In vivo screens using a selective CRISPR antigen removal lentiviral vector system reveal immune dependencies in renal cell carcinoma

•CRISPR vector components are immunogenic and cause rejection in mouse cancer models•The SCAR vector system allows removal of immunogens from CRISPR-edited cells•SCAR enables in vivo screening of antigen-sensitive cell lines•Renal cell carcinoma requires autophagy and MHC class I expression for immune evasion CRISPR-Cas9 genome engineering has increased the pace of discovery for immunology and cancer biology, revealing potential therapeutic targets and providing insight into mechanisms underlying resistance to immunotherapy. However, endogenous immune recognition of Cas9 has limited the applicability of CRISPR technologies in vivo. Here, we characterized immune responses against Cas9 and other expressed CRISPR vector components that cause antigen-specific tumor rejection in several mouse cancer models. To avoid unwanted immune recognition, we designed a lentiviral vector system that allowed selective CRISPR antigen removal (SCAR) from tumor cells. The SCAR system reversed immune-mediated rejection of CRISPR-modified tumor cells in vivo and enabled high-throughput genetic screens in previously intractable models. A pooled in vivo screen using SCAR in a CRISPR-antigen-sensitive renal cell carcinoma revealed resistance pathways associated with autophagy and major histocompatibility complex class I (MHC class I) expression. Thus, SCAR presents a resource that enables CRISPR-based studies of tumor-immune interactions and prevents unwanted immune recognition of genetically engineered cells, with implications for clinical applications. CRISPR-Cas9 genome engineering has increased the pace of discovery for immunology and cancer biology, revealing potential therapeutic targets and providing insight into mechanisms underlying resistance to immunotherapy. However, endogenous immune recognition of Cas9 has limited the applicability of CRISPR technologies in vivo. Here, we characterized immune responses against Cas9 and other expressed CRISPR vector components that cause antigen-specific tumor rejection in several mouse cancer models. To avoid unwanted immune recognition, we designed a lentiviral vector system that allowed selective CRISPR antigen removal (SCAR) from tumor cells. The SCAR system reversed immune-mediated rejection of CRISPR-modified tumor cells in vivo and enabled high-throughput genetic screens in previously intractable models. A pooled in vivo screen using SCAR in a CRISPR-antigen-sensitive renal cell carcinoma revealed resistance pathways associated with autophagy and major histocompatibility complex class I (MHC class I) expression. Thus, SCAR presents a resource that enables CRISPR-based studies of tumor-immune interactions and prevents unwanted immune recognition of genetically engineered cells, with implications for clinical applications. The CRISPR-Cas9 genome engineering system is a powerful tool for studying cell biology. CRISPR can interrogate the function of genes and pathways in complex biological systems (Jinek et al., 2012Jinek M. Chylinski K. Fonfara I. Hauer M. Doudna J.A. Charpentier E. A programmable dual-RNA-guided DNA endonuclease in adaptive bacterial immunity.Science. 2012; 337: 816-821Crossref PubMed Scopus (7874) Google Scholar; Cong et al., 2013Cong L. Ran F.A. Cox D. Lin S. Barretto R. Habib N. Hsu P.D. Wu X. Jiang W. Marraffini L.A. Zhang F. Multiplex genome engineering using CRISPR/Cas systems.Science. 2013; 339: 819-823Crossref PubMed Scopus (8856) Google Scholar; Hsu et al., 2013Hsu P.D. Scott D.A. Weinstein J.A. Ran F.A. Konermann S. Agarwala V. Li Y. Fine E.J. Wu X. Shalem O. et al.DNA targeting specificity of RNA-guided Cas9 nucleases.Nat. Biotechnol. 2013; 31: 827-832Crossref PubMed Scopus (2710) Google Scholar; Mali et al., 2013Mali P. Yang L. Esvelt K.M. Aach J. Guell M. DiCarlo J.E. Norville J.E. Church G.M. RNA-guided human genome engineering via Cas9.Science. 2013; 339: 823-826Crossref PubMed Scopus (5800) Google Scholar; Doench et al., 2014Doench J.G. Hartenian E. Graham D.B. Tothova Z. Hegde M. Smith I. Sullender M. Ebert B.L. Xavier R.J. Root D.E. Rational design of highly active sgRNAs for CRISPR-Cas9-mediated gene inactivation.Nat. Biotechnol. 2014; 32: 1262-1267Crossref PubMed Scopus (825) Google Scholar) and edit genomes for clinical applications such as gene therapy and chimeric antigen receptor (CAR) T cell therapies (Eyquem et al., 2017Eyquem J. Mansilla-Soto J. Giavridis T. van der Stegen S.J. Hamieh M. Cunanan K.M. Odak A. Gönen M. Sadelain M. Targeting a CAR to the TRAC locus with CRISPR/Cas9 enhances tumour rejection.Nature. 2017; 543: 113-117Crossref PubMed Scopus (752) Google Scholar; Liu et al., 2017Liu X. Zhang Y. Cheng C. Cheng A.W. Zhang X. Li N. Xia C. Wei X. Liu X. Wang H. CRISPR-Cas9-mediated multiplex gene editing in CAR-T cells.Cell Res. 2017; 27: 154-157Crossref PubMed Scopus (169) Google Scholar). CRISPR has made pooled genetic screens feasible for a range of biological questions (Doench et al., 2014Doench J.G. Hartenian E. Graham D.B. Tothova Z. Hegde M. Smith I. Sullender M. Ebert B.L. Xavier R.J. Root D.E. Rational design of highly active sgRNAs for CRISPR-Cas9-mediated gene inactivation.Nat. Biotechnol. 2014; 32: 1262-1267Crossref PubMed Scopus (825) Google Scholar; Shalem et al., 2014Shalem O. Sanjana N.E. Hartenian E. Shi X. Scott D.A. Mikkelsen T.S. Heckl D. Ebert B.L. Root D.E. et al.Genome-scale CRISPR-Cas9 knockout screening in human cells.PLoS Genet. 2014; 4: e9Google Scholar; Wang et al., 2014Wang T. Wei J.J. Sabatini D.M. Lander E.S. Genetic screens in human cells using the CRISPR-Cas9 system.Science. 2014; 343: 80-84Crossref PubMed Scopus (1687) Google Scholar; Chen et al., 2015Chen S. Sanjana N.E. Zheng K. Shalem O. Lee K. Shi X. Scott D.A. Song J. Pan J.Q. Weissleder R. et al.Genome-wide CRISPR screen in a mouse model of tumor growth and metastasis.Cell. 2015; 160: 1246-1260Abstract Full Text Full Text PDF PubMed Scopus (493) Google Scholar). For cancer immunology, in vitro CRISPR screens have identified genes that modulate T cell or natural killer (NK) cell cytotoxicity (Patel et al., 2017Patel S.J. Sanjana N.E. Kishton R.J. Eidizadeh A. Vodnala S.K. Cam M. Gartner J.J. Jia L. Steinberg S.M. Yamamoto T.N. et al.Identification of essential genes for cancer immunotherapy.Nature. 2017; 548: 537-542Crossref PubMed Scopus (391) Google Scholar; Kearney et al., 2018Kearney C.J. Vervoort S.J. Hogg S.J. Ramsbottom K.M. Freeman A.J. Lalaoui N. Pijpers L. Michie J. Brown K.K. Knight D.A. et al.Tumor immune evasion arises through loss of TNF sensitivity.Sci. Immunol. 2018; 3: eaar3451Crossref PubMed Scopus (101) Google Scholar; Pan et al., 2018Pan D. Kobayashi A. Jiang P. Ferrari de Andrade L. Tay R.E. Luoma A.M. Tsoucas D. Qiu X. Lim K. Rao P. et al.A major chromatin regulator determines resistance of tumor cells to T cell-mediated killing.Science. 2018; 359: 770-775Crossref PubMed Scopus (348) Google Scholar; Shifrut et al., 2018Shifrut E. Carnevale J. Tobin V. Roth T.L. Woo J.M. Bui C.T. Li P.J. Diolaiti M.E. Ashworth A. Marson A. Genome-wide CRISPR screens in primary human T cells reveal key regulators of immune function.Cell. 2018; 175: 1958-1971.e15Abstract Full Text Full Text PDF PubMed Scopus (159) Google Scholar; Freeman et al., 2019Freeman A.J. Vervoort S.J. Ramsbottom K.M. Kelly M.J. Michie J. Pijpers L. Johnstone R.W. Kearney C.J. Oliaro J. Natural killer cells suppress t cell-associated tumor immune evasion.Cell Rep. 2019; 28: 2784-2794.e5Abstract Full Text Full Text PDF PubMed Scopus (36) Google Scholar; Lawson et al., 2020Lawson K.A. Sousa C.M. Zhang X. Kim E. Akthar R. Caumanns J.J. Yao Y. Mikolajewicz N. Ross C. Brown K.R. et al.Functional genomic landscape of cancer-intrinsic evasion of killing by T cells.Nature. 2020; 586: 120-126Crossref PubMed Scopus (64) Google Scholar; Williams et al., 2020Williams J.B. Li S. Higgs E.F. Cabanov A. Wang X. Huang H. Gajewski T.F. Tumor heterogeneity and clonal cooperation influence the immune selection of IFN-γ-signaling mutant cancer cells.Nat. Commun. 2020; 11: 602Crossref PubMed Scopus (42) Google Scholar). In vivo genetic screens enable discovery of genes that regulate tumor interactions with the intact host immune system and tumor microenvironment in contexts such as immunotherapy (Manguso et al., 2017Manguso R.T. Pope H.W. Zimmer M.D. Brown F.D. Yates K.B. Miller B.C. Collins N.B. Bi K. LaFleur M.W. Juneja V.R. et al.In vivo CRISPR screening identifies Ptpn2 as a cancer immunotherapy target.Nature. 2017; 547: 413-418Crossref PubMed Scopus (443) Google Scholar; Dong et al., 2019Dong M.B. Wang G. Chow R.D. Ye L. Zhu L. Dai X. Park J.J. Kim H.R. Errami Y. Guzman C.D. et al.Systematic immunotherapy target discovery using genome-scale in vivo CRISPR screens in CD8 T cells.Cell. 2019; 178: 1189-1204.e23Abstract Full Text Full Text PDF PubMed Scopus (78) Google Scholar; LaFleur et al., 2019LaFleur M.W. Nguyen T.H. Coxe M.A. Yates K.B. Trombley J.D. Weiss S.A. Brown F.D. Gillis J.E. Coxe D.J. Doench J.G. et al.A CRISPR-Cas9 delivery system for in vivo screening of genes in the immune system.Nat. Commun. 2019; 10: 1668Crossref PubMed Scopus (44) Google Scholar; Li et al., 2020Li F. Huang Q. Luster T.A. Hu H. Zhang H. Ng W.L. Khodadadi-Jamayran A. Wang W. Chen T. Deng J. et al.In vivo epigenetic CRISPR screen identifies Asf1a as an immunotherapeutic target in Kras-mutant lung adenocarcinoma.Cancer Discov. 2020; 10: 270-287Crossref PubMed Scopus (54) Google Scholar). However, recognition of CRISPR components by the immune system is a limitation to the use of this technology in vivo (Wang et al., 2015Wang D. Mou H. Li S. Li Y. Hough S. Tran K. Li J. Yin H. Anderson D.G. Sontheimer E.J. et al.Adenovirus-mediated somatic genome editing of Pten by CRISPR/Cas9 in mouse liver in spite of Cas9-specific immune responses.Hum. Gene Ther. 2015; 26: 432-442Crossref PubMed Scopus (206) Google Scholar; Chew et al., 2016Chew W.L. Tabebordbar M. Cheng J.K. Mali P. Wu E.Y. Ng A.H. Zhu K. Wagers A.J. Church G.M. A multifunctional AAV-CRISPR-Cas9 and its host response.Nat. Methods. 2016; 13: 868-874Crossref PubMed Scopus (308) Google Scholar; Hu et al., 2018Hu J. Schokrpur S. Archang M. Hermann K. Sharrow A.C. Khanna P. Novak J. Signoretti S. Bhatt R.S. Knudsen B.S. et al.A non-integrating lentiviral approach overcomes Cas9-induced immune rejection to establish an immunocompetent metastatic renal cancer model.Mol. Ther. Methods Clin. Dev. 2018; 9: 203-210Abstract Full Text Full Text PDF PubMed Scopus (18) Google Scholar; Schmidts et al., 2019Schmidts A. Ormhøj M. Choi B.D. Taylor A.O. Bouffard A.A. Scarfò I. Larson R.C. Frigault M.J. Gallagher K. Castano A.P. et al.Rational design of a trimeric APRIL-based CAR-binding domain enables efficient targeting of multiple myeloma.Blood Adv. 2019; 3: 3248-3260Crossref PubMed Scopus (39) Google Scholar), although the scope of this problem, especially for mouse models, has not been well characterized. CRISPR applications such as genetic screens require stable expression of the bacterial endonuclease Cas9 and a single-guide RNA (sgRNA) (Jinek et al., 2012Jinek M. Chylinski K. Fonfara I. Hauer M. Doudna J.A. Charpentier E. A programmable dual-RNA-guided DNA endonuclease in adaptive bacterial immunity.Science. 2012; 337: 816-821Crossref PubMed Scopus (7874) Google Scholar; Cong et al., 2013Cong L. Ran F.A. Cox D. Lin S. Barretto R. Habib N. Hsu P.D. Wu X. Jiang W. Marraffini L.A. Zhang F. Multiplex genome engineering using CRISPR/Cas systems.Science. 2013; 339: 819-823Crossref PubMed Scopus (8856) Google Scholar; Mali et al., 2013Mali P. Yang L. Esvelt K.M. Aach J. Guell M. DiCarlo J.E. Norville J.E. Church G.M. RNA-guided human genome engineering via Cas9.Science. 2013; 339: 823-826Crossref PubMed Scopus (5800) Google Scholar), which are usually delivered to target cells via lentiviral transduction with vectors that include antibiotic resistance markers, fluorescent proteins, or other reporters (Shalem et al., 2014Shalem O. Sanjana N.E. Hartenian E. Shi X. Scott D.A. Mikkelsen T.S. Heckl D. Ebert B.L. Root D.E. et al.Genome-scale CRISPR-Cas9 knockout screening in human cells.PLoS Genet. 2014; 4: e9Google Scholar; Hartenian and Doench, 2015Hartenian E. Doench J.G. Genetic screens and functional genomics using CRISPR/Cas9 technology.FEBS J. 2015; 282: 1383-1393Crossref PubMed Scopus (60) Google Scholar; Joung et al., 2017Joung J. Konermann S. Gootenberg J.S. Abudayyeh O.O. Platt R.J. Brigham M.D. Sanjana N.E. Zhang F. Genome-scale CRISPR-Cas9 knockout and transcriptional activation screening.Nat. Protoc. 2017; 12: 828-863Crossref PubMed Scopus (364) Google Scholar; Canver et al., 2018Canver M.C. Haeussler M. Bauer D.E. Orkin S.H. Sanjana N.E. Shalem O. Yuan G.C. Zhang F. Concordet J.P. Pinello L. Integrated design, execution, and analysis of arrayed and pooled CRISPR genome-editing experiments.Nat. Protoc. 2018; 13: 946-986Crossref PubMed Scopus (35) Google Scholar). The expression of foreign elements can present challenges when these cells are introduced in immunocompetent hosts. In preclinical studies of CRISPR-based gene therapy, Cas9-specific antibodies and T cell responses were elicited in mice following intramuscular electroporation or adeno-associated virus (AAV)-mediated delivery of Cas9 (Chew et al., 2016Chew W.L. Tabebordbar M. Cheng J.K. Mali P. Wu E.Y. Ng A.H. Zhu K. Wagers A.J. Church G.M. A multifunctional AAV-CRISPR-Cas9 and its host response.Nat. Methods. 2016; 13: 868-874Crossref PubMed Scopus (308) Google Scholar). In humans, antibodies and T cells reactive against Cas9 are detected in a majority of healthy blood donors (Charlesworth et al., 2019Charlesworth C.T. Deshpande P.S. Dever D.P. Camarena J. Lemgart V.T. Cromer M.K. Vakulskas C.A. Collingwood M.A. Zhang L. Bode N.M. et al.Identification of preexisting adaptive immunity to Cas9 proteins in humans.Nat. Med. 2019; 25: 249-254Crossref PubMed Scopus (333) Google Scholar; Schmidts et al., 2019Schmidts A. Ormhøj M. Choi B.D. Taylor A.O. Bouffard A.A. Scarfò I. Larson R.C. Frigault M.J. Gallagher K. Castano A.P. et al.Rational design of a trimeric APRIL-based CAR-binding domain enables efficient targeting of multiple myeloma.Blood Adv. 2019; 3: 3248-3260Crossref PubMed Scopus (39) Google Scholar; Wagner et al., 2019Wagner D.L. Amini L. Wendering D.J. Burkhardt L.M. Akyüz L. Reinke P. Volk H.D. Schmueck-Henneresse M. High prevalence of Streptococcus pyogenes Cas9-reactive T cells within the adult human population.Nat. Med. 2019; 25: 242-248Crossref PubMed Scopus (157) Google Scholar). Foreign protein expression within tumors can alter tumor-immune interactions. For example, the ovalbumin-derived peptide SIINFEKL is frequently engineered on cancer cells to provoke an anti-tumor immune response or to study antigen specificity within the immune compartment. Common components of viral vectors, such as the fluorescent marker GFP or the puromycin resistance protein (puromycin acetyltransferase [PAC]), can also induce an immune response when expressed on transplanted cells (Stripecke et al., 1999Stripecke R. Carmen Villacres M. Skelton D. Satake N. Halene S. Kohn D. Immune response to green fluorescent protein: implications for gene therapy.Gene Ther. 1999; 6: 1305-1312Crossref PubMed Scopus (283) Google Scholar; Chen et al., 2014Chen R. Bélanger S. Frederick M.A. Li B. Johnston R.J. Xiao N. Liu Y.C. Sharma S. Peters B. Rao A. et al.In vivo RNA interference screens identify regulators of antiviral CD4(+) and CD8(+) T cell differentiation.Immunity. 2014; 41: 325-338Abstract Full Text Full Text PDF PubMed Scopus (66) Google Scholar). However, immune recognition of CRISPR-vector-associated antigens may not be problematic in all model systems and has previously been identified ad hoc. Strong anti-vector immune responses are of particular concern when stable lentiviral transduction is experimentally required in vivo, as in genetic screens where cell perturbations are identified with a heritable barcode (Shalem et al., 2014Shalem O. Sanjana N.E. Hartenian E. Shi X. Scott D.A. Mikkelsen T.S. Heckl D. Ebert B.L. Root D.E. et al.Genome-scale CRISPR-Cas9 knockout screening in human cells.PLoS Genet. 2014; 4: e9Google Scholar; Hartenian and Doench, 2015Hartenian E. Doench J.G. Genetic screens and functional genomics using CRISPR/Cas9 technology.FEBS J. 2015; 282: 1383-1393Crossref PubMed Scopus (60) Google Scholar; Joung et al., 2017Joung J. Konermann S. Gootenberg J.S. Abudayyeh O.O. Platt R.J. Brigham M.D. Sanjana N.E. Zhang F. Genome-scale CRISPR-Cas9 knockout and transcriptional activation screening.Nat. Protoc. 2017; 12: 828-863Crossref PubMed Scopus (364) Google Scholar; Canver et al., 2018Canver M.C. Haeussler M. Bauer D.E. Orkin S.H. Sanjana N.E. Shalem O. Yuan G.C. Zhang F. Concordet J.P. Pinello L. Integrated design, execution, and analysis of arrayed and pooled CRISPR genome-editing experiments.Nat. Protoc. 2018; 13: 946-986Crossref PubMed Scopus (35) Google Scholar). In such cases, there are few available solutions beyond replacing problematic immunogens on a case-by-case basis with customized vectors (Chen et al., 2014Chen R. Bélanger S. Frederick M.A. Li B. Johnston R.J. Xiao N. Liu Y.C. Sharma S. Peters B. Rao A. et al.In vivo RNA interference screens identify regulators of antiviral CD4(+) and CD8(+) T cell differentiation.Immunity. 2014; 41: 325-338Abstract Full Text Full Text PDF PubMed Scopus (66) Google Scholar), which can be costly and time-consuming. Thus, a more complete analysis of the extent of CRISPR vector immunogenicity and a more generalized solution would be broadly useful for cancer biology and immunology. Here, we characterized the CRISPR-vector-antigen-directed immune response that causes tumor rejection in several murine cancer models. We found immune responses against not only Cas9 epitopes but also other vector components such as antibiotic resistance genes. These responses altered the tumor microenvironment and artificially enhanced the efficacy of immune checkpoint blockade, preventing the study of immune responses against natural tumor antigens. We designed a screening-compatible selective CRISPR antigen removal (SCAR) lentiviral vector system that allowsremoval of immunogenic vector components after genome editing. Unlike other removable or transient delivery methods, the SCAR system left integrated sgRNA barcodes intact, which is crucial for readout of pooled screens. SCAR prevented the immune response to vector-encoded antigens and enabled an in vivo screen in Renca renal cell carcinoma, a model for which CRISPR-Cas9 expression causes tumor rejection in immunocompetent syngeneic mice (Hu et al., 2018Hu J. Schokrpur S. Archang M. Hermann K. Sharrow A.C. Khanna P. Novak J. Signoretti S. Bhatt R.S. Knudsen B.S. et al.A non-integrating lentiviral approach overcomes Cas9-induced immune rejection to establish an immunocompetent metastatic renal cancer model.Mol. Ther. Methods Clin. Dev. 2018; 9: 203-210Abstract Full Text Full Text PDF PubMed Scopus (18) Google Scholar). In contrast to previous screens in B16 melanoma, in vivo screens in Renca revealed major histocompatibility complex class I (MHC class I)- and autophagy-mediated suppressive mechanisms, highlighting the diversity of pathways used by tumors to evade immunity. Importantly, the SCAR system will enable functional genomics for immuno-oncology discovery in a wider range of preclinical models that more accurately represent the biology of human cancer. Understanding the immune impact of CRISPR-Cas9 expression in mouse cancer models is critical for performing and interpreting in vivo experiments. To assess the effect of these components on murine tumor cells in vivo, we transduced several cancer models with the widely used two-vector CRISPR system shown in Figure 1A. In this system, expression of S. pyogenes (sp)Cas9 is driven from the elongation factor-1-alpha (EF1a) promoter (pLX311-Cas9; Addgene #96924), and the U6-sgRNA cassette is delivered on a second vector with phosphoglycerate kinase (PGK)-promoter-driven puromycin resistance (pXPR_024, shown here), a design similar to commonly used sgRNA vectors (see lentiGuide-puro [Addgene #52963] or pRDA_118 [Addgene #133459]) (Shalem et al., 2014Shalem O. Sanjana N.E. Hartenian E. Shi X. Scott D.A. Mikkelsen T.S. Heckl D. Ebert B.L. Root D.E. et al.Genome-scale CRISPR-Cas9 knockout screening in human cells.PLoS Genet. 2014; 4: e9Google Scholar; Wang et al., 2014Wang T. Wei J.J. Sabatini D.M. Lander E.S. Genetic screens in human cells using the CRISPR-Cas9 system.Science. 2014; 343: 80-84Crossref PubMed Scopus (1687) Google Scholar; Joung et al., 2017Joung J. Konermann S. Gootenberg J.S. Abudayyeh O.O. Platt R.J. Brigham M.D. Sanjana N.E. Zhang F. Genome-scale CRISPR-Cas9 knockout and transcriptional activation screening.Nat. Protoc. 2017; 12: 828-863Crossref PubMed Scopus (364) Google Scholar; Doench, 2018Doench J.G. Am I ready for CRISPR? A user’s guide to genetic screens.Nat. Rev. Genet. 2018; 19: 67-80Crossref PubMed Scopus (140) Google Scholar; Ishizuka et al., 2019Ishizuka J.J. Manguso R.T. Cheruiyot C.K. Bi K. Panda A. Iracheta-Vellve A. Miller B.C. Du P.P. Yates K.B. Dubrot J. et al.Loss of ADAR1 in tumours overcomes resistance to immune checkpoint blockade.Nature. 2019; 565: 43-48Crossref PubMed Scopus (209) Google Scholar). Consistent with our previous findings (Ishizuka et al., 2019Ishizuka J.J. Manguso R.T. Cheruiyot C.K. Bi K. Panda A. Iracheta-Vellve A. Miller B.C. Du P.P. Yates K.B. Dubrot J. et al.Loss of ADAR1 in tumours overcomes resistance to immune checkpoint blockade.Nature. 2019; 565: 43-48Crossref PubMed Scopus (209) Google Scholar), B16 melanoma tumors expressing these vectors grew progressively and did not exhibit an enhanced response to anti-PD-1 immunotherapy (Figure S1A). However, vector-transduced CT26 colon carcinoma and YUMMER1.7 Braf/Pten melanoma cells were spontaneously rejected in immunocompetent mice (Figure 1B, p < 0.05; Figure 1C, p < 0.01; Student’s t test). While vector-transduced KPC pancreatic adenocarcinomas were not rejected, tumors had reduced growth and an exaggerated response to anti-PD-1 checkpoint blockade compared to unmodified controls (Figure 1D; p < 0.01, Student’s t test). As vector-transduced tumors exhibited no growth defect compared to unmodified control tumors in immunodeficient NSG (non-obese diabetic [NOD] severe combined immunodeficiency [SCID] interleukin-2 receptor g [IL2Rg] null) mice, tumor rejection was immune-dependent rather than cell-intrinsic (Figure 1E). These tumor growth kinetics suggested that the presence of Cas9 and sgRNA vectors altered tumor-immune interactions and rendered immunotherapy-resistant cells artificially sensitive to treatment. Other groups have reported that antibiotic resistance genes such as PAC could be immunogenic and that their removal from viral vectors prevents aberrant immune responses (Chen et al., 2014Chen R. Bélanger S. Frederick M.A. Li B. Johnston R.J. Xiao N. Liu Y.C. Sharma S. Peters B. Rao A. et al.In vivo RNA interference screens identify regulators of antiviral CD4(+) and CD8(+) T cell differentiation.Immunity. 2014; 41: 325-338Abstract Full Text Full Text PDF PubMed Scopus (66) Google Scholar). However, we also observed immune-dependent rejection using similar vectors with hygromycin B phosphotransferase (HBP) and blasticidin S deaminase (BSD) resistance markers (Figures S1B and S1E). This pattern suggested that Cas9 was the dominant rejection antigen or that additional resistance genes were immunogenic. In an attempt to eliminate Cas9 as a rejection antigen, we transduced the D4m3a Braf/Pten melanoma cell line with a doxycycline-inducible vector system that drives the expression of Cas9 from the tetracycline response element, which allows Cas9 to be expressed only in the presence of doxycycline in vitro. Vector-transduced tumors grew significantly more slowly than unmodified tumors in vivo and were rejected following treatment with PD-1 blocking antibody, while unmodified cells grew progressively and had a transient response to anti-PD-1 (Figure S1F; p < 0.01, Student’s t test). This suggested that additional vector components can elicit immune responses or that even very low levels of Cas9 expression are sufficient for tumor rejection. To determine which of the vectors drove immunogenicity in vivo, we transduced CT26 and YUMMER1.7 cells with either pLX311-Cas9 (which also encodes BSD) or the pXPR_024 sgRNA vector (which encodes PAC) and implanted them into wild-type (WT) mice with or without anti-PD-1. Both YUMMER1.7 and CT26 cells expressing the Cas9 vector were spontaneously rejected in WT mice, even without the addition of anti-PD-1 (Figures 1F and S1G; p < 0.01 and p < 0.001, respectively, Student’s t test). However, CT26 cells were insensitive to the expression of pXPR_024 and grew in WT mice at the same rate as unmodified cells, regardless of treatment (Figure S1G, right panels in blue), while YUMMER1.7 cells expressing pXPR_024 grew more slowly than unmodified controls in WT mice and were rejected following PD-1 blockade (Figure 1F, right panels). We obtained similar results with Cas9 and sgRNA vector variants that encode HBP or BSD, respectively (Figures S1H and S1I). To determine the immunogenic potential of vector-encoded genes, we used NetMHCpan-4.0 (Jurtz et al., 2017Jurtz V. Paul S. Andreatta M. Marcatili P. Peters B. Nielsen M. NetMHCpan-4.0: improved peptide-MHC Class I interaction predictions integrating eluted ligand and peptide binding affinity data.J. Immunol. 2017; 199: 3360-3368Crossref PubMed Scopus (521) Google Scholar) to generate H2-K and H2-D epitope binding affinity predictions for all 8-, 9-, and 10-mer Cas9 peptides, as well as for BSD, PAC, and HBP antibiotic resistance proteins, and compared them with the epitope binding predictions for the known model antigens ovalbumin, Pmel, and GFP. We identified high-affinity (predicted binding affinity <500 nM) MHC-class-I-binding peptides for vector-encoded PAC, HBP, and BSD, but the strongest predicted binders were identified from Cas9 (Figure 2A; Table S1). Five Cas9-derived peptides were predicted to bind to H2-Kb with higher affinity than SIINFEKL, the immunodominant ovalbumin T cell epitope that is frequently used as a model antigen (Figure 2A). To confirm that tumor-infiltrating CD8+ T cells were specifically recognizing vector-derived antigens, we generated peptide-pentamer complexes for two of the strongest predicted Cas9 epitopes for H2-Db and H2-Kd (Table S1). We observed an expansion of pentamer-positive, Cas9-specific T cells in YUMMER1.7 (Figures 2B and S2A; p < 0.05, Student’s t test) and CT26 tumors (Figure S2B; p < 0.01, Student’s t test). To determine whether observed anti-tumor immune responses were directed primarily against Cas9, we used B6J.129(Cg)-Gt(ROSA)26Sortm1.1(CAG-cas9∗,-EGFP)Fezh/J (Cas9) mice, which constitutively express Cas9 in all cells and thus have central immune tolerance to Cas9. YUMMER1.7 cells transduced with pLX311-Cas9 and implanted into Cas9 mice had an intermediate growth phenotype. Although a greater percentage of tumors grew progressively in Cas9 mice relative to WT mice (Figure 2C versus Figure 1F), two out of five vector-transduced tumors regressed, while all of the unmodified tumors grew progressively (Figure 2C). This mixed result suggested that although anti-Cas9 immunity is a major factor in rejection of vector-transduced tumors, BSD was still immunogenic, and immune tolerance to Cas9 alone could not fully restore normal tumor-immune interactions. This is consistent with our data showing that YUMMER1.7 cells transduced with an sgRNA vector encoding BSD (pXPR_055) had slower growth and were rejected after PD-1 blockade in WT mice (Figure S1I; p < 0.05, Student’s t test). Our observation of enhanced responses to PD-1 blockade, MHC class I peptide binding predictions, and pentamer staining suggested a role for CD8+ T cells in mediating anti-vector immunity (Chew et al., 2016Chew W.L. Tabebordbar M. Cheng J.K. Mali P. Wu E.Y. Ng A.H. Zhu K. Wagers A.J. Church G.M. A multifunctional AAV-CRISPR-Cas9 and its host response.Nat. Methods. 2016; 13: 868-874Crossref PubMed Scopus (308) Google Scholar). We depleted CD8+ T cells in mice implanted with pLX311-Cas9 CT26 or YUMMER1.7 tumors and found that pLX311-Cas9 tumors grew progressively, suggesting that tumor rejection requires CD8+ T cells (Figure 2D; p < 0.01, Student’s t test). However, growth was attenuated compared to unmodified cells (Figure 2D; p < 0.001, Student’s t test). To further test for Cas9-vector-specific responses, we performed a memory rechallenge experiment using two different Cas9-expressing tumor models. Although MC38 tumors transduced with pLX311-Cas9 grew progressively in naive C57BL/6 mice (Figure 2E), we hypothesized that the growth of MC38 pLX311-Cas9 tumors would be inhibited in mice with immune memory against pLX311-Cas9. To generate mice with pLX311-Cas9-specific memory, but not MC38-specific memory, we implanted YUMMER1.7 melanoma transduced with pLX311-Cas9 into C57BL/6 mice. Consistent with data shown in Figure 2B, 29 of 40 tumors were spontaneously rejected (Figure 2E). Two months after tumor rejection, we rechallenged the cured mice with unmodified or pLX311-Cas9 MC38 tumors (Figure 2E). In the mice that had previously rejected pLX311-Cas9 YUMMER1.7 tumors, pLX311-Cas9 MC38 tumors grew significantly more slowly than their unmodified counterparts, and three of nine tumors were rejected (p < 0.001, Student’s t test). In naive mice, both unmodified and pLX311-Cas9-expressing MC38 tumors grew rapidly (Figure 2E). We performed a similar experiment using two different tumor models in the BALB/c strain. We first challenged mice with pLX311-Cas9 CT26 tumors and then, after tumor rejection, rechallenged cured mice with the 4T1 breast cancer model with or without pLX311-Cas9 (Figure