Decoding Heterogenous Single-cell Perturbation Responses.

Understanding diverse responses of individual cells to the same perturbation is central to many biological and biomedical problems. Current methods, however, do not precisely quantify the strength of perturbation responses and, more importantly, reveal new biological insights from heterogeneity in responses. Here we introduce the perturbation-response score (PS), based on constrained quadratic optimization, to quantify diverse perturbation responses at a single-cell level. Applied to single-cell transcriptomes of large-scale genetic perturbation datasets (e.g., Perturb-seq), PS outperforms existing methods for quantifying partial gene perturbation responses. In addition, PS presents two major advances. First, PS enables large-scale, single-cell-resolution dosage analysis of perturbation, without the need to titrate perturbation strength. By analyzing the dose-response patterns of over 2,000 essential genes in Perturb-seq, we identify two distinct patterns, depending on whether a moderate reduction in their expression induces strong downstream expression alterations. Second, PS identifies intrinsic and extrinsic biological determinants of perturbation responses. We demonstrate the application of PS in contexts such as T cell stimulation, latent HIV-1 expression, and pancreatic cell differentiation. Notably, PS unveiled a previously unrecognized, cell-type-specific role of coiled-coil domain containing 6 (CCDC6) in guiding liver and pancreatic lineage decisions, where CCDC6 knockouts drive the endoderm cell differentiation towards liver lineage, rather than pancreatic lineage. The PS approach provides an innovative method for dose-to-function analysis and will enable new biological discoveries from single-cell perturbation datasets.

Internal checkpoint regulates T cell neoantigen reactivity and susceptibility to PD1 blockade.

BACKGROUND: Adoptive transfer of tumor-infiltrating lymphocytes (TIL) fails to consistently elicit tumor rejection. Manipulation of intrinsic factors that inhibit T cell effector function and neoantigen recognition may therefore improve TIL therapy outcomes. We previously identified the cytokine-induced SH2 protein (CISH) as a key regulator of T cell functional avidity in mice. Here, we investigate the mechanistic role of CISH in regulating human T cell effector function in solid tumors and demonstrate that CRISPR/Cas9 disruption of CISH enhances TIL neoantigen recognition and response to checkpoint blockade. METHODS: Single-cell gene expression profiling was used to identify a negative correlation between high CISH expression and TIL activation in patient-derived TIL. A GMP-compliant CRISPR/Cas9 gene editing process was developed to assess the impact of CISH disruption on the molecular and functional phenotype of human peripheral blood T cells and TIL. Tumor-specific T cells with disrupted Cish function were adoptively transferred into tumor-bearing mice and evaluated for efficacy with or without checkpoint blockade. FINDINGS: CISH expression was associated with T cell dysfunction. CISH deletion using CRISPR/Cas9 resulted in hyper-activation and improved functional avidity against tumor-derived neoantigens without perturbing T cell maturation. Cish knockout resulted in increased susceptibility to checkpoint blockade in vivo. CONCLUSIONS: CISH negatively regulates human T cell effector function, and its genetic disruption offers a novel avenue to improve the therapeutic efficacy of adoptive TIL therapy. FUNDING: This study was funded by Intima Bioscience, U.S. and in part through the Intramural program CCR at the National Cancer Institute.

A toolbox for class I HDACs reveals isoform specific roles in gene regulation and protein acetylation.

The class I histone deacetylases are essential regulators of cell fate decisions in health and disease. While pan- and class-specific HDAC inhibitors are available, these drugs do not allow a comprehensive understanding of individual HDAC function, or the therapeutic potential of isoform-specific targeting. To systematically compare the impact of individual catalytic functions of HDAC1, HDAC2 and HDAC3, we generated human HAP1 cell lines expressing catalytically inactive HDAC enzymes. Using this genetic toolbox we compare the effect of individual HDAC inhibition with the effects of class I specific inhibitors on cell viability, protein acetylation and gene expression. Individual inactivation of HDAC1 or HDAC2 has only mild effects on cell viability, while HDAC3 inactivation or loss results in DNA damage and apoptosis. Inactivation of HDAC1/HDAC2 led to increased acetylation of components of the COREST co-repressor complex, reduced deacetylase activity associated with this complex and derepression of neuronal genes. HDAC3 controls the acetylation of nuclear hormone receptor associated proteins and the expression of nuclear hormone receptor regulated genes. Acetylation of specific histone acetyltransferases and HDACs is sensitive to inactivation of HDAC1/HDAC2. Over a wide range of assays, we determined that in particular HDAC1 or HDAC2 catalytic inactivation mimics class I specific HDAC inhibitors. Importantly, we further demonstrate that catalytic inactivation of HDAC1 or HDAC2 sensitizes cells to specific cancer drugs. In summary, our systematic study revealed isoform-specific roles of HDAC1/2/3 catalytic functions. We suggest that targeted genetic inactivation of particular isoforms effectively mimics pharmacological HDAC inhibition allowing the identification of relevant HDACs as targets for therapeutic intervention.

Adeno-associated Virus (AAV) Capsid Chimeras with Enhanced Infectivity Reveal a Core Element in the AAV Genome Critical for both Cell Transduction and Capsid Assembly.

Adeno-associated viruses (AAV) have attracted significant attention in the field of gene and cell therapy due to highly effective delivery of therapeutic genes into human cells. The ability to generate recombinant AAV vectors compromised of unique or substituted protein sequences has led to the development of capsid variants with improved therapeutic properties. Seeking novel AAV vectors capable of enhanced transduction for therapeutic applications, we have developed a series of unique capsid variants termed AAV X-Vivo (AAV-XV) derived from chimeras of AAV12 VP1/2 sequences and the VP3 sequence of AAV6. These AAV variants showed enhanced infection of human primary T cells, hematopoietic stem cells, and neuronal cell lines over wildtype parental viruses, and superiority over AAV6 for genomic integration of DNA sequences by AAV alone or in combination with CRISPR gene editing. AAV-XV variants demonstrate transduction efficiency equivalent to AAV6 at multiplicities of infection 2 logs lower, enabling T cell engineering at low AAV doses. The protein coding sequence of these novel AAV chimeras revealed disruptions within the assembly-activating protein (AAP) which likely accounted for observed lower virus yield. A series of genome alterations, reverting the AAP sequence back to wildtype AAV6, had a negative impact on the enhanced transduction seen with AAV-VX, indicating overlapping functions within this sequence for both viral assembly and effective T cell transduction. Our findings show these AAV-XV variants are highly efficient at cell transduction at low dose and demonstrates the importance of the AAP coding region in both viral particle assembly and cell infection.IMPORTANCE A major hurdle to the therapeutic potential of AAV in gene therapy lies in achieving clinically meaningful AAV doses, and secondarily, ability to manufacture commercially viable titers of AAV to support this. By virtue of neutralizing antibodies against AAV that impede patient repeat-dosing, the dose of AAV for in vivo gene delivery has been high, which has resulted in unfortunate recent safety concerns and deaths in patients given higher-dose AAV gene therapy. We have generated new AAV variants possessing unique combinations of capsid proteins for gene and cell therapy applications termed AAV-XV, which have high levels of cell transduction and gene delivery at lower MOI. Furthermore, we demonstrate a novel finding, and an important consideration for recombinant AAV design, that a region of the AAV genome encoding the capsid viral protein and AAP is critical for both virus yield and the enhancement of infection/transduction.

Biological plasticity rescues target activity in CRISPR knock outs.

Gene knock outs (KOs) are efficiently engineered through CRISPR-Cas9-induced frameshift mutations. While the efficiency of DNA editing is readily verified by DNA sequencing, a systematic understanding of the efficiency of protein elimination has been lacking. Here we devised an experimental strategy combining RNA sequencing and triple-stage mass spectrometry to characterize 193 genetically verified deletions targeting 136 distinct genes generated by CRISPR-induced frameshifts in HAP1 cells. We observed residual protein expression for about one third of the quantified targets, at variable levels from low to original, and identified two causal mechanisms, translation reinitiation leading to N-terminally truncated target proteins or skipping of the edited exon leading to protein isoforms with internal sequence deletions. Detailed analysis of three truncated targets, BRD4, DNMT1 and NGLY1, revealed partial preservation of protein function. Our results imply that systematic characterization of residual protein expression or function in CRISPR-Cas9-generated KO lines is necessary for phenotype interpretation.

A small cassette enables conditional gene inactivation by CRISPR/Cas9.

The availability of CRISPR/Cas9 technology has enabled the rapid establishment of gene knockouts in many cell types and even whole organisms. However, conditional inactivation of essential genes remains a challenge. We devised an approach named DECAI (DEgradation based on Cre-regulated- Artificial Intron). It utilizes a small cassette of just 201 nucleotides that is inserted into the coding exon of a target gene using CRISPR/Cas9 technology and homology-directed repair. As its sequence is derived from an artificial intron, the cassette is removed by the splicing machinery and thus leaves no trace in the "off-state". Upon activation with Cre recombinase ("on-state"), the intron is crippled and the target gene is disrupted by a series of stop codons. We exemplify the utility of this approach on several non-essential and essential human genes. Clones bearing the conditional knockout cassette are recovered at frequencies above 5% and cassette function can be traced at the genomic DNA and the mRNA level. Importantly, cassette activation leads to loss of gene expression as judged by flow cytometry, Western blot or immunofluorescence. Altogether, this highlights the broad utility of the approach for conditional gene inactivation and suggests that this tool could be used to study the loss-of-function phenotypes of essential genes.

Efficient generation and reversion of chromosomal translocations using CRISPR/Cas technology.

BACKGROUND: Chromosomal translocations are a hallmark of cancer cells and give rise to fusion oncogenes. To gain insight into the mechanisms governing tumorigenesis, adequate model cell lines are required. RESULTS: We employ the versatile CRISPR/Cas system to engineer cell lines in which chromosomal translocations are either generated de novo (CD74-ROS1) or existing translocations are reverted back to the original configuration (BCR-ABL1). To this end, we co-apply two guide RNAs to artificially generate two breakpoints and screen for spontaneous fusion events by PCR. CONCLUSIONS: The approach we use is efficient and delivers clones bearing translocationsin a predictable fashion. Detailed analysis suggests that the clones display no additional undesired alterations, implying that the approach is robust and precise.

Parallel reverse genetic screening in mutant human cells using transcriptomics.

Reverse genetic screens have driven gene annotation and target discovery in model organisms. However, many disease-relevant genotypes and phenotypes cannot be studied in lower organisms. It is therefore essential to overcome technical hurdles associated with large-scale reverse genetics in human cells. Here, we establish a reverse genetic approach based on highly robust and sensitive multiplexed RNA sequencing of mutant human cells. We conduct 10 parallel screens using a collection of engineered haploid isogenic cell lines with knockouts covering tyrosine kinases and identify known and unexpected effects on signaling pathways. Our study provides proof of concept for a scalable approach to link genotype to phenotype in human cells, which has broad applications. In particular, it clears the way for systematic phenotyping of still poorly characterized human genes and for systematic study of uncharacterized genomic features associated with human disease.

A generic strategy for CRISPR-Cas9-mediated gene tagging.

Genome engineering has been greatly enhanced by the availability of Cas9 endonuclease that can be targeted to almost any genomic locus using so called guide RNAs (gRNAs). However, the introduction of foreign DNA sequences to tag an endogenous gene is still cumbersome as it requires the synthesis or cloning of homology templates. Here we present a strategy that enables the tagging of endogenous loci using one generic donor plasmid. It contains the tag of interest flanked by two gRNA recognition sites that allow excision of the tag from the plasmid. Co-transfection of cells with Cas9, a gRNA specifying the genomic locus of interest, the donor plasmid and a cassette-specific gRNA triggers the insertion of the tag by a homology-independent mechanism. The strategy is efficient and delivers clones that display a predictable integration pattern. As showcases we generated NanoLuc luciferase- and TurboGFP-tagged reporter cell lines.

Megabase-scale deletion using CRISPR/Cas9 to generate a fully haploid human cell line.

Near-haploid human cell lines are instrumental for genetic screens and genome engineering as gene inactivation is greatly facilitated by the absence of a second gene copy. However, no completely haploid human cell line has been described, hampering the genetic accessibility of a subset of genes. The near-haploid human cell line HAP1 contains a single copy of all chromosomes except for a heterozygous 30-megabase fragment of Chromosome 15. This large fragment encompasses 330 genes and is integrated on the long arm of Chromosome 19. Here, we employ a CRISPR/Cas9-based genome engineering strategy to excise this sizeable chromosomal fragment and to efficiently and reproducibly derive clones that retain their haploid state. Importantly, spectral karyotyping and single-nucleotide polymorphism (SNP) genotyping revealed that engineered-HAPloid (eHAP) cells are fully haploid with no gross chromosomal aberrations induced by Cas9. Furthermore, whole-genome sequence and transcriptome analysis of the parental HAP1 and an eHAP cell line showed that transcriptional changes are limited to the excised Chromosome 15 fragment. Together, we demonstrate the feasibility of efficiently engineering megabase deletions with the CRISPR/Cas9 technology and report the first fully haploid human cell line.

A reversible gene trap collection empowers haploid genetics in human cells.

Knockout collections are invaluable tools for studying model organisms such as yeast. However, there are no large-scale knockout collections of human cells. Using gene-trap mutagenesis in near-haploid human cells, we established a platform to generate and isolate individual 'gene-trapped cells' and used it to prepare a collection of human cell lines carrying single gene-trap insertions. In most cases, the insertion can be reversed. This growing library covers 3,396 genes, one-third of the expressed genome, is DNA-barcoded and allows systematic screens for a wide variety of cellular phenotypes. We examined cellular responses to TNF-α, TGF-ß, IFN-γ and TNF-related apoptosis-inducing ligand (TRAIL), to illustrate the value of this unique collection of isogenic human cell lines.

Experimental characterization of the human non-sequence-specific nucleic acid interactome.

BACKGROUND: The interactions between proteins and nucleic acids have a fundamental function in many biological processes, including gene transcription, RNA homeostasis, protein translation and pathogen sensing for innate immunity. While our knowledge of the ensemble of proteins that bind individual mRNAs in mammalian cells has been greatly augmented by recent surveys, no systematic study on the non-sequence-specific engagement of native human proteins with various types of nucleic acids has been reported. RESULTS: We designed an experimental approach to achieve broad coverage of the non-sequence-specific RNA and DNA binding space, including methylated cytosine, and tested for interaction potential with the human proteome. We used 25 rationally designed nucleic acid probes in an affinity purification mass spectrometry and bioinformatics workflow to identify proteins from whole cell extracts of three different human cell lines. The proteins were profiled for their binding preferences to the different general types of nucleic acids. The study identified 746 high-confidence direct binders, 139 of which were novel and 237 devoid of previous experimental evidence. We could assign specific affinities for sub-types of nucleic acid probes to 219 distinct proteins and individual domains. The evolutionarily conserved protein YB-1, previously associated with cancer and drug resistance, was shown to bind methylated cytosine preferentially, potentially conferring upon YB-1 an epigenetics-related function. CONCLUSIONS: The dataset described here represents a rich resource of experimentally determined nucleic acid-binding proteins, and our methodology has great potential for further exploration of the interface between the protein and nucleic acid realms.

Viral immune modulators perturb the human molecular network by common and unique strategies.

Viruses must enter host cells to replicate, assemble and propagate. Because of the restricted size of their genomes, viruses have had to evolve efficient ways of exploiting host cell processes to promote their own life cycles and also to escape host immune defence mechanisms. Many viral open reading frames (viORFs) with immune-modulating functions essential for productive viral growth have been identified across a range of viral classes. However, there has been no comprehensive study to identify the host factors with which these viORFs interact for a global perspective of viral perturbation strategies. Here we show that different viral perturbation patterns of the host molecular defence network can be deduced from a mass-spectrometry-based host-factor survey in a defined human cellular system by using 70 innate immune-modulating viORFs from 30 viral species. The 579 host proteins targeted by the viORFs mapped to an unexpectedly large number of signalling pathways and cellular processes, suggesting yet unknown mechanisms of antiviral immunity. We further experimentally verified the targets heterogeneous nuclear ribonucleoprotein U, phosphatidylinositol-3-OH kinase, the WNK (with-no-lysine) kinase family and USP19 (ubiquitin-specific peptidase 19) as vulnerable nodes in the host cellular defence system. Evaluation of the impact of viral immune modulators on the host molecular network revealed perturbation strategies used by individual viruses and by viral classes. Our data are also valuable for the design of broad and specific antiviral therapies.

SAMHD1 is a nucleic-acid binding protein that is mislocalized due to aicardi-goutières syndrome-associated mutations.

Aicardi-Goutières syndrome (AGS) is a rare inherited autoimmune disease caused by mutations in genes encoding the RNase H2 subunits A, B, and C; the DNase three prime repair exonuclease 1 (TREX1); and sterile alpha motif (SAM) domain and HD domain-containing protein 1 (SAMHD1). Using unbiased affinity purification coupled to protein mass spectrometry, we identify SAMHD1 as a nucleic-acid-binding protein displaying a preference for RNA over DNA. In contrast to TREX1 and the RNase H2 complex, SAMHD1 has no obvious nuclease activity. In addition, interrogating truncation mutants of SAMHD1 observed in AGS patients, we map the nucleic-acid-binding domain to residues 164-442, thus overlapping with the HD domain. Furthermore, we show that although wild-type SAMHD1 displays almost exclusive nuclear localization, 11 of 12 SAMHD1 mutants show at least partial mislocalization to the cytosol. Overall, these data suggest that SAMHD1 has a role in the nucleus that, if disrupted by mutation, leads to cytosolic accumulation of SAMHD1 and autoimmune disease.

Functional dissection of the TBK1 molecular network.

TANK-binding kinase 1 (TBK1) and inducible IκB-kinase (IKK-i) are central regulators of type-I interferon induction. They are associated with three adaptor proteins called TANK, Sintbad (or TBKBP1) and NAP1 (or TBKBP2, AZI2) whose functional relationship to TBK1 and IKK-i is poorly understood. We performed a systematic affinity purification-mass spectrometry approach to derive a comprehensive TBK1/IKK-i molecular network. The most salient feature of the network is the mutual exclusive interaction of the adaptors with the kinases, suggesting distinct alternative complexes. Immunofluorescence data indicated that the individual adaptors reside in different subcellular locations. TANK, Sintbad and NAP1 competed for binding of TBK1. The binding site for all three adaptors was mapped to the C-terminal coiled-coil 2 region of TBK1. Point mutants that affect binding of individual adaptors were used to reconstitute TBK1/IKK-i-deficient cells and dissect the functional relevance of the individual kinase-adaptor edges within the network. Using a microarray-derived gene expression signature of TBK1 in response virus infection or poly(I∶C) stimulation, we found that TBK1 activation was strictly dependent on the integrity of the TBK1/TANK interaction.

IFIT1 is an antiviral protein that recognizes 5'-triphosphate RNA.

Antiviral innate immunity relies on the recognition of microbial structures. One such structure is viral RNA that carries a triphosphate group on its 5' terminus (PPP-RNA). By an affinity proteomics approach with PPP-RNA as the 'bait', we found that the antiviral protein IFIT1 (interferon-induced protein with tetratricopeptide repeats 1) mediated binding of a larger protein complex containing other IFIT family members. IFIT1 bound PPP-RNA with nanomolar affinity and required the arginine at position 187 in a highly charged carboxy-terminal groove of the protein. In the absence of IFIT1, the growth and pathogenicity of viruses containing PPP-RNA was much greater. In contrast, IFIT proteins were dispensable for the clearance of pathogens that did not generate PPP-RNA. On the basis of this specificity and the great abundance of IFIT proteins after infection, we propose that the IFIT complex antagonizes viruses by sequestering specific viral nucleic acids.

Initial characterization of the human central proteome.

BACKGROUND: On the basis of large proteomics datasets measured from seven human cell lines we consider their intersection as an approximation of the human central proteome, which is the set of proteins ubiquitously expressed in all human cells. Composition and properties of the central proteome are investigated through bioinformatics analyses. RESULTS: We experimentally identify a central proteome comprising 1,124 proteins that are ubiquitously and abundantly expressed in human cells using state of the art mass spectrometry and protein identification bioinformatics. The main represented functions are proteostasis, primary metabolism and proliferation. We further characterize the central proteome considering gene structures, conservation, interaction networks, pathways, drug targets, and coordination of biological processes. Among other new findings, we show that the central proteome is encoded by exon-rich genes, indicating an increased regulatory flexibility through alternative splicing to adapt to multiple environments, and that the protein interaction network linking the central proteome is very efficient for synchronizing translation with other biological processes. Surprisingly, at least 10% of the central proteome has no or very limited functional annotation. CONCLUSIONS: Our data and analysis provide a new and deeper description of the human central proteome compared to previous results thereby extending and complementing our knowledge of commonly expressed human proteins. All the data are made publicly available to help other researchers who, for instance, need to compare or link focused datasets to a common background.

The hepatitis C virus non-structural NS5A protein impairs both the innate and adaptive hepatic immune response in vivo.

The role of hepatitis C virus (HCV) protein non-structural (NS) 5A in HCV-associated pathogenesis is still enigmatic. To investigate the in vivo role of NS5A for viral persistence and virus-associated pathogenesis a transgenic (Tg) mouse model was established. Mice with liver-targeted NS5A transgene expression were generated using the albumin promoter. Alterations in the hepatic immune response were determined by Western blot, infection by lymphocytic choriomeningitis virus (LCMV), and using transient NS3/4A Tg mice generated by hydrodynamic injection. Cytotoxic T lymphocyte (CTL) activity was investigated by the Cr-release assay. The stable NS5A Tg mice did not reveal signs of spontaneous liver disease. The intrahepatic immunity was disrupted in the NS5A Tg mice as determined by clearance of LCMV infection or transiently NS3/4A Tg hepatocytes in vivo. This impaired immunity was explained by a reduced induction of interferon beta, 2',5'-OAS, and PKR after LCMV infection and an impairment of the CTL-mediated elimination of NS3-expressing hepatocytes. In conclusion, these data indicate that in the present transgenic mouse model, NS5A does not cause spontaneous liver disease. However, we discovered that NS5A could impair both the innate and the adaptive immune response to promote chronic HCV infection.

The TLR-independent DNA recognition pathway in murine macrophages: Ligand features and molecular signature.

Recognition of foreign DNA by cytosolic innate immune receptors triggers the production of IFN-beta. However, it is unclear whether different types of DNA ligands are recognized by similar receptors and whether the resulting response is distinct from the endosomal TLR response. To address these questions, we compared the two most commonly used types of DNA ligands (IFN-stimulatory DNA (ISD) and poly(dAdT)) and assessed the minimal structural requirements for stimulatory capacity in RAW264.7 cells. Gene expression signatures and competition experiments suggest that ISD and poly(dAdT) are qualitatively indistinguishable and differ from the CpG-containing oligonucleotides triggering the TLR9 pathway. Structure - activity relationship analyses revealed that a minimal length of two helical turns is sufficient for ISD-mediated IFN-beta induction, while phosphorylation at the 5'-end is dispensable. Altogether, our data suggest that, in murine macrophages, only one major cytosolic DNA recognition pathway is operational.

An orthogonal proteomic-genomic screen identifies AIM2 as a cytoplasmic DNA sensor for the inflammasome.

Cytoplasmic DNA triggers activation of the innate immune system. Although 'downstream' signaling components have been characterized, the DNA-sensing components remain elusive. Here we present a systematic proteomics screen for proteins that associate with DNA, 'crossed' to a screen for transcripts induced by interferon-beta, which identified AIM2 as a candidate cytoplasmic DNA sensor. AIM2 showed specificity for double-stranded DNA. It also recruited the inflammasome adaptor ASC and localized to ASC 'speckles'. A decrease in AIM2 expression produced by RNA-mediated interference impaired DNA-induced maturation of interleukin 1beta in THP-1 human monocytic cells, which indicated that endogenous AIM2 is required for DNA recognition. Reconstitution of unresponsive HEK293 cells with AIM2, ASC, caspase-1 and interleukin 1beta showed that AIM2 was sufficient for inflammasome activation. Our data suggest that AIM2 is a cytoplasmic DNA sensor for the inflammasome.