Stephen Tirrell

Comparison of the Predictive Accuracy of DNA Array-Based Multigene Classifiers across cDNA Arrays and Affymetrix GeneChips

We examined how well differentially expressed genes and multigene outcome classifiers retain their class-discriminating values when tested on data generated by different transcriptional profiling platforms. RNA from 33 stage I-III breast cancers was hybridized to both Affymetrix GeneChip and Millennium Pharmaceuticals cDNA arrays. Only 30% of all corresponding gene expression measurements on the two platforms had Pearson correlation coefficient r >or= 0.7 when UniGene was used to match probes. There was substantial variation in correlation between different Affymetrix probe sets matched to the same cDNA probe. When cDNA and Affymetrix probes were matched by basic local alignment tool (BLAST) sequence identity, the correlation increased substantially. We identified 182 genes in the Affymetrix and 45 in the cDNA data (including 17 common genes) that accurately separated 91% of cases in supervised hierarchical clustering in each data set. Cross-platform testing of these informative genes resulted in lower clustering accuracy of 45 and 79%, respectively. Several sets of accurate five-gene classifiers were developed on each platform using linear discriminant analysis. The best 100 classifiers showed average misclassification error rate of 2% on the original data that rose to 19.5% when tested on data from the other platform. Random five-gene classifiers showed misclassification error rate of 33%. We conclude that multigene predictors optimized for one platform lose accuracy when applied to data from another platform due to missing genes and sequence differences in probes that result in differing measurements for the same gene.

Phase I Study of the Novel Investigational NEDD8-Activating Enzyme Inhibitor Pevonedistat (MLN4924) in Patients with Relapsed/Refractory Multiple Myeloma or Lymphoma

Purpose: Evaluate the safety, pharmacokinetic profile, pharmacodynamic effects, and antitumor activity of the first-in-class investigational NEDD8-activating enzyme (NAE) inhibitor pevonedistat (TAK-924/MLN4924) in patients with relapsed/refractory lymphoma or multiple myeloma. Experimental Design: Patients with relapsed/refractory myeloma (n = 17) or lymphoma (n = 27) received intravenous pevonedistat 25 to 147 mg/m2 on days 1, 2, 8, 9 (schedule A; n = 27) or 100 to 261 mg/m2 on days 1, 4, 8, 11 (schedule B; n = 17) of 21-day cycles. Results: Maximum tolerated doses were 110 mg/m2 (schedule A) and 196 mg/m2 (schedule B). Dose-limiting toxicities included febrile neutropenia, transaminase elevations, muscle cramps (schedule A), and thrombocytopenia (schedule B). Common adverse events included fatigue and nausea. Common grade ≥3 events were anemia (19%; schedule A), and neutropenia and pneumonia (12%; schedule B). Clinically significant myelosuppression was uncommon. There were no treatment-related deaths. Pevonedistat pharmacokinetics exhibited a biphasic disposition phase and approximate dose-proportional increases in systemic exposure. Consistent with the short mean elimination half-life of approximately 8.5 hours, little-to-no drug accumulation in plasma was seen after multiple dosing. Pharmacodynamic evidence of NAE inhibition included increased skin levels of CDT-1 and NRF-2 (substrates of NAE-dependent ubiquitin ligases), and increased NRF-2-regulated gene transcript levels in whole blood. Pevonedistat–NEDD8 adduct was detected in bone marrow aspirates, indicating pevonedistat target engagement in the bone marrow compartment. Three lymphoma patients had partial responses; 30 patients achieved stable disease. Conclusions: Pevonedistat demonstrated anticipated pharmacodynamic effects in the clinical setting, a tolerable safety profile, and some preliminary evidence that may be suggestive of the potential for activity in relapsed/refractory lymphoma. Clin Cancer Res; 22(1); 34–43. ©2015 AACR.

Phase I assessment of new mechanism-based pharmacodynamic biomarkers for MLN8054, a small-molecule inhibitor of Aurora A kinase

The mitotic kinase Aurora A is an important therapeutic target for cancer therapy. This study evaluated new mechanism-based pharmacodynamic biomarkers in cancer patients in two phase I studies of MLN8054, a small-molecule inhibitor of Aurora A kinase. Patients with advanced solid tumors received MLN8054 orally for 7 consecutive days in escalating dose cohorts, with skin and tumor biopsies obtained before and after dosing. Skin biopsies were evaluated for increased mitotic cells within the basal epithelium. Tumor biopsies were assessed for accumulation of mitotic cells within proliferative tumor regions. Several patients in the highest dose cohorts showed marked increases in the skin mitotic index after dosing. Although some tumors exhibited increases in mitotic cells after dosing, others displayed decreases, a variable outcome consistent with dual mechanisms of mitotic arrest and mitotic slippage induced by antimitotics in tumors. To provide a clearer picture, mitotic cell chromosome alignment and spindle bipolarity, new biomarkers of Aurora A inhibition that act independently of mitotic arrest or slippage, were assessed in the tumor biopsies. Several patients, primarily in the highest dose cohorts, had marked decreases in the percentage of mitotic cells with aligned chromosomes and bipolar spindles after dosing. Evidence existed for an exposure-effect relationship for mitotic cells with defects in chromosome alignment and spindle bipolarity that indicated a biologically active dose range. Outcomes of pharmacodynamic assays from skin and tumor biopsies were concordant in several patients. Together, these new pharmacodynamic assays provide evidence for Aurora A inhibition by MLN8054 in patient skin and tumor tissues.

Reproducibility of Gene Expression Signature–Based Predictions in Replicate Experiments

Purpose: The goals of this analysis were to (a) determine concordance of gene expression results from replicate experiments, (b) examine prediction agreement of multigene predictors on replicate data, and (c) assess the robustness of prediction results in the face of noise. Patients and Methods: Affymetrix U133A gene chips were used for gene expression profiling of 97 fine-needle aspiration biopsies from breast cancer. Thirty-five cases were profiled in replicates: 17 within the same laboratory, 11 in two different laboratories, and 15 to assess manual and robotic labeling. We used data from 62 cases to develop 111 distinct pharmacogenomic predictors of response to therapy. These were tested on cases profiled in duplicates to determine prediction agreement and accuracy. To evaluate the robustness of the pharmacogenomic predictors, we also introduced random noise into the informative genes in one half of the replicates. Results: The average concordance correlation coefficient was 0.978 (range, 0.96-0.99) for intralaboratory replicates, 0.962 (range, 0.94-0.98) for between-laboratory replicates, and 0.971 (range, 0.93-0.99) for manual versus robotic labeling. The mean % prediction agreement on replicate data was 97% (95% CI, 0.96-0.98; SD, 0.006), 92% (95% CI, 0.90-0.93; SD, 0.009), and 94% (95% CI, 0.92-0.95; SD, 0.008) for support vector machines, diagonal linear discriminant analysis, and k-nearest neighbor prediction methods, respectively. Mean accuracy in the test set was 77% (95% CI, 0.74-0.79; SD, 0.014), 66% (95% CI, 0.63-0.73; SD, 0.015), and 64% (95% CI, 0.60-0.67; SD, 0.016), respectively. Conclusion: Gene expression results obtained with Affymetrix U133A chips are highly reproducible within and across two high-volume laboratories. Pharmacogenomic predictions yielded >90% agreement in replicate data.

Transcriptional Responses to Estrogen and Progesterone in Mammary Gland Identify Networks Regulating p53 Activity

Estrogen and progestins are essential for mammary growth and differentiation but also enhance the activity of the p53 tumor suppressor protein in the mammary epithelium. However, the pathways by which these hormones regulate p53 activity are unknown. Microarrays were used to profile the transcriptional changes within the mammary gland after administration of either vehicle, 17beta-estradiol (E), or progesterone (P) individually and combined (EP). Treatment with EP yielded 1182 unique genes that were differentially expressed compared to the vehicle-treated group. Although 30% of genes were responsive to either E or P individually, combined treatment with both EP had a synergistic effect accounting for 60% of the differentially regulated genes. Analysis of protein-protein interactions identified p53, RelA, Snw1, and Igfals as common targets of genes regulated by EP. RelA and p53 form hubs within a network connected by genes that are regulated by EP and that may coordinate the competing functions of RelA and p53 in proliferation and survival of cells. Induction of early growth response 1 (Egr1) and Stratifin (Sfn) (also known as 14-3-3sigma) by EP was confirmed by reverse transcription-quantitative PCR and shown to be p53 independent. In luciferase reporter assays, Egr1 was shown to enhance transcriptional activation by p53 and inhibit nuclear factor kappaB activity. These results identify a gene expression network that provides redundant activation of RelA to support proliferation as well as sensitize p53 to ensure proper surveillance and integration of their competing functions through factors such as Egr1, which both enhance p53 and inhibit RelA.

Applications of Pathology-Assisted Image Analysis of Immunohistochemistry-Based Biomarkers in Oncology

Immunohistochemistry-based biomarkers are commonly used to understand target inhibition in key cancer pathways in preclinical models and clinical studies. Automated slide-scanning and advanced high-throughput image analysis software technologies have evolved into a routine methodology for quantitative analysis of immunohistochemistry-based biomarkers. Alongside the traditional pathology H-score based on physical slides, the pathology world is welcoming digital pathology and advanced quantitative image analysis, which have enabled tissue- and cellular-level analysis. An automated workflow was implemented that includes automated staining, slide-scanning, and image analysis methodologies to explore biomarkers involved in 2 cancer targets: Aurora A and NEDD8-activating enzyme (NAE). The 2 workflows highlight the evolution of our immunohistochemistry laboratory and the different needs and requirements of each biological assay. Skin biopsies obtained from MLN8237 (Aurora A inhibitor) phase 1 clinical trials were evaluated for mitotic and apoptotic index, while mitotic index and defects in chromosome alignment and spindles were assessed in tumor biopsies to demonstrate Aurora A inhibition. Additionally, in both preclinical xenograft models and an acute myeloid leukemia phase 1 trial of the NAE inhibitor MLN4924, development of a novel image algorithm enabled measurement of downstream pathway modulation upon NAE inhibition. In the highlighted studies, developing a biomarker strategy based on automated image analysis solutions enabled project teams to confirm target and pathway inhibition and understand downstream outcomes of target inhibition with increased throughput and quantitative accuracy. These case studies demonstrate a strategy that combines a pathologist’s expertise with automated image analysis to support oncology drug discovery and development programs.

The beagle dog MicroRNA tissue atlas: identifying translatable biomarkers of organ toxicity

BackgroundMicroRNAs (miRNA) are varied in length, under 25 nucleotides, single-stranded noncoding RNA that regulate post-transcriptional gene expression via translational repression or mRNA degradation. Elevated levels of miRNAs can be detected in systemic circulation after tissue injury, suggesting that miRNAs are released following cellular damage. Because of their remarkable stability, ease of detection in biofluids, and tissue specific expression patterns, miRNAs have the potential to be specific biomarkers of organ injury. The identification of miRNA biomarkers requires a systematic approach: 1) determine the miRNA tissue expression profiles within a mammalian species via next generation sequencing; 2) identify enriched and/or specific miRNA expression within organs of toxicologic interest, and 3) in vivo validation with tissue-specific toxicants. While miRNA tissue expression has been reported in rodents and humans, little data exists on miRNA tissue expression in the dog, a relevant toxicology species. The generation and evaluation of the first dog miRNA tissue atlas is described here.ResultsAnalysis of 16 tissues from five male beagle dogs identified 106 tissue enriched miRNAs, 60 of which were highly enriched in a single organ, and thus may serve as biomarkers of organ injury. A proof of concept study in dogs dosed with hepatotoxicants evaluated a qPCR panel of 15 tissue enriched miRNAs specific to liver, heart, skeletal muscle, pancreas, testes, and brain. Dogs with elevated serum levels of miR-122 and miR-885 had a correlative increase of alanine aminotransferase, and microscopic analysis confirmed liver damage. Other non-liver enriched miRNAs included in the screening panel were unaffected. Eli Lilly authors created a complimentary Sprague Dawely rat miRNA tissue atlas and demonstrated increased pancreas enriched miRNA levels in circulation, following caerulein administration in rat and dog.ConclusionThe dog miRNA tissue atlas provides a resource for biomarker discovery and can be further mined with refinement of dog genome annotation. The 60 highly enriched tissue miRNAs identified within the dog miRNA tissue atlas could serve as diagnostic biomarkers and will require further validation by in vivo correlation to histopathology. Once validated, these tissue enriched miRNAs could be combined into a powerful qPCR screening panel to identify organ toxicity during early drug development.

A small-molecule inhibitor of the ubiquitin activating enzyme for cancer treatment

The ubiquitin–proteasome system (UPS) comprises a network of enzymes that is responsible for maintaining cellular protein homeostasis. The therapeutic potential of this pathway has been validated by the clinical successes of a number of UPS modulators, including proteasome inhibitors and immunomodulatory imide drugs (IMiDs). Here we identified TAK-243 (formerly known as MLN7243) as a potent, mechanism-based small-molecule inhibitor of the ubiquitin activating enzyme (UAE), the primary mammalian E1 enzyme that regulates the ubiquitin conjugation cascade. TAK-243 treatment caused depletion of cellular ubiquitin conjugates, resulting in disruption of signaling events, induction of proteotoxic stress, and impairment of cell cycle progression and DNA damage repair pathways. TAK-243 treatment caused death of cancer cells and, in primary human xenograft studies, demonstrated antitumor activity at tolerated doses. Due to its specificity and potency, TAK-243 allows for interrogation of ubiquitin biology and for assessment of UAE inhibition as a new approach for cancer treatment.

KRAS Genotype Correlates with Proteasome Inhibitor Ixazomib Activity in Preclinical In Vivo Models of Colon and Non-Small Cell Lung Cancer: Potential Role of Tumor Metabolism

In non-clinical studies, the proteasome inhibitor ixazomib inhibits cell growth in a broad panel of solid tumor cell lines in vitro. In contrast, antitumor activity in xenograft tumors is model-dependent, with some solid tumors showing no response to ixazomib. In this study we examined factors responsible for ixazomib sensitivity or resistance using mouse xenograft models. A survey of 14 non-small cell lung cancer (NSCLC) and 6 colon xenografts showed a striking relationship between ixazomib activity and KRAS genotype; tumors with wild-type (WT) KRAS were more sensitive to ixazomib than tumors harboring KRAS activating mutations. To confirm the association between KRAS genotype and ixazomib sensitivity, we used SW48 isogenic colon cancer cell lines. Either KRAS-G13D or KRAS-G12V mutations were introduced into KRAS-WT SW48 cells to generate cells that stably express activated KRAS. SW48 KRAS WT tumors, but neither SW48-KRAS-G13D tumors nor SW48-KRAS-G12V tumors, were sensitive to ixazomib in vivo. Since activated KRAS is known to be associated with metabolic reprogramming, we compared metabolite profiling of SW48-WT and SW48-KRAS-G13D tumors treated with or without ixazomib. Prior to treatment there were significant metabolic differences between SW48 WT and SW48-KRAS-G13D tumors, reflecting higher oxidative stress and glucose utilization in the KRAS-G13D tumors. Ixazomib treatment resulted in significant metabolic regulation, and some of these changes were specific to KRAS WT tumors. Depletion of free amino acid pools and activation of GCN2-eIF2α-pathways were observed both in tumor types. However, changes in lipid beta oxidation were observed in only the KRAS WT tumors. The non-clinical data presented here show a correlation between KRAS genotype and ixazomib sensitivity in NSCLC and colon xenografts and provide new evidence of regulation of key metabolic pathways by proteasome inhibition.

A Sensitive IHC Method for Monitoring Autophagy-Specific Markers in Human Tumor Xenografts

Objective. Use of tyramide signal amplification (TSA) to detect autophagy biomarkers in formalin fixed and paraffin embedded (FFPE) xenograft tissue. Materials and Methods. Autophagy marker regulation was studied in xenograft tissues using Amp HQ IHC and standard IHC methods. Results. The data demonstrate the feasibility of using high sensitivity TSA IHC assays to measure low abundant autophagy markers in FFPE xenograft tissue.