Introduction: Pancreatic adenocarcinoma is an aggressive malignancy that carries a poor overall prognosis. New therapeutic approaches are urgently needed. Oncolytic virotherapy is an innovative treatment strategy in which conditionally replicative adenoviruses (CRAds) are genetically modified to efficiently infect and specifically lyse targeted tumor cells while sparing normal cells. We previously demonstrated that by modifying the knob domain of the adenovirus with a chimeric 5/3 knob, we could increase efficiency of infectivity in pancreatic cancer cells, independent of the CAR pathway. However, whether we can restrict oncolytic activity specifically to pancreatic cancers was not addressed. In this study, we incorporated a CXCR4 tumor specific promoter to restrict oncolysis to cells that overexpress CXCR4 (i.e. human pancreatic cancers). Methods: The human pancreatic cancer lines CFPAC-1 and PANC-1 and a rat breast cancer line MAT BIII were obtained from ATCC and cultured in DMEM containing 10% FBS. MAT BIII is known not to overexpress CXCR4 and therefore serves as our negative control. Expression of CXCR4 was determined by Western blot analysis. To determine infection efficiency, cells were infected with increasing titers of unmodified and fiber modified Ads expressing the green fluorescence protein (GFP) marker. GFP expression was determined by flow cytometry. Oncolytic activity was determined by infecting the cells with increasing doses of Ad5/3-CXCR4 CRAds and evaluating the percentage of adherent live cells by crystal violet staining. Results: As expected, CFPAC-1 and PANC-1 pancreatic cancer cells highly expressed CXCR4 whereas MAT BIII cells did not. The serotype 3 fiber-modified Ad (Ad5/3) was most efficient in infecting CFPAC-1 and PANC-1 cells compared to an Ad containing an unmodified fiber. The CRAd construct that has both the CXCR4 promoter and the Ad5/3 knob (Ad5/3-CXCR4) demonstrated highly effective oncolytic activity against CFPAC-1 and PANC-1, but not against MAT BIII cells. Conclusions: The novel Ad5/3-CXCR4 CRAd was highly efficient and specific in infecting and lysing human pancreatic cancer cells. Further development and testing of this novel construct should be considered for potential therapy against pancreatic cancer.

Introduction: Oncolytic virotherapy using conditionally replicative adenoviruses (CRAds) represents a promising therapeutic approach in the treatment of pancreatic adenocarcinomas. Genetically modified CRAds can be designed to specifically infect and lyse targeted tumor cells while sparing normal cells. One such modification is to replace the knob domain of the serotype 5 adenovirus vector with a serotype 3 chimeric knob (Ad5/3) to increase efficiency of infectivity, and to incorporate a CXCR4 tumor specific promoter (Ad5/3-CXCR4) to restrict oncolysis to cells that overexpress CXCR4 (i.e. pancreatic tumors). We previously demonstrated the efficacy of such an approach against the CFPAC-1 and PANC-1 human pancreatic cancer cell lines in vitro. The purpose of this study is to evaluate the role of this novel CRAd against human pancreatic xenograft tumors in SCID mice. Methods: CFPAC-1 cells were intradermally injected into the flanks of 40 SCID mice (2.4 × 106 cells/side in 10 μl PBS). The 40 SCID mice were divided into 4 groups of 10 mice. Once palpable tumors were established (approximately 10 mm2), then 5 × 109 viral particles of Ad5/3-CXCR4, wild-type human serotype 5 adenovirus (positive control), replication deficient green fluorescent protein expressing adenovirus (negative control), or phosphate buffered saline (no virus control) in 30 μl were injected intratumorally each day for three consecutive days. The tumors were measured every other day for twenty-five days. The physical condition and behavior of the mice were assessed and recorded daily using an established body condition scale based on appearance, natural behavior, provoked behavior, and weight changes. ANOVA 1-way was used to compare the mean tumor size among the groups. Results: Ad5/3-CXCR4 CRAds significantly suppressed tumor growth compared with the other treatment groups (p=0.001). The toxicity profile demonstrated that mice in the Ad5/3-CXCR4 group and PBS group had the best physical and behavioral profiles, whereas mice treated with control wild-type and GFP viruses had the worst physical and behavioral profiles. Conclusions: The novel Ad5/3-CXCR4 adenoviruses not only exhibited a potent antitumor effect against xenograft human pancreatic cancers in SCID mice, but also demonstrated no observable host toxicity. Thus, Ad5/3-CXCR4 serves as a promising candidate for clinical treatment of pancreatic adenocarcinoma.

Background: Hypoxia is a common characteristic of many aggressive tumors including pancreatic cancer. Under hypoxia, cancer has been observed to be more aggressive. The chemokine receptor CXCR4 has been linked to the invasion and metastasis of many types of cancers. How CXCR4 expression is regulated under the hypoxic condition in pancreatic cancer is still not clear. We hypothesize that in human pancreatic cancer cell line CFPAC-1, CXCR4 is overexpressed under hypoxia and this overexpression is mediated by the hypoxia-inducible factor-1α (HIF-1α). Materials and Methods: Human pancreatic cancer cell line CFPAC-1 and HeLa cells (positive control) were obtained from ATTC. Both cell lines were exposed to variable hypoxic (1% O2) and normoxic conditions over specified times (0 to 24 hrs). Nuclear and whole cell lysates of the samples were taken and analyzed by Western blot for CXCR4 and HIF-1α expressions. The cells were then exposed to varying doses of YC-1, a drug known to inhibit HIF-1α expression, under normoxia and 24 hr hypoxia. HIF-1α and CXCR4 expression were then analyzed as previously described. Results: Baseline CXCR4 overexpression was observed for HeLa and CFPAC-1 cells under normoxia. However, longer exposure to hypoxia yielded increasing levels of HIF-1α and CXCR4 for both cell lines. Treatment of both with YC-1 yielded decreasing level of HIF-1α protein levels which corresponded to a reduction of CXCR4 expression in both HeLa and CFPAC-1 cell lines. Conclusions: CXCR4 and HIF-1α overexpression were both increased under hypoxia. Hypoxia induces HIF-1α which mediates CXCR4 expression. These novel findings may be used to design target specific therapy against pancreatic cancers. View Large Image Figure Viewer

Introduction: African-American women with breast cancer have a worse outcome when compared to Caucasian women. Triple-receptor negative breast cancer (TNBC) is biologically aggressive and is associated with a worse prognosis. Since TNBC is more prevalent among African-American women, it has been postulated that this contributes to worse outcomes in African-American women. To understand the impact of race/ethnicity on outcome for TNBC, socioeconomic status (SES) needs to be controlled. We examined the impact of race/ethnicity on a cohort of patients of low SES who have TNBC. Methods: A prospective database created in 1998 collected data on 786 patients with Stage 0-III breast. Of these, 202 patients had TNBC (26%). Primary endpoints were cancer recurrence and death. ZIP code-based income tract and financial data from our computer services were used to assess SES. Data were analyzed using Kaplan-Meier survival analysis, log-rank tests, Cox Proportional hazard regression, chi square test, and t-tests. A p-value≤0.05 was considered statistically significant. Results: Of the 468 African-Americans (60%) in the database, 138 had TNBC; 64 of 318 Caucasians had TNBC. Mean age at diagnosis was 49 years. Eighty-eight percent of patients had an annual income of ≤ $20,000, and 79% were either free care or Medicaid. The 5-yr overall survival was 72% for African-American women versus 75% for Caucasian women (p=0.93). On multivariate analysis, race/ethnicity had an impact on disease-free survival (p=0.027) but not on overall survival (p=0.9813). Conclusion: In a predominantly indigent population, race/ethnicity had no impact on overall survival for patients with triple-receptor negative breast cancer.

Introduction: CXCR4 is a G-protein coupled receptor that has been linked with metastasis in several cancers, including breast cancer. We recently demonstrated that high CXCR4 level in primary tumors of patients with breast cancer had a prognostic significance. What is not known, however, is whether CXCR4 can predict response to chemotherapy. We hypothesize that patients whose tumors had a low CXCR4 overexpression level following neoadjuvant chemotherapy will have a lower recurrence rate than those whose tumors had persistently high CXCR4 levels. Methods: Seventeen locally advanced breast cancer patients were accrued, and tumor specimens were obtained before and after neoadjuvant therapy. CXCR4 levels were identified and quantified by Western blots and the results were quantified against 1 μg of protein from HeLa cells. The primary end-point was cancer recurrence. Statistical tests utilized include Kaplan-Meier survival analysis and log-rank test. A p-value≤0.05 was considered significant. Results: Based on our previous work, low CXCR4 overexpression was defined as≤6-fold elevation and high overexpression was defined as>6-fold elevation over HeLa cells. Of 17 locally advanced breast cancers evaluated, 10 tumors (59%) remained in the low group, 5 tumors (29%) had a reduction from high CXCR4 overexpression to low overexpression, and 2 tumors (12%) maintained a high CXCR4 overexpression after neoadjuvant therapy. With a median follow-up of 28 months, patients whose tumors remained at a high CXCR4 overexpression level after neoadjuvant therapy had a significantly higher rate of cancer recurrence (p=0.0068). Conclusions: CXCR4 was a predictive molecular marker of response to neoadjuvant chemotherapy for patients with locally advanced breast cancer. Patients whose tumors had a persistently high CXCR4 overexpression level after neoadjuvant therapy are at a significant risk for recurrence, and therefore, should be targeted for more intensive and/or novel therapy.

Background: Due to advances in viral design, oncolytic adenoviruses have emerged as a promising approach for treatment of human cancer. Development of this technology is impeded by two major obstacles: one being poor tropism, and another being a lack of specificity for cancerous cells. It has become possible to target these viruses to preferentially infect cancer cells by modifying the fiber protein of the viral capsid, making a serotype 5/3 chimera, thereby changing tropism. It is also now possible to render the replication of the virus conditional with tumor selective expression of the E1A gene using promoters that are active in cancer cells but not in normal tissue, such as the CXCR4 promotor region. It is vital in the preclinical evaluation of these strategies that the most stringent demonstrations of efficacy be utilized. The tissue slice model system is thought to be a stringent model system for preclinical evaluation of adenovirus therapies, as the cells here in the tissue-slice model system have morphology and a phenotype that more closely resembles that seen in the clinical setting than cell-line based culture can provide, and it has been shown to have utility in the evaluation of viral infectivity and replication. Evaluations of efficiency of cell death have previously been limited within this model system. This obstacle can be circumvented due to the emerging understanding of the mechanism of adenoviral induced cell death. In this study, we evaluate the efficacy of viral infection, replication, and induction of cell death with a tropism-modified conditionally replicative adenovirus in a breast cancer tissue slice model system. Methods: The slices were infected with a tropism-modified conditionally replicative adenovirus, and a wild-type adenovirus for comparison. The tissues used were evaluated for the expression of viral target proteins using immunohistochemistry. Efficiency of infection was evaluated using fluorescent microscopy, as the viruses used have been modified to express red fluorescent protein. Replication of the viruses was evaluated with quantitative real-time PCR to assay viral genome copy number, a indicator for comparison of the number of virions replicated. Results: Infection and replication of our tropism modified conditionally replicative virus has been observed in breast cancer tissue slice model system and is comparative to wild type virus. Conclusions: We have shown that our tropism modified conditionally replicative oncolytic adenovirus can infect and replicate in breast cancer tissue slices, which may be an important preclinical indicator for its therapeutic utility.

e22192 Introduction: CXCR4 is a chemokine receptor that is linked to breast cancer metastasis. The putative mechanism involves activation of the HER-2/PI3K/PTEN/Akt/mammalian target of rapamycin (mTOR) pathway, resulting in enhanced synthesis of downstream proteins such as CXCR4 and eukaryotic initiation factor 4E (eIF4E). Overexpression of eIF4E has been shown to be an independent predictor of breast cancer outcome. Because HER-2/mTOR/CXCR4 pathway has mainly been observed in vitro, it deserves scrutiny to determine whether it is also seen in human breast cancers. We hypothesize that HER-2/mTOR axis is not significantly associated with CXCR4 regulation in human breast cancers. Method: 107 stages 1 - 3 breast cancers from a prospective breast cancer database were analyzed. Samples were quantified for mTOR, CXCR4, and eIF4E levels by Western Blot. mTOR, and eIF4E levels were quantified as fold-over benign breast tissues and CXCR4 levels were quantified as fold-over Hela cells. CXCR4 overexpression was defined as low (< 6.0 fold) or high (≥ 6.0 fold), and eIF4E expression level was separated into tertiles. Primary end points were cancer recurrence and death. Statistical analyses performed were independent student t-test, Spearman correlation, Kaplan-Meier survival curves, log-rank test, and Cox proportional hazard model. Results: All 59 HER-2 (-) and 48 HER-2 (+) breast cancer specimens overexpressed CXCR4 (mean=6.18 ±4.4), mTOR (mean=1.24 ±1.58), and eIF4E (mean=9.53 ±6.42). Unlike in vitro data, there was no difference in mean CXCR4 expression level between HER-2 (-) and HER-2 (+) tumors (mean=6.55, 5.72, respectively) (p=0.33). While mTOR correlated significantly with eIF4E level (p=0.0006, r = 0.34), it did not correlate with CXCR4 level (p=0.713, r=0.036). Only nodal status (p=0.04) and eIF4E level (p=0.03) were independent predictors of breast cancer death. Conclusions: We demonstrate that CXCR4 regulation in human breast cancers is more complex than previously thought. There was no correlation between mTOR and CXCR4 levels, nor was there a difference in mean CXCR4 overexpression levels between HER-2 (+) and HER-2 (-) tumors. Pathways regulating CXCR4 expression warrant further investigation. No significant financial relationships to disclose.

“Triple-negative” (estrogen receptor [ER], progesterone receptor [PR], and HER2-negative) breast tumors are typically high grade and portend a higher risk for relapse. Eukaryotic Initiation Factor 4E (eIF4E) overexpression in tumor specimens has been reported to be a predictor for relapse in breast cancer independent of nodal status. In this prospective study. eIF4E elevation in breast cancer specimens was studied to determine outcome in “triple-negative” breast cancer patients.

OBJECTIVE:In a prospective trial, to determine if eIF4E overexpression in breast cancer specimens is correlated with VEGF elevation, increased tumor microvessel density (MVD) counts, and a worse clinical outcome irrespective of nodal status.SUMMARY AND BACKGROUND DATA:In vitro, the overexpression of eukaryotic initiation factor 4E (eIF4E) up-regulates the translation of mRNAs with long 5'-untranslated regions (5'-UTRs). One such gene product is the vascular endothelial growth factor (VEGF).METHODS:A total of 114 stage I to III breast cancer patients were prospectively accrued and followed with a standardized clinical surveillance protocol. Cancer specimens were quantified for eIF4E, VEGF, and MVD. Outcome endpoints were cancer recurrence and cancer-related death.RESULTS:eIF4E overexpression was found in all cancer specimens (mean +/- SD, 12.5 +/- 7.6-fold). Increasing eIF4E overexpression correlated with increasing VEGF elevation (r = 0.24, P = 0.01, Spearman's coefficient), and increasing MVD counts (r = 0.35, P < 0.0002). Patients whose tumor had high eIF4E overexpression had shorter disease-free survival (P = 0.004, log-rank test) and higher cancer-related deaths (P = 0.002) than patients whose tumors had low eIF4E overexpression. Patients with high eIF4E had a hazard ratio for cancer recurrence and cancer-related death of 1.8 and 2.1 times that of patients with low eIF4E (respectively, P = 0.009 and P = 0.002, Cox proportional hazard model).CONCLUSIONS:In breast cancer patients, increasing eIF4E overexpression in the cancer specimens correlates with higher VEGF levels and MVD counts. Patients whose tumors had high eIF4E overexpression had a worse clinical outcome, independent of nodal status. Thus, eIF4E overexpression in breast cancer appears to predict increased tumor vascularity and perhaps cancer dissemination by hematogenous means.