higher incidence rate in blacks is attributable to a greater prevalence of oncogenic viral DNA in prostatic tissues. Viral L1 and E6 gene sequences were polymerase chain reaction (PCR) amplified in archival tissues from 56 prostate cancer cases and 42 hyperplastic controls. L1 amplimers were hybridized by dot blot to HPV L1 generic probes, as were E6 amplimers to E6 probes specific for HPV 6, 11, 16, 18, 31, 33, and 45. 12.5% of cases and 9.5% of controls were HPV positive by L1 hybridization (age/race adjusted odds ratio = 1.66, 95% confidence interval = 0.33, 8.37). Four of 52 (7.7%) blacks were HPV positive compared to 7 of 46 (15.2%) whites. However, none of the L1-positive samples hybridized to the E6 type-specific probes, and positive results were not replicable using a broader spectrum of PCR primers and probes. These data suggest that HPV infection is not a significant risk factor for prostate cancer and does not explain the excess cancer risk in blacks.

High frequency of mutator phenotype in human prostatic adenocarcinoma

Mutator phenotype of nucleotide repeats has been implicated to be involved in human cancer and other diseases. This type of instability may be the direct result of DNA replication and/or repair errors. To examine mutator phenotype during the development of human prostate cancer, we undertook this study to screen 57 patients with prostatic adenocarcinoma for possible mutator phenotype at 18 microsatellite marker loci on 12 chromosomes (3p, 5q, 6p, 7p, 8p, 10q, 11p, 13q, 16q, 17p, 18q and Xq). Overall, in 37 of 57 patients, we have found positive mutator phenotype in at least one of the loci analysed. A significantly greater number of cases were found to be positive for this phenotype among the poorly differentiated than the moderately- and well-differentiated prostatic adenocarcinomas. Our data suggest that mutator phenotype may play an important role in the development and progression of human prostate cancer.

Immunohistologic evaluation of invasion-associated proteases in breast carcinoma

Immunostaining of two invasion-associated proteolytic enzymes, cathepsin D (CD) and urokinase-type plasminogen activator (uPA), was assessed in cryostat sections of 86 stage-heterogeneous breast carcinomas using monoclonal antibodies. Most tumors displayed a focal and/or heterogeneous staining pattern. Overall, staining was more frequent in host-derived stromal and inflammatory cells (uPA 54%, CD 89%) than neoplastic epithelium per se (uPA 24%, CD 70%). Intense (i.e., 2+) stromal, but not neoplastic, CD was significantly correlated with nodal or systematic metastases (node negative--10% versus node positive/systemic--33%, p = 0.04). Further, cumulative staining of more than one enzyme (CD + uPA) or more than one tumor component (stroma + epithelium) correlated with metastatic disease (no metastases--35% versus metastatic--72%, p = 0.005). Neither stromal nor epithelial CD alone was significantly correlated with short-term recurrence free survival, however additive CD staining (i.e., stromal + epithelial) was strongly predictive, overall (both + -75% recurred versus both weak/negative--16% recurred, p = 0.0004) and in node positive patients (p = 0.02). We conclude that (a) enzymes putatively mediating extracellular matrix dissolution may be derived from multiple sources and (b) the metastatic capacity and/or clinical aggressiveness of breast carcinomas may reflect overall proteolytic enzyme expression, suggesting that cooperative enzyme interaction may be required for invasive growth and/or metastasis.