Is the crowd better as an assistant or a replacement in ontology engineering? An exploration through the lens of the Gene Ontology.

Biomedical ontologies contain errors. Crowdsourcing, defined as taking a job traditionally performed by a designated agent and outsourcing it to an undefined large group of people, provides scalable access to humans. Therefore, the crowd has the potential to overcome the limited accuracy and scalability found in current ontology quality assurance approaches. Crowd-based methods have identified errors in SNOMED CT, a large, clinical ontology, with an accuracy similar to that of experts, suggesting that crowdsourcing is indeed a feasible approach for identifying ontology errors. This work uses that same crowd-based methodology, as well as a panel of experts, to verify a subset of the Gene Ontology (200 relationships). Experts identified 16 errors, generally in relationships referencing acids and metals. The crowd performed poorly in identifying those errors, with an area under the receiver operating characteristic curve ranging from 0.44 to 0.73, depending on the methods configuration. However, when the crowd verified what experts considered to be easy relationships with useful definitions, they performed reasonably well. Notably, there are significantly fewer Google search results for Gene Ontology concepts than SNOMED CT concepts. This disparity may account for the difference in performance - fewer search results indicate a more difficult task for the worker. The number of Internet search results could serve as a method to assess which tasks are appropriate for the crowd. These results suggest that the crowd fits better as an expert assistant, helping experts with their verification by completing the easy tasks and allowing experts to focus on the difficult tasks, rather than an expert replacement.

Distinct eRF3 requirements suggest alternate eRF1 conformations mediate peptide release during eukaryotic translation termination.

Organisms that use the standard genetic code recognize UAA, UAG, and UGA as stop codons, whereas variant code species frequently alter this pattern of stop codon recognition. We previously demonstrated that a hybrid eRF1 carrying the Euplotes octocarinatus domain 1 fused to Saccharomyces cerevisiae domains 2 and 3 (Eo/Sc eRF1) recognized UAA and UAG, but not UGA, as stop codons. In the current study, we identified mutations in Eo/Sc eRF1 that restore UGA recognition and define distinct roles for the TASNIKS and YxCxxxF motifs in eRF1 function. Mutations in or near the YxCxxxF motif support the cavity model for stop codon recognition by eRF1. Mutations in the TASNIKS motif eliminated the eRF3 requirement for peptide release at UAA and UAG codons, but not UGA codons. These results suggest that the TASNIKS motif and eRF3 function together to trigger eRF1 conformational changes that couple stop codon recognition and peptide release during eukaryotic translation termination.

Noninvasive molecular imaging of c-Myc activation in living mice.

The cytoplasmic Myc protein (c-Myc) regulates various human genes and is dysregulated in many human cancers. Phosphorylation mediates the protein activation of c-Myc and is essential for the function of this transcription factor in normal cell behavior and tumor growth. To date, however, the targeting of Myc as a therapeutic approach for cancer treatment has been achieved primarily at the nonprotein level. We have developed a molecular imaging sensor for noninvasive imaging of c-Myc activity in living subjects using a split Firefly luciferase (FL) complementation strategy to detect and quantify the phosphorylation-mediated interaction between glycogen synthase kinase 3beta (GSK3beta) and c-Myc. This sensor system consists of two fusion proteins, GSK 35-433-CFL and NFL-c-Myc, in which specific fragments of GSK3beta and c-Myc are fused with C-terminal and N-terminal fragments of the split FL, respectively. The sensor detects phosphorylation-specific GSK3beta-c-Myc interaction, the imaging signal of which correlates with the steady-state and temporal regulation of c-Myc phosphorylation in cell culture. The sensor also detects inhibition of c-Myc activity via differential pathways, allowing noninvasive monitoring of c-Myc-targeted drug efficacy in intact cells and living mice. Notably, this drug inhibition is detected before changes in tumor size are apparent in mouse xenograft and liver tumor models. This reporter system not only provides an innovative way to investigate the role of functional c-Myc in normal and cancer-related biological processes, but also facilitates c-Myc-targeted drug development by providing a rapid quantitative approach to assessing cancer response to therapy in living subjects.

Connection between stop codon reassignment and frequent use of shifty stop frameshifting.

Ciliated protozoa of the genus Euplotes have undergone genetic code reassignment, redefining the termination codon UGA to encode cysteine. In addition, Euplotes spp. genes very frequently employ shifty stop frameshifting. Both of these phenomena involve noncanonical events at a termination codon, suggesting they might have a common cause. We recently demonstrated that Euplotes octocarinatus peptide release factor eRF1 ignores UGA termination codons while continuing to recognize UAA and UAG. Here we show that both the Tetrahymena thermophila and E. octocarinatus eRF1 factors allow efficient frameshifting at all three termination codons, suggesting that UGA redefinition also impaired UAA/UAG recognition. Mutations of the Euplotes factor restoring a phylogenetically conserved motif in eRF1 (TASNIKS) reduced programmed frameshifting at all three termination codons. Mutation of another conserved residue, Cys124, strongly reduces frameshifting at UGA while actually increasing frameshifting at UAA/UAG. We will discuss these results in light of recent biochemical characterization of these mutations.

Eukaryotic ribosomal RNA determinants of aminoglycoside resistance and their role in translational fidelity.

Recent studies of prokaryotic ribosomes have dramatically increased our knowledge of ribosomal RNA (rRNA) structure, functional centers, and their interactions with antibiotics. However, much less is known about how rRNA function differs between prokaryotic and eukaryotic ribosomes. The core decoding sites are identical in yeast and human 18S rRNAs, suggesting that insights obtained in studies with yeast rRNA mutants can provide information about ribosome function in both species. In this study, we examined the importance of key nucleotides of the 18S rRNA decoding site on ribosome function and aminoglycoside susceptibility in Saccharomyces cerevisiae cells expressing homogeneous populations of mutant ribosomes. We found that residues G577, A1755, and A1756 (corresponding to Escherichia coli residues G530, A1492, and A1493, respectively) are essential for cell viability. We also found that residue G1645 (A1408 in E. coli) and A1754 (G1491 in E. coli) both make significant and distinct contributions to aminoglycoside resistance. Furthermore, we found that mutations at these residues do not alter the basal level of translational accuracy, but influence both paromomycin-induced misreading of sense codons and readthrough of stop codons. This study represents the most comprehensive mutational analysis of the eukaryotic decoding site to date, and suggests that many fundamental features of decoding site function are conserved between prokaryotes and eukaryotes.

Distinct paths to stop codon reassignment by the variant-code organisms Tetrahymena and Euplotes.

The reassignment of stop codons is common among many ciliate species. For example, Tetrahymena species recognize only UGA as a stop codon, while Euplotes species recognize only UAA and UAG as stop codons. Recent studies have shown that domain 1 of the translation termination factor eRF1 mediates stop codon recognition. While it is commonly assumed that changes in domain 1 of ciliate eRF1s are responsible for altered stop codon recognition, this has never been demonstrated in vivo. To carry out such an analysis, we made hybrid proteins that contained eRF1 domain 1 from either Tetrahymena thermophila or Euplotes octocarinatus fused to eRF1 domains 2 and 3 from Saccharomyces cerevisiae. We found that the Tetrahymena hybrid eRF1 efficiently terminated at all three stop codons when expressed in yeast cells, indicating that domain 1 is not the sole determinant of stop codon recognition in Tetrahymena species. In contrast, the Euplotes hybrid facilitated efficient translation termination at UAA and UAG codons but not at the UGA codon. Together, these results indicate that while domain 1 facilitates stop codon recognition, other factors can influence this process. Our findings also indicate that these two ciliate species used distinct approaches to diverge from the universal genetic code.

Advances in food allergy oral immunotherapy: toward tolerance.

The incidence of food allergy, a disease characterized by adverse immune responses that can render common foods life-threatening, is rising. Yet our current standard of care is simply avoidance of allergenic foods and administration of emergency medications upon accidental exposure. Significant advances have been made in food allergy oral immunotherapy, which is emerging as a potential preventive and curative treatment for this disease. The fundamental strategy of oral immunotherapy is to mitigate adverse immune responses to allergenic food proteins through repeated exposure; reduced reactivity to food allergens (desensitization) often results, but the establishment of sustained immune unresponsiveness or of permanent resolution (tolerance) is not certain. This review examines exciting recent developments in oral immunotherapy for food allergy.

Relating hepatocellular carcinoma tumor samples and cell lines using gene expression data in translational research.

Cancer cell lines are used extensively to study cancer biology and to test hypotheses in translational research. The relevance of cell lines is dependent on how closely they resemble the tumors being studied. Relating tumors and cell lines, and recognizing their similarities and differences are thus very important for translational research. Rapid advances in genomics have led to the generation of large volumes of genomic and transcriptomic data for a diverse set of primary cancer samples, normal tissue samples and cancer cell lines. Hepatocellular Carcinoma (HCC) is one of the most common tumors worldwide, with high occurrence in Asia and sub-Saharan regions. The current effective treatments of HCC remain limited. In this work, we compared the gene expression measurements of 200 HCC tumor samples from The Cancer Genome Atlas and over 1000 cancer cell lines including 25 HCC cancer cell lines from Cancer Cell Line Encyclopedia. We showed that the HCC tumor samples correlate closely with HCC cell lines in comparison to cell lines derived from other tumor types. We further demonstrated that the most commonly used HCC cell lines resemble HCC tumors, while we identified nearly half of the cell lines that do not resemble primary tumors. Interestingly, a substantial number of genes that are critical for disease development or drug response are either expressed at low levels or absent among highly correlated cell lines; additional attention should be paid to these genes in translational research. Our study will be used to guide the selection of HCC cell lines and pinpoint the specific genes that are differentially expressed in either tumors or cell lines.

A systematic assessment of linking gene expression with genetic variants for prioritizing candidate targets.

Gene expression and disease-associated variants are often used to prioritize candidate genes for target validation. However, the success of these gene features alone or in combination in the discovery of therapeutic targets is uncertain. Here we evaluated the effectiveness of the differential expression (DE), the disease-associated single nucleotide polymorphisms (SNPs) and the combination of the two in recovering and predicting known therapeutic targets across 56 human diseases. We demonstrate that the performance of each feature varies across diseases and generally the features have more recovery power than predictive power. The combination of the two features, however, has significantly higher predictive power than each feature alone. Our study provides a systematic evaluation of two common gene features, DE and SNPs, for prioritization of candidate targets and identified an improved predictive power of coupling these two features.

Extracting Computational and Semantic Features from Portable Chest X-rays for Diagnosis of Acute Respiratory Distress Syndrome.

Acute respiratory distress syndrome (ARDS) is a severe inflammatory lung disease with high mortality risk. Development of new and effective therapies for ARDS has been slow due to a lack of precision in its diagnostic criteria. We report preliminary research to extract computational and semantic features directly from chest X-ray images that are used to train machine learning classifiers. Our approach demonstrates the feasibility of using machine learning to identify radiographic criteria that are more consistent and accurate for the diagnosis of ARDS.

A c-Myc activation sensor-based high-throughput drug screening identifies an antineoplastic effect of nitazoxanide.

Deregulation of c-Myc plays a central role in the tumorigenesis of many human cancers. Yet, the development of drugs regulating c-Myc activity has been challenging. To facilitate the identification of c-Myc inhibitors, we developed a molecular imaging sensor-based high-throughput screening (HTS) system. This system uses a cell-based assay to detect c-Myc activation in a HTS format, which is established from a pure clone of a stable breast cancer cell line that constitutively expresses a c-Myc activation sensor. Optimization of the assay performance in the HTS format resulted in uniform and robust signals at the baseline. Using this system, we conducted a quantitative HTS against approximately 5,000 existing bioactive compounds from five different libraries. Thirty-nine potential hits were identified, including currently known c-Myc inhibitors. There are a few among the top potent hits that are not known for anti-c-Myc activity. One of these hits is nitazoxanide, a thiazolide for treating human protozoal infections. Validation of nitazoxanide in different cancer cell lines revealed a high potency for c-Myc inhibition with IC50 ranging between 10 and 500 nmol/L. Oral administration of nitazoxanide in breast cancer xenograft mouse models significantly suppressed tumor growth by inhibition of c-Myc and induction of apoptosis. These findings suggest a potential of nitazoxanide to be repurposed as a new antitumor agent for inhibition of c-Myc-associated neoplasia. Our work also demonstrated the unique advantage of molecular imaging in accelerating discovery of drugs for c-Myc-targeted cancer therapy.

Preclinical efficacy of the anti-hepatocyte growth factor antibody ficlatuzumab in a mouse brain orthotopic glioma model evaluated by bioluminescence, PET, and MRI.

PURPOSE: Ficlatuzumab is a novel therapeutic agent targeting the hepatocyte growth factor (HGF)/c-MET pathway. We summarize extensive preclinical work using this agent in a mouse brain orthotopic model of glioblastoma. EXPERIMENTAL DESIGN: Sequential experiments were done using eight- to nine-week-old nude mice injected with 3 × 10(5) U87 MG (glioblastoma) cells into the brain. Evaluation of ficlatuzumab dose response for this brain tumor model and comparison of its response to ficlatuzumab and to temozolamide were conducted first. Subsequently, various small-animal imaging modalities, including bioluminescence imaging (BLI), positron emission tomography (PET), and MRI, were used with a U87 MG-Luc 2 stable cell line, with and without the use of ficlatuzumab, to evaluate the ability to noninvasively assess tumor growth and response to therapy. ANOVA was conducted to evaluate for significant differences in the response. RESULTS: There was a survival benefit with ficlatuzumab alone or in combination with temozolamide. BLI was more sensitive than PET in detecting tumor cells. Fluoro-D-thymidine (FLT) PET provided a better signal-to-background ratio than 2[(18)F]fluoro-2-deoxy-d-glucose (FDG) PET. In addition, both BLI and FLT PET showed significant changes over time in the control group as well as with response to therapy. MRI does not disclose any time-dependent change. Also, the MRI results showed a temporal delay in comparison to the BLI and FLT PET findings, showing similar results one drug cycle later. CONCLUSIONS: Targeting the HGF/c-MET pathway with the novel agent ficlatuzumab appears promising for the treatment of glioblastoma. Various clinically applicable imaging modalities including FLT, PET, and MRI provide reliable ways of assessing tumor growth and response to therapy. Given the clinical applicability of these findings, future studies on patients with glioblastoma may be appropriate.

MYC phosphorylation, activation, and tumorigenic potential in hepatocellular carcinoma are regulated by HMG-CoA reductase.

MYC is a potential target for many cancers but is not amenable to existing pharmacologic approaches. Inhibition of 3-hydroxy-3-methylglutaryl-coenzyme A reductase (HMG-CoA reductase) by statins has shown potential efficacy against a number of cancers. Here, we show that inhibition of HMG-CoA reductase by atorvastatin (AT) blocks both MYC phosphorylation and activation, suppressing tumor initiation and growth in vivo in a transgenic model of MYC-induced hepatocellular carcinoma (HCC) as well as in human HCC-derived cell lines. To confirm specificity, we show that the antitumor effects of AT are blocked by cotreatment with the HMG-CoA reductase product mevalonate. Moreover, by using a novel molecular imaging sensor, we confirm that inhibition of HMG-CoA reductase blocks MYC phosphorylation in vivo. Importantly, the introduction of phosphorylation mutants of MYC at Ser62 or Thr58 into tumors blocks their sensitivity to inhibition of HMG-CoA reductase. Finally, we show that inhibition of HMG-CoA reductase suppresses MYC phosphorylation through Rac GTPase. Therefore, HMG-CoA reductase is a critical regulator of MYC phosphorylation, activation, and tumorigenic properties. The inhibition of HMG-CoA reductase may be a useful target for the treatment of MYC-associated HCC as well as other tumors.