Tumor immune evasion: insights from CRISPR screens and future directions

Despite the clinical success of cancer immunotherapies including immune checkpoint blockade and adoptive cellular therapies across a variety of cancer types, many patients do not respond or ultimately relapse; however, the molecular underpinnings of this are not fully understood. Thus, a system‐level understating of the routes to tumor immune evasion is required to inform the design of the next generation of immunotherapy approaches. CRISPR screening approaches have proved extremely powerful in identifying genes that promote tumor immune evasion or sensitize tumor cells to destruction by the immune system. These large‐scale efforts have brought to light decades worth of fundamental immunology and have uncovered the key immune‐evasion pathways subverted in cancers in an acquired manner in patients receiving immune‐modulatory therapies. The comprehensive discovery of the main pathways involved in immune evasion has spurred the development and application of novel immune therapies to target this process. Although successful, conventional CRISPR screening approaches are hampered by a number of limitations, which obfuscate a complete understanding of the precise molecular regulation of immune evasion in cancer. Here, we provide a perspective on screening approaches to interrogate tumor‐lymphocyte interactions and their limitations, and discuss further development of technologies to improve such approaches and discovery capability.


Introduction
The human immune system is capable of recognizing and eliminating cells that have become cancerous.Immune cells known as killer T cells, particularly CD8+ T cells, are often the target of immunotherapies due to their ability to recognize cancer-specific antigens on the surface of cancer cells and destroy these cells.Immunotherapy, including adoptive cellular therapies (ACTs) and immune checkpoint blockade (ICB) approaches collectively aim to harness a patient's immune system to fight a cancer.Indeed, immunotherapies have revolutionized medical oncology and are now recognized as one of the core pillars of cancer treatment alongside surgery, chemotherapy, and radiation therapy and are rapidly emerging as a first-line treatment for several advanced cancers [1][2][3].This can be broadly divided into two arms: immune checkpoint inhibitors (ICIs) and ACT.ICIs bolster a pre-existing immune response by blocking endogenous immune checkpoints that normally dampen T-cell function in immune-suppressive tumor microenvironments.Upon checkpoint blockade, T cells more efficiently recognize and destroy cancer cells [4,5].Examples of checkpoint proteins found on T cells and cancer cells include PD-1/PD-L1 and CTLA-4/B7-1/B7-2 [6], which have demonstrated remarkable success in some cancer patients including melanoma and nonsmall cell lung cancer (NSLC) [7][8][9].However, primary resistance frequently occurs in other cancer types and individuals who respond to treatment early on eventually relapse due to acquired resistance [10].In contrast to ICIs, ACT and bispecific T-cell engagers are engineered to redirect the immune response to tumor epitopes/antigens [11][12][13].ACTs can be deployed in different ways, including engineered T-cell receptor (TCR) therapy and chimeric antigen receptor (CAR) T-cell therapy.These approaches rely on isolating a patients' immune cells, genetically engineering them to better target the cancer, then reinfusing them back into the patient [14].While such approaches have enjoyed much success in hematological cancers, success in solid cancers is poor and patient relapse occurs frequently.Importantly, clustered regularly interspaced short palindromic repeats (CRISPR) technology has enabled rapid and high-throughput identification of avenues leading to tumor escape from the immune system and potential therapeutic strategies to circumvent it.Indeed, such approaches including CRISPR/Cas9 knockout screens and base-editing screens have identified genes and the mutations that drive immune evasion in patients, emphasizing the power of applying diverse Cas enzyme-based screening technologies to cancer immunology.

CRISPR/Cas technology
The accomplishment of the Human Genome Project in the early 2000s has provided a fundamental genetic human blueprint, offering accessible and unlimited information to decipher the inner working machinery of human biology.Many years after its completion, we still have not discovered the function of overwhelming number of genes and more importantly, how their dysregulations promote aberrant biological processes underlying various human diseases.One common way to discover a gene function is by generating loss-of-function mutation and investigating the resulting phenotype.First, large-scale genetic screens were conceived with the discovery of short hairpin RNA (shRNA)mediated gene knockdown, built on the discovery of natural RNAi in Caenorhabditis elegans [15].ShRNA screens are based on introducing a library of shRNAs into cells and select or screen for a particular phenotype, such as resistance to a drug, increased cell growth, or cell death.
The power of genetic screens invited scientists to utilize other technologies, such as zinc-finger nucleases (ZFNs) [16][17][18][19][20][21], transcription activator-like effector nucleases (TALENs) [22,23], and finally, CRISPR [24][25][26][27][28], as the most modular and utilized gene-editing techniques.TALENs, in particular, are based on prokaryotic TAL effector proteins, that bind DNA in a sequence-specific manner and a fused FokI nuclease domain, that introduces DNA double-strand breaks, resulting in locus-specific insertions and deletions.TALENs, due to their design complexity, have not been as popular for broad genetic screens and are rather used for smaller focused genetic screens.
CRISPR was originally discovered as unique DNA repeat elements that contribute to archaeal and bacterial adaptive immune responses toward invading foreign DNA [26,27,29].Over the last decade, it has become an indispensable tool for biological research with CRISPR/Cas9 system extensively utilized as a genomic editing technique.The system consists of an endonuclease Cas enzyme that acts as a 'molecular scissor', which can induce double-strand DNA break (DSB) at a specific location dictated by a complementary guide RNA.This DSB can be repaired through homology-directed repair or nonhomologous end-joining, of which the latter commonly causes gene disruption [28].CRISPR Cas9 application is wide, ranging from successful gene knockout/knock-in in numerous model organisms in vitro and in vivo, to high-throughput bulk and single-cell genetic screens that can efficiently identify genes and pathways associated with biological phenotypes [30].CRISPR screens have many advantages over other genome editing techniques including siRNA that commonly induces incomplete gene inactivation leading to ambiguous results or TALEN and ZFN that are both costly and inefficient due to their reliant on protein-DNA interaction for sequence specificity [31].Since its discovery, CRISPR has expanded its application beyond genome editing through the development of Cas variants.For example, the catalytically inactivate Cas9, dCas9, can be paired with different effector parts such as VP64 and KRAB to activate or inhibit transcription respectively [32].Instead of recognizing dsDNA, other endonucleases of the Cas13 family, such as RfxCas13b/d, can bind to ssRNA thereby, inhibiting translation without altering the genomic sequence [33].Here, we will discuss the flexibility and various aspects of CRISPR/Cas technology that have been implemented to elucidate the pathways driving tumor immune evasion.

Application of CRISPR screens for immunotherapy target discovery
Genome-scale CRISPR screens for immune discovery often involve a two-cell-type co-culture system where a sgRNA library is introduced into a tumor cell line which is engineered to express Cas9.This approach has now been conducted across many tumor cell lines derived from various tissue types [34][35][36].Typically, these screens involve an antigen/TCR matched system, where clonally expanded T cells are co-cultured with tumor cells either expressing an endogenous antigen or are engineered to do so.Such screens have been executed using mouse transgenic TCR CD8+ T cells, that is, OT-I or Pmel and cell lines engineered to express the corresponding antigen on major histocompatibility complex class I (MHC-I).This approach has also been conducted in human cancer cell lines where primary human T cells are transduced with a redirected TCR vector, designed to react with an endogenous peptide displayed on the appropriate human leukocyte antigen (HLA) molecule, for example, NY-ESO1 or MAGE.There are important factors to consider when conducting such screens.Firstly, the level of selection pressure can be toggled through titration of the effector to target (E : T) ratio.A high E : T ratio is more suitable for identifying factors, that when lost, provide protection from killing by CD8+ T cells.Conversely, lower E : T ratios, and therefore, less selection pressure, is amenable to identifying factors that when deleted, sensitize to killing by CD8+ T cells.Furthermore, the level of selection pressure can be managed through several rounds of exposure to the effector cells after the surviving tumor cells have re-populated the in vitro culture.Such approaches can also be performed in vivo; however, this approach comes with several technical obstacles and drawbacks which will be discussed in further detail.

Mechanisms of tumor cell-intrinsic immune evasion from CD8+ T cells
Broadly speaking, three main pathways, which when disrupted, lead to immune evasion have emerged from these approaches; tumor necrosis factor (TNF)induced cell death, interferon signaling to MHC-I expression, and antigen processing and presentation (Fig. 1).Speaking to the power of these screens, loss of antigen presentation is a known tumor immune evasion mechanism in patients that fail to respond to ICB [37,38].Indeed, multiple components of the antigen presentation machinery include B2M, TAP1, TAP2, and TAPBP, and have been identified across a variety of genome-wide CRISPR screens using CD8+ T-cell selection pressure [34,36,39].
Multiple components of the interferon-gamma pathway (JAK1, JAK2, IFNGR1, IFNGR2, STAT1, STAT2, and APLNR) have been identified in screens for increased resistance to CD8+ T-cell killing [34,35,39,40].Indeed, interferon-gamma is known to be important for antitumor immunity through promoting the expression of MHC-I.Consistent with this, patients have been identified with JAK1/2 mutations that fail to respond to ICB [38], speaking to the power of such screening approaches.In the context of sensitization, the BAF complex components Prbm1, Arid2, and Brd7 have been identified as epigenetic suppressors of interferon-gamma-induced genes and thus loss of Prbm1 sensitized to ICB in preclinical models [36].Similarly, Manguso et al. [40] identified Ptpn2 as a negative regulator of interferon-gamma signaling strength, and thus Ptpn2 loss sensitized to CD8+ T-cell killing.Furthermore, the tumor cell autophagy machinery has emerged as regulators of immune evasion through protecting from the cytotoxic effects of TNF and IFNc [41,42].
Perhaps the most surprising key pathway that emerged from CRISPR screening in this context was the importance of TNF-induced cell death.Until recently, the importance of TNF-induced cell death as a CD8+ T-cell effector mechanism was relatively unappreciated.Indeed, knockout of multiple TNF-induced cell death pathway members, including Tnfrsf1a, Tnfrsf1b, Casp8, and Tradd, robustly protected from T-cell-mediated killing in vitro and in vivo [34,43,44].It is important to note that TNF serves predominantly as a pro-survival, pro-inflammatory cytokine, triggering NfKb activation and synthesis of multiple cytokines and chemokines to amplify inflammation [45].However, multiple CRISPR screens have identified that several genes within the TNF pathway that are targeted to sensitize tumor cells to T-cell-derived TNF-induced cell death (Traf2, Ripk1, Ciap1/2, Rbck1, and Rnf31) [39,44,46,47].Importantly, it has now been demonstrated that small molecule inhibitors of such proteins, for example, Smac mimetics that trigger cIAP1/2 degradation, can be utilized to sensitize tumor cells to TNF-induced cell death and improve response to ICB [44,48].Similarly, small molecule inhibitors of the LUBAC component HOIP sensitize to CD8+ T-cell killing [46].It is interesting to note that other death receptor pathways including Fas/ CD95 and TRAIL have not been identified in these screens, highlighting the importance of TNF as a CTL effector molecule.This is not to say that other death receptor pathways are dispensable, since a genomewide screen with the Brunello library identified both FADD and its downstream effector BID as a key mediator for CD-19-specific CAR T-cell killing of Nalm6 cells [49].Simultaneously, deletion of either FAS-L or TRAIL in T cells significantly reduced CART19 cytotoxicity, whereas the effect of antibody  4) (NBR1 and ATGs) enhances the levels of MHC-I and improves antitumor immunity.Finally, loss of death receptor signaling components (TNF, CAS8, and TRAF2) impairs the extrinsic apoptotic pathway induced by T cells, further enabling immune evasion by cancer cells.CAS8, caspase-8; IFNc, interferon-gamma; JAK1, Janus kinase 1; MHC-1, major histocompatibility complex-I; NBR1, neighbor of BRCA1 gene 1 protein; STAT1, signal transducer and activator of transcription 1; TAP1, transporter associated with antigen processing 1; TNF, tumor necrosis factor; TRAF2, TNF receptor-associated factor 2; b2M, b2 microglobulin.
treatment against TNF was mild [49], suggesting that the primary tumor cell death pathway induced by T cells is cell type and context dependent.
Recently, in a seminal paper by Martin et al., the authors utilize CRISPR screens both in vitro and in vivo, using transplantable tumor models in both immune-competent and immune-deficient (NSG) mice simultaneously.The authors screened a sgRNA library of 7500 putative druggable target genes and identified a marked enrichment for the loss of tumor suppressor genes (TSGs), only in the presence of an intact immune system [50].The findings suggest a major role for the immune system in driving tumor evolution and that the adaptive immune system is a major driver of selection for TSG inactivation.

IFNc signaling in tumor immune evasion: a double-edged sword
It should be noted that recent evidence has revealed that IFNc signaling and expression of classical and nonclassical MHC-I molecules can act as a double-edge sword in tumor-lymphocyte interactions in vivo.Using in vivo genome-scale CRISPR screens, Dubrot et al. [35] reported that loss of IFNc signaling sensitizes many transplantable cancer models to immunity.The immune inhibitory effects of tumor intrinsic IFNc signaling were shown to occur through two mechanisms.Firstly, IFNc-induced tumor upregulation of classical MHC class I inhibited natural killer cell-mediated antitumor immunity.Secondly, IFNc-induced expression of nonclassical MHC-I molecule Qa-1b inhibited CD8+ T-cell-driven immunity through the NKG2A/CD94 receptor.Similarly, a study by Benci et al. [51] found that when tumor IFNc signaling is blocked, it reduces interferon-stimulated genes (ISGs) in cancer cells but increases ISGs in immune cells by boosting IFNG produced by exhausted T cells.In settings where tumors have high immunogenicity, these exhausted T cells can drive tumor control.In tumors with MHC-I loss, exhausted CD8+ T cells use IFNc to drive innate immune cell maturation, including ILC1 populations.Thus, by blocking an inhibitory circuit involving PD1 and TRAIL, inhibiting tumor IFNG signaling promotes innate immune killing.Taken together, this study suggests that the interplay of IFNc signaling in cancer cells and immune cells establishes a regulatory relationship that limits both adaptive and innate immune killing.Furthermore, by inhibiting the IFN-I signaling in cancer cells, another recent study from the same group reported that the response to antiprogrammed cell death protein 1 (PD1) therapy can be restored.This restoration occurs through increased IFNc production in immune cells, which promotes interactions between dendritic cells and CD8+ T cells and leads to the expansion of T cells into effector-like states rather than exhausted states [52].Similarly, prolonged interferon signaling orchestrates PDL1-dependent and PDL1independent resistance to ICB through acquired STAT1-driven epigenetic changes that drives expression of ISGs and ligands for multiple T-cell inhibitory receptors [53].Thus, the role of IFNc is complex and is likely to be context dependent, with variation in the relative contribution to immune evasion likely to depend on the tumor model used for CRISPR screening.

Immune evasion and sensitization to alternative killer cells
Unlike CD8+ T cells, NK cells can efficiently recognize and kill tumors with low MHC-1 expression or expression of stress ligands such as those that ligate NKG2D or CD112/CD155 [54,55].One-way tumor cells evade NK cell recognition through the loss of expression of various ligands activating NK cells including NK cell receptor DNAM-1 ligand CD155 [54] and non-MHC ligand CD48 [56].Consistent with inhibitory effect of MHC-I on NK cells, various in vitro CRISPR screens have confirmed the loss of antigen presentation (HLA-A/B/C/E, TAP1, and TAP2) and interferon-c-signaling pathways (IFNcR2, JAK2, and STAT1) enhance NK cells induced killing [57][58][59].Although inhibiting IFNc sounds like an attractive strategy [51,53], the application should proceed with caution as IFNc loss can promote tumor immune evasion from CD8+ T cells [34].Since both CD8+ T and NK cells can utilize TNF to induce tumor cell apoptosis, targeting TNF signaling pathway, as mentioned above, appears as a more attractive strategy [43].It will be interesting to use CRISPR-based approaches to interrogate tumor intrinsic resistance mechanisms to alternative cell types with lytic activity such as gamma delta T cells, MAIT cells, and cytotoxic CD4+ T cells; such studies are yet to be performed but may uncover novel immunotherapy approaches through harnessing alternative cells of the immune system and engineered cells such as CAR T cells and re-directed TCR T cells.Certainly, to date, CD8+ T cells have been the focus of CRISPR screens to elucidate immune evasion mechanisms.

CRISPR screening for lymphocyte intrinsic functions
Numerous CRISPR screens conducted across multiple cancer cell lines have successfully identified cellintrinsic tumor evasion mechanism, albeit with some limitations (Fig. 2).More importantly, only a handful of studies have conducted such screens to dissect primary immune cell function.Examples include genomewide CRISPR screens in Th2 CD4+ T cells, dendritic cells, and CD8+ T cells [60,61].This is likely because lenti/retroviral transduction, the commonly used method for delivering CRISPR/Cas components, still faces various technical barriers in introducing genes into primary immune cells.Indeed, this method of delivery usually renders low transduction efficiency resulting from antiviral activities of CD8+ T and NK cells as well as the difficulties of expressing larger sized proteins such as Cas9.While direct delivery of Cas9 ribonucleoprotein and pretranscribed sgRNA via electroporation have been favorable for single gene knockout in both T and NK cells [62], this method is less Fig. 2. Limitations and potential solutions associated with CRISPR/Cas9 screening for investigation of immune evasion mechanisms from cytotoxic T cells.Three major challenges are addressed, including the difficulty of screening lethal genes (1), the dominance of certain 'known' genes in the hit list, insufficient representation of sgRNA in vivo (2), and low transduction efficiency for T cells (3).
feasible for CRISPR screens due to the requirement of sgRNA pool library introduction and sequencing identification via the integrated DNA barcode.Most notably, easy-to-transfect PiggyBac transposon system may provide an alternative delivery system as it allows insertion of large DNA cargo for up to 100 kb and by extension, Cas9 and sgRNA simultaneously from one vector [63].Alternatively, an sgRNA cassette delivered using the Sleeping Beauty (SB) transposon [64] and a nonintegrating adeno-associated virus (AAV) vector, has successfully uncovered novel insights on the previously underappreciated gene targets in CD8+ T cells within mouse models of glioblastoma [65,66].The transposon system may have advantages over randomized integration of genes with lenti/retroviral delivery.The PiggyBac system, for example, favors DNA cargo insertion into 'TTAA' sequences located in the euchromatin of mammalian chromosomes, of which it can easily be removed.This allows reversible CRISPR application, which is useful for dCas9 and Cas13 familygene-mediated transcription modulation [30,33,67].

Genome-wide versus targeted CRISPR libraries
Generally, in vitro CRISPR screens to identify resistance and sensitization mechanisms to killer cell attack have interrogated genome-wide sgRNA libraries, due to the relative ease of representing libraries containing up to 150 000 sgRNAs with at least 2509 coverage.Certainly, this approach allows the elucidation of a gene network/pathway hierarchy where the relative importance of such factors for T-cell interactions can be assessed across the genome.A caveat to this approach, however, is that loss of particular genes or pathways can dominate resistance or sensitization mechanisms, hampering the ability to detect more subtle mechanisms which may indeed be relevant biologically and in a clinical context.For example, loss of ability to present antigen through genetic deletion of B2m will of course dominate screening results in a positive enrichment screen.Similarly, if the cell line used for screening is exquisitely sensitive to the cytotoxic effects of TNF, for example murine MC38 cells, deletion of members of the TNF-induced cell death pathway will likely dominate results and mask identification of other factors at play [34].In this regard, it is important to consider if the level of cytokines present in an in vitro, antigen-TCR-matched co-culture system accurately re-capitulate CTL-derived cytokine-meditated selection pressure in vivo.To circumvent this issue in the context of domination by Ifng-mediated effects, Vredevoogd et al. [44] employed an elegant approach where a genome-scale screen upon MART1 T-cell selection pressure was performed in tumor cells on an interferongamma receptor 1 (IFNGR1) knockout background.Indeed, it was found that TNF signaling dominated the IFNc-independent CD8 T-cell-associated tumor vulnerability landscape.In this regard, it would be interesting to conduct a similar screen on a IFNG receptor and TNF receptor double knockout background to identify secondary mechanisms of immune evasion/sensitization.
An alternative approach to avoid 'hit' domination is to screen targeted, smaller, compartmentalized sgRNA libraries.For example, sgRNA libraries can be designed that target particular gene families, for example, kinases, phosphatases, transcription factors, epigenetic factors, and trans-plasma membrane proteins.This approach has been utilized less frequently and offers an excellent opportunity for novel immunotherapy target identification.An elegant example of such an approach is a study by Dervovic et al. [68], where they utilize a targeted CRISPR library of 573 genes associated with altered cytotoxicity in human cancers to study immune-modulatory cancer genes, directly in murine lung cancer models.Here, the authors introduce the lentiviral sgRNA library into tumor-prone mice (Kras G12D or Braf V600E ) at P2, followed by OT-I CD8+ T-cell-mediated selection pressure.They identify the known immune evasion factors Stat1 and Serpinb9 and also Adam2 as an immune modulator.Indeed, ADAM2 was shown to restrain interferon and TNF cytokine signaling causing reduced presentation of tumor-associated antigens.The elegance of this approach lies in the maintenance of proper tumor tissue architecture and microenvironment, which may be disrupted due to wound healing responses when using the popular transplantable, subcutaneous tumor models.

In vitro versus in vivo CRISPR screens: limitations and solutions
While in vitro CRISPR screens with CD8+ T-cell cocultures have been immensely successful in identifying many factors involved in tumor immune evasion and sensitization, such screens are biased toward uncovering genes regulating antigen presentation via MHC-I, and they usually do not take into account of the complex interaction of the adaptive immune system within tumor microenvironment in vivo [50].CRISPR screens in vivo, however, come with the caveat of several technical challenges.Firstly, achieving sufficient representation of genome-wide gRNAs in an in vivo tumor model is challenging.To achieve at least 2509 representation of gRNAs would require implanting up to 25 million tumor cells into an animal, which is not compatible with most syngeneic tumor models.Furthermore, this does not account for the engraftment rate of the tumor cells, which may be highly variable and unknown across divergent syngeneic models.One potential way to circumvent this issue is to use multiple mice to sufficiently represent the sgRNA library; however, this is a costly approach, requires pooling of tumor genomic DNA from multiple mice, which introduces heterogeneity, and the experiment is not internally controlled.Thus, smaller, targeted CRISPR library screening is more achievable and appropriate in vivo, as first demonstrated by Manguso et al. [40], identifying Ptpn2 as a negative regulator of IFNcinduced MHC-I expression.The caveat of this approach is that bias is inherently introduced through selection of the genes to be screened.In this regard, one possible approach to screen the entire genome in vivo could be to split the genome up into smaller subpools and screen them systematically.An alternative approach to circumvent engraftment issues has been demonstrated by LaFleur et al. [69] using bone marrow chimeras, a technology called CHIME: CHimeric IMmune Editing.Here, hematopoietic stem cells from Cas9+ mice are transduced with sgRNAs and used to reconstitute the immune system.This technology could in theory be used to screen immune evasion mechanisms of hematopoietic malignancies.It should be noted genome-wide CRISPR screens have been performed in vivo in immunodeficient animals, as this may improve the tumor engraftment rate [70].Here, the GeCKOv2 KO lentiviral library was introduced into SUM159 cells and 30 million cells were subcutaneously transplanted into six immunodeficient NOD scid gamma (NSG) mice.Thirty days posttransplantation, tumor samples were processed through next-generation sequencing and the pooled sgRNA abundance and distribution, following in vivo selective pressure.Indeed, genomic samples contained an average sgRNA library representation of 99-97%, indicating sufficient library coverage.Certainly however, such approaches remain a challenge for immune discovery where immune-competent mice are required and engraftment rates likely to be considerably lower.Such an approach may be useful for interrogation of sensitization/resistance to ACTs where the use of immunodeficient animals is more suitable.

Other limitations of CRISPR/Cas9 screens for immune discovery
One major obstacle to performing Cas9-based screens in vivo is that Cas9 itself is inherently immunogenetic, triggering an endogenous immune response to the cells engineered to express it.To circumvent this issue, Cas9-positive cells can be transferred into cas9+ mice, which would be tolerized to cas9 antigens.This is particularly important when performing adoptive cell transfer experiment, such as primary Cas9+ CD8+ T cells.Another option is to transiently introduce recombinant cas9 protein via electroporation, into a population of cells carrying a sgRNA library.While this has been performed [71], in an approach termed single guide RNA (sgRNA) lentiviral infection with Cas9 protein electroporation (SLICE), the knockout efficiently compared with stably integrated Cas9 remains unclear.
Another major caveat to CRISPR screening is that essential genes for cell viability cannot be screened for their involvement in a particular biological process.It is estimated that 10% of genes are essential for cell viability [72]; thus, 'genome-wide' screens are in fact not genome wide; they are screens that interrogate a subset of the genome that is nonessential for viability.Cas13 is an enzyme that specifically targets and degrades RNA molecules, rather than genomic DNA, and thus, is a reversible and tunable system.The level of knockdown can be toggled by altering the expression of Cas13, or by having the Cas13 or sgRNA vector inducible.This would allow transient and reversible knockdown of transcripts (which may facilitate survival of cells where an essential gene has been targeted).While genome-wide Cas13 screens have been performed [73], this approach has yet to be utilized for discovery in cancer immunology or other immune contexts.We believe this approach will uncover important findings in the context of tumor immune evasion in the near future.

Bulk versus single-cell CRISPR screening: a comparative perspective
In recent years, the advancement of high-throughput screening methodologies using CRISPR technology has transformed our understanding of genetic functional landscapes.Single-cell CRISPR screening is the next evolution in pooled CRISPR screens.Conventional bulk CRISPR screens can mainly assay for simple phenotypes, such as cellular viability or reporter readouts, and are limited to single-cell types or require flow cytometry sorting of cell types.It enables affordable large-scale, genome-wide screens, providing valuable insight into genetic landscapes.However, it falls short in offering high-resolution data at the individual cell level, which is especially important when investigating various immune factors within diverse immune cell populations and cancer cells.
The advances in single-cell RNA sequencing and the parallel readout of CRISPR guide RNA and transcriptome on a single-cell level allow for a much richer phenotypic readout while maintaining high scalability and provides a more nuanced view, allowing for massively parallel functional genomics readouts.This technique grants a deeper understanding of regulatory networks, by identifying multiple potential targets upstream and downstream of impacted pathways.Single-cell screens also circumvent viability or reporter discrimination, another advantage over the bulk method.However, they are limited by the significant cost required for genome-wide coverage, which depends on negative or positive selection and can range from 100 to 1000 cells per gene [30].The finer the granularity desired from each gene perturbation, the more cells/reads-per-cell required, making this approach quite costly.
Moreover, single-cell screening presents challenges in sample preparation and requires considerable technical expertise.Notably, the bioinformatics analysis in single-cell screens is often labor-intensive and lacks well-established protocols.Furthermore, single-cell CRISPR screening technologies are still in their infancy.A large portion of the published papers focus on assessing data quality, technology, reproducibility, and robustness of the method, illustrating the need for continued refinement and standardization of these techniques.Indeed, reading out CRISPR guide RNAs has evolved from indirect barcode detection within a polyadenylated transcript captured together with the rest of the transcriptome (Perturb-Seq and CRISP-Seq) to a readout of a direct copy of the protospacer into a polyadenylated transcript (CROP-Seq) and direct capture of the functional CRISPR guide RNA (109 Genomics 5 0 CRISPR kit) [74][75][76].
Interestingly, bulk screening often serves as a foundation for single-cell screens.The data from bulk screens can guide the optimization of single-cell screens, offering insight into which guides are effective and which ones are not.This allows researchers to refine their approach and focus their resources on a narrower range of guides in single-cell screening.In addition, bulk screens help in determining the expected heterogeneity of guides at certain time points, informing how many cells to collect and optimize selection pressure regarding timing and dosage.
In conclusion, while both bulk and single-cell CRISPR screening have significantly advanced our understanding of genetic functional landscapes, each approach has its unique challenges and opportunities.The continued evolution and optimization of these techniques, individually and in concert, will undoubtedly provide further depth and nuance to our understanding of genomic interactions and the complex biological processes they regulate.

Technological advancements poised to improve CRISPR screens for immune discovery
Immune screens have been at the forefront of decoding intricate interactions within biological systems, offering promising perspectives for disease treatment and prevention.With the ongoing evolution of CRISPR screening technologies (Fig. 3A), we can envision several advancements that could augment the power of immune screens in the near future.
A major advancement lies in the integration of multiomic readouts with single-cell CRISPR (scCRISPR) screens.Combining genomic, transcriptomic, epigenetic, and proteomic data from individual cells can provide a holistic view of cellular processes under different genetic perturbations (Fig. 3B).Epigenetic readouts can offer additional layers of information beyond gene sequences, such as DNA accessibility, histone modification, and DNA binding proteins, providing valuable insights into gene regulation and functional noncoding regions.Simultaneously, proteomic profiling using techniques like CITE-Seq can help distinguish cell subtypes within immune cells.Fig. 3. Cutting-edge technological advancements to revolutionize the landscape of CRISPR screens for immune discovery.(A) The application of CRISPR/Cas goes far beyond traditional gene knockout as it now encompasses multilevel perturbations, including transcriptional regulation, epigenetic modifications, and precise genome modulation.Transcriptional regulation techniques such as CRISPR-activation and CRISPR-interference utilize modified Cas9 protein that is attached to specific regulatory elements to selectively up-or downregulate both coding and noncoding genes.Notably, the use of Cas13, another variant of Cas protein, enables the inactivation of gene expression by directly cleaving mRNA.Modified Cas9 protein, when attached to other regulatory elements, can modify the epigenome by editing DNA methylation or histone marks.Finally, Base-editing (BE) techniques directly modify specific nucleotides in the genome by either altering cytosine to thymine (CBE) or adenine to guanine (ABE).(B) Integration of multiomics readouts with single-cell CRISPR (scCRISPR) screen offers a powerful tool to decode complex cellular processes under different genetic perturbations (from A).Forefront sequencing technologies for genomics, epigenetics, and transcriptomics readouts are highlighted, including ones that analyze chromosome structure, chromatin accessibility, histone, and DNA modifications as well as RNA expression.Proteomic and transcriptomic readouts as CITE-Seq can be used to identify unique immune cell subsets.(C) Combinatorial screen (one in T cells and one in tumor cells) and spatial mapping techniques can unravel intricate genetic interactions between various types of immune cells and also cancer cells within the tumor microenvironment.
Along with multiomic readouts, the spatial context of cellular interactions can be leveraged to boost the power of immune screens (Fig. 3C).The advent of spatial transcriptomics allows the mapping of gene expression within the tissue microenvironment, leading to the discovery of novel cellular interactions and dependencies.Focusing on the tumor microenvironment and the interactions between T cells and other cell types such as tumor cells and stromal cells can offer critical insights into immune response, evasion, and tumorigenesis [77].
Combinatorial CRISPR screens, another promising direction, can help to decipher complex genetic interactions [78].By concurrently screening two librariesone in T cells and one in tumor cells-it was recently shown that we can probe the cross-talk between these two key players in cancer immunity [79].This combinatorial approach could elucidate novel cooperative or antagonistic gene interactions involved in cancer progression and immune evasion.Alternatively, single-cell CRISPR screening using CRISPR arrays, targeting multiple genes and small CRISPR libraries transduced at higher multiplicity of infection (MOI), thus achieving multiple guides per cell, can support investigations into combinatorial effects of gene perturbations.
CRISPR/Cas technology advancements directly feed into innovative CRISPR screening applications, such as ORF (Open Reading Frame) screens and baseediting screens, which add a new dimension to immune screens.The discovery of novel enzymes, such as the Fanzor [80] and transposases [64], offers new tools for genome manipulation.Incorporating these enzymes into screening protocols could improve the efficiency of gene editing and introduce new capabilities.In summary, these anticipated technological advancements are set to enhance the power of immune screens.By integrating multiomics, spatial context, combinatorial CRISPR screens, and novel enzymatic tools, we can anticipate a deeper understanding of complex cellular interactions and, ultimately, better-informed therapeutic strategies.
In conclusion, CRISPR-based applications have dramatically accelerated our understanding of tumor-Tcell interactions and the factors that promote tumor immune evasion or indeed, sensitize them to T-cellmediated killing.While there are several limitations and challenges relating to these approaches, technological advances including improved gRNA/Cas delivery methods and the increased availability of single-cellbased readouts for CRISPR-based screens, further important discoveries are likely to arise, shedding more light on the tumor-T-cell immunity cycle.

1392
The FEBS Journal 291 (2024) 1386-1399 ª 2023 The Authors.The FEBS Journal published by John Wiley & Sons Ltd on behalf of Federation of European Biochemical Societies.