Gene-Based Natural Killer Cell Therapies for the Treatment of Pediatric Hematologic Malignancies

BACKGROUND

Pediatric leukemias and lymphomas comprise the most common subset of pediatric cancers.^1,2 The diagnosis of pediatric leukemia encompasses diverse clinical and biological diseases. While patients with pediatric acute lymphoblastic leukemia (ALL) have a greater than 80% chance of cure,³ other blood cancers have poorer prognoses. Novel therapies for acute myeloid leukemia (AML), infant ALL, adolescent/young adult ALL, and relapsed or refractory ALL are urgently needed. With advanced genetic and molecular profiling of these diseases, there is now the opportunity to develop targeted therapies with more favorable safety profiles and improved efficacy.

T-cell adoptive transfer is one type of targeted therapy that has had great success in the treatment of high-risk ALL. A multicenter clinical trial of treatment with the anti-CD19 chimeric antigen receptor (CAR) T-cell therapy, tisagenlecleucel, demonstrated an overall remission rate within 3 months of 81%,⁴ leading to FDA-approval for therapy of relapsed pre-B ALL.⁵ Though this clinical success suggests promise for all pediatric blood cancers, expanded use of CAR-T cell therapy also has clear limitations. Treatment with CAR-T cells has the risk of severe toxicities including Cytokine Release Syndrome (CRS), Immune Effector Cell Associated Neurotoxicity Syndrome (ICANS), and CAR-associated Hemophagocytic Lymphohistiocytosis (carHLH).^6,7 Further, tisagenlecleucel and all other commercial CAR-T cell products are derived from autologous hematopoietic cell collections, which can be challenging to perform in heavily pretreated patients with active disease. Per-patient CAR-T cell manufacture is also time-consuming and costly.⁸ While the collection of healthy donor T cells can be considered, the use of human leukocyte antigen (HLA)-mismatched allogeneic products carries the serious risk of graft-versus-host disease (GVHD).⁹ Natural Killer (NK) cells are alternate lymphocyte effector cells that are also potent killers. NK cells have the potential for ex vivo expansion and storage as an “off the shelf” therapy. While activated NK cells secrete cytokines that contribute to the inflammatory milieu, they do not directly cause graft versus host disease^9,10 and have an overall less severe side effect profile than allogeneic T-cell products.^11-15 Because of these favorable characteristics, there is expanded focus on the study and development of NK cells as immunotherapeutics targeted to cancer.

Natural Killer Cell Biology

NK cells are lymphocytes of the innate immune system that serve a critical function in host defense against viral infections and malignancy.¹⁶ They constitute 5% to 15% of circulating lymphocytes and are found in the peripheral blood as well as lymphoid and nonlymphoid organs such as the spleen, lung, and liver.^15,17,18 NK cells arise from CD34+ hematopoietic stem cells and progress through developmental stages defined by the expression of surface receptors in response to cytokines including Interleukin (IL)-2, IL-7, and IL-15.^15,19 NK cells in the peripheral blood are subdivided into two primary categories: CD56^brightCD16^dim/− cells are phenotypically less mature with the capability to produce higher levels of inflammatory cytokines, while CD56^dimCD16⁺ NK cells are more mature with greater cytotoxic potential.^15,18,19 Most NK cells in peripheral circulation fit the latter CD56^dim profile, with the CD56^bright subset representing <15% of the total circulating cell population,^15,18 though recent single cell analyses have revealed additional diversity.^20,21 NK cells perform their immune function through four major pathways: (1) secretion of cytokines and chemokines, (2) direct cytotoxicity, (3) target killing through induced apoptosis, and (4) antibody-dependent cellular cytotoxicity.²² NK cells are activated following the integration of stimulatory and inhibitory signals without the need for antigen processing and presentation or HLA matching.^22,23 Major histocompatibility complex class I (MHC I) molecules are found on all healthy nucleated cells and provide protection from NK cell targeting through the ligation of the killer immunoglobulin-like (KIR) family of inhibitory NK cell receptors.^15,24 Cells infected with virus and those that are malignant typically downregulate MHC I expression^25-27 and upregulate NK cell activating ligands, thereby tipping the scale in favor of NK cell activation and killing.^15,28

Unique Challenges to Natural Killer Cell Immunotherapy

Despite these promising attributes, there are challenges to the clinical translation of NK cell-based therapies. These include nonstandardized clinical-grade ex vivo expansion and historical reports of poor genetic engineering.^15,29 Other challenges include limited in vivo persistence with a peripheral half-life of approximately 7 to 10 days,^30-32 limited trafficking to and infiltration of the tumor microenvironment (TME),³³ and tumor immunoevasion.^34-36 Genetic modification of NK cells may overcome these challenges to optimize innate cytotoxicity.

STARTING MATERIAL FOR NATURAL KILLER CELL IMMUNOTHERAPY

NK cells for clinical application can be derived from multiple sources. Peripheral blood (PB) is the best-studied, given ease of collection. NK cells can be purified after blood apheresis with CD3+ cell depletion and/or CD56+ selection.^37,38 Because NK cells represent only 10% to 15% of circulating PB cells, the expansion of peripheral blood-derived NK cells (PB-NKs) on a clinical scale is achieved by culture with supplemental cytokine and/or feeder cells.^39-41 NK cell expansion using feeder cells was first described by Campana and colleagues,⁴² and consists of K562 cells engineered to express ligands that trigger NK cell activation (ex. membrane-bound IL15, IL21, and 4-1BBL).^40,43,44 PB-NKs highly express a repertoire of activating receptors and are functionally mature, with cytotoxic capacity.⁴⁵ PB-NKs also express higher levels of KIRs than other sources. This is particularly important for NK cell functionality, since KIR expression has a central role in NK cell education and licensing.^46-48 One of the theoretic limitations to the use of PB-NKs as cancer immunotherapy is reportedly poor success with genetic modification.^49,50 However, we and others have shown high levels of vector transduction of PB-NKs.^37,38,51

NK cells can also be differentiated from induced pluripotent stem cells (iPSCs) understandardized culture conditions.^52,53 The production of NK cells from iPSCs (iPSC-NKs) requires more expertise than peripheral blood NK cell selection.⁵³ The iPSC platform produces a homogeneous NK cell population. NK cell iPSC derivation allows for multiple genetic modifications while cells are relatively undifferentiated, including CAR-engineering and genetic deletion.^54,55 iPSC-NKs have a similar, but more immature phenotypic profile to PB-NKs, with higher expression of the inhibitory receptor NKG2A.⁵² Moreover, iPSC-NKs exhibit lower KIR expression.⁵² Nevertheless, iPSC-NKs have promise as anticancer immunotherapy^54-56 and can be developed as “off-the-shelf” cellular banks, with well-defined NK cell therapy products generated on large scale, readily available for adoptive transfer.

Cancerous NK cell lines have also been established over the years as sources of NK cell therapy products (NK-92,⁵⁷ NK-101,⁵⁸ NK-YS,⁵⁹ NKL,⁶⁰ NKG,⁶¹ KHYG-1,⁶² and others). The NK-92 lymphoma cell line has been most widely used as a platform for genetic modification and subsequent adoptive transfer. NK-92 produce high levels of granzymes, perforins, and other death-inducing molecules (FAS-L, TRAIL) that exert potent cytotoxicity against cancer cells.⁶³ NK-92 adoptive transfer (with or without CAR modification) has been shown to be safe, but with limited efficacy in the context of both hematologic^64-66 and solid tumors.^67,68 The use of aneuploid transformed cells with multiple cytogenetic abnormalities as anticancer therapy mandates preinfusion irradiation, which has the expected deleterious effect on in vivo expansion, persistence, and thus efficacy.⁶⁹ Moreover, while NK-92 cells generally share a receptor repertoire with PB-NK cells, they lack CD16 (FcγRIII) and thus the capacity for antibody-dependent cellular cytotoxicity (ADCC).^57,70 NK-92 also lack activating receptors, for example, NKp44 and almost all of the KIRs, with the exception of KIR2DL4.⁶³

Umbilical cord blood (UCB) is another source for the generation of “off-the-shelf” NK cell products. UCB is relatively easy to isolate⁷¹ and contains a similar percentage of NK cells as PB.^48,72 However, the small total blood volume in each CB unit makes acquiring satisfactory NK cell numbers for clinical use challenging. Similar to PB, UCB is heterogeneous, and NK cell purification is needed. Freshly isolated UCB-NKs are characterized by an immature immune phenotype with lower surface expression of NKG2C, CD57, adhesion molecules (CD2, CD11α, CD18, DNAM-1),⁷³ CD62 L (LN homing receptor), and KIRs (CD158a, CD158b, and others) and higher expression of 2B4, CXCR4 (BM homing receptor), and NKG2A, compared to PB counterparts.^48,74,75 Studies comparing granzyme and perforin expression in UCB- and PB-NK cells have reported contradictory results.^75,76 The cytotoxic capacity of UCB-NK is lower than PB-NK, yet this difference can be partially mitigated with cytokine stimulation.^48,75,76 Clinical trials testing adoptive transfer of unmodified or CAR-expressing UCB-NK as hematologic malignancy treatment have shown that this approach is safe, without GVHD or other toxicities, but with limited antitumor effect.^77,78

METHODS OF GENETIC ENGINEERING IN NATURAL KILLER CELLS

Though simple in concept, efficient genetic modification of NK cells is a challenge. Early attempts at NK cell engineering have reported low gene transfer efficiency using both viral-based and nonviral methods.²⁹ Recent optimization of gene transfer strategies has been more successful with an improved safety profile (Fig. 1).⁷⁹

Fig. 1.

Methods of Genetic Engineering in NK Cells. (A) Viral-mediated, including γ-retroviral and lentiviral vectors. (B) Nonviral mediated, including electroporation, Charge-altering releasable transporters, and transposon systems. (C) Targeted knockdown and knock-in, including zinc-finger nucleases (ZFN), transcription activator-like nucleases (TALENs), and CRISPR/Cas9 systems as NK cell engineering methods.

Viral-mediated natural killer cell engineering

NK cell modification with viral vectors is now associated with high-efficiency gene transfer. Viral vectors are readily manufactured and can stably integrate genetic material into the host genome.⁸⁰ The Retroviridae family are the most commonly used vectors for gene therapy applications, including NK cell genetic modification.⁷⁹ The Retroviridae family includes seven members: α-, β-, γ-, δ-, and ε-retroviruses, spumaviruses, and lentiviruses.⁷⁹ Of these, γ-retroviruses and lentiviruses are most often used in cellular gene transfer for clinical application.^79,80 Retroviral genetic modification includes the reverse transcription of the viral RNA genome into double-stranded DNA and permanent integration into the host genome facilitated by viral protein integrase encoded by the pol gene. Retroviral transduction has evolved to use replication-incompetent retroviruses as a safety mechanism to limit self-replication and infectivity.⁸⁰

In γ-retrovirus vectors, the γ retroviral coding sequences are replaced by the transgene of interest and this recombinant plasmid as well as helper plasmids encoding viral capsid proteins, replication enzymes, and envelope glycoproteins (gag, pol, and env) are cotransfected into a virus packaging cell line where the recombinant plasmid is synthesized and packaged into viral particles. A limitation of successful γ-retroviral transduction is the requirement for a disrupted nuclear membrane in actively dividing target cells.^80,81 Despite this obstacle, γ-retroviral vectors have demonstrated high rates of transduction and persistence of transgene expression in gene therapy applications, including the use of NK cells.

Lentivirus-based vectors similarly involve the replacement of essential viral genes with the transgene of interest and also use helper plasmids to encode gag, pol, and env. Unlike γ-retroviral vectors, lentivirus-based vectors additionally require the incorporation of the rev gene, which encodes the Rev protein and enhances nuclear export and expression of gag-pol transcripts through the binding of the rev responsive element (RRE).^80,82 Lentivirus-based vectors further contain a central polypurine tract (cPTT)/central termination signal (CTS) that functions to facilitate nuclear import of the preintegration complex after target cell uptake of the virion.^80,83 As a result, lentiviral vector transduction is not dependent on active cellular division, which notably expands the repertoire of potential target cells. Despite this reach, the use of lentiviral-based vectors for NK and other hematopoietic cell transduction has not resulted in high percentages of transduced cells, largely because of limited cellular entry. Lentivirus is standardly pseudotyped with vesicular stomatitis virus g-protein (VSV-G), which binds LDL-R on human cells. NK cells express LDL-R at low levels, making transduction efficiency low.⁸⁴ The use of lineage-specific promotors and cell-type-specific lentiviral vectors using envelope proteins from viruses such as RD114 or baboon retrovirus are potential ways to improve the efficacy and specificity of lentiviral vectors.^85-87

Though viral-based cell engineering can effectively achieve stable transgenic expression, without genomic targeting this method is inherently nonspecific and therefore associated with the risk of insertional mutagenesis. γ-retroviral vectors integrate with proximity to cellular gene promoters, transcription start sites, and CpG islands. This can result in neoplastic transformation if genomic insertion occurs at a protooncogenic site.^80,88,89

Nonviral natural killer cell engineering

Nonviral NK cell modification can circumvent some potential pitfalls of viral engineering. The most used nonviral gene delivery method is electroporation, which is a simple and cost-effective strategy that can introduce nucleic acid into a target cell with high efficiency. Electroporation generates small, temporary pores in the cell and nuclear membrane by electrical pulsing, which allows for the transfer of charged nucleic acid molecules.^90,91 The primary downside to electroporation is the toxicity that can lead to high rates of cell death.⁹² Successful NK cell transfection via electroporation has been described using primary NK cells with resultant efficient gene transfer.^90,91,93

Alternative methods for the delivery of charged nucleotides across a nonpolar cell membrane include the use of cationic polymers, lipid nanoparticles, and more recently charge-altering releasable transporters. Charge-altering releasable transporters are multiblock polycationic oligomers that noncovalently complex with polyanionic mRNA to facilitate its delivery across the cellular membrane.⁹⁴ Charge-altering releasable transporters have been used to transfect resting primary NK cells with mRNA to generate cytotoxic human anti-CD19-CAR NK cells with efficient transfection, viability, and preservation of the NK cell phenotype.⁹⁵ A primary limitation to the use of charge-altering releasable transporter-mediated gene delivery is the transient nature of mRNA mediated cellular modification.95.

Another method for stable, nonviral genomic modification involves the use of DNA transposons, which are naturally occurring repetitive DNA sequences that can jump from one location in the genome to another.^91,96 This natural phenomenon can be harnessed for gene editing by using a transposon vector containing the gene of interest flanked by terminal inverted repeats (TIRs), and a transposase enzyme that targets the TIRs to excise the gene of interest from the original site and stably reintegrate into chromosomal sites.^91,97 DNA transposons have been effectively used to engineer human cells, including NK cells.^91,97,98 The most extensively studied transposon systems for clinical gene transfer are Sleeping Beauty (SB), piggyback (PB), and TcBuster, which have been used as platforms for delivery of CARs into hematopoietic cells.⁹¹ Human iPSC-derived NK cells have been successfully engineered to express CARs using the PB transposon system with enhanced antigen-specific NK cell activity observed against mesothelin-expressing tumors in vitro and in vivo.⁵⁵ PB and SB systems have also been used to successfully express a panel of mesothelin directed CARs in NK-92 cells.⁵⁵ As with any nonspecific gene transfer, transposon-mediated chromosomal integration has the risk of causing insertional mutagenesis. A recent phase I clinical trial of donor-derived CD19-targeting CAR T-cells engineered using the PB transposon system reported the development of malignant CAR-expressing CD4+ T-cell lymphomas in 2 out of 10 treated patients.^99,100 Transcriptomic analysis of malignant tissue confirmed transgene promoter-driven transcriptional upregulation of surrounding regions, background genomic copy number variations, and high transgene copy number per cell.¹⁰⁰ This outcome highlights the need for careful surveillance and monitoring of patients receiving novel engineered cell therapy products, particularly when nontargeted gene engineering methodology is used.

Targeted knockdown and knock-in gene modification strategies

Targeted gene editing can also be achieved using zinc-finger nucleases (ZFN), transcription activator-like nucleases (TALENs), and (CRISPR)/CRISPR-associated protein 9 (Cas-9) systems, often in combination with adeno-associated virus (AAV) vectors. ZFN and TALEN manipulate protein-DNA interactions to target specific genomic loci for editing and require expertise in protein engineering.¹⁰¹ CRISPR-Cas9 uses RNA-guided DNA recognition, with guide RNA designed with homology to the genetic locus of interest.¹⁰¹ CRISPR-Cas9 can be used to reliably knockout target genes or to knock-in genetic modifications with targeted integration of a homologous recombination repair template.^91,102 Efficient CRISPR-Cas9 mediated knockout of inhibitory signaling pathways in primary NK cells was recently described with efficiency reaching 90%.⁹³ The CRISPR-Cas9 system was subsequently used in combination with recombinant AAV serotype 6 (rAAV6) for delivery of a homologous recombination template DNA for an efficient knock-in strategy.⁹³ As an alternative, other groups have packaged Cas9 and guide RNAs into ribonucleoprotein (RNP) complexes for high gene-editing efficiency in NK cells.^91,103 Related gene-editing techniques include Base editors (BEs) that fuse a catalytically inactive Cas9 protein to a DNA deaminase domain with the goal of precise introduction of a targeted single nucleotide change without double-strand breaks or DNA donor molecules.^91,104

The development of successful NK cell gene-based therapies is rooted in an effective gene-editing strategy. Research investigating viral and nonviral delivery platforms with or without targeted knockdown or knock-in mechanisms aims to optimize the efficiency and safety profile of genetic engineering of NK cells. Regardless of the platform selected, each described genetic modification strategy has evidence of successful NK Cell modification and thus, can be considered to engineer enhanced effector cell function and immunotherapeutic efficacy.

THERAPEUTIC APPLICATIONS OF GENETICALLY MODIFIED NATURAL KILLER CELLS

Modulation of surface receptor expression

Chimeric antigen receptor-natural killer cells demonstrate target-specific activation

CAR-T cell therapy has revolutionized the treatment landscape for pediatric hematologic malignancies and broadened the developmental scope to include alternate tumor types and effector cell starting materials. NK cell CARs can be designed to target diverse antigens for specific NK cell activation by using unique extracellular single-chain variable fragments (scFvs) linked to combinations of intracellular costimulatory domains (Fig. 2A). A robust body of preclinical work demonstrates that NK cells can effectively be engineered to express CARs, and this has been comprehensively reviewed elsewhere.^105-110 These preclinical investigations have laid the ground for clinical translation, and a number of CAR-NK cell therapy trials are underway. As of November 2021, 26 CAR-NK cell cancer treatment trials are registered on clinicaltrials.gov (Table 1). Two of these have published results. One example is an actively recruiting phase I/II trial (NCT03056339) led by the University of Texas MD Anderson Cancer Center. In this ongoing study, HLA-mismatched anti-CD19 UCB CAR-NK cells are administered to patients with relapsed or refractory CD19-positive hematologic malignancies. In a report describing the initial cohort of 11 patients, the treatment was well-tolerated without the development of CRS, ICANS, or GVHD. Lymphodepleting chemotherapy was given prior to CAR-NK cell infusion (1 × 10⁵, 1 × 10⁶, or 1 × 10⁷ CAR-NK cells per kilogram of body weight), with 7 of 11 patients achieving a complete remission.⁷⁸ A separate phase I clinical trial conducted in Suzhou, Jiangsu, China evaluated the safety of CD33 CAR-NK cells in patients with relapsed and refractory AML (NCT02944162). A third-generation lentiviral CAR construct (αCD33.CD28.4-1BB) was used for NK-92 cell transduction in this study. Three patients with relapsed and refractory AML treated with salvage chemotherapy were infused with three escalating doses of irradiated CD33-CAR NK-92 cells. No adverse effects or durable clinical efficacy were recorded.⁶⁶

Fig. 2.

NK Cell Engineering for Therapeutic Application. Schematic depicting (A) Expression of CARs targeting TAA to enhance cytotoxicity and NK cell activation. (B) Constitutive secretion or (C) membrane-tethered cytokine can sustain NK cell activation and persistence. (D) Enhanced surface expression of chemokine receptors can mediate NK cell localization to tumor along chemokine gradients. (E) Engager molecules (BiKEs or TriKEs) can be combined with engineered CD16 to direct powerful ADCC of tumor cells. Incorporation of IL15 in the small molecule can further support persistence.

Table 1.

Clinical trials testing the use of CAR-NK cells as anticancer immunotherapy

Target	Agent	NCT	Year Posted	Phase	Status	Tumor	NK Source	Sponsor	Vector
CD19	CAR-NK019	NCT04887012	2021	I	Recruiting	B-cell NHL	iPSC	Second Affiliated Hospital, School of Medicine, Zhejiang University, China	Lentiviral
CD19	Anti-CD19 CAR NK Cells	NCT04639739	2020	Early I	Planned	• NHL	Unknown	Xinqiao Hospital of Chongqing, China	Unknown
CD22	Anti-CD22 CAR-NK Cells	NCT03692767	2018	Early I	Unknown	• Refractory B-cell Lymphoma	Unknown	Allife Medical Science and Technology Co., Ltd., China	Unknown
CD19	Anti-CD19 CAR NK Cells	NCT03690310	2018	Early I	Unknown	• Refractory B-cell Lymphoma	Unknown	Allife Medical Science and Technology Co., Ltd., China	Unknown
CD33	Anti-CD33 CAR NK Cells	NCT05008575	2021	I	Planned	• AML	Unknown	Xinqiao Hospital of Chongqing, China	Unknown
Mesothelin	Anti-Mesothelin CAR-NK Cells	NCT03692637	2018	Early I	Unknown	• Epithelial Ovarian Cancer	Unknown	Allife Medical Science and Technology Co., Ltd., China	Unknown
NKG2D	CAR-NK cells targeting NKG2D ligands	NCT03415100	2018	I	Unknown	• Metastatic Solid Tumors	PB-NK	The Third Affiliated Hospital of Guangzhou Medical University, China	mRNA electroporation
PSMA	Anti-PSMA CAR NK Cells	NCT03692663	2018	Early I	Planned	Castration-Resistant Prostate Cancer	Unknown	Allife Medical Science and Technology Co., Ltd., China	Unknown
BCMA	Anti-BCMA CAR-NK Cells	NCT05008536	2021	Early I	Planned	Refractory Multiple Myeloma	UCB-NK	Xinqiao Hospital of Chongqing, China	Unknown
ROBO1	ROBO1 CAR-NK Cells	NCT03940820	2019	I/II	Recruiting	Solid Tumors	NK-92	Asclepius Technology Company Group (Suzhou) Co., Ltd., China	Lentiviral
BCMA	BCMA CAR-NK 92 Cells	NCT03940833	2019	I/II	Planned	Multiple Myeloma	NK-92	Asclepius Technology Company Group (Suzhou) Co., Ltd., China	Lentiviral
PD-L1	PD-L1 t-haNK	NCT04847466	2021	II	Planned	Gastroesophageal Junction Cancers; Advanced HNSCC	NK-92	National Cancer Institute, United States	Unknown
CD19, CD22	Anti-CD19/CD22 CAR NK CElls	NCT03824964	2019	Early I	Unknown	Refractory B-Cell Lymphoma	Unknown	Allife Medical Science and Technology Co., Ltd., China	Unknown
CD19	CAR-NK-CD19 Cells	NCT04796675	2021	I	Recruiting	ALL, CLL, NHL	CB-NK	Wuhan Union Hospital, China	Retroviral
CD19	NKX019	NCT05020678	2021	I	Recruiting	B-ALL, CLL, NHL	PB-NK	Nkarta Inc., United States	Unknown
CD33	Anti-CD33 CAR-NK cells	NCT02944162	2016	I/II	Unknown	AML	NK-92	PersonGen BioTherapeutics (Suzhou) Co., Ltd., China	Lentiviral
CD19	PCAR-119	NCT02892695	2016	I/II	Unknown	Leukemia, Lymphoma	NK-92	PersonGen BioTherapeutics (Suzhou) Co., Ltd., China	Lentiviral
NKG2D	NKX101	NCT04623944	2021	I	Recruiting	AML, MDS	PB-NK	Nkarta Inc., United States	Unknown
ROBO1	BiCAR-NK cells	NCT03941457	2019	I/II	Recruiting	Pancreatic Cancer	NK-92	Asclepius Technology Company Group (Suzhou) Co., Ltd., China	Lentiviral
CD19	CAR.CD19-CD28-zeta-2A-iCasp9-IL15-transduced CB-NK cells	NCT03579927	2018	I/II	Withdrawn	B-cell lymphoma	UCB, CB-NK	M.D. Anderson Cancer Center	Retroviral
CD19	iC9/CAR.19/IL15-Transduced CB-NK Cells	NCT03056339	2017	I/II	Active	B-Lymphoid Malignancies, ALL, CLL, NHL	UCB, CB-NK	M.D. Anderson	Retroviral
CD70	CAR.70/IL15-transduced CB-NK cells	NCT05092451	2021	I/II	Recruiting	B-cell Lymphoma, MDS, AML	CB-NK	M.D. Anderson	Retroviral
CD19	FT596	NCT04555811	2020	I	Recruiting	NHL	iPSC	Masonic Cancer Center, University of Minnesota	Lentiviral
CD19	FT596	NCT04245722	2020	I	Recruiting	B-cell lymphoma, CLL	iPSC	Fate Therapeutics	Lentiviral
ErbB2	NK-92/5.28.z	NCT03383978	2017	I	Recruiting	HER2-positive Glioblastoma	NK-92	Johann Wolfgang Goethe University Hospital	Lentiviral
CD19	CAR-ITNK	NCT04747093	2021	I/II	Recruiting	B cell leukemia/lymphoma	T-cells	Nanfang Hospital of Southern Medical University, China	Unknown

Abbreviations: ALL, acute lymphoblastic leukemia. acute lymphocytic leukemia; AML, acute myeloid leukemia; CLL, chronic lymphocytic leukemia; iPSC, induced pluripotent stem cells; MDS, myelodysplastic syndrome; NHL, Non-Hodgkin lymphoma; PB-NK, Peripheral blood-derived NK cells; UCB, CB-NK, umbilical, cord blood-derived CAR-engineered NK cells.

Downregulation of inhibitory receptors enhances natural killer cell activation

An alternative approach to cell surface receptor modulation is the downregulation of innate NK inhibitory receptors to bias NK cells toward activation. Cancer cells can block immune responses by targeting inhibitory receptors on the surface of NK cells. NKG2A is one such receptor that dimerizes with CD94 to bind HLA-E molecules on tumor cells, which dampens NK cell activity.¹¹¹ Consequently, RNAi-mediated inhibition of NKG2A expression has improved NK cell activity in the preclinical study. NKG2A silencing can enhance NK cell cytotoxicity by up to 40%.¹¹² Silencing of inhibitory receptors using CRISPR/Cas9, ZFN, or TALEN is likely to result in similarly amplified NK cell antitumor functionality.

Enhancing natural killer cell activity through the modulation of cytokine signaling

Natural killer cell in vivo persistence can be sustained with cytokine stimulation

The success of adoptive NK cell transfer relies to a great extent on the persistence of the cells.¹¹ Mature adoptively transferred NK cells have a limited lifespan in vivo with a peripheral half-life of about 7 to 10 days.^30,32 This short lifespan has certain advantages including: decreased risk of prolonged on target, off-tumor effects,⁴⁵ decreased immunogenicity, reduced risk of cytokine release syndrome and neurotoxicity,¹¹⁰ and the opportunity to safely redose NK cells for continued effect. Even so, clinical studies have demonstrated the important role of persistence and expansion for clinical efficacy.^11,113-115 In order to prolong NK cell survival, exogenous cytokine administration with recombinant IL2, IL15, or IL15 “superagonists” has been used in clinical trials.^116-121 However, cytokine infusion is associated with systemic toxicities that are dose- and therapeutic route-dependent.^118-121 Localized cytokine delivery to support NK cell persistence may be safer, as first studied using NK-92 (Fig. 2B). An animal model of liver metastases paired with NK-92 treatment was used to show that NK-92 stably expressing and secreting IL2 had better antitumor control than NK-92 alone.¹²² Similarly, UCB CAR-NK cells engineered to secrete IL15 prolonged the survival of mice engrafted with CD19 positive tumors.¹²³ Safety of IL15-secreting CAR-NK cells was tested in human patients in the MD Anderson phase I clinical trial (NCT03056339) using UCB-derived CD19-CAR.IL15 NK cells described above.⁷⁸ The trial found that patients who had a response to therapy had a significantly higher early expansion of CAR-NK cells and that early expansion and low-level persistence of CAR-NK cells for at least 12 months after infusion was related to the inclusion of IL-15 in the CAR construct.⁷⁸ Our group has also investigated transgenic IL15 in PB-NK cells designed for the treatment of hematological malignancies. As expected, constitutive IL15 secretion enhances NK and CAR-NK cell activation and supports in vivo persistence. However, we found that treatment with IL15-secreting NK cells caused lethal toxicity in one AML xenograft model.³⁸ Membrane-bound IL15 or the IL15/IL15 R complex can as well augment NK cell persistence and may be a safer alternative to soluble IL15 (Fig. 2C).⁴³ To illustrate, oncolytic virus-mediated local IL15/IL15 Rα (OV-IL15 C) secretion has been used to enhance PB-CAR-NK persistence and in vivo functionality.¹²⁴ Similarly, the IL15/IL15 Rα fusion protein, a CD19-CAR, and a high-affinity, noncleavable CD16 (hnCD16) have together been expressed in iPSC-NKs, resulting in enhanced functionality in preclinical CD19+ malignancy models.¹²⁵ This product, FT596, is now being investigated in a phase 1 clinical trial for the treatment of B-cell lymphoma and chronic lymphocytic leukemia (NCT04245722).

Natural killer cell modification can direct tumor site homing

Surface expression of chemokine receptors has been tested as a method to enhance NK cell tumor trafficking (Fig. 2D). CXCR2 expression in the YTS NK cell line and CXCR4 in PB-NKs can promote NK cell trafficking to renal cell carcinoma (RCC) and CXCL12-secreting glioblastoma (GB), respectively.^126,127 Electroporation of primary human NK cells with gain-of-function CXCR4_R334X mRNA upregulated CXCR4 on the NK cell surface and supported NK cell trafficking to bone marrow in a xenograft model.¹²⁸ Engineered CCR7 expression on NK cells can promote migration toward CCL19 in lymph nodes.¹²⁹ Primary NK cells transfected with mRNAs encoding for CXCR1 and an NKG2D-CAR can be efficiently directed toward IL8-secreting cancers.¹³⁰ Moreover, combinational immunotherapy utilizing an oncolytic virus encoding CCL5 and NK cells engineered to express the CCR5 receptor can improve NK cell tumor infiltration.¹³¹

Manipulating natural killer cell interaction with small molecule engagers

BiKEs enhance natural killer cell activation and antibody-dependent cellular cytotoxicity

Preclinical and clinical studies support the use of bispecific small molecular engagers to stabilize the NK cell-target cell immunologic synapse and improve antitumor cytolysis (Fig. 2E). BiKEs that simultaneously target a tumor-specific antigen and CD16 (FcγRIII), the low-affinity Fc receptor,^132,133 are capable of stimulating ADCC. For example, a CD16xCD33 BiKE tested in patients with myelodysplastic syndromes (MDS) can successfully activate primary patient NK cells with specific degranulation and cytokine production against CD33+ MDS targets and can reverse myeloid-derived suppressor cell (MDSC)-mediated immunosuppression of NK cells through targeted CD33+ MDSC lysis.¹³⁴ Secondly, the use of a CD30xCD16 A BiKE (AFM13) targeting CD30 on tumor cells and the CD16 A isoform on NK cells has been extensively tested in patients with CD30+ malignancies with notable antitumor responses, a good safety profile, and dose-dependent NK cell CD69 upregulation.^135-137

With these encouraging results, optimizing ADCC by combining BiKEs with genetically modified NK cells holds promise. Engineered expression of high-affinity Fc receptor isoforms to enhance ADCC is being studied by multiple groups.^138-142 For example, FT516 is a CD19-targeted CAR-NK cell product derived from a clonal iPSC line, further engineered to include a high-affinity, noncleavable CD16.¹⁴¹ Preliminary results from a phase I trial of FT516 (NCT04023071) used in patients with relapsed/refractory B-cell lymphoma evaluated the safety of infusion in combination with the monoclonal CD20 antibody, rituximab. Patients received 3 days of lymphodepleting chemotherapy, followed by 1 dose of rituximab, and 3 weekly infusions of FT516 administered with systemic IL-2. Six patients with a median age of 65.5 years enrolled in the trial, all heavily pretreated, 5 of whom completed the therapy. Escalating doses were evaluated including 3 million cells/dose (2 patients), 90 million cells/dose (3 patients), and 300 million cells/dose (1 patients). No CRS, ICANS, GVHD, or other serious side effect were observed in 5 treated patients. Three patients treated at > 90 million cells/dose achieved an objective response.¹⁴³

Trispecific killer engagers can complement antibody-dependent cellular cytotoxicity with cytokine driven stimulation

Similar to BiKEs, trispecific killer engagers (TriKEs) are composed of an antigen-specific scFv linked to the scFvs of two other antibodies of different specificities or to the scFv of one other antibody and a cytokine. TriKEs have also been designed to augment ADCC by engaging CD16 along with tumor-specific antigens.¹⁴⁴ Newer generations of TriKEs incorporate the IL-15 cytokine to augment NK cell function. This, coupled with CD16 engineering modalities described above may further enhance the specificity and persistence of antitumor NK cell activation. Additional NK cell manipulations discussed above such as the downregulation of inhibitory receptors can be employed to further enhance TriKE immunotherapy, and combinatorial manipulation remains an area of active study.

SUMMARY

NK cells are a powerful tool for targeted immunotherapy. They harbor a range of innate cytotoxic mechanisms and can be genetically engineered to enhance their function. This can be conducted by modulating cell surface receptor expression, by manipulating stimulatory cytokines and chemokines, and by optimizing engagement with targeted small molecules (Table 2). NK cells are subject to genetic modification using viral and nonviral vectors and can be isolated or derived from different primary sources. This diversity leaves much option for further study and discovery. While early clinical experiences underline their promise, the clinical development of engineered NK cell therapy remains in its infancy with current reports limited to small cohorts of treated patients. There is a need for further discovery and subsequent translation into large-scale clinical trials to truly detail the role of these effector cells in the expand-ing armamentarium of immunotherapies for blood cancers. The success of genetically engineered NK cell therapy is dependent on the complete understanding of the factors that drive NK cell activation, immune synapse formation, and target cell killing. There remains much to be discovered given the potential for these remarkable innate effector cells to be effective anticancer immunotherapeutics.

Table 2.

Therapeutic applications for genetically modified NK cells

Therapeutic Modality	Mechanism of Action	Clinical Translation
Modulation of Surface Receptor Expression
CAR-NK	CARs with specificity for tumor antigens are genetically introduced via viral or nonviral vectors for stable NK cell surface expression.	Several105-110 (see Table 1)
NKG2A Silencing	RNAi-mediated inhibition of NKG2A expression, an innate inhibitory NK cells receptor. Other applicable methods of gene silencing include CRISPR/Cas9, ZFN, and TALEN.	Preclinical112
Modulation of Cytokine Signaling
IL2-secreting NK cells	NK-92 cells genetically engineered to stably secrete IL2, an important stimulatory cytokine for NK cell proliferation and functionality.	Preclinical122
IL15-secreting CAR-NK cells	UCB CAR-NK cells genetically engineered to secrete IL15, an important stimulatory cytokine for NK cell proliferation and functionality. PB-NK cells genetically engineered to constitutively secrete transgenic IL15 to enhance NK cell activation and persistence.	UCB-derived CD19-CAR.IL15 NK cells, phase I/II clinical trial (NCT03056339)78,123 Preclinical38
IL15/IL15 R complex expressing CAR-NK cells	PB-NK cells genetically engineered to locally secrete IL15/IL15 Rα to enhance NK cell persistence and functionality. iPSC-NK cells genetically engineered to express the IL15/IL15 Rα fusion protein, a CD19-targeting CAR, and a high-affinity, noncleavable CD16 for enhanced functionality in CD19+ malignancies.	Preclinical124 iPSC-derived CD19 CAR NK cell with a high-affinity, noncleavable CD16, and recombinant fusion of IL15 and IL15RF (FT596) as monotherapy or in combination with rituximab or obinutuzumab, phase I clinical trial (NCT04245722)125
Chemokine receptor expression on NK cells (ex. CXCR2, CXCR4, CCR7, CXCR1, CCR5)	NK cells are engineered to express gain of function chemokine receptors to support NK cell tumor trafficking.	Preclinical126-131
Small Molecule Engagers
BiKEs	Genetically modified NK cells engineered to express high-affinity CD16 Fc receptor isoform to enhance ADCC used in combination with BiKEs, targeting CD16 on NK cells and tumor antigens on target cells.	iPSC-derived CD19 CAR NK cell with a high-affinity, noncleavable CD16 (FT516) in combination with rituximab, phase I clinical trial (NCT04023071)138-143
TriKEs	TriKEs incorporating IL-15 cytokine for enhanced NK cells function used together with NK cells genetically modified to express high-affinity CD16 Fc receptors, for further enhanced specify and persistence of antitumor NK cell activation.	Preclinical144

Abbreviations: ADCC, antibody-dependent cellular cytotoxicity; BiKE, bispecific killer engager; iPSC, induced pluripotent stem cell; PB, peripheral blood; TALEN, transcription activator-like nucleases; TriKE, trispecifc killer engager; UCB, umbilical cord blood; ZFN, zinc-finger nucleases.

KEY POINTS.

• NK cells are lymphocytes of the innate immune system with powerful intrinsic cytotoxic mechanisms that can be further enhanced by genetic engineering.

• NK cells for therapeutic use are derived from numerous sources, including peripheral blood, cord blood, pluripotent stem cells, embryonic stem cells, and transformed NK cell lines.

• Genetic modification of NK cells has been studied using viral and nonviral vectors for nontargeted and targeted genomic editing. Retroviral vectors have been optimized for safety and efficiency and are the preferred vehicle for ex vivo genetic engineering.

• NK cell activity can be enhanced through the modification of cell surface receptors, manipulation of the inflammatory and suppressive cytokine milieu, and directed evasion of regulatory mechanisms.

• Combination with small molecular engagers can strengthen NK cell targeting and activation.

CLINICS CARE POINTS.

• NK cells are an option for “off-the-shelf” cellular therapy with a favorable safety profile and limited risk of graft-versus-host disease with allogeneic products.

• NK cells can be genetically modified using both genomic targeted and nontargeted techniques. There is a risk of insertional mutagenesis without targeting, thus there is a need for close patient monitoring following engineered NK cell infusion.

• NK-cell-based immunotherapy is made more effective by the presence of a tumor antigen for CAR or BiKE/TriKE targeting and mechanisms that promote NK cell persistence and activation.