p53 Hinders CRISPR/Cas9-Mediated Targeted Gene Disruption in Memory CD8 T Cells In Vivo Free

This article is featured in Top Reads, p.1979

The CRISPR/Cas9 system is being increasingly used to edit mammalian germline sequences in cell lines and primary cells, to study gene functions, as well as to sustainably alter these cells genetically and functionally (1). Although CRISPR/Cas9 has been used rather efficiently in various primary human and mouse cells to undertake targeted gene disruption, its competency has been particularly limited in immune cells, including T cells. The inability to deliver plasmids encoding Cas9 and single-guide RNA (sgRNAs) to primary T cells as reliably and efficiently as in other cell types may have contributed to this. Nevertheless, the increasing use of Cas9-sgRNA ribonucleoprotein complex transfections, viral vectors, and better transfection methods in general have helped make significant progress in ablating target genes in primary human and mouse T cells (2–5).

The potential to efficiently induce genetic modifications in primary T cells and return these cells to the host for analysis is a powerful experimental approach, with additional implications for immunotherapy (6–8). Although CRISPR/Cas9–mediated targeted gene disruption in T cells currently requires ex vivo activation and prolonged maintenance in culture, this approach has helped modify CD8 T cell function to aid in immunotherapy, as well as to investigate the contributions of the targeted genes in CD8 T cell function. As a testament to these advancements, multiple scientific and clinical trials are underway to manipulate T cell genomes (7, 9), and a recent study provided results from a phase 1 trial using CRISPR/Cas9 to both delete genes and add back specific TCR sequences to T cells that were then expanded in vitro and infused into patients (10). However, few studies have moved beyond the demonstration of gene editing to actually study the biology of the gene-edited T cells after their transfusion into a recipient animal or human participant, despite the tremendous research and translational significance of this method (7, 11). The physiological perturbations associated with in vitro culture and activation of CD8 T cells, which are currently indispensable to facilitate CRISPR/Cas9–mediated ablation of genes, are also a caveat in learning the biology of these cells, even after their subsequent adoptive transfer into the host (2–4). Although recent advances have helped overcome some of the above shortcomings in the mouse model, all of these attempts were also directed at modifying naive CD8 T cells, to the best of our knowledge (12). Given that already expanded populations of Ag-experienced memory CD8 T cells could be the primary target of personalized immunotherapy, and that our understanding of the transcriptional basis of T cell memory establishment, maintenance, and recall with repeated Ag exposure is still elementary, we attempted to undertake CRISPR/Cas9–mediated targeted gene disruption in primary memory CD8 T cells. If targeted gene disruption can be achieved in memory CD8 T cells, without ex vivo culture or activation of the CD8 T cells, it will help investigate the contributions of individual genes to the development and maintenance of function of subsequent memory populations in vivo, as well as in mounting recall CD8 T cell responses.

To address these issues, we undertook CRISPR/Cas9–mediated targeted gene disruption in memory CD8 T cells in mice. Unexpectedly, we found that CRISPR/Cas9 gene editing in memory CD8 T cells precluded their proliferation in vivo—apparently, because of p53-mediated CD8 T cell apoptosis in response to the DNA strand breaks. Temporarily subduing p53 function to offer a window of opportunity for DNA repair helped achieve CRISPR/Cas9–mediated target gene disruption in memory CD8 T cells in vivo, after adoptive transfer and restimulation in recipient mice. We in this study provide a roadmap for functional alteration of primary, Ag-experienced memory CD8 T cells by CRISPR/Cas9–mediated targeted gene disruption.

The C57BL/6 (B6; Thy1.2/1.2) mice were purchased from the National Cancer Institute (Frederick, MD) and maintained in the animal facilities at the University of Iowa at the appropriate biosafety level. P14 (obtained from Michael J. Bevan) and OT-I (Jackson Laboratory) TCR-transgenic mice (Thy1.1/1.1) were bred and maintained at the University of Iowa (Iowa City, IA). OT-I mice were crossed with B6 eGFP mice (Jackson Laboratory) at the University of Iowa to yield OT-I-eGFP Thy1.1 B6 mice. Lymphocytic choriomeningitis virus (LCMV)–Armstrong strain was injected i.p. (2 × 10⁵ PFUs). Attenuated (ActA-deficient) or virulent Listeria monocytogenes expressing chicken OVA (Lm-Ova) (13) was injected i.v.

To generate primary memory CD8 T cells, 5 × 10⁴ naive CD8 T cells enriched from the spleen of naive P14 or OT-I eGFP mice (CD90.1) were adoptively transferred into B6 recipients (CD90.2) i.v., followed by infection with LCMV-Armstrong or Lm-Ova. At the indicated time point, recipient spleens were harvested and homogenized, and single-cell suspensions were prepared. The CD90.1 CD8 T cells were negatively enriched using the Mouse CD8a T Cell Isolation Kit and CD90.2 microbeads (Miltenyi Biotec), following the manufacturer’s protocols with >98% purity.

CD8 T cells transfected with the pX458 plasmid that encodes recombinant Cas9 (rCas9) from Streptococcus pyogenes with a 2A-EGFP tag (catalog no. 48138; Addgene). This pX458 plasmid was cloned with the indicated target gene–specific sgRNAs (please see below for sgRNA sequences). Transfections employing the “nickase mutant” used the pX461 plasmid that encodes nCas9 (D10A nickase mutant) from S. pyogenes with 2A-EGFP (catalog no. 48140; Addgene) (14) and cloned with the indicated target gene–specific sgRNAs.

In vitro synthesis of sgRNAs was performed using the manufacturer’s instructions for the MEGAscript T7 Kit (Thermo Fisher Scientific). The reaction products were purified using alcohol-precipitation, resuspended in water, and quantified using a spectrometer. The ribonucleoprotein (RNP) complexes were generated and used as described in detail before (3). In short, RNP complexes were generated in under 10 μl of RNP reaction buffer by coincubating 30 μg of sgRNA with 10 μg of rCas9 protein (Integrated DNA Technologies) at room temperature for 15 min.

Nucleofection of CD8 T cells were performed following the manufacturer’s (Lonza) protocol, and as described in detail elsewhere (3). In short, 1 × 10⁶ CD8 T cells were resuspended in 100 μl of nucleofector buffer (Lonza) along with the RNP complex, with or without the p53siRNA (sc-29436; Santa Cruz Biotechnology) in a glass cuvette (Lonza). Nucleofector 2b device (Lonza) with Nucleofector program X-001 was used. Following nucleofection, cell suspension was removed from cuvette, placed in Eppendorf tube with 100 μl of RPMI 1640 (Life Technologies) containing 10% FCS (Life Technologies). After incubation at 37°C for 10 min, the CD8 T cells were then placed in culture containing the T cell growth media (Lonza) or transferred i.v. into recipient mouse.

Blood was collected from the mice and the RBC lysed using Vitalyse (CytoMedical Design Group). The cells were then washed in RPMI 1640 with 10% FCS and stimulated for 3.5 h in the presence of brefeldin A (BioLegend), 5 ng/ml PMA (Sigma-Aldrich), and 500 ng/ml ionomycin (Sigma-Aldrich) as described in detailed before (15). The cells were subsequently stained for surface markers and intracellular cytokines after membrane permeabilization (Fix Perm Kit; BD Biosciences) before analysis by flow cytometry.

The cells were stained with the following Abs: CD8 (clone 53–6.7; eBioscience), CD4 (clone GK1.5; eBioscience), CD90.1 (clone OX-7; eBioscience), CD90.2 (clone 53–2.1; eBioscience), PD-1 (clone J43; BioLegend), KLRG1 (clone 2F1; BioLegend), p53 (clone pAb 240; Novus Biologicals), TNF (clone MP6-XT22; BioLegend), and IFN-γ (clone XMG1.2; BioLegend) at the appropriate dilutions and with compatible fluorochromes. The lymphocyte gates in the samples were plotted directly to examine CD8 T cell populations depicted in the flow plots.

In vitro proliferation assay in CD8 T cells was performed as described in detail previously (4). In short, CD8 T cells were stained with CellTrace Violet (Invitrogen) as per the manufacturers protocol, before transferring them to a 96-well U-bottom plate in T cell growth media (Lonza). The indicated doses of the IL-7 (catalog no. 200-07; PeproTech) and IL-15 (catalog no. 210-15; PeproTech) were added into the culture media and T cell proliferation was determined at various time points by dye dilution using flow cytometry.

CD8 T cells were harvested from mouse periphery at 530 d postinfection. CD8 T cell single-cell clones were obtained by limiting dilution in 96-well plate as described in detail before (3). Following proliferation of the clones, the CD8 T cells were harvested, and DNA was extracted, and the indicated gene was sequenced at the Iowa Institute of Human Genetics.

Naive IFNgKO mice (Jackson Laboratory) were adoptively transferred with Cas9-IfngsgRNA–transfected 6 × 10⁵ OT-I CD8 T cells i.v. Control mice either received no cells or cells transfected with only rCas9 or rCas9.CtrlsgRNA. The next day, the recipient mice were challenged with 4 × 10³ CFUs of virulent Lm-Ova i.v. Four days after the challenge, liver and spleen were harvested and assessed for bacterial burden (4).

L. monocytogenes infection burdens in mice were determine as described in detail before (16). In short, the livers and spleens obtained from the Lm-Ova challenged mice were weighed, homogenized in antibiotic-free RPMI 1640 media (Life Technologies) containing 10% FCS (Life Technologies), with a mechanical disruptor. Serial dilutions were performed, and aliquots were placed on 5-cm petri dishes containing lysogeny broth ampicillin agar. The Lm-Ova bacteria are transgenically ampicillin resistant. Plates were incubated at 37°C, 5% CO₂ overnight, and colonies were counted by hand the next morning to calculate the bacterial burden per gram of tissue.

L. monocytogenes tissue burdens were determined as described in detail elsewhere (17). Total tissue DNA was extracted using phenol/chloroform, and quantitative PCR was performed with the following primers: hlyA-177-F (5′-TGCAAGTCCTAAGACGCCA-3′) and hlyA-177-R (5′-CACTGCATCTCCGTGGTATACTAA-3′). The cycle threshold ratios were normalized against the housekeeping gene for mouse TNF as the reference, using the following primers: TNF-5241 (5′-TCCCTCTCATCAGTTCTATGGCCCA-3′) and TNF-5411 (5′-CAGCAAGCATCTATGCACTTAGACCCC-3′).

The following sgRNA templates were used: Klrg1 sgRNA: 5′-CCTTACATTTCCGGACAACC-3′ (crispr.mit.edu), eGFP sgRNA: 5′-GGTGGTGCAGATGAACTTCA-3′ (18), p53 sgRNA: 5′-AGGAGCTCCTGACACTCGGA-3′ (crispr.mit.edu), Pdcd1 sgRNA: 5′-GACACACGGCGCAATGACAG-3′ (3), and IFNg sgRNA: 5′-GGCTTTCAATGACTGTGCCG-3′, 5′-GGCTTTGCAGCTCTTCCTCA-3′ (crispr.mit.edu).

Successful CRISPR/Cas9 editing and subsequent evaluation of bona fide memory CD8 T cells remains an unrealized yet, potentially, a fruitful experimental and translational goal. To achieve this goal, we generated memory TCR-transgenic CD8 T cells specific to the LCMV GP33-41 Ag (P14 cells, CD90.1) in CD90.2 B6 mice, as depicted in Fig. 1A. The memory P14 CD8 T cells were negatively enriched, transfected with the plasmid (pX458)-encoding Cas9, eGFP (to identify transfected cells), and the indicated target gene–specific sgRNA (14) before adoptively transferring them to congenically distinct recipient mice. The recipient mice were subsequently challenged with LCMV to induce recall responses from the transfected P14 cells (Supplemental Fig. 1A). The adoptively transferred memory P14 cell population expanded and expressed vector-derived eGFP when transfected with the “empty” (no sgRNA template encoded) pX458 plasmid (Supplemental Fig. 1A). As a proof of concept, we chose to disrupt the Klrg1 gene in memory P14 cells because KLRG1 deficiency does not influence CD8 T cell proliferation or function in mice (19). Based on this, we expected the Klrg1sgRNA-encoded pX458 (pX458-Klrg1sgRNA) transfected P14 cells to proliferate like those with the control pX458 empty plasmid. To our surprise, pX458-Klrg1sgRNA–transfected memory P14 cells failed to undergo secondary expansion and accumulate in the recipient mice after challenge with LCMV (Supplemental Fig. 1B). We next targeted the Pdcd1 gene (encoding PD-1) for disruption in memory P14 cells, following the same approach. PD-1 deficiency does not disrupt cognate Ag–mediated proliferation of memory CD8 T cells (20). Nevertheless, we failed to detect secondary expansion and accumulation of the pX458-Pdcd1sgRNA–transfected P14 CD8 cell population (Supplemental Fig. 1B). Thus, although transfection of memory P14 cells with pX458 lacking sgRNA did not hinder their secondary expansion and accumulation, merely cotransfecting them with pX458 and Klrg1- or Pdcd1-targeting sgRNAs (but not scrambled sgRNA; data not shown) prevented their proliferation in the recipient mice challenged with LCMV (Supplemental Fig. 1C). Cas9 nuclease functions by inducing double-strand breaks in the genomic DNA (14). To determine if the double-strand breaks in genomic DNA itself may have been responsible for the failure of CD8 T cells proliferation, we induced single-strand breaks in the P14 CD8 T cell DNA using a “nickase” mutant Cas9 (nCas9)–encoding pX461 plasmid (14) and the Klrg1sgRNA. The nCas9-Klrg1sgRNA–transfected P14 CD8 T cells also failed to proliferate in the recipient mice (Supplemental Fig. 1D).

Target gene disruption and transfection efficiencies in cells can be greatly enhanced by substituting Cas9-encoding plasmids with ribonucleoprotein complexes (RNPs) consisting of rCas9 and the target gene–specific sgRNA (3, 5). We thus cotransfected memory P14 cells with RNPs consisting of rCas9 and in vitro–transcribed Klrg1sgRNA or the scrambled control sgRNA before adoptively transferring them into recipient mice and challenging them with LCMV (Fig. 1A). In this study also, the adoptively transferred P14 cells failed to expand and accumulate following Klrg1 targeted gene disruption (Supplemental Fig. 1E). Taken together, these data show that CRISPR/Cas9–mediated DNA breaks in memory P14 cells prevent their expansion and accumulation in response to LCMV infection in mice.

To rule out the possibility that T cells with a particular transgenic TCR might influence memory CD8 T cell proliferation in CRISPR/Cas9-mediated targeted gene disruption, we next used chicken OVA Ag–specific TCR-transgenic OT-I CD8 T cells (which also coexpressed eGFP and OT-IeGFP) for targeted gene disruption (Fig. 1A). As with the P14 CD8 T cells, Klrg1 gene disrupted memory OT-IeGFP CD8 T cells failed to expand and accumulate in the recipient mice when challenged with Lm-Ova (Fig. 1B). Similar results were obtained when we targeted genes that are known to serve specific functions in CD8 T cells: the Ifng gene for cytotoxicity (Fig. 1B) or the Pdcd1 gene for coinhibition (data not shown). Neither of these gene products are known to directly influence the proliferative abilities of memory CD8 T cells. Notably, disrupting the exogenous and nonessential eGFP gene using CRIPSR/Cas9 also prevented the expansion and accumulation memory OT-IeGFP CD8 T cells in the recipient mice (Fig. 1B), indicating that genomic DNA damage itself may be responsible for this defect rather than the targeting of any specific gene. It is noteworthy that CRISPR/Cas9-mediated gene disruption also hinders the cytokine-driven proliferation of the memory CD8 T cells in vitro. rCas9.Klrg1sgRNA RNP, but not rCas9 itself (data not shown) or the rCas9.CtrlsgRNA RNP–transfected memory OT-IeGFP cells failed to proliferate in response to IL-7 and IL-15 (Supplemental Fig. 2A). Additionally, letting the adoptively transferred memory OT-IeGFP CD8 T cells “rest” in vitro (or in vivo in the recipient mice; data not shown) to undergo DNA repair for up to 10 d after transfection with rCas9.Klrg1sgRNA RNP also did not rescue their inability to proliferate on cognate Ag encounter (Supplemental Fig. 2B). These data suggested that targeted gene disruption using CRISPR/Cas9 hinders the proliferation of memory CD8 T cells in response to cytokines or Ag-stimulation.

DNA damage is known to elicit cell cycle arrest and apoptosis in cells (21, 22). Although there are many molecular components in the DNA damage response pathway, the proapoptotic transcription factor p53 is an integral part of this response (22, 23). T cells provisionally downregulate p53 expression to facilitate Ag-specific proliferative responses in vitro (24), as well as to cope with double-strand breaks introduced by TCR recombination during development (25). It is noteworthy that proliferating T cells are very sensitive to DNA damage responses and undergo rapid cell death on detecting induced DNA damages (26, 27). We thus hypothesized that memory CD8 T cells that undergo CRISPR/Cas9-mediated DNA breaks might undergo apoptotic death in response to proliferative cues in vivo, possibly mediated by p53 (28). It is known that human pluripotent stem cells that express lower p53 levels are more amenable to CRISPR/Cas9-mediated targeted gene disruption (29, 30). Considering these notions, we chose to undertake CRISPR/Cas9-mediated targeted disruption of the p53 gene itself in memory CD8 T cells to possibly rescue the proliferating cells from undergoing apoptotic death in response to the DNA damage introduced by editing. Indeed, memory OT-IeGFP cells transfected with rCas9.p53sgRNA RNP expanded and accumulated in the recipient mice after challenge with Lm-Ova (Fig. 1B). The rCas9.p53sgRNA RNP–transfected memory OT-I CD8 T cell population that expanded in the recipient mice also showed a concurrent loss of p53 expression (Fig. 1C). Taken together, these data indicated that the p53-mediated apoptotic response to DNA damage may be responsible for our inability to perform CRISPR/Cas9-mediated targeted gene disruption and amplify memory CD8 T cell responses by Ag restimulation in vivo.

The above data also suggested that we could achieve targeted gene disruption via CRISPR/Cas9 in memory CD8 T cells if the p53 function was compromised. However, p53 is a critical tumor suppressor gene and is associated with tumor development (31). In addition, T cells lacking the p53 gene gave rise to spontaneous T cell lymphomas in mice (32). Considering the risk, concurrent p53 gene disruption may not be a feasible approach to achieving other specific gene deletions in memory CD8 T cells. Nevertheless, we surmised that a temporary loss of p53 function would present the window of opportunity to repair the CRISPR/Cas9–induced DNA breaks in memory CD8 T cells without the long-term risks associated with the perpetual loss of p53 function. Hence, we tested if we could achieve targeted gene disruption in memory CD8 T cells, under the protection of a temporary knockdown of p53 transcription factor using p53siRNA.

To test this hypothesis, we first assessed whether knockdown of p53 could rescue proliferation of transgenic T cells in which the nonessential eGFP gene had been disrupted. As depicted in Fig. 2A, we enriched memory OT-IeGFP CD8 T cells generated in donor mice, cotransfected them with rCas9.eGFPsgRNA, rCas9 or rCas9.CtrlsgRNA (data not shown), and p53siRNA before adoptively transferring them to congenically distinct recipient mice. The transferred OT-IeGFP CD8 T cells expanded and accumulated after challenge with Lm-Ova and established long-term (530 d) persistence in the recipient mice (Fig. 2B). A substantial proportion of rCas9.eGFPsgRNA transfected memory OT-IeGFP CD8 T cells also exhibited loss of eGFP expression and this loss of eGFP was sustained in their memory phase (Fig. 2C). The gene disruption was also retained on secondary expansion of these rCas9-eGFPsgRNA memory OT-IeGFP CD8 T cells (Fig. 2B, 2C). These data suggested that CRISPR/Cas9-mediated targeted gene disruption can be achieved in Ag-experienced CD8 T cells in vivo under temporarily subdued p53 function.

To test if we could employ CRISPR/Cas9 to functionally alter memory CD8 T cells, we next targeted the Ifng gene, which is critical for the ability of CD8 T cells to protect against many infections, including L. monocytogenes infection (33, 34). As with the targeted gene disruption of the eGFP gene, we adoptively transferred memory OT-I cells obtained from donor mice into recipient mice after cotransfection with rCas9.IfngsgRNA, rCas9 or rCas9.CtrlsgRNA (data not shown), and p53siRNA, before challenging with Lm-Ova to drive their cognate Ag–mediated expansion. Whereas the rCas9.IfngsgRNA transfected memory OT-I CD8 T cells expanded and accumulated (Fig. 3A) in response to Lm-Ova infection, a substantial proportion of these cells showed loss of IFN-γ production in response to ex vivo stimulation. In contrast, TNF, a nontargeted cytokine produced by CD8 T cells after stimulation, was not reduced in rCas9.IfngsgRNA–transfected memory OT-I CD8 T cells, indicating that these cells were viable and responsive to stimulation (Supplemental Fig. 3). As with the eGFP gene disruption, we observed that the rCas9.IfngsgRNA–transfected memory OT-I CD8 T cells underwent secondary memory expansion after being adoptively transferred to recipient mice, which were subsequently challenged with Lm-Ova (Fig. 3A), and remained unable to produce IFN-γ (Fig. 3B).

To determine if the CRISPR/Cas9-mediated loss of IFN-γ in the OT-I CD8 T cells resulted in a perceivable functional deficiency, we determined the ability of the gene-edited OT-I CD8 T cells to protect from a challenge with virulent Lm-Ova in a model that is absolutely dependent on memory CD8 T cell–derived IFN-γ (34). Memory OT-I CD8 T cells cotransfected with rCas9.IfngsgRNA or rCas9 and p53siRNA were adoptively transferred into IFN-γ–deficient (IfngKO) recipient mice. These mice and control IfngKO mice were subsequently challenged with Lm-Ova infection (Fig. 4A) and the resulting bacterial loads determined in the spleen and liver. Mice that received rCas9 + p53siRNA–transfected OT-I (which maintained IFN-γ expression) exhibited reduced bacterial numbers in both the spleen and livers. In sharp contrast, the mice that received OT-I CD8 T cells transfected with rCas9.IfngsgRNA and p53siRNA failed to control the Lm-Ova infection, yielding similar bacterial numbers as the infected naive controls (Fig. 4B). These data indicated that CRISPR/Cas9-mediated targeted gene disruption can be used to functionally alter memory CD8 T cells.

In this study, we demonstrate that Ag-experienced CD8 T cells fail to proliferate in response to cognate Ag encounter in vivo, after undergoing CRISPR/Cas9-mediated targeted gene disruption. We show that this defect is dependent on the proapoptotic transcription factor p53, which possibly drives the DNA damage response resulting from the DNA strand breaks introduced by the CRISPR/Cas9 system. We also show that disengaging the proapoptotic pathway either by ablating the p53 gene itself or temporarily subduing p53 function using p53siRNA may help create a window of opportunity for the CD8 T cells to undertake DNA damage repair without undergoing programmed cell death. This approach makes CRISPR/Cas9-mediated targeted gene disruption possible in vivo in memory CD8 T cells. Our approach also circumvents the necessity to expose the primary memory CD8 T cells obtained from a patient to nonspecific activation or lengthy in vitro culture and cytokine-induced nonspecific expansion spanning multiple days before returning back by adoptive transfer (10). The CD8 T cells obtained from mice were adoptively transferred to recipient mice after CD8 T cell enrichment and Cas9-sgRNA + p53siRNA transfection in under 20 min. Of note, naive CD8 T cells were recently genetically modified with CRISPR/Cas9 in vivo without the need for altering their p53 expression (12). The underlying differences in the responses of naive and memory CD8 T cells to CRISPR/Cas9 gene editing remain to be elucidated, but our results suggest that subduing p53 function may be a requirement specific to memory CD8 T cells. The minimal in vitro manipulation coupled with the rapid retransfusion time in our approach makes it a more biologically pertinent means to investigate CD8 T cell memory, and offers an opportunity to significantly improve upon the current CRISPR/Cas9–based immunotherapeutic approaches.

The ability to undertake single-gene or gene family ablation in primary cells may be one of the most significant contributions of the CRISPR/cas9 system to basic biomedical sciences (1). Targeted gene disruption in T cells would help us investigate the genetic and transcriptional bases of T cell–mediated immunity, establishment, and maintenance of long-term memory, localization, tissue residence, recall responses etc., and help design better prophylactic and immunotherapeutic approaches to using it (35–37). Although whole-gene knockout mice and the Cre-lox system have been used to investigate the relevance of various genes in T cell memory (38–41), the ability to ablate genes after memory formation would help segregate their roles in the various aspects of memory formation, maintenance, recall, etc. Although inducible Cre-lox systems have partially filled this gap, the practical and technical limitations of the methodology (42, 43) have hindered a high-throughput investigation of the genetic basis of T cell memory. It is well established that T cell memory function is reflective of its transcriptional signature (44) and these transcriptional signatures vary between various iterations of memory CD8 T cells, depending on the frequency of Ag exposures (45). In addition to helping achieve single-gene disruption in memory CD8 T cells, the CRISPR/Cas9 technology can be harnessed to alter transcriptional signatures of memory CD8 T cells by undertaking multigene knockouts (10, 18, 46–48) at any stage of memory formation or maintenance. The genetically altered cells can be specifically identified by their targeted phenotypic changes or by single-cell sequencing. Coexpression of fluorescent markers with CRISPR/Cas9 and target gene–specific sgRNA in plasmid transfection-based strategies (e.g., using pX458) can help identify the transfected cells using flow cytometry. Alternatively, disruption of a fluorescent protein or inert surface marker gene (e.g., CD90.1) in CD8 T cells using specific sgRNAs encoded in plasmids with CRISPR/Cas9 and target gene–specific sgRNA may also be used. Thus, enabling CRISPR/Cas9 gene editing of memory CD8 T cells by transient p53 silencing should open new avenues to explore and exploit the protective capacity of these cells.

CD8 T cells are also at the core of modern-day cancer therapy (49). CRIPSR/Cas9 technology has helped alter the CD8 T cell genome to ablate or express specific genes, to enhance their ability to proliferate, detect, and kill tumor cells (10). However, these approaches are currently limited to using CRISPR/Cas9 technology to undertake genetic alteration of Ag inexperienced (naive) CD8 T cells. This approach also requires transgenic replacement of native TCRs of autologous T cells, with known tumor Ag–specific TCRs, to promote tumor cell targeting after transfusion into a patient. Although these CD8 T cells can be concurrently targeted with CRISPR/Cas9-mediated disruption of genes promoting T cell exhaustion (e.g., Pdcd1), adequate expansion of these cells to generate transfusable numbers requires prolonged in vitro maintenance and nonspecific, cytokine-driven propagation (10). The ability to undertake targeted gene disruption in Ag-experienced memory CD8 T cells will simplify this process greatly. Hypothetically, enriched Ag–experienced, tumor-specific memory CD8 T cells from a patient can be used for CRISPR/Cas9-mediated targeted gene disruption to restore their function under the aegis of p53siRNA before transfusing it back into the patient. The memory CD8 T cells would undergo Ag re-encounter–driven proliferation in vivo and target the tumor cells. Although this application needs to be tested rigorously in humans, our study in mice provides the proof of principle for the ability to collect Ag-specific memory CD8 T cells, functionally alter them by targeted gene disruption using CRISPR/Cas9, and transfer it back into a recipient with tangible functional consequences as intended.

In conclusion, we, in this study, demonstrate in vivo CRISPR/Cas9-mediated targeted gene disruption in Ag-experienced, Ag-specific memory CD8 T cells. Although DNA damage produced by targeted gene disruption precipitates apoptotic cell death, preventing Ag-driven proliferation of these CD8 T cells in vivo, temporarily inhibiting p53-mediated apoptotic responses helps rescue these CD8 T cells. This facilitates reliable targeted gene disruption and functional transformation of memory CD8 T cells in vivo. We believe this is the first demonstration of CRISPR/Cas9-mediated gene editing in memory CD8 T cells.

We thank Dr. Duo Peng (Harvard University) for insights and suggestions during the course of this study. We thank Drs. Vladimir Badovinac and Scott Anthony (University of Iowa) as well as Kim Klonowski and Rick Tarleton (University of Georgia) for helpful suggestions on the manuscript, the University of Iowa vivarium staff, and the University of Georgia Center for Tropical and Emerging Global Diseases Flow Cytometry Core Facility for help.

The authors have no financial conflicts of interest.