Limited agreement of independent RNAi screens for virus-required host genes owes more to false-negative than false-positive factors

Systematic review and meta-analysis of genome-wide pooled CRISPR screens to identify host factors involved in influenza A virus infection

The host-virus interactome is increasingly recognized as an important research field to discover new therapeutic targets to treat influenza. Multiple pooled genome-wide CRISPR-Cas screens have been reported to identify new pro- and antiviral host factors of the influenza A virus. However, at present, a comprehensive summary of the results is lacking. We performed a systematic review of all reported CRISPR studies in this field in combination with a meta-analysis using the algorithm of meta-analysis by information content (MAIC). Two ranked gene lists were generated based on evidence in 15 proviral and 4 antiviral screens. Enriched pathways in the proviral MAIC results were compared to those of a prior array-based RNA interference (RNAi) meta-analysis. The top 50 proviral MAIC list contained genes whose role requires further elucidation, such as the endosomal ion channel TPCN1 and the kinase WEE1. Moreover, MAIC indicated that ALYREF, a component of the transcription export complex, has antiviral properties, whereas former knockdown experiments attributed a proviral role to this host factor. CRISPR-Cas-pooled screens displayed a bias toward early-replication events, whereas the prior RNAi meta-analysis covered early and late-stage events. RNAi screens led to the identification of a larger fraction of essential genes than CRISPR screens. In summary, the MAIC algorithm points toward the importance of several less well-known pathways in host-influenza virus interactions that merit further investigation. The results from this meta-analysis of CRISPR screens in influenza A virus infection may help guide future research efforts to develop host-directed anti-influenza drugs.

IMPORTANCE

Viruses rely on host factors for their replication, whereas the host cell has evolved virus restriction factors. These factors represent potential targets for host-oriented antiviral therapies. Multiple pooled genome-wide CRISPR-Cas screens have been reported to identify pro- and antiviral host factors in the context of influenza virus infection. We performed a comprehensive analysis of the outcome of these screens based on the publicly available gene lists, using the recently developed algorithm meta-analysis by information content (MAIC). MAIC allows the systematic integration of ranked and unranked gene lists into a final ranked gene list. This approach highlighted poorly characterized host factors and pathways with evidence from multiple screens, such as the vesicle docking and lipid metabolism pathways, which merit further exploration.

Pathogen-driven CRISPR screens identify TREX1 as a regulator of DNA self-sensing during influenza virus infection

Host:pathogen interactions dictate the outcome of infection, yet the limitations of current approaches leave large regions of this interface unexplored. Here, we develop a novel fitness-based screen that queries factors important during the middle to late stages of infection. This is achieved by engineering influenza virus to direct the screen by programming dCas9 to modulate host gene expression. Our genome-wide screen for pro-viral factors identifies the cytoplasmic DNA exonuclease TREX1. TREX1 degrades cytoplasmic DNA to prevent inappropriate innate immune activation by self-DNA. We reveal that this same process aids influenza virus replication. Infection triggers release of mitochondrial DNA into the cytoplasm, activating antiviral signaling via cGAS and STING. TREX1 metabolizes the DNA, preventing its sensing. Collectively, these data show that self-DNA is deployed to amplify innate immunity, a process tempered by TREX1. Moreover, they demonstrate the power and generality of pathogen-driven fitness-based screens to pinpoint key host regulators of infection.

Pathogen-driven CRISPR screens identify TREX1 as a regulator of DNA self-sensing during influenza virus infection

Intracellular pathogens interact with host factors, exploiting those that enhance replication while countering those that suppress it. Genetic screens have begun to define the host:pathogen interface and establish a mechanistic basis for host-directed therapies. Yet, limitations of current approaches leave large regions of this interface unexplored. To uncover host factors with pro-pathogen functions, we developed a novel fitness-based screen that queries factors important during the middle-to-late stages of infection. This was achieved by engineering influenza virus to direct the screen by programing dCas9 to modulate host gene expression. A genome-wide screen identified the cytoplasmic DNA exonuclease TREX1 as a potent pro-viral factor. TREX1 normally degrades cytoplasmic DNA to prevent inappropriate innate immune activation by self DNA. Our mechanistic studies revealed that this same process functions during influenza virus infection to enhance replication. Infection triggered release of mitochondrial DNA into the cytoplasm, activating antiviral signaling via cGAS and STING. TREX1 metabolized the mitochondrial DNA preventing its sensing. Collectively, these data show that self-DNA is deployed to amplify host innate sensing during RNA virus infection, a process tempered by TREX1. Moreover, they demonstrate the power and generality of pathogen driven fitness-based screens to pinpoint key host regulators of intracellular pathogens.

Sequence-Specific Features of Short Double-Strand, Blunt-End RNAs Have RIG-I- and Type 1 Interferon-Dependent or -Independent Anti-Viral Effects

Pathogen-associated molecular patterns, including cytoplasmic DNA and double-strand (ds)RNA trigger the induction of interferon (IFN) and antiviral states protecting cells and organisms from pathogens. Here we discovered that the transfection of human airway cell lines or non-transformed fibroblasts with 24mer dsRNA mimicking the cellular micro-RNA (miR)29b-1* gives strong anti-viral effects against human adenovirus type 5 (AdV-C5), influenza A virus X31 (H3N2), and SARS-CoV-2. These anti-viral effects required blunt-end complementary RNA strands and were not elicited by corresponding single-strand RNAs. dsRNA miR-29b-1* but not randomized miR-29b-1* mimics induced IFN-stimulated gene expression, and downregulated cell adhesion and cell cycle genes, as indicated by transcriptomics and IFN-I responsive Mx1-promoter activity assays. The inhibition of AdV-C5 infection with miR-29b-1* mimic depended on the IFN-alpha receptor 2 (IFNAR2) and the RNA-helicase retinoic acid-inducible gene I (RIG-I) but not cytoplasmic RNA sensors MDA5 and ZNFX1 or MyD88/TRIF adaptors. The antiviral effects of miR29b-1* were independent of a central AUAU-motif inducing dsRNA bending, as mimics with disrupted AUAU-motif were anti-viral in normal but not RIG-I knock-out (KO) or IFNAR2-KO cells. The screening of a library of scrambled short dsRNA sequences identified also anti-viral mimics functioning independently of RIG-I and IFNAR2, thus exemplifying the diverse anti-viral mechanisms of short blunt-end dsRNAs.

False and true positives in arthropod thermal adaptation candidate gene lists

Genome-wide studies are prone to false positives due to inherently low priors and statistical power. One approach to ameliorate this problem is to seek validation of reported candidate genes across independent studies: genes with repeatedly discovered effects are less likely to be false positives. Inversely, genes reported only as many times as expected by chance alone, while possibly representing novel discoveries, are also more likely to be false positives. We show that, across over 30 genome-wide studies that reported Drosophila and Daphnia genes with possible roles in thermal adaptation, the combined lists of candidate genes and orthologous groups are rapidly approaching the total number of genes and orthologous groups in the respective genomes. This is consistent with the expectation of high frequency of false positives. The majority of these spurious candidates have been identified by one or a few studies, as expected by chance alone. In contrast, a noticeable minority of genes have been identified by numerous studies with the probabilities of such discoveries occurring by chance alone being exceedingly small. For this subset of genes, different studies are in agreement with each other despite differences in the ecological settings, genomic tools and methodology, and reporting thresholds. We provide a reference set of presumed true positives among Drosophila candidate genes and orthologous groups involved in response to changes in temperature, suitable for cross-validation purposes. Despite this approach being prone to false negatives, this list of presumed true positives includes several hundred genes, consistent with the "omnigenic" concept of genetic architecture of complex traits.

SIGNAL: A web-based iterative analysis platform integrating pathway and network approaches optimizes hit selection from genome-scale assays

Hit selection from high-throughput assays remains a critical bottleneck in realizing the potential of omic-scale studies in biology. Widely used methods such as setting of cutoffs, prioritizing pathway enrichments, or incorporating predicted network interactions offer divergent solutions yet are associated with critical analytical trade-offs. The specific limitations of these individual approaches and the lack of a systematic way by which to integrate their rankings have contributed to limited overlap in the reported results from comparable genome-wide studies and costly inefficiencies in secondary validation efforts. Using comparative analysis of parallel independent studies as a benchmark, we characterize the specific complementary contributions of each approach and demonstrate an optimal framework to integrate these methods. We describe selection by iterative pathway group and network analysis looping (SIGNAL), an integrated, iterative approach that uses both pathway and network methods to optimize gene prioritization. SIGNAL is accessible as a rapid user-friendly web-based application (https://signal.niaid.nih.gov). A record of this paper's transparent peer review is included in the Supplemental information.

Antibody profiling of patients with prostate cancer reveals differences in antibody signatures among disease stages

BACKGROUND

Previous studies of prostate cancer autoantibodies have largely focused on diagnostic applications. So far, there have been no reports attempting to more comprehensively profile the landscape of prostate cancer-associated antibodies. Specifically, it is unknown whether the quantity of antibodies or the types of proteins recognized change with disease progression.

METHODS

A peptide microarray spanning the amino acid sequences of the gene products of 1611 prostate cancer-associated genes was synthesized. Serum samples from healthy male volunteers (n=15) and patients with prostate cancer (n=85) were used to probe the array. These samples included patients with various clinical stages of disease: newly diagnosed localized prostate cancer (n=15), castration-sensitive non-metastatic prostate cancer (nmCSPC, n=40), castration-resistant non-metastatic prostate cancer (n=15) and castration-resistant metastatic disease (n=15). The patients with nmCSPC received treatment with either standard androgen deprivation therapy (ADT) or an antitumor DNA vaccine encoding prostatic acid phosphatase. Serial sera samples from these individuals were also used to probe the array, to secondarily determine whether this approach could be used to detect treatment-related changes.

RESULTS

We demonstrated that this peptide array yielded highly reproducible measurements of serum IgG levels. We found that the overall number of antibody responses did not increase with disease burden. However, the composition of recognized proteins shifted with clinical stage of disease. Our analysis revealed that the largest difference was between patients with castration-sensitive and castration-resistant disease. Patients with castration-resistant disease recognized more proteins associated with nucleic acid binding and gene regulation compared with men in other groups. Our longitudinal data showed that treatments can elicit antibodies detectable by this array, and notably vaccine-treated patients developed increased responses to more proteins over the course of treatment than did ADT-treated patients.

CONCLUSIONS

This study represents the largest survey of prostate cancer-associated antibodies to date. We have been able to characterize the classes of proteins recognized by patients and determine how they change with disease burden. Our findings further demonstrate the potential of this platform for measuring antigen spread and studying responses to immunomodulatory therapies.

Experimental Approaches to Identify Host Factors Important for Influenza Virus

An ever-expanding toolkit of experimental methods provides the means to discover and characterize host factors important for influenza virus. Here, we describe common methods for investigating genetic relationships and physical interactions between virus and host. A comprehensive knowledge of host:virus interactions is key to understanding how influenza virus exploits the host cell and to potentially identify vulnerabilities that may be manipulated to prevent or treat disease.

Human Gene Functional Network-Informed Prediction of HIV-1 Host Dependency Factors

Identification of HIV-1 HDFs remains a crucial step to understand the complicated relationships between human and HIV-1. To complement the experimental identification of HDFs, we have implemented an existing network-based gene discovery strategy to predict HDFs from the human genome. The core idea of the proposed method is that the rich information deposited in host gene functional networks can be effectively utilized to infer the potential HDFs. We hope the proposed prediction method could further guide hypothesis-driven experimental efforts to interrogate human–HIV-1 relationships and provide new hints for the development of antiviral drugs to combat HIV-1 infection.

Human immunodeficiency virus type 1 (HIV-1) depends on a class of host proteins called host dependency factors (HDFs) to facilitate its infection. So far experimental efforts have detected a certain number of HDFs, but the gene inventory of HIV-1 HDFs remains incomplete. Here, we implemented an existing network-based gene discovery strategy to predict HIV-1 HDFs. First, an encoding scheme based on a publicly available human tissue-specific gene functional network (GIANT; http://giant.princeton.edu/) was designed to convert each human gene into a 25,825-dimensional feature vector. Then, a random forest-based predictive model was trained on a data set containing 868 known HDFs and 1,736 non-HDFs. Through 5-fold cross-validation, an independent test, and comparison with one existing method, the proposed prediction method consistently revealed accurate and competitive performance. The highlight of our method should be ascribed to the introduction of the GIANT encoding scheme, which contains rich information regarding gene interactions. By merging known HDFs and genome-wide HDF prediction results, network analysis was conducted to catch the common patterns of HDFs in the context of the GIANT network. Interestingly, HDFs reveal significantly lower betweenness than HIV-1-interacting human proteins (i.e., HIV targets). In the meantime, the functional roles of HDFs were also examined by mapping all the HDF candidates into human protein complexes. Especially, we observed the frequent co-occurrence of HDFs and HIV targets at the protein complex level. Collectively, we hope the proposed prediction method not only can accelerate the HDF identification and antiviral drug target discovery, but also can provide some mechanistic insights into human-virus relationships.

IMPORTANCE Identification of HIV-1 HDFs remains a crucial step to understand the complicated relationships between human and HIV-1. To complement the experimental identification of HDFs, we have implemented an existing network-based gene discovery strategy to predict HDFs from the human genome. The core idea of the proposed method is that the rich information deposited in host gene functional networks can be effectively utilized to infer the potential HDFs. We hope the proposed prediction method could further guide hypothesis-driven experimental efforts to interrogate human–HIV-1 relationships and provide new hints for the development of antiviral drugs to combat HIV-1 infection.

System-Based Approaches to Delineate the Antiviral Innate Immune Landscape

Viruses pose substantial challenges for society, economy, healthcare systems, and research. Their distinctive pathologies are based on specific interactions with cellular factors. In order to develop new antiviral treatments, it is of central importance to understand how viruses interact with their host and how infected cells react to the virus on a molecular level. Invading viruses are commonly sensed by components of the innate immune system, which is composed of a highly effective yet complex network of proteins that, in most cases, mediate efficient virus inhibition. Central to this process is the activity of interferons and other cytokines that coordinate the antiviral response. So far, numerous methods have been used to identify how viruses interact with cellular processes and revealed that the innate immune response is highly complex and involves interferon-stimulated genes and their binding partners as functional factors. Novel approaches and careful experimental design, combined with large-scale, high-throughput methods and cutting-edge analysis pipelines, have to be utilized to delineate the antiviral innate immune landscape at a global level. In this review, we describe different currently used screening approaches, how they contributed to our knowledge on virus-host interactions, and essential considerations that have to be taken into account when planning such experiments.

The E3 Ubiquitin Ligase Mind Bomb 1 Controls Adenovirus Genome Release at the Nuclear Pore Complex

Adenoviruses (AdVs) cause respiratory, ocular, and gastrointestinal tract infection and inflammation in immunocompetent people and life-threatening disease upon immunosuppression. AdV vectors are widely used in gene therapy and vaccination. Incoming particles attach to nuclear pore complexes (NPCs) of post-mitotic cells, then rupture and deliver viral DNA (vDNA) to the nucleus or misdeliver to the cytosol. Our genome-wide RNAi screen in AdV-infected cells identified the RING-type E3 ubiquitin ligase Mind bomb 1 (Mib1) as a proviral host factor for AdV infection. Mib1 is implicated in Notch-Delta signaling, ciliary biogenesis, and RNA innate immunity. Mib1 depletion arrested incoming AdVs at NPCs. Induced expression of full-length but not ligase-defective Mib1 in knockout cells triggered vDNA uncoating from NPC-tethered virions, nuclear import, misdelivery of vDNA, and vDNA expression. Mib1 is an essential host factor for AdV uncoating in human cells, and it provides a new concept for licensing virion DNA delivery through the NPC.

Augmenting subnetwork inference with information extracted from the scientific literature

Many biological studies involve either (i) manipulating some aspect of a cell or its environment and then simultaneously measuring the effect on thousands of genes, or (ii) systematically manipulating each gene and then measuring the effect on some response of interest. A common challenge that arises in these studies is to explain how genes identified as relevant in the given experiment are organized into a subnetwork that accounts for the response of interest. The task of inferring a subnetwork is typically dependent on the information available in publicly available, structured databases, which suffer from incompleteness. However, a wealth of potentially relevant information resides in the scientific literature, such as information about genes associated with certain concepts of interest, as well as interactions that occur among various biological entities. We contend that by exploiting this information, we can improve the explanatory power and accuracy of subnetwork inference in multiple applications. Here we propose and investigate several ways in which information extracted from the scientific literature can be used to augment subnetwork inference. We show that we can use literature-extracted information to (i) augment the set of entities identified as being relevant in a subnetwork inference task, (ii) augment the set of interactions used in the process, and (iii) support targeted browsing of a large inferred subnetwork by identifying entities and interactions that are closely related to concepts of interest. We use this approach to uncover the pathways involved in interactions between a virus and a host cell, and the pathways that are regulated by a transcription factor associated with breast cancer. Our experimental results demonstrate that these approaches can provide more accurate and more interpretable subnetworks. Integer program code, background network data, and pathfinding code are available at https://github.com/Craven-Biostat-Lab/subnetwork_inference

There is a multitude of publicly available databases that contain information about biological entities (i.e., genes, proteins, and other small molecules) as well as information about how these entities interact together. However, these databases are often incomplete. There is a wealth of information present in the text of the scientific literature that is not yet available in these databases. Using tools that mine the scientific literature we are able to extract some of this potentially relevant information. In this work we show how we can use publicly available databases in conjunction with the information extracted from the scientific literature to infer the networks that are involved in specific biological processes, such as viral replication and cancer tumor growth.

A dual controllability analysis of influenza virus-host protein-protein interaction networks for antiviral drug target discovery

BACKGROUND

Host factors of influenza virus replication are often found in key topological positions within protein-protein interaction networks. This work explores how protein states can be manipulated through controllability analysis: the determination of the minimum manipulation needed to drive the cell system to any desired state. Here, we complete a two-part controllability analysis of two protein networks: a host network representing the healthy cell state and an influenza A virus-host network representing the infected cell state. In this context, controllability analyses aim to identify key regulating host factors of the infected cell's progression. This knowledge can be utilized in further biological analysis to understand disease dynamics and isolate proteins for study as drug target candidates.

RESULTS

Both topological and controllability analyses provide evidence of wide-reaching network effects stemming from the addition of viral-host protein interactions. Virus interacting and driver host proteins are significant both topologically and in controllability, therefore playing important roles in cell behavior during infection. Functional analysis finds overlap of results with previous siRNA studies of host factors involved in influenza replication, NF-kB pathway and infection relevance, and roles as interferon regulating genes. 24 proteins are identified as holding regulatory roles specific to the infected cell by measures of topology, controllability, and functional role. These proteins are recommended for further study as potential antiviral drug targets.

CONCLUSIONS

Seasonal outbreaks of influenza A virus are a major cause of illness and death around the world each year with a constant threat of pandemic infection. This research aims to increase the efficiency of antiviral drug target discovery using existing protein-protein interaction data and network analysis methods. These results are beneficial to future studies of influenza virus, both experimental and computational, and provide evidence that the combination of topology and controllability analyses may be valuable for future efforts in drug target discovery.

Network-Guided Discovery of Influenza Virus Replication Host Factors

Integrating virus-host interactions with host protein-protein interactions, we have created a method using these established network practices to identify host factors (i.e., proteins) that are likely candidates for antiviral drug targeting. We demonstrate that interaction cascades between host proteins that directly interact with viral proteins and host factors that are important to influenza virus replication are enriched for signaling and immune processes. Additionally, we show that host proteins that interact with viral proteins are in network locations of power. Finally, we demonstrate a new network methodology to predict novel host factors and validate predictions with an siRNA screen. Our results show that integrating virus-host proteins interactions is useful in the identification of antiviral drug target candidates.

The positions of host factors required for viral replication within a human protein-protein interaction (PPI) network can be exploited to identify drug targets that are robust to drug-mediated selective pressure. Host factors can physically interact with viral proteins, be a component of virus-regulated pathways (where proteins do not interact with viral proteins), or be required for viral replication but unregulated by viruses. Here, we demonstrate a method of combining human PPI networks with virus-host PPI data to improve antiviral drug discovery for influenza viruses by identifying target host proteins. Analysis shows that influenza virus proteins physically interact with host proteins in network positions significant for information flow, even after the removal of known abundance-degree bias within PPI data. We have isolated a subnetwork of the human PPI network that connects virus-interacting host proteins to host factors that are important for influenza virus replication without physically interacting with viral proteins. The subnetwork is enriched for signaling and immune processes distinct from those associated with virus-interacting proteins. Selecting proteins based on subnetwork topology, we performed an siRNA screen to determine whether the subnetwork was enriched for virus replication host factors and whether network position within the subnetwork offers an advantage in prioritization of drug targets to control influenza virus replication. We found that the subnetwork is highly enriched for target host proteins—more so than the set of host factors that physically interact with viral proteins. Our findings demonstrate that network positions are a powerful predictor to guide antiviral drug candidate prioritization.

The conserved protective cyclic AMP-phosphodiesterase function PDE4B is expressed in the adenoma and adjacent normal colonic epithelium of mammals and silenced in colorectal cancer

Conservation over three mammalian genera-the mouse, rat, and human-has been found for a subset of the transcripts whose level differs between the adenoma and normal epithelium of the colon. Pde4b is one of the triply conserved transcripts whose level is enhanced both in the colonic adenoma and in the normal colonic epithelium, especially adjacent to adenomas. It encodes the phosphodiesterase PDE4B, specific for cAMP. Loss of PDE4B function in the ApcMin/+ mouse leads to a significant increase in the number of colonic adenomas. Similarly, Pde4b-deficient ApcMin/+ mice are hypersensitive to treatment by the inflammatory agent DSS, becoming moribund soon after treatment. These observations imply that the PDE4B function protects against ApcMin-induced adenomagenesis and inflammatory lethality. The paradoxical enhancement of the Pde4b transcript in the adenoma versus this inferred protective function of PDE4B can be rationalized by a feedback model in which PDE4B is first activated by early oncogenic stress involving cAMP and then, as reported for frank human colon cancer, inactivated by epigenetic silencing.

Incorporating interaction networks into the determination of functionally related hit genes in genomic experiments with Markov random fields

Motivation

Incorporating gene interaction data into the identification of ‘hit’ genes in genomic experiments is a well-established approach leveraging the ‘guilt by association’ assumption to obtain a network based hit list of functionally related genes. We aim to develop a method to allow for multivariate gene scores and multiple hit labels in order to extend the analysis of genomic screening data within such an approach.

Results

We propose a Markov random field-based method to achieve our aim and show that the particular advantages of our method compared with those currently used lead to new insights in previously analysed data as well as for our own motivating data. Our method additionally achieves the best performance in an independent simulation experiment. The real data applications we consider comprise of a survival analysis and differential expression experiment and a cell-based RNA interference functional screen.

Availability and implementation

We provide all of the data and code related to the results in the paper.

Supplementary information

Supplementary data are available at Bioinformatics online.

A Systems Survey of Progressive Host-Cell Reorganization during Rotavirus Infection

Pathogen invasion is often accompanied by widespread alterations in cellular physiology, which reflects the hijacking of host factors and processes for pathogen entry and replication. Although genetic perturbation screens have revealed the complexity of host factors involved for numerous pathogens, it has remained challenging to temporally define the progression of events in host cell reorganization during infection. We combine high-confidence genome-scale RNAi screening of host factors required for rotavirus infection in human intestinal cells with an innovative approach to infer the trajectory of virus infection from fixed cell populations. This approach reveals a comprehensive network of host cellular processes involved in rotavirus infection and implicates AMPK in initiating the development of a rotavirus-permissive environment. Our work provides a powerful approach that can be generalized to order complex host cellular requirements along a trajectory of cellular reorganization during pathogen invasion.

Benchmarking selected computational gene network growing tools in context of virus-host interactions

Several available online tools provide network growing functions where an algorithm utilizing different data sources suggests additional genes/proteins that should connect an input gene set into functionally meaningful networks. Using the well-studied system of influenza host interactions, we compare the network growing function of two free tools GeneMANIA and STRING and the commercial IPA for their performance of recovering known influenza A virus host factors previously identified from siRNA screens. The result showed that given small (~30 genes) or medium (~150 genes) input sets all three network growing tools detect significantly more known host factors than random human genes with STRING overall performing strongest. Extending the networks with all the three tools significantly improved the detection of GO biological processes of known host factors compared to not growing networks. Interestingly, the rate of identification of true host factors using computational network growing is equal or better to doing another experimental siRNA screening study which could also be true and applied to other biological pathways/processes.

Multiplex shRNA Screening of Germ Cell Development by in Vivo Transfection of Mouse Testis

Spermatozoa are one of the few mammalian cell types that cannot be fully derived in vitro, severely limiting the application of modern genomic techniques to study germ cell biology. The current gold standard approach of characterizing single-gene knockout mice is slow as generation of each mutant line can take 6–9 months. Here, we describe an in vivo approach to rapid functional screening of germline genes based on a new nonsurgical, nonviral in vivo transfection method to deliver nucleic acids into testicular germ cells. By coupling multiplex transfection of short hairpin RNA (shRNA) constructs with pooled amplicon sequencing as a readout, we were able to screen many genes for spermatogenesis function in a quick and inexpensive experiment. We transfected nine mouse testes with a pilot pool of RNA interference (RNAi) against well-characterized genes to show that this system is highly reproducible and accurate. With a false negative rate of 18% and a false positive rate of 12%, this method has similar performance as other RNAi screens in the well-described Drosophila model system. In a separate experiment, we screened 26 uncharacterized genes computationally predicted to be essential for spermatogenesis and found numerous candidates for follow-up studies. Finally, as a control experiment, we performed a long-term selection screen in neuronal N2a cells, sampling shRNA frequencies at five sequential time points. By characterizing the effect of both libraries on N2a cells, we show that our screening results from testis are tissue-specific. Our calculations indicate that the current implementation of this approach could be used to screen thousands of protein-coding genes simultaneously in a single mouse testis. The experimental protocols and analysis scripts provided will enable other groups to use this procedure to study diverse aspects of germ cell biology ranging from epigenetics to cell physiology. This approach also has great promise as an applied tool for validating diagnoses made from medical genome sequencing, or designing synthetic biological sequences that can act as potent and highly specific male contraceptives.

Making the cut: improved ranking and selection for large-scale inference

Identifying leading measurement units from a large collection is a common inference task in various domains of large-scale inference. Testing approaches, which measure evidence against a null hypothesis rather than effect magnitude, tend to overpopulate lists of leading units with those associated with low measurement error. By contrast, local maximum likelihood (ML) approaches tend to favor units with high measurement error. Available Bayesian and empirical Bayesian approaches rely on specialized loss functions that result in similar deficiencies. We describe and evaluate a generic empirical Bayesian ranking procedure that populates the list of top units in a way that maximizes the expected overlap between the true and reported top lists for all list sizes. The procedure relates unit-specific posterior upper tail probabilities with their empirical distribution to yield a ranking variable. It discounts high-variance units less than popular non-ML methods and thus achieves improved operating characteristics in the models considered.

Combined Gene Expression and RNAi Screening to Identify Alkylation Damage Survival Pathways from Fly to Human

Alkylating agents are a key component of cancer chemotherapy. Several cellular mechanisms are known to be important for its survival, particularly DNA repair and xenobiotic detoxification, yet genomic screens indicate that additional cellular components may be involved. Elucidating these components has value in either identifying key processes that can be modulated to improve chemotherapeutic efficacy or may be altered in some cancers to confer chemoresistance. We therefore set out to reevaluate our prior Drosophila RNAi screening data by comparison to gene expression arrays in order to determine if we could identify any novel processes in alkylation damage survival. We noted a consistent conservation of alkylation survival pathways across platforms and species when the analysis was conducted on a pathway/process level rather than at an individual gene level. Better results were obtained when combining gene lists from two datasets (RNAi screen plus microarray) prior to analysis. In addition to previously identified DNA damage responses (p53 signaling and Nucleotide Excision Repair), DNA-mRNA-protein metabolism (transcription/translation) and proteasome machinery, we also noted a highly conserved cross-species requirement for NRF2, glutathione (GSH)-mediated drug detoxification and Endoplasmic Reticulum stress (ER stress)/Unfolded Protein Responses (UPR) in cells exposed to alkylation. The requirement for GSH, NRF2 and UPR in alkylation survival was validated by metabolomics, protein studies and functional cell assays. From this we conclude that RNAi/gene expression fusion is a valid strategy to rapidly identify key processes that may be extendable to other contexts beyond damage survival.

Functional Genomic Strategies for Elucidating Human-Virus Interactions: Will CRISPR Knockout RNAi and Haploid Cells?

Over the last several years a wealth of transformative human–virus interaction discoveries have been produced using loss-of-function functional genomics. These insights have greatly expanded our understanding of how human pathogenic viruses exploit our cells to replicate. Two technologies have been at the forefront of this genetic revolution, RNA interference (RNAi) and random retroviral insertional mutagenesis using haploid cell lines (haploid cell screening), with the former technology largely predominating. Now the cutting edge gene editing of the CRISPR/Cas9 system has also been harnessed for large-scale functional genomics and is poised to possibly displace these earlier methods. Here we compare and contrast these three screening approaches for elucidating host–virus interactions, outline their key strengths and weaknesses including a comparison of an arrayed multiple orthologous RNAi reagent screen to a pooled CRISPR/Cas9 human rhinovirus 14–human cell interaction screen, and recount some notable insights made possible by each. We conclude with a brief perspective on what might lie ahead for the fast evolving field of human–virus functional genomics.

LEGO: a novel method for gene set over-representation analysis by incorporating network-based gene weights

Pathway or gene set over-representation analysis (ORA) has become a routine task in functional genomics studies. However, currently widely used ORA tools employ statistical methods such as Fisher’s exact test that reduce a pathway into a list of genes, ignoring the constitutive functional non-equivalent roles of genes and the complex gene-gene interactions. Here, we develop a novel method named LEGO (functional Link Enrichment of Gene Ontology or gene sets) that takes into consideration these two types of information by incorporating network-based gene weights in ORA analysis. In three benchmarks, LEGO achieves better performance than Fisher and three other network-based methods. To further evaluate LEGO’s usefulness, we compare LEGO with five gene expression-based and three pathway topology-based methods using a benchmark of 34 disease gene expression datasets compiled by a recent publication, and show that LEGO is among the top-ranked methods in terms of both sensitivity and prioritization for detecting target KEGG pathways. In addition, we develop a cluster-and-filter approach to reduce the redundancy among the enriched gene sets, making the results more interpretable to biologists. Finally, we apply LEGO to two lists of autism genes, and identify relevant gene sets to autism that could not be found by Fisher.

USP9X inhibition promotes radiation-induced apoptosis in non-small cell lung cancer cells expressing mid-to-high MCL1

BACKGROUND AND PURPOSE

Radiotherapy (RT) is vital for the treatment of locally advanced non-small cell lung cancer (NSCLC), yet its delivery is limited by tolerances of adjacent organs. We sought therefore to identify and characterize gene targets whose inhibition may improve RT.

MATERIALS AND METHODS

Whole genome pooled shRNA cytotoxicity screens were performed in A549 and NCI-H460 using a retroviral library of 74,705 sequences. Cells were propagated with or without daily radiation Monday-Friday. Radiosensitization by top differential dropout hits was assessed by clonogenic assays. Apoptosis was assessed using a caspase 3/7 cell-based activity assay and by annexin V-FITC and PI staining. MCL1 expression was assessed by qPCR and Western blotting.

RESULTS

USP9X, a deubiquitinase, was a top hit among druggable gene products. WP1130, a small molecule USP9X inhibitor, showed synergistic cytotoxicity with IR. MCL1, an anti-apoptotic protein deubiquitinated by USP9X, decreased with USP9X inhibition and IR. This was accompanied by increases in caspase 3/7 activity and apoptosis. In a panel of NSCLC lines, MCL1 and USP9X protein and gene expression levels were highly correlated. Lines showing high levels of MCL1 expression were the most sensitive to USP9X inhibition.

CONCLUSIONS

These data support the use of MCL1 expression as a predictive biomarker for USP9X inhibitors in NSCLC therapy.

Specific induction of endogenous viral restriction factors using CRISPR/Cas-derived transcriptional activators

Whereas several mammalian proteins can restrict the replication of HIV-1 and other viruses, these are often not expressed in relevant target cells. A potential method to inhibit viral replication might therefore be to use synthetic transcription factors to induce restriction factor expression. In particular, mutants of the RNA-guided DNA binding protein Cas9 that have lost their DNA cleavage activity could be used to recruit transcription activation domains to specific promoters. However, initial experiments revealed only weak activation unless multiple promoter-specific single guide RNAs (sgRNAs) were used. Recently, the recruitment of multiple transcription activation domains by a single sgRNA, modified to contain MS2-derived stem loops that recruit fusion proteins consisting of the MS2 coat protein linked to transcription activation domains, was reported to induce otherwise silent cellular genes. Here, we demonstrate that such "synergistic activation mediators" can induce the expression of two restriction factors, APOBEC3G (A3G) and APOBEC3B (A3B), in human cells that normally lack these proteins. We observed modest activation of endogenous A3G or A3B expression using single sgRNAs but high expression when two sgRNAs were used. Whereas the induced A3G and A3B proteins both blocked infection by an HIV-1 variant lacking a functional vif gene by inducing extensive dC-to-dU editing, only the induced A3B protein inhibited wild-type HIV-1. These data demonstrate that Cas9-derived transcriptional activators have the potential to be used for screens for endogenous genes that affect virus replication and raise the possibility that synthetic transcription factors might prove clinically useful if efficient delivery mechanisms could be developed.

Virus-Host Interactions: From Unbiased Genetic Screens to Function

Deciphering the many interactions that occur between a virus and host cell over the course of infection is paramount to understanding mechanisms of pathogenesis and to the future development of antiviral therapies. Over the past decade, researchers have started to understand these complicated relationships through the development of methodologies, including advances in RNA interference, proteomics, and the development of genetic tools such as haploid cell lines, allowing high-throughput screening to identify critical contact points between virus and host. These advances have produced a wealth of data regarding host factors hijacked by viruses to promote infection, as well as antiviral factors responsible for subverting viral infection. This review highlights findings from virus-host screens and discusses our thoughts on the direction of screening strategies moving forward.

Non-POU Domain-Containing Octamer-Binding Protein Negatively Regulates HIV-1 Infection in CD4(+) T Cells

HIV-1 interacts with numerous cellular proteins during viral replication. Identifying such host proteins and characterizing their roles in HIV-1 infection can deepen our understanding of the dynamic interplay between host and pathogen. We previously identified non-POU domain-containing octamer-binding protein (NonO or p54nrb) as one of host factors associated with catalytically active preintegration complexes (PIC) of HIV-1 in infected CD4(+) T cells. NonO is involved in nuclear processes including transcriptional regulation and RNA splicing. Although NonO has been identified as an HIV-1 interactant in several recent studies, its role in HIV-1 replication has not been characterized. We investigated the effect of NonO on the HIV-1 life cycle in CD4(+) T cell lines and primary CD4(+) T cells using single-cycle and replication-competent HIV-1 infection assays. We observed that short hairpin RNA (shRNA)-mediated stable NonO knockdown in a CD4(+) Jurkat T cell line and primary CD4(+) T cells did not affect cell viability or proliferation, but enhanced HIV-1 infection. The enhancement of HIV-1 infection in Jurkat T cells correlated with increased viral reverse transcription and gene expression. Knockdown of NonO expression in Jurkat T cells modestly enhanced HIV-1 gag mRNA expression and Gag protein synthesis, suggesting that viral gene expression and RNA regulation are the predominantly affected events causing enhanced HIV-1 replication in NonO knockdown (KD) cells. Furthermore, overexpression of NonO in Jurkat T cells reduced HIV-1 single-cycle infection by 41% compared to control cells. Our data suggest that NonO negatively regulates HIV-1 infection in CD4(+) T cells, albeit it has modest effects on early and late stages of the viral life cycle, highlighting the importance of host proteins associated with HIV-1 PIC in regulating viral replication.

An integrative approach for a network based meta-analysis of viral RNAi screens

BACKGROUND

Big data is becoming ubiquitous in biology, and poses significant challenges in data analysis and interpretation. RNAi screening has become a workhorse of functional genomics, and has been applied, for example, to identify host factors involved in infection for a panel of different viruses. However, the analysis of data resulting from such screens is difficult, with often low overlap between hit lists, even when comparing screens targeting the same virus. This makes it a major challenge to select interesting candidates for further detailed, mechanistic experimental characterization.

RESULTS

To address this problem we propose an integrative bioinformatics pipeline that allows for a network based meta-analysis of viral high-throughput RNAi screens. Initially, we collate a human protein interaction network from various public repositories, which is then subjected to unsupervised clustering to determine functional modules. Modules that are significantly enriched with host dependency factors (HDFs) and/or host restriction factors (HRFs) are then filtered based on network topology and semantic similarity measures. Modules passing all these criteria are finally interpreted for their biological significance using enrichment analysis, and interesting candidate genes can be selected from the modules.

CONCLUSIONS

We apply our approach to seven screens targeting three different viruses, and compare results with other published meta-analyses of viral RNAi screens. We recover key hit genes, and identify additional candidates from the screens. While we demonstrate the application of the approach using viral RNAi data, the method is generally applicable to identify underlying mechanisms from hit lists derived from high-throughput experimental data, and to select a small number of most promising genes for further mechanistic studies.

Identification of HNRNPK as regulator of hepatitis C virus particle production

Hepatitis C virus (HCV) is a major cause of chronic liver disease affecting around 130 million people worldwide. While great progress has been made to define the principle steps of the viral life cycle, detailed knowledge how HCV interacts with its host cells is still limited. To overcome this limitation we conducted a comprehensive whole-virus RNA interference-based screen and identified 40 host dependency and 16 host restriction factors involved in HCV entry/replication or assembly/release. Of these factors, heterogeneous nuclear ribonucleoprotein K (HNRNPK) was found to suppress HCV particle production without affecting viral RNA replication. This suppression of virus production was specific to HCV, independent from assembly competence and genotype, and not found with the related Dengue virus. By using a knock-down rescue approach we identified the domains within HNRNPK required for suppression of HCV particle production. Importantly, HNRNPK was found to interact specifically with HCV RNA and this interaction was impaired by mutations that also reduced the ability to suppress HCV particle production. Finally, we found that in HCV-infected cells, subcellular distribution of HNRNPK was altered; the protein was recruited to sites in close proximity of lipid droplets and colocalized with core protein as well as HCV plus-strand RNA, which was not the case with HNRNPK variants unable to suppress HCV virion formation. These results suggest that HNRNPK might determine efficiency of HCV particle production by limiting the availability of viral RNA for incorporation into virions. This study adds a new function to HNRNPK that acts as central hub in the replication cycle of multiple other viruses.

Multitask learning of signaling and regulatory networks with application to studying human response to flu

Reconstructing regulatory and signaling response networks is one of the major goals of systems biology. While several successful methods have been suggested for this task, some integrating large and diverse datasets, these methods have so far been applied to reconstruct a single response network at a time, even when studying and modeling related conditions. To improve network reconstruction we developed MT-SDREM, a multi-task learning method which jointly models networks for several related conditions. In MT-SDREM, parameters are jointly constrained across the networks while still allowing for condition-specific pathways and regulation. We formulate the multi-task learning problem and discuss methods for optimizing the joint target function. We applied MT-SDREM to reconstruct dynamic human response networks for three flu strains: H1N1, H5N1 and H3N2. Our multi-task learning method was able to identify known and novel factors and genes, improving upon prior methods that model each condition independently. The MT-SDREM networks were also better at identifying proteins whose removal affects viral load indicating that joint learning can still lead to accurate, condition-specific, networks. Supporting website with MT-SDREM implementation: http://sb.cs.cmu.edu/mtsdrem

To understand why some flu strains are more virulent than others, researchers attempt to profile and model the molecular human response to these strains and identify similarities and differences between the resulting models. So far, the modeling and analysis part has been done independently for each strain and the results contrasted in a post-processing step. Here we present a new method, termed MT-SDREM, that simultaneously models the response to all strains allowing us to identify both, the core response elements that are shared among the strains, and factors that are uniquely activated or repressed by individual strains. We applied this method to study the human response to three flu strains: H1N1, H3N2 and H5N1. As we show, the method was able to correctly identify several common and known factors regulating immune response to such strains and also identified unique factors for each of the strains. The models reconstructed by the simultaneous analysis method improved upon those generated by methods that model each strain response separately. Our joint models can be used to identify strain specific treatments as well as treatments that are likely to be effective against all three strains.

Genome-wide shRNA screening identifies host factors involved in early endocytic events for HIV-1-induced CD4 down-regulation

Background

Down-modulation of the CD4 receptor is one of the hallmarks of HIV-1 infection and it is believed to confer a selective replicative advantage to the virus in vivo. This process is mainly mediated by three viral proteins: Env, Vpu and Nef. To date, the mechanisms that lead to CD4 depletion from the surface of infected cells during HIV-1 infection are still only partially characterized. In this study, we sought to identify and characterize cellular host factors in HIV-1-induced CD4 down-modulation.

Results

To identify host factors involved in CD4 down-regulation, we used a whole genome-targeting shRNA lentiviral library in HeLa CD4+ cells expressing Nef as an inducer of CD4 down-modulation. We identified 55 genes, mainly encoding for proteins involved in various steps of clathrin-mediated endocytosis. For confirmation and further selection of the hits we performed several rounds of validation, using individual shRNA lentiviral vectors with a different target sequence for gene knock-down in HIV-1-infected T cells. By this stringent validation set-up, we could demonstrate that the knock-down of DNM3 (dynamin 3), SNX22 (sorting nexin 22), ATP6AP1 (ATPase, H+ Transporting, Lysosomal Accessory Protein 1), HRBL (HIV-Rev binding protein Like), IDH3G (Isocitrate dehydrogenase), HSP90B1 (Heat shock protein 90 kDa beta member 1) and EPS15 (Epidermal Growth Factor Receptor Pathway Substrate 15) significantly increases CD4 levels in HIV-infected SupT1 T cells compared to the non-targeting shRNA control. Moreover, EPS15, DNM3, IDH3G and ATP6AP1 knock-down significantly decreases HIV-1 replication in T cells.

Conclusions

We identified seven genes as cellular co-factors for HIV-1-mediated CD4 down-regulation in T cells. The knock-down of four out of seven of these genes also significantly reduces HIV-1 replication in T cells. Next to a role in HIV-mediated CD4 down-regulation, these genes might however affect HIV-1 replication in another way. Our findings give insights in the HIV-1-mediated CD4 down-regulation at the level of the plasma membrane and early endosomes and identify four possible new HIV-1 replication co-factors.

Electronic supplementary material

The online version of this article (doi:10.1186/s12977-014-0118-4) contains supplementary material, which is available to authorized users.

Comprehensive identification of host modulators of HIV-1 replication using multiple orthologous RNAi reagents

RNAi screens have implicated hundreds of host proteins as HIV-1 dependency factors (HDFs). While informative, these early studies overlap poorly due to false positives and false negatives. To ameliorate these issues, we combined information from the existing HDF screens together with new screens performed with multiple orthologous RNAi reagents (MORR). In addition to being traditionally validated, the MORR screens and the historical HDF screens were quantitatively integrated by the adaptation of an established analysis program, RIGER, for the collective interpretation of each gene’s phenotypic significance. False positives were addressed by the removal of poorly expressed candidates through gene expression filtering, as well as with GESS, which identifies off-target effects. This workflow produced a quantitatively integrated network of genes that modulate HIV-1 replication. We further investigated the roles of GOLGI49, SEC13, and COG in HIV-1 replication. Collectively, the MORR-RIGER method minimized the caveats of RNAi screening and improved our understanding of HIV-1–host cell interactions.

A High-Content RNAi Screen Identifies Ubiquitin Modifiers That Regulate TNF-Dependent Nuclear Accumulation of NF-κB

The mammalian tumor necrosis factor (TNF) cytokine is a central mediator of inflammatory events. Recent studies revealed a number of complex and sophisticated interactions between the TNF pathway and the enzymatic activities encoded by ubiquitin ligases and deubiquitylation enzymes. However, very little is known about the identity of the ubiquitin pathway members that control the extent of ubiquitylation in TNF responses. To address this deficit, we conducted an unbiased, high-content screen of the human ubiquitin pathway for gene products that control defining features of the cellular response to TNF. In particular, we sought to identify ubiquitin modifying enzymes that alter the ability of TNF to regulate the nuclear accumulation of nuclear factor kappa B. In this screen, we identified and validated several novel regulators of the TNF pathway. We believe these regulators constitute potential targets for pharmacological interventions that manipulate TNF-dependent inflammation.

Genome-wide analysis of host factors in nodavirus RNA replication

Flock House virus (FHV), the best studied of the animal nodaviruses, has been used as a model for positive-strand RNA virus research. As one approach to identify host genes that affect FHV RNA replication, we performed a genome-wide analysis using a yeast single gene deletion library and a modified, reporter gene-expressing FHV derivative. A total of 4,491 yeast deletion mutants were tested for their ability to support FHV replication. Candidates for host genes modulating FHV replication were selected based on the initial genome-wide reporter gene assay and validated in repeated Northern blot assays for their ability to support wild type FHV RNA1 replication. Overall, 65 deletion strains were confirmed to show significant changes in the replication of both FHV genomic RNA1 and sub-genomic RNA3 with a false discovery rate of 5%. Among them, eight genes support FHV replication, since their deletion significantly reduced viral RNA accumulation, while 57 genes limit FHV replication, since their deletion increased FHV RNA accumulation. Of the gene products implicated in affecting FHV replication, three are localized to mitochondria, where FHV RNA replication occurs, 16 normally reside in the nucleus and may have indirect roles in FHV replication, and the remaining 46 are in the cytoplasm, with functions enriched in translation, RNA processing and trafficking.