Identification of the gamma irradiation dose applied to ground beef that reduces Shiga toxin producing Escherichia coli but has no impact on consumer acceptance

The aims of the present study were: a) to estimate the minimal dose of gamma irradiation required to reduce 5 log CFU/g of native O157 and non-O157 Shiga toxin-producing Escherichia coli population in ground beef samples inoculated with high inoculum; b) to assess its effectiveness in samples with low inoculum and 3) to evaluate consumer acceptance. Based on the results, 1 kGy was estimated as the minimal dose of gamma irradiation required to reduce 5 log CFU/g of STEC in ground beef. However, when samples with low inoculum level were subjected to 1 kGy, 3.9% of the samples were positive for stx and eae genes after an enrichment step. Consumer acceptance analysis was carried out with samples subjected to 2.5 kGy and no significant differences were found between irradiated and control samples. Therefore, 2.5 kGy was identified as the gama irradiation dose that reduces STEC but has no impact on consumer acceptance of ground beef.

Abstract P4-06-07: Activity of nivolumab alone or in combination with targeted therapies in a humanized BLT-mouse model of human breast cancer

Background: Recent advances in the field of cancer immunotherapy have increased demand for reliable preclinical models to inform patient selection and rational drug combination strategies. The development of the bone marrow-liver-thymus (BLT) mouse may provide the opportunity to study the complex interactions of human tumor and host immune systems in vivo. Other models are limited by the rapid onset of graft versus host disease (GVHD) and a lack of orderly maturation and trafficking of human T and B cells. In BLT mice, implantation of human fetal liver and thymus fragments beneath the kidney capsule of NSG (NOD/SCID/IL-2RΥ-/-) mice followed by engraftment of matched ex vivoexpanded CD34+ cells supports the production of an almost complete human immune system. In this study, we used this model to assess the efficacy of the anti-PD-1 therapeutic antibody, nivolumab, in combination with targeted therapeutics in specific breast cancer sub-types.

Materials and Methods: For triple negative breast cancer (TNBC), the activity of nivolumab was assessed in combination with the PARP1/2 inhibitor, talazoparib, in humanized BLT mice. Xenografts were established by subcutaneous injection of 5.0x106MDA-231 (TNBC) cells. Mice (n=5) were randomized into treatment groups as follows; 1) Vehicle control (PBS), 2) nivolumab (10mg/kg QW), 3) talazoparib (0.33mg/kg Q5/2D) and 4) nivolumab+talazoparib. After 21 days of treatment, tumor tissue, serum and PBMCs were collected for biomarker analysis.

Results: Successful reconstitution of mature human T and B cells was confirmed in BLT mice 12-weeks post engraftment of donor tissue and CD34+ hematopoietic stem cells. MDA-231 cells injected subcutaneously into the flank of these mice formed palpable tumors (150-200mm3) within 9 days of injection. For vehicle control treated mice, tumors grew (2.5-fold) throughout the 21-day study. Single agent nivolumab induced significant tumor growth inhibition (TGI) relative to vehicle control treated mice at Day 21. Single agent talazoparib also induced comparable levels of TGI as did the combination of nivolumab plus talazoparib. Nivolumab treated mice continued to gain weight throughout the study without overt signs of toxicity. Reversible weight loss was observed in the talazoparib and combination treated arms. Overt signs of GVHD were not observed in any of these animals. Preliminary tissue analysis identified high levels of cell surface PD-L1 protein in control treated MDA-231 xenografts. Further analysis of the treated tumors will provide valuable insight into the mechanism of action of this class of molecule. We are establishing xenograft models of hormone receptor (ER+) positive breast cancer to measure the activity of nivolumab+/-CDK4/6-inhibition in humanized BLT mice and these data will also be presented.

Discussion: The data presented here highlight the potential of the PD-1 antibody nivolumab to have activity in TNBC. Furthermore, these findings illustrate the potential of the humanized BLT-mouse to model responses to immune check-point in the preclinical setting. Expanded use of this model may help to identify response biomarkers and inform design of combination therapies using immune oncology molecules and approved targeted therapies.

Citation Format: O'Brien NA, Luo T, Ayala R, Salgar S, Conklin D, McDermott M, Kitchen S, Rezek V, Horak C, Dugan U, Slamon DJ. Activity of nivolumab alone or in combination with targeted therapies in a humanized BLT-mouse model of human breast cancer [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P4-06-07.

Establishing a complementary diagnostic for anti-PD-1 immune checkpoint inhibitor therapy

The role of tumor PD-L1 expression was investigated across the nivolumab clinical development program. Phase III nivolumab trials have shown that patients with tumors not expressing PD-L1 may benefit; therefore, testing is not required to select patients for therapy. The Dako PD-L1 IHC 28-8 pharmDx assay may be used to determine tumor PD-L1 expression as a complementary and informative test.

Abstract PD07-02: βIII Tubulin Expression by Immunohistochemistry (IHC) and mRNA in a Randomized Phase 2 Neoadjuvant Breast Cancer (BC) Study of Ixabepilone (ixa) Compared to Paclitaxel (pac)

Background: Overexpression of βIII tubulin has been shown to correlate with poor prognosis and reduced response to taxanes. Ixa exhibits activity in βIII overexpressing tumor models and BC patients (pts) with high βIII mRNA levels. We report the prevalence of βIII by a validated IHC assay and the correlation of βIII protein and mRNA expression in BC tissue samples. Methods: In this neoadjuvant BC study CA163-100, 295 patients were randomized 1:1 to ixa (40 mg/m2 IV x 4 cycles) or pac (80 mg/m2 IV wkly x 12) after all pts received 4 cycles of doxorubicin (60 mg/m2) plus cyclophosphamide (600 mg/m2). Pre-chemotherapy formalin-fixed paraffin-embedded pt tissue sections were assessed for intensity and percentage of tumor cell staining of βIII by IHC. RNA was extracted from additional pre-chemotherapy core needle biopsy specimens and profiled for gene expression using Affymetrix HG-U133Av2 gene chips. Tumors were classified as basal-like, Her2-enriched, luminal-like A and B and normal-like using a 50-gene subtype predictor. In addition, Her2 and hormone receptor status was used for tumor subtyping. Results: βIII staining data were obtained for 261 pts, of which 247 were randomized to either ixa or pac arms of study CA163-100. The prevalence of βIII positivity was 39.1% (102/261) using the prespecified cut-off of ≥50% (2+, 3+) positive cells. The distribution of IHC scores was bimodal. 64% of specimens classified as βIII negative had no observable staining, while 77% of βIII positive specimens had a majority of tumor cells staining at the highest intensity (3+). IHC and gene expression profiling data were available for 235 pts and a correlation between βIII tubulin protein and mRNA expression was observed. The relative βIII mRNA expression levels were significantly higher in pts classified as βIII positive vs. βIII negative by IHC (P<0.0001) and 80% of tumors classified as βIII tubulin positive by IHC, had high relative TUBB3 mRNA levels, using a median cut-off. An association between βIII IHC status and breast cancer subtypes was also seen with 69% (62/90) of βIII positive specimens also classified as basal-like and conversely, 55% (62/113) of basal-likes classified as βIII positive. Nearly half (48%) the specimens in this cohort were basal-like, a classification that was quite concordant with triple negative status and as such, triple negative tumors were also enriched for βIII positivity. Conclusion: βIII is reliably measured by a validated IHC assay, providing a bimodal distribution of scores in this BC study. Correlative analyses of βIII status by BC subtype suggest that βIII positivity is significantly associated with aggressive basal-like/TN breast cancers. The clinical utility of βIII as a potential tool to select patients who may benefit from ixa in this study will be presented separately.

Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr PD07-02.

Abstract PD07-01: Phase 2 Study of Ixabepilone Versus Paclitaxel as Neoadjuvant Therapy for Early Stage Breast Cancer with Comparative Biomarker Analysis

Background: Anthracyclines (A) and taxanes (T) are standard neoadjuvant treatments for breast cancer (BC), achieving pathologic complete response (pCR) rates of 20-30% in unselected patient (pt) populations. Ixabepilone (ixa) is approved for treatment of metastatic BC: plus capecitabine (Cap) in pts progressing after A and T or as monotherapy after progression on A, T and Cap. Prior data suggest that overexpression of βIII tubulin is associated with resistance to paclitaxel (P) while activity of ixa was unaffected. We present the first randomized comparison of neoadjuvant ixa and P in early stage BC. Primary objectives were to estimate pathologic complete response rate (pCR) in the overall population and in biomarker-defined populations. Methods Pts with early stage BC were biopsied for immunohistochemical (IHC) and mRNA biomarker analyses prior to chemotherapy. Following 4 cycles of doxorubicin/cyclophosphamide (AC), pts were randomized to either every 3 week ixa (40mg/m2:4 cycles) or weekly P (80mg/m2:12 doses). Post-therapy surgery and pathological reports were used to assess pCR. Baseline βIII expression was assessed via a standardized IHC assay (Dako, CA) and predefined single gene mRNA markers (including TUBB3, CAPG, TACC3) were assessed via Affymetrix gene expression profilling. The pCR rate and cutoff for biomarker positivity were estimated using a cross-validation method. Secondary endpoints in the study included clinical objective response rate and safety. Results Pts (N=384) were enrolled in 15 countries: 313 pts were treated with AC. 295 pts were randomized; 289 were treated with either ixa or P. Of these, 247 (123, ixa; 124, P) had βIII IHC data and 231 (114, ixa; 117, P) had both pathologic and βIII IHC data. Baseline characteristics were balanced between arms, including triple negative pts (TN, 49%). The pCR rate in all randomized pts was 24.3% (90% CI: 18.6-30.8) with ixa and 25.2% (90% CI: 19.4-31.7) with P. pCR rates were similar regardless of sub-group.

Table 1. pCR rates

No significant interaction was observed between βIII expression and treatment arms (logistic regression analysis). Secondary efficacy measures were consistent with the pCR results. No clinically meaningful differences in efficacy endpoints were noted between ixa and P with mRNA markers (TUBB3, CAPG and TACC3; analyses are ongoing with others). The safety profiles of ixa and P were similar, including incidence of peripheral neuropathy (Grade 3/4: ixa 4.1% vs P 3.5%). An exception was greater neutropenia with ixa (Grade 3/4; 41.3% versus 8.4% with P) although there was no difference in the rate of febrile neutropenia (0.7%). Summary Overall, the results indicated that ixa had similar efficacy to P when measured by pCR in the neoadjuvant BC setting. No clinically meaningful differences were noted in the efficacy profile of ixa compared to P across the subsets analyzed. Ixa or P following AC was well tolerated with similar safety profiles.

K-Ras mutation (mut), EGFR-related, and exploratory markers as response predictors of cetuximab in first-line advanced NSCLC: Retrospective analyses of the BMS099 trial

8021

Background: This phase III study investigated cetuximab (C) plus taxane/carboplatin (TC) in 1st-line advanced NSCLC. Progression-free survival (PFS) was not significantly different with the addition of C to TC; response rate (RR) was significantly higher; median overall survival (OS) was longer, with a difference (not statistically significant) similar in magnitude to the significant OS improvement from FLEX (cisplat/vinorelb±C). Potential patient-selection markers were analyzed, including mut in K-Ras, EGFR mut and gene copy number (GCN), EGFR protein expression, and gene expression profiles. Methods: Chemonaïve patients (pts) with stage IIIB/IV NSCLC (any histology or EGFR expression status) were randomized to TC±C. K-Ras gene exon 2, and EGFR exons 18–21 were sequenced bidirectionally using formalin-fixed paraffin embedded tissue-derived genomic DNA. EGFR GCN was measured by fluorescence in situ hybridization (FISH) and EGFR protein expression by immunohistochemistry (IHC). Exploratory expression profiling studies using Affymetrix GeneChip are ongoing. Analyses were blinded to clinical outcome. The relationships between biomarker status and efficacy were assessed with log-rank tests per study arm for treatment-specific effects, and across the total evaluable population. Results: Samples from 225/676 randomized pts were available; 35/202 (17%) had K-Ras mut, 17/167 (10%) had EGFR mut, 54/104 (52%) were EGFR FISH+, 131/148 (89%) were EGFR IHC+. There were no significant treatment-specific correlations between any biomarker assessed and PFS, OS or RR, except superior OS with TC for the EGFR FISH+ group (P=0.03). With all pts combined, OS was longer in the EGFR mut subset vs wt (P=0.09); PFS was longer in EGFR IHC- pts vs IHC+ (P=0.048). Results from gene expression profiling studies, including EGFR ligands, will be presented. Conclusions: There was no significant correlation between patient response to C and any molecular marker evaluated to date (K-RAS mut, EGFR mut, EGFR IHC, EGFR FISH), exploratory analyses are ongoing. Additional predictive-marker studies are needed to optimize cetuximab therapeutic use in NSCLC.

[Table: see text]