A 3D lung lesion variational autoencoder.

In this study, we develop a 3D beta variational autoencoder (beta-VAE) to advance lung cancer imaging analysis, countering the constraints of conventional radiomics methods. The autoencoder extracts information from public lung computed tomography (CT) datasets without additional labels. It reconstructs 3D lung nodule images with high quality (structural similarity: 0.774, peak signal-to-noise ratio: 26.1, and mean-squared error: 0.0008). The model effectively encodes lesion sizes in its latent embeddings, with a significant correlation with lesion size found after applying uniform manifold approximation and projection (UMAP) for dimensionality reduction. Additionally, the beta-VAE can synthesize new lesions of varying sizes by manipulating the latent features. The model can predict multiple clinical endpoints, including pathological N stage or KRAS mutation status, on the Stanford radiogenomics lung cancer dataset. Comparisons with other methods show that the beta-VAE performs equally well in these tasks, suggesting its potential as a pretrained model for predicting patient outcomes in medical imaging.

Performance of alternative manual and automated deep learning segmentation techniques for the prediction of benign and malignant lung nodules.

Purpose: We aim to evaluate the performance of radiomic biopsy (RB), best-fit bounding box (BB), and a deep-learning-based segmentation method called no-new-U-Net (nnU-Net), compared to the standard full manual (FM) segmentation method for predicting benign and malignant lung nodules using a computed tomography (CT) radiomic machine learning model. Materials and Methods: A total of 188 CT scans of lung nodules from 2 institutions were used for our study. One radiologist identified and delineated all 188 lung nodules, whereas a second radiologist segmented a subset (n=20) of these nodules. Both radiologists employed FM and RB segmentation methods. BB segmentations were generated computationally from the FM segmentations. The nnU-Net, a deep-learning-based segmentation method, performed automatic nodule detection and segmentation. The time radiologists took to perform segmentations was recorded. Radiomic features were extracted from each segmentation method, and models to predict benign and malignant lung nodules were developed. The Kruskal-Wallis and DeLong tests were used to compare segmentation times and areas under the curve (AUC), respectively. Results: For the delineation of the FM, RB, and BB segmentations, the two radiologists required a median time (IQR) of 113 (54 to 251.5), 21 (9.25 to 38), and 16 (12 to 64.25) s, respectively (p=0.04). In dataset 1, the mean AUC (95% CI) of the FM, RB, BB, and nnU-Net model were 0.964 (0.96 to 0.968), 0.985 (0.983 to 0.987), 0.961 (0.956 to 0.965), and 0.878 (0.869 to 0.888). In dataset 2, the mean AUC (95% CI) of the FM, RB, BB, and nnU-Net model were 0.717 (0.705 to 0.729), 0.919 (0.913 to 0.924), 0.699 (0.687 to 0.711), and 0.644 (0.632 to 0.657). Conclusion: Radiomic biopsy-based models outperformed FM and BB models in prediction of benign and malignant lung nodules in two independent datasets while deep-learning segmentation-based models performed similarly to FM and BB. RB could be a more efficient segmentation method, but further validation is needed.

Topological data analysis of thoracic radiographic images shows improved radiomics-based lung tumor histology prediction.

Topological data analysis provides tools to capture wide-scale structural shape information in data. Its main method, persistent homology, has found successful applications to various machine-learning problems. Despite its recent gain in popularity, much of its potential for medical image analysis remains undiscovered. We explore the prominent learning problems on thoracic radiographic images of lung tumors for which persistent homology improves radiomic-based learning. It turns out that our topological features well capture complementary information important for benign versus malignant and adenocarcinoma versus squamous cell carcinoma tumor prediction while contributing less consistently to small cell versus non-small cell-an interesting result in its own right. Furthermore, while radiomic features are better for predicting malignancy scores assigned by expert radiologists through visual inspection, we find that topological features are better for predicting more accurate histology assessed through long-term radiology review, biopsy, surgical resection, progression, or response.

Machine Learning Radiomics Model for Early Identification of Small-Cell Lung Cancer on Computed Tomography Scans.

PURPOSE: Small-cell lung cancer (SCLC) is the deadliest form of lung cancer, partly because of its short doubling time. Delays in imaging identification and diagnosis of nodules create a risk for stage migration. The purpose of our study was to determine if a machine learning radiomics model can detect SCLC on computed tomography (CT) among all nodules at least 1 cm in size. MATERIALS AND METHODS: Computed tomography scans from a single institution were selected and resampled to 1 × 1 × 1 mm. Studies were divided into SCLC and other scans comprising benign, adenocarcinoma, and squamous cell carcinoma that were segregated into group A (noncontrast scans) and group B (contrast-enhanced scans). Four machine learning classification models, support vector classifier, random forest (RF), XGBoost, and logistic regression, were used to generate radiomic models using 59 quantitative first-order and texture Imaging Biomarker Standardization Initiative compliant PyRadiomics features, which were found to be robust between two segmenters with minimum Redundancy Maximum Relevance feature selection within each leave-one-out-cross-validation to avoid overfitting. The performance was evaluated using a receiver operating characteristic curve. A final model was created using the RF classifier and aggregate minimum Redundancy Maximum Relevance to determine feature importance. RESULTS: A total of 103 studies were included in the analysis. The area under the receiver operating characteristic curve for RF, support vector classifier, XGBoost, and logistic regression was 0.81, 0.77, 0.84, and 0.84 in group A, and 0.88, 0.87, 0.85, and 0.81 in group B, respectively. Nine radiomic features in group A and 14 radiomic features in group B were predictive of SCLC. Six radiomic features overlapped between groups A and B. CONCLUSION: A machine learning radiomics model may help differentiate SCLC from other lung lesions.

A meta-learning approach for genomic survival analysis.

RNA sequencing has emerged as a promising approach in cancer prognosis as sequencing data becomes more easily and affordably accessible. However, it remains challenging to build good predictive models especially when the sample size is limited and the number of features is high, which is a common situation in biomedical settings. To address these limitations, we propose a meta-learning framework based on neural networks for survival analysis and evaluate it in a genomic cancer research setting. We demonstrate that, compared to regular transfer-learning, meta-learning is a significantly more effective paradigm to leverage high-dimensional data that is relevant but not directly related to the problem of interest. Specifically, meta-learning explicitly constructs a model, from abundant data of relevant tasks, to learn a new task with few samples effectively. For the application of predicting cancer survival outcome, we also show that the meta-learning framework with a few samples is able to achieve competitive performance with learning from scratch with a significantly larger number of samples. Finally, we demonstrate that the meta-learning model implicitly prioritizes genes based on their contribution to survival prediction and allows us to identify important pathways in cancer.

Rational design and identification of immuno-oncology drug combinations.

BACKGROUND: Clinical trials investigating immuno-oncology (IO) drug combinations are largely based on empiricism or limited non-clinical evaluations. This study identified the current combination IO drug clinical trials and investigated how tumour molecular profiling can help rationalise IO drug combinations. METHODS: IO targets were identified via PubMed search and expert opinion. IO drugs were compiled by searching the National Cancer Institute Drug Dictionary and pharmaceutical pipelines, August 2016. Combination IO trials were obtained by searching doublet IO drug combinations in www.clinicaltrials.gov from September to November 2016. IO target gene expressions were extracted from The Cancer Genome Atlas (TCGA) data set and compared with normal tissues from the Genotype-Tissue Expression database. Differentially expressed genes for each cancer were determined using the Wilcoxon rank-sum test, and p-values were corrected for multiple testing. RESULTS: In total, 178 IO targets were identified; 90 targets have either regulatory approved or investigational therapeutics. In total, 410 combination trials involving ≥2 IO drugs were identified: skin (n = 102) and genitourinary (n = 41) malignancies have the largest number of combination IO trials; 109 trials involved >2 disease sites. Summative patient accrual estimates among all trials are 71,345. Trials combining cytotoxic T lymphocyte antigen 4 (CTLA4) with programmed cell death protein 1 (n = 79) and CTLA4 with programmed cell death ligand 1 (n = 44) are the most common. Gene expression data from TCGA were mined to extract the 178 IO targets in 9089 tumours originating from 19 cancer types. IO target expression-clustered heatmap analysis identified several promising drug combinations. CONCLUSION: Our review highlights the great interest in combination IO clinical trials. Our analysis can enrich IO combination therapy selection.

Antitumor activity of the polo-like kinase inhibitor, TAK-960, against preclinical models of colorectal cancer.

BACKGROUND: Polo-like kinase 1 (Plk1) is a serine/threonine kinase that is a key regulator of multiple stages of mitotic progression. Plk1 is upregulated in many tumor types including colorectal cancer (CRC) and portends a poor prognosis. TAK-960 is an ATP-competitive Plk1 inhibitor that has demonstrated efficacy across a broad range of cancer cell lines, including CRC. In this study, we investigated the activity of TAK-960 against a large collection of CRC models including 55 cell lines and 18 patient-derived xenografts. METHODS: Fifty-five CRC cell lines and 18 PDX models were exposed to TAK-960 and evaluated for proliferation (IC50) and Tumor Growth Inhibition Index, respectively. Additionally, 2 KRAS wild type and 2 KRAS mutant PDX models were treated with TAK-960 as single agent or in combination with cetuximab or irinotecan. TAK-960 mechanism of action was elucidated through immunoblotting and cell cycle analysis. RESULTS: CRC cell lines demonstrated a variable anti-proliferative response to TAK-960 with IC50 values ranging from 0.001 to > 0.75 µmol/L. Anti-proliferative effects were sustained after removal of drug. Following TAK-960 treatment a highly variable accumulation of mitotic (indicating cell cycle arrest) and apoptotic markers was observed. Cell cycle analysis demonstrated that TAK-960 treatment induced G2/M arrest and polyploidy. Six out of the eighteen PDX models responded to single agent TAK-960 therapy (TGII< 20). The addition of TAK-960 to standard of care chemotherapy resulted in largely additive antitumor effects. CONCLUSION: TAK-960 is an active anti-proliferative agent against CRC cell lines and PDX models. Collectively, these data suggest that TAK-960 may be of therapeutic benefit alone or in combination with other agents, although future work should focus on the development of predictive biomarkers and hypothesis-driven rational combinations.

Targeting the protein ubiquitination machinery in melanoma by the NEDD8-activating enzyme inhibitor pevonedistat (MLN4924).

Background The neddylation pathway conjugates NEDD8 to cullin-RING ligases and controls the proteasomal degradation of specific proteins involved in essential cell processes. Pevonedistat (MLN4924) is a selective small molecule targeting the NEDD8-activating enzyme (NAE) and inhibits an early step in neddylation, resulting in DNA re-replication, cell cycle arrest and death. We investigated the anti-tumor potential of pevonedistat in preclinical models of melanoma. Methods Melanoma cell lines and patient-derived tumor xenografts (PDTX) treated with pevonedistat were assessed for viability/apoptosis and tumor growth, respectively, to identify sensitive/resistant models. Gene expression microarray and gene set enrichment analyses were performed in cell lines to determine the expression profiles and pathways of sensitivity/resistance. Pharmacodynamic changes in treated-PDTX were also characterized. Results Pevonedistat effectively inhibited cell viability (IC50 < 0.3 µM) and induced apoptosis in a subset of melanoma cell lines. Sensitive and resistant cell lines exhibited distinct gene expression profiles; sensitive models were enriched for genes involved in DNA repair, replication and cell cycle regulation, while immune response and cell adhesion pathways were upregulated in resistant models. Pevonedistat also reduced tumor growth in melanoma cell line xenografts and PDTX with variable responses. An accumulation of pevonedistat-NEDD8 adduct and CDT1 was observed in sensitive tumors consistent with its mechanism of action. Conclusions This study provided preclinical evidence that NAE inhibition by pevonedistat has anti-tumor activity in melanoma and supports the clinical benefits observed in recent Phase 1 trials of this drug in melanoma patients. Further investigations are warranted to develop rational combinations and determine predictive biomarkers of pevonedistat.

BatchQC: interactive software for evaluating sample and batch effects in genomic data.

Sequencing and microarray samples often are collected or processed in multiple batches or at different times. This often produces technical biases that can lead to incorrect results in the downstream analysis. There are several existing batch adjustment tools for '-omics' data, but they do not indicate a priori whether adjustment needs to be conducted or how correction should be applied. We present a software pipeline, BatchQC, which addresses these issues using interactive visualizations and statistics that evaluate the impact of batch effects in a genomic dataset. BatchQC can also apply existing adjustment tools and allow users to evaluate their benefits interactively. We used the BatchQC pipeline on both simulated and real data to demonstrate the effectiveness of this software toolkit. AVAILABILITY AND IMPLEMENTATION: BatchQC is available through Bioconductor: http://bioconductor.org/packages/BatchQC and GitHub: https://github.com/mani2012/BatchQC CONTACT: wej@bu.eduSupplementary information: Supplementary data are available at Bioinformatics online.

Antitumor activity of the aurora a selective kinase inhibitor, alisertib, against preclinical models of colorectal cancer.

BACKGROUND: The Aurora kinases are a family of serine/threonine kinases comprised of Aurora A, B, and C which execute critical steps in mitotic and meiotic progression. Alisertib (MLN8237) is an investigational Aurora A selective inhibitor that has demonstrated activity against a wide variety of tumor types in vitro and in vivo, including CRC. RESULTS: CRC cell lines demonstrated varying sensitivity to alisertib with IC50 values ranging from 0.06 to > 5 umol/L. Following exposure to alisertib we observed a decrease in pAurora A, B and C in four CRC cell lines. We also observed an increase in p53 and p21 in a sensitive p53 wildtype cell line in contrast to the p53 mutant cell line or the resistant cell lines. The addition of alisertib to standard CRC treatments demonstrated improvement over single agent arms; however, the benefit was largely less than additive, but not antagonistic. METHODS: Forty-seven CRC cell lines were exposed to alisertib and IC50s were calculated. Twenty-one PDX models were treated with alisertib and the Tumor Growth Inhibition Index was assessed. Additionally, 5 KRAS wildtype and mutant PDX models were treated with alisertib as single agent or in combination with cetuximab or irinotecan, respectively. CONCLUSION: Alisertib demonstrated anti-proliferative effects against CRC cell lines and PDX models. Our data suggest that the addition of alisertib to standard therapies in colorectal cancer if pursued clinically, will require further investigation of patient selection strategies and these combinations may facilitate future clinical studies.

PharmacoGx: an R package for analysis of large pharmacogenomic datasets.

UNLABELLED: Pharmacogenomics holds great promise for the development of biomarkers of drug response and the design of new therapeutic options, which are key challenges in precision medicine. However, such data are scattered and lack standards for efficient access and analysis, consequently preventing the realization of the full potential of pharmacogenomics. To address these issues, we implemented PharmacoGx, an easy-to-use, open source package for integrative analysis of multiple pharmacogenomic datasets. We demonstrate the utility of our package in comparing large drug sensitivity datasets, such as the Genomics of Drug Sensitivity in Cancer and the Cancer Cell Line Encyclopedia. Moreover, we show how to use our package to easily perform Connectivity Map analysis. With increasing availability of drug-related data, our package will open new avenues of research for meta-analysis of pharmacogenomic data. AVAILABILITY AND IMPLEMENTATION: PharmacoGx is implemented in R and can be easily installed on any system. The package is available from CRAN and its source code is available from GitHub. CONTACT: bhaibeka@uhnresearch.ca or benjamin.haibe.kains@utoronto.ca SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Antitumor activity of a potent MEK inhibitor, TAK-733, against colorectal cancer cell lines and patient derived xenografts.

BACKGROUND: CRC is a significant cause of cancer mortality, and new therapies are needed for patients with advanced disease. TAK-733 is a highly potent and selective investigational novel MEK allosteric site inhibitor. MATERIALS AND METHODS: In a preclinical study of TAK-733, a panel of CRC cell lines were exposed to varying concentrations of the agent for 72 hours followed by a sulforhodamine B assay. Twenty patient-derived colorectal cancer xenografts were then treated with TAK-733 in vivo. Tumor growth inhibition index (TGII) was assessed to evaluate the sensitivity of the CRC explants to TAK-733 while linear regression was utilized to investigate the predictive effects of genotype on the TGII of explants. RESULTS: Fifty-four CRC cell lines were exposed to TAK-733, while 42 cell lines were deemed sensitive across a broad range of mutations. Eighty-two percent of the cell lines within the sensitive subset were BRAF or KRAS/NRAS mutant, whereas 80% of the cell lines within the sensitive subset were PIK3CA WT. Twenty patient-derived human tumor CRC explants were then treated with TAK-733. In total, 15 primary human tumor explants were found to be sensitive to TAK-733 (TGII ≤ 20%), including 9 primary human tumor explants that exhibited tumor regression (TGII > 100%). Explants with a BRAF/KRAS/NRAS mutant and PIK3CA wild-type genotype demonstrated increased sensitivity to TAK-733 with a median TGII of -6%. MEK-response gene signatures also correlated with responsiveness to TAK-733 in KRAS-mutant CRC. CONCLUSIONS: The MEK inhibitor TAK-733 demonstrated robust antitumor activity against CRC cell lines and patient-derived tumor explants. While the preclinical activity observed in this study was considerable, single-agent efficacy in the clinic has been limited in CRC, supporting the use of these models in an iterative manner to elucidate resistance mechanisms that can guide rational combination strategies.

Combined inhibition of MEK and Aurora A kinase in KRAS/PIK3CA double-mutant colorectal cancer models.

Aurora A kinase and MEK inhibitors induce different, and potentially complementary, effects on the cell cycle of malignant cells, suggesting a rational basis for utilizing these agents in combination. In this work, the combination of an Aurora A kinase and MEK inhibitor was evaluated in pre-clinical colorectal cancer models, with a focus on identifying a subpopulation in which it might be most effective. Increased synergistic activity of the drug combination was identified in colorectal cancer cell lines with concomitant KRAS and PIK3CA mutations. Anti-proliferative effects were observed upon treatment of these double-mutant cell lines with the drug combination, and tumor growth inhibition was observed in double-mutant human tumor xenografts, though effects were variable within this subset. Additional evaluation suggests that degree of G2/M delay and p53 mutation status affect apoptotic activity induced by combination therapy with an Aurora A kinase and MEK inhibitor in KRAS and PIK3CA mutant colorectal cancer. Overall, in vitro and in vivo testing was unable to identify a subset of colorectal cancer that was consistently responsive to the combination of a MEK and Aurora A kinase inhibitor.

Antitumor activity of the MEK inhibitor TAK-733 against melanoma cell lines and patient-derived tumor explants.

The goal of this study was to investigate the activity of the selective MEK1/2 inhibitor TAK-733 in both melanoma cell lines and patient-derived melanoma xenograft models. In vitro cell proliferation assays using the sulforhodamine B assay were conducted to determine TAK-733 potency and melanoma responsiveness. In vivo murine modeling with eleven patient-derived melanoma explants evaluated daily dosing of TAK-733 at 25 or 10 mg/kg. Immunoblotting was performed to evaluate on-target activity and downstream inhibition by TAK-733 in both in vitro and in vivo studies. TAK-733 demonstrated broad activity in most melanoma cell lines with relative resistance observed at IC50 > 0.1 µmol/L in vitro. TAK-733 also exhibited activity in 10 out of 11 patient-derived explants with tumor growth inhibition ranging from 0% to 100% (P < 0.001-0.03). Interestingly, BRAF(V600E) and NRAS mutational status did not correlate with responsiveness to TAK-733. Pharmacodynamically, pERK was suppressed in sensitive cell lines and tumor explants, confirming TAK-733-mediated inhibition of MEK1/2, although the demonstration of similar effects in the relatively resistant cell lines and tumor explants suggests that escape pathways are contributing to melanoma survival and proliferation. These data demonstrate that TAK-733 exhibits robust tumor growth inhibition and regression against human melanoma cell lines and patient-derived xenograft models, suggesting that further clinical development in melanoma is of scientific interest. Particularly interesting is the activity in BRAF wild-type models, where current approved therapy such as vemurafenib has been reported not to be active.

Overcoming IGF1R/IR resistance through inhibition of MEK signaling in colorectal cancer models.

PURPOSE: Results from clinical trials involving resistance to molecularly targeted therapies have revealed the importance of rational single-agent and combination treatment strategies. In this study, we tested the efficacy of a type 1 insulin-like growth factor receptor (IGF1R)/insulin receptor (IR) tyrosine kinase inhibitor, OSI-906, in combination with a mitogen-activated protein (MAP)-ERK kinase (MEK) 1/2 inhibitor based on evidence that the MAP kinase pathway was upregulated in colorectal cancer cell lines that were resistant to OSI-906. EXPERIMENTAL DESIGN: The antiproliferative effects of OSI-906 and the MEK 1/2 inhibitor U0126 were analyzed both as single agents and in combination in 13 colorectal cancer cell lines in vitro. Apoptosis, downstream effector proteins, and cell cycle were also assessed. In addition, the efficacy of OSI-906 combined with the MEK 1/2 inhibitor selumetinib (AZD6244, ARRY-142886) was evaluated in vivo using human colorectal cancer xenograft models. RESULTS: The combination of OSI-906 and U0126 resulted in synergistic effects in 11 of 13 colorectal cancer cell lines tested. This synergy was variably associated with apoptosis or cell-cycle arrest in addition to molecular effects on prosurvival pathways. The synergy was also reflected in the in vivo xenograft studies following treatment with the combination of OSI-906 and selumetinib. CONCLUSIONS: Results from this study demonstrate synergistic antiproliferative effects in response to the combination of OSI-906 with an MEK 1/2 inhibitor in colorectal cancer cell line models both in vitro and in vivo, which supports the rational combination of OSI-906 with an MEK inhibitor in patients with colorectal cancer. Clin Cancer Res; 19(22); 6219-29. ©2013 AACR.

Rational combination of a MEK inhibitor, selumetinib, and the Wnt/calcium pathway modulator, cyclosporin A, in preclinical models of colorectal cancer.

PURPOSE: The mitogen-activated protein kinase (MAPK) pathway is a crucial regulator of cell proliferation, survival, and resistance to apoptosis. MEK inhibitors are being explored as a treatment option for patients with KRAS-mutant colorectal cancer who are not candidates for EGFR-directed therapies. Initial clinical results of MEK inhibitors have yielded limited single-agent activity in colorectal cancer, indicating that rational combination strategies are needed. EXPERIMENTAL DESIGN: In this study, we conducted unbiased gene set enrichment analysis and synthetic lethality screens with selumetinib, which identified the noncanonical Wnt/Ca++ signaling pathway as a potential mediator of resistance to the MEK1/2 inhibitor selumetinib. To test this, we used shRNA constructs against relevant WNT receptors and ligands resulting in increased responsiveness to selumetinib in colorectal cancer cell lines. Further, we evaluated the rational combination of selumetinib and WNT pathway modulators and showed synergistic antiproliferative effects in in vitro and in vivo models of colorectal cancer. RESULTS: Importantly, this combination not only showed tumor growth inhibition but also tumor regression in the more clinically relevant patient-derived tumor explant (PDTX) models of colorectal cancer. In mechanistic studies, we observed a trend toward increased markers of apoptosis in response to the combination of MEK and WntCa(++) inhibitors, which may explain the observed synergistic antitumor effects. CONCLUSIONS: These results strengthen the hypothesis that targeting both the MEK and Wnt pathways may be a clinically effective rational combination strategy for patients with metastatic colorectal cancer.

Association of the epithelial-to-mesenchymal transition phenotype with responsiveness to the p21-activated kinase inhibitor, PF-3758309, in colon cancer models.

The p21-activated kinase (PAK) family of serine/threonine kinases, which are overexpressed in several cancer types, are critical mediators of cell survival, motility, mitosis, transcription, and translation. In the study presented here, we utilized a panel of colorectal cancer (CRC) cell lines to identify potential biomarkers of sensitivity or resistance that may be used to individualize therapy to the PAK inhibitor PF-03758309. We observed a wide range of proliferative responses in the CRC cell lines exposed to PF-03758309, this response was recapitulated in other phenotypic assays such as anchorage-independent growth, three-dimensional (3D) tumor spheroid formation, and migration. Interestingly, we observed that cells most sensitive to PF-03758309 exhibited up-regulation of genes associated with a mesenchymal phenotype (CALD1, VIM, ZEB1) and cells more resistant had an up-regulation of genes associated with an epithelial phenotype (CLDN2, CDH1, CLDN3, CDH17) allowing us to derive an epithelial-to-mesenchymal transition (EMT) gene signature for this agent. We assessed the functional role of EMT-associated genes in mediating responsiveness to PF-3758309, by targeting known genes and transcriptional regulators of EMT. We observed that suppression of genes associated with the mesenchymal phenotype conferred resistance to PF-3758309, in vitro and in vivo. These results indicate that PAK inhibition is associated with a unique response phenotype in CRC and that further studies should be conducted to facilitate both patient selection and rational combination strategies with these agents.

Utilization of quantitative in vivo pharmacology approaches to assess combination effects of everolimus and irinotecan in mouse xenograft models of colorectal cancer.

PURPOSE: The PI3K/AKT/mTOR pathway is frequently dysregulated in cancers and inhibition of mTOR has demonstrated the ability to modulate pro-survival pathways. As such, we sought to determine the ability of the mTOR inhibitor everolimus to potentiate the antitumor effects of irinotecan in colorectal cancer (CRC). EXPERIMENTAL DESIGN: The combinatorial effects of everolimus and irinotecan were evaluated in vitro and in vivo in CRC cell lines harboring commonly found mutations in PIK3CA, KRAS and/or BRAF. Pharmacokinetically-directed dosing protocols of everolimus and irinotecan were established and used to assess the in vivo antitumor effects of the agents. At the end of treatment, 3-6 tumors per treatment arm were harvested for biomarker analysis by NMR metabolomics. RESULTS: Everolimus and irinotecan/SN38 demonstrated synergistic anti-proliferative effects in multiple CRC cell lines in vitro. Combination effects of everolimus and irinotecan were determined in CRC xenograft models using clinically-relevant dosing protocols. Everolimus demonstrated significant tumor growth inhibition alone and when combined with irinotecan in HT29 and HCT116 tumor xenografts. Metabolomic analysis showed that HT29 tumors were more metabolically responsive than HCT116 tumors. Everolimus caused a decrease in glycolysis in both tumor types whilst irinotecan treatment resulted in a profound accumulation of lipids in HT29 tumors indicating a cytotoxic effect. CONCLUSIONS: Quantitative analysis of tumor growth and metabolomic data showed that the combination of everolimus and irinotecan was more beneficial in the BRAF/PIK3CA mutant HT29 tumor xenografts, which had an additive effect, than the KRAS/PIK3CA mutant HCT116 tumor xenografts, which had a less than additive effect.

Predictive biomarkers of sensitivity to the aurora and angiogenic kinase inhibitor ENMD-2076 in preclinical breast cancer models.

PURPOSE: The Aurora kinases are a family of conserved serine-threonine kinases with key roles in mitotic cell division. As with other promising anticancer targets, patient selection strategies to identify a responsive subtype will likely be required for successful clinical development of Aurora kinase inhibitors. The purpose of this study was to evaluate the antitumor activity of the Aurora and angiogenic kinase inhibitor ENMD-2076 against preclinical models of breast cancer with identification of candidate predictive biomarkers. EXPERIMENTAL DESIGN: Twenty-nine breast cancer cell lines were exposed to ENMD-2076 and the effects on proliferation, apoptosis, and cell-cycle distribution were evaluated. In vitro activity was confirmed in MDA-MB-468 and MDA-MB-231 triple-negative breast cancer xenografts. Systematic gene expression analysis was used to identify up- and downregulated pathways in the sensitive and resistant cell lines, including within the triple-negative breast cancer subset. RESULTS: ENMD-2076 showed antiproliferative activity against breast cancer cell lines, with more robust activity against cell lines lacking estrogen receptor expression and those without increased HER2 expression. Within the triple-negative breast cancer subset, cell lines with a p53 mutation and increased p53 expression were more sensitive to the cytotoxic and proapoptotic effects of ENMD-2076 exposure than cell lines with decreased p53 expression. CONCLUSIONS: ENMD-2076 exhibited robust anticancer activity against models of triple-negative breast cancer and the candidate predictive biomarkers identified in this study may be useful in selecting patients for Aurora kinase inhibitors in the future.

A short survey of computational analysis methods in analysing ChIP-seq data.

Chromatin immunoprecipitation followed by massively parallel next-generation sequencing (ChIP-seq) is a valuable experimental strategy for assaying protein-DNA interaction over the whole genome. Many computational tools have been designed to find the peaks of the signals corresponding to protein binding sites. In this paper, three computational methods, ChIP-seq processing pipeline (spp), PeakSeq and CisGenome, used in ChIP-seq data analysis are reviewed. There is also a comparison of how they agree and disagree on finding peaks using the publically available Signal Transducers and Activators of Transcription protein 1 (STAT1) and RNA polymerase II (PolII) datasets with corresponding negative controls.