The Next Generation of Cellular Immunotherapy: CAR-NK Cells

Abstract

The advent of chimeric antigen receptor (CAR) engineering has led to the development of powerful cellular therapies for cancer. CAR-T cell based treatments have had notable clinical success, but logistical issues and associated toxicities are recognized limitations. There is an emerging interest in utilizing other immune effector cell types for CAR therapy. Natural killer cells are part of the innate immune system and these lymphocytes play a major role in immunosurveillance and anti-tumor immune responses. Incorporating chimeric antigen receptors into natural killer cells provides the opportunity to harness and enhance their innate cytotoxic potential towards malignancies. In this review we discuss the production of CAR engineered natural killer cells, highlight data on their preclinical and clinical efficacy, and examine the obstacles and strategies to overcome them.

Introduction

Cellular therapies for cancer have emerged as powerful treatment strategies that have helped revolutionize the immunotherapy field. The advent of chimeric antigen receptor (CAR) engineering, which was first applied in the 1990s for use in T cells with the goal of combining the beneficial anticancer characteristics of T cells with antigen specificity, has proven to be a successful example of personalized medicine.¹ Clinical trials have shown particular promise in the use of CAR-T cells for the treatment of hematological B-cell malignancies² and there are now five FDA-approved CAR-T cell therapies for hematologic cancers.^3–9 Despite the impressive advances in the cellular therapy field over the past decade there are still major limitations and obstacles to the use of CAR-T cell based therapeutics. These drawbacks include: time-consuming production, the cumbersome logistics involving generation of autologous products for each patient, high cost, and a significant toxicity profile.¹⁰ Allogeneic CAR-T cells have been used to address some of these challenges, however this approach presents its own challenges, including graft versus host disease (GVHD) which poses a significant risk even when accounting for human leukocyte antigen (HLA) matching.¹¹ Limiting side-effects include the substantial risk of cytokine release syndrome (CRS) and immune effector cell associated neurotoxicity syndrome (ICANS).^12,13 The challenges to more widespread adoption of CAR-T cells in the clinic have spurred investigation of alternative immune effector cell populations as potential vehicles for CARs. In recent years, natural killer cells have emerged as promising candidates as an alternative to modified T cells.

Natural killer (NK) cells were first discovered in mice in 1975 and were shown to selectively kill leukemia cells.¹⁴ These lymphocytes are components of the innate immune system, and can eliminate cancerous cells without the need for prior sensitization, as their function and activity are regulated solely by the balance of signals received from activating and inhibitory receptors.¹⁵ NK cells mediate target cell lysis through several distinct mechanisms such as cytotoxic granule exocytosis, upregulation of death ligand expression (i.e., Fas and TRAIL), and production of cytokines such as interferon gamma (IFN-γ).¹⁶ NK cells are also capable of mediating antibody-dependent cell-mediated cytotoxicity (ADCC). In ADCC, monoclonal antibodies that target antigens expressed by tumor cells will recruit NK cells through FcγR-Fc interactions leading to NK cell activation and target cell lysis.¹⁷ NK cells have multiple advantages over T cells for CAR-based immunotherapy. As NK cell activation is triggered by their innate receptors, the addition of a CAR will redirect towards antigen-specific targets but NK cell-mediated cytotoxicity will be preserved even in the event of immune escape mechanisms such as target antigen loss or downregulation. Importantly, potential adverse events due to NK cell-based therapies are less likely due to their limited persistence in vivo and CRS, ICANS, and GVHD have not emerged as toxicities.¹⁸ Ultimately, CAR-NK cells are promising as the next generation of cellular therapy because they do not require HLA compatibility, are associated with minimal side-effects, and can be produced on a large scale from several sources to make “off-the-shelf” treatment a possibility. In this review we will briefly summarize CAR-NK cell design and discuss the progress being made in the development of NK cells as platforms for CAR-based therapies. We will focus on the challenges to CAR-NK cell therapy and the recent advances in improving their clinical efficacy.

CAR-NK Cell Production

Commonly used sources of functional NK cells for CAR-NK production include the NK-92 cell line, adult peripheral blood (AB), umbilical cord blood (CB), and induced pluripotent stem cells (iPSCs).^19–21 Although autologous NK cells can be utilized to generate CAR-NK cells, patient sourced NK cells have limited efficacy against their cancer and enhancing their clinical utility with CAR engineering may be challenging.²² Depending on their origin, harvested NK cells have varying maturation stages, viability, and anti-tumor responses. The advantages and disadvantages associated with each source of NK cells are summarized in Figure 1. Some of the major benefits of utilizing NK-92 cells is they are highly cytotoxic, express relatively low levels of inhibitory receptors, and large numbers of these NK cells can quickly and easily be generated in vitro.^23,24 However, as they were originally derived from a patient with a NK cell lymphoma, there is a small risk of tumor engraftment and irradiation prior to infusion is a safety requirement to prevent expansion. Unfortunately, this process negatively impacts their in vivo proliferation and persistence, and both factors have been shown to be crucial for the success of adoptive cellular therapy.²⁵ An additional drawback is that NK-92 cells lack expression of the Fc-recognizing CD16 domain and are therefore unable to trigger ADCC without further manipulation. One of the downsides of using primary NK cells isolated from AB or CB is their in vitro culture and expansion is much more labor-intensive. However, both AB and CB have been shown to be an effective source of NK cells for CAR.²⁶ Large numbers of NK cells can be obtained from samples of CB compared to AB; however, in contrast to AB, CAR-NK cells derived from CB progenitors mostly present an immature phenotype and have inferior cytotoxic capabilities prior to ex-vivo expansion.²⁷ Generating clinical doses of CAR-NK cells using CB and AB NK cells is still a challenge and requires ex-vivo expansion of the cells. This ex-vivo expansion can be done through several methods.²⁸ This includes coculturing with IL-2, IL-12, IL-15, IL-18 and/or IL-21.²⁸ This method is relatively easy and has the benefit of not requiring feeder cells. However, it requires 4–10 weeks to reach clinically relevant numbers, which is significantly longer than utilizing feeder cells.²⁹ CB and AB derived CAR-NK cells can also be expanded using universal antigen presenting cells (uAPC).³⁰ AB and CB cells can be cocultured with the leukemia cell line K562, which is frequently transduced with membrane bound cytokines. Coculturing NK cells with HLA deficient K562 cells expressing CD48, 4–1BBL, and membrane-bound IL-21 causes the expansion of CAR-transduced CB and AB cells by >900-fold in only 2 weeks.³⁰ NK cells derived from iPSCs offer some of the same benefits of NK-92, in that large numbers of mature and functional NK cells can be generated with relative ease.³¹ In contrast to NK-92 cells, iPSCs do not require irradiation and have also been shown to be effective upon CAR modification albeit with a less mature phenotype, reminiscent of CB-derived cells.³² Additionally, cytokine induced memory like (CIML) NK cells have been developed by activating AB cells with IL-12, IL-15, and IL-18.³³ These cells exhibit potent antitumor responses and have been utilized in the treatment of leukemia, inducing complete remission in these patients.³⁴ Recently, these cells have been transduced with CAR constructs targeted against CD19 and have effectively controlled lymphoma burden in vivo.³⁵ These memory like CAR-NK cells exhibited increased degranulation and specific killing compared to conventional CAR-NK cells.³⁵

Figure 1.

Advantages and Disadvantages of CAR-NK Cell Sources

Created with BioRender.com

A CAR construct is comprised of three components: an ectodomain responsible for antigen-recognition, a transmembrane linker, and an intracellular signaling domain. The extracellular antigen receptor portion is generally composed of a single-chain variable fragment (scFv). scFvs can target unique or overexpressed tumor cell antigens and CAR constructs are capable of targeting both surface proteins and HLA peptide complexes.³⁶ When a CAR expressing NK cell binds the target antigen, the intracellular signaling domain transmits an activation signal, leading to cytokine secretion and lysis of the target cell.³⁷ In first-generation CARs, the signaling domain was CD3ζ alone.³⁸ Second-generation CARs added a single costimulatory domain such as CD28 or 4–1BB, while third-generation CARs contain two or more distinct costimulatory domains. Future generations of CARs are being developed with the goal of further modulating and enhancing the activation of CAR-engineered cells to kill malignant cells.³⁹ A fourth generation of CARs was recently described, where the immune effector cells were endowed with two transgenic products, the CAR and a transgenic payload, such as the release of IL-7, IL-12, IL-15, IL-18, IL-23 or some combination.^40,41 CAR constructs that implemented NK-specific 2B4, DAP-10, or DAP-12 co-stimulatory domains instead of CD28 or 4–1BB were characterized as having improved cytotoxicity and higher IFN-γ secretion.⁴² These CAR constructs are depicted in Figure 2. In order to introduce the CAR construct into a cell, genetically engineered viruses are used. CAR-NK cells are commonly generated through lentiviral or alpharetroviral transduction, with some reports of transduction efficiency well over 90% for both NK-92 and primary NK cells.⁴³ Lentiviruses are capable of infecting both dividing and non-dividing cells and can be used for transduction of a wider variety of cell types including quiescent stem cells, providing an advantage over retroviral vectors.⁴⁴ For a more detailed discussion of CAR-NK design and engineering see a review by Gong et al.⁴⁵ CAR constructs enable NK cells to conjugate target cells more rapidly than unmodified primary NK cells or NK-92 cells, thereby enhancing their cytotoxic potential.⁴⁶ In addition to the targeted cytotoxicity of CAR-NK cells, they may also still function through NK cell receptors in a CAR-independent manner. Some of the promising CAR targets under investigation with NK cells include CD38, CD20, mesothelin, CD22, CD19, HER2, and GD2.⁴⁷ The best candidate antigens for use in CAR therapy will be minimally expressed in normal tissues, thus carrying a lower risk of off-target effects and maximizing clinical benefit while minimizing the risk of harmful toxicities to patients.

Figure 2.

Generations of CAR Constructs

Created with BioRender.com

Pre-Clinical and Clinical Data on CAR-NK Cell Therapy

Pre-clinical data on the use of CAR-NK cells are very promising and have demonstrated efficacy for a variety of cancers and target antigens. CAR-NK cells derived from multiple sources are under extensive study for use in hematological malignancies and solid tumors.⁴⁸ It is worth noting that although primary NK cells are an effective source for CAR modification, the majority of pre-clinical work has employed the NK-92 cell line. Second and third generation CAR-NK cells derived from NK-92 cells have shown to be effective in pre-clinical work targeting a range of antigens and cell types.⁴⁹ CAR-NK-92 cells have enhanced cytotoxicity against T cell acute lymphoblastic leukemia and CD19+ leukemias and lymphomas, as well as HER2 overexpressing glioblastoma.^50–52 Multiple studies have confirmed activity of CAR-NK-92 cells targeting ovarian cancer antigens in in vitro and in vivo pre-clinical models.^53–55 Lastly, CAR-NK-92 also have pre-clinical efficacy when used to treat CS1 expressing multiple myeloma and prostate cancer.^56,57 CAR-NK cells derived from the KHYG-1 have also been shown to have enhanced cytotoxic capabilities against glioblastoma cells expressing EGFRvIII.⁵⁸ While there are relatively fewer studies employing CAR-NK cells sourced from AB, CB, or iPSCs; in vitro models and in vivo experiments have also demonstrated promising anti-tumor effects.⁵⁹

While there have now been hundreds of clinical trials for CAR-T therapy, CAR-NK cell based treatments have only recently begun clinical investigation. There are currently only three published clinical trials on the use of CAR-NK therapy. In those clinical trials, the administration of CAR-NK cells were found to be safe and effective for patients with multiple myeloma, acute myeloid leukemia, and CD19+ lymphoid tumors.^60–62 In a trial of eleven patients with either chronic lymphocytic leukemia (CLL) or non-Hodgkin’s lymphoma (NHL) who had previously received a median of four lines of therapy, a single infusion of anti-CD19 CAR-NK cells led to objective responses in eight patients, of whom 7 had a complete response within the first month after treatment.⁶³ No serious adverse events were observed in any of the patients. Despite the limited number of published clinical trials on CAR-NK therapy, additional pilot clinical studies have achieved notable responses. Using an RNA electroporation approach, transient expression of a NKG2D RNA CAR significantly augmented the cytolytic activity of NK cells in vitro and was piloted as a treatment for two patients with chemotherapy-refractory metastatic colorectal cancer.⁶⁴ Malignant cells in ascites fluid were almost undetectable after treatment and the observed volume of ascites was also significantly lower post-treatment with no recorded serious adverse effects. A third patient with metastatic colorectal cancer in the liver was treated by direct injection into liver lesions, which resulted in significant reduction in tumor burden. There are currently 32 CAR-NK based clinical trials registered as completed or active in NIH.gov (Table 1), with a majority ongoing in China. In summary, pre-clinical and early clinical data support further investigation of CAR-NK cell based therapeutics for a wide range of malignancies.

Table 1.

Completed or Active CAR-NK Clinical Trials

NCT Number	NK Cell Source	Target	Indication	Status
Hematological Malignancies
NCT00995137	Haplo-identical donor	CD19	ALL	Completed
NCT02892695	NK-92	CD19	CD19+ Leukemia/Lymphoma	Phase 1/2, Recruiting
NCT02944162	NK-92	CD33	AML	Phase 1/2, Recruiting
NCT03940833	NK-92	BCMA	Multiple Myeloma	Phase 1/2, Recruiting
NCT02742727	NK-92	CD7	CD7+ Leukemia/Lymphoma	Phase 1/2, Recruiting
NCT03056339	Umbilical and Cord Blood	CD19	B Cell Malignancies	Phase 1/2, Not yet recruiting
NCT05092451	Cord Blood	CD70	B Cell Lymphoma, AML	Phase 1/2, Not yet recruiting
NCT04887012	Haplo-identical donor	CD19	B Cell NHL	Phase 1, Recruiting
NCT04796688	AT19	CD19	B Cell Malignancies	Phase 1, Recruiting
NCT05008536	Umbilical and Cord Blood	BCMA	Multiple Myeloma	Phase 1, Recruiting
NCT04796675	Cord Blood	CD19	B Cell Malignancies	Phase 1, Recruiting
NCT05020678	Allogeneic Donor	CD19	B Cell Malignancies	Phase 1, Recruiting
NCT04623944	Allogeneic Donor	NKG2D	AML	Phase 1, Recruiting
NCT05182073	iPSC	BCMA	Multiple Myeloma	Phase 1, Recruiting
NCT04245722	iPSC	CD19	B Cell Lymphoma, CLL	Phase 1, Recruiting
NCT04555811	iPSC	CD19	B Cell Lymphoma	Phase 1, Recruiting
NCT05247957	Cord Blood	NKG2D	AML	Phase 1, Recruiting
NCT05215015	Undisclosed	CD33/CLL1	AML	Phase 1, Recruiting
NCT05008575	Undisclosed	CD33	AML	Phase 1, Recruiting
NCT05379647	Allogeneic Donor	CD19	B Cell Malignancies	Phase 1. Recruiting
NCT03824951	iPSC	CD19	B Cell Lymphoma	Phase 1, Not yet recruiting
NCT03824964	iPSC	CD19/CD22	B Cell Lymphoma	Phase 1, Not yet recruiting
NCT03692767	iPSC	CD22	B Cell Lymphoma	Phase 1, Not yet recruiting
NCT03690310	iPSC	CD19	B Cell Lymphoma	Phase 1, Not yet recruiting
NCT04639739	Undisclosed	CD19	NHL	Phase 1, Not yet recruiting
Solid Tumors
NCT04390399	t-haNK (NK-92 derived)	PD-L1	Pancreatic Cancer	Phase 2, Recruiting
NCT04847466	t-haNK (NK-92 derived)	PD-L1	Gastroesophageal Junction (GEJ) Cancers, Advanced HNSCC	Phase 2, Recruiting
NCT02839954	Undisclosed	MUC1	MUC1+ Solid Tumors	Phase 1/2, Recruiting
NCT03940820	Undisclosed	ROBO1	Solid Tumors	Phase 1/2, Recruiting
NCT03931720	Undisclosed	ROBO1	Solid Tumors	Phase 1/2, Recruiting
NCT03941457	Undisclosed	ROBO1	Pancreatic Cancer	Phase 1/2, Recruiting
NCT03415100	Allogeneic Donor	NKG2D	Metastatic Solid Tumors	Phase 1, Recruiting
NCT03383978	NK-92	HER2	Glioblastoma	Phase 1, Recruiting
NCT05213195	Undisclosed	NKG2D	Metastatic Colorectal Cancer	Phase 1, Recruiting
NCT05194709	Undisclosed	5T4	Solid tumors	Phase 1, Recruiting
NCT03692663	iPSC	PSMA	Prostate Cancer	Phase 1, Not yet recruiting
NCT03692637	iPSC	Mesothelin	Ovarian Cancer	Phase 1, Not yet recruiting
NCT04050709	t-haNK (NK-92 derived)	PD-L1	Locally advanced or metastatic solid tumors	Phase 1, Not yet recruiting

NHL, non-Hodgkin lymphoma; CLL, chronic lymphocytic leukemia; AML, acute myelogenous leukemia, HNSCC, head and neck squamous cell carcinoma

Challenges to CAR-NK Therapy and Strategies to Overcome Resistance

Pre-clinical and clinical data support the potential of CAR-NK cell therapies for cancer, however, a number of challenges to their clinical application must be addressed for that potential to be realized. First, the recent recognition of advantages that NK cells have over T cells as CAR engineering vehicles has led to the rapid development of CAR-NK therapeutics. Yet, the CAR constructs currently in use were designed for building CAR-T cells and recent evidence that the location of CAR binding epitopes modulates NK cell ability to bind antigen and become activated suggests such constructs may be suboptimal for application in NK cells.⁶⁵ Designing CAR constructs specifically for NK cell activation and cytotoxicity could greatly improve efficacy. Another challenge is that, in contrast to T cells, NK cells are very sensitive to the freeze-thaw process. The survival rate and cytotoxic capabilities of CAR-NK cells are significantly reduced by thawing procedures, but functional activity may be restored by incubation with interleukin-2 (IL-2).⁶⁶ Identification of ideal cryopreservation and recovery protocols will be necessary before CAR-NK cell therapy is viable as an “off-the-shelf” treatment.

One of the major barriers to CAR-NK cell therapy is the lack of persistence in vivo in the absence of cytokine support. While the lack of long-term proliferation in patients is desirable for safety, it is a limiting factor for efficacy. IL-2 and interleukin-15 (IL-15) are the two most commonly used cytokines to enhance the viability and prolong the persistence of adoptively NK cell therapies. The use of IL-2 is associated with substantial side effects and IL-15, which does not support Treg expansion and has a more favorable toxicity profile, is preferred.⁶⁷ New CAR-NK cell engineering strategies incorporate genes in order to ectopically express cytokines such as IL-15 to improve the persistence of these cells in vivo.⁶⁸ In order to address safety concerns, a suicide gene was also included which was able to quickly eliminate the CAR-modified NK cells. Furthermore, adoptive cellular therapies for solid tumors depend on proper homing and trafficking to the tumor site.⁶⁹ Several studies have examined ways to engineer CAR-NK cells to make them more effective at infiltrating solid tumors. For example, NK cell homing to lymph nodes can be enhanced by the addition of the chemokine receptor CCR7 via trogocytosis.⁷⁰ Similarly, electroporation with mRNA coding for CCR7 could improve NK cell migration towards lymph nodes that express the chemokine CCL19.⁷¹ Additionally, NK cells transduced with a CXCR2-encoding viral vector had augmented infiltration into renal cell carcinoma tumors that expressed CXCR2 cognate ligands.⁷² CXCR1, an interleukin-8 receptor, was found to be downregulated upon ex-vivo expansion of primary NK cells sourced from peripheral blood. After restoring expression through mRNA transfection, NK cells expressing an NKG2D CAR showed enhanced migration to and infiltration of solid tumors in mice with ovarian cancer xenografts.⁷³ CAR-NK cells targeting EGFRvIII modified to overexpress the chemokine receptor CXCR4 showed greater tumor infiltration of glioblastoma secreting CXCL12/SDF-1α and likely enhanced migration to these tumors as well.⁷⁴ Innovative strategies to overcome these challenges to CAR-NK cell therapy are being explored in numerous pre-clinical studies with planned evaluation of efficacy in upcoming clinical trials.

In addition to the above limitations of CAR-NK therapy, CAR-NK cells must also overcome tumor cell-based resistance mechanisms to NK cells. Resistance to NK cell therapy is complex and involves a multitude of factors that may be unique to individual patients.⁷⁵ The addition of both first and second-generation CAR constructs targeting CD19 to NK cells has been shown to help overcome resistance to NK-killing in B-cell malignancies.^76,77 These results demonstrate that the addition of CAR constructs alone may help overcome resistance. More novel CAR-NK cell based strategies aim to specifically tip the balance between NK cell activating and inhibitory receptor signals toward tumor-specific activation. For example, knocking out genes that code for proteins that transmit inhibitory signals may also enhance CAR-NK cell cytotoxicity and help overcome NK-cell resistance. Shp-2 knock-out in YT (NK-like) cells enhanced CAR-mediated cytotoxicity towards human prostate cancer cells resistant to NK-killing.⁷⁸ Deletion of CISH, which codes for cytokine-inducible Src homology 2–containing protein and acts as an important cytokine checkpoint in NK cells, also resulted in greater expression of granzyme B, perforin, as well as other cytotoxicity and activation markers. CAR-NK cells with CISH KO constitutively expressing IL-15 doubled persistence time in vivo, enhanced metabolic fitness, and exhibited enhanced killing of CD19 positive lymphoma cells.⁷⁹ Campana et al. designed a CAR construct that incorporated the two key signaling molecules, DAP10 and CD3ζ, and found markedly enhanced cytotoxic capabilities.⁸⁰ Similarly, integration of the signaling lymphocyte activation molecule-related receptor 2B4 into CARs significantly enhanced NK cell activation and overcame NK cell resistance of target autologous leukemia cells.⁸¹ Additional resistance mechanisms that must be overcome involve the tumor microenvironment (TME). Hypoxia is one feature commonly observed in solid tumors that has a marked effect on NK cell function.⁸² Hypoxia has been shown to downregulate key NK activating receptors such as NKp46, NKp30, and NKG2D, resulting in decreased degranulation capability.⁸³ Immunosuppressive cytokines such as TGF-β, which has been shown to suppress NK cell production of IFN-γ, reduce surface expression of NKp30 and NKG2D, and inhibit NK cell function are also found in the TME.^84–86 Tumor-infiltrating NK cells isolated from patients with breast cancer have shown decreased expression of molecules associated with NK cytotoxicity including CD57, perforin, and granzyme B.⁸⁷ CAR-NK cells can be engineered to render them resistant to the action of TGF-β. One promising strategy is the use of CRISPRCas9 technology to delete the TGF-β receptor 2 gene (TGFβR2) in primary human NK cells before CAR transduction, which renders them resistant to TGF-β mediated immunosuppression and enhanced their effector activity against a model of acute myeloid leukemia, as well as glioblastoma.^88,89 In a similar strategy, CAR-NK cells were modified to express a dominant-negative TGF-β receptor, which binds TGF-β with high affinity but does not transduce downstream signaling, mitigating its suppressive effects and restored NK cell cytotoxicity.⁹⁰ In order to effectively treat malignancies, especially solid tumors, CAR-NK cells must include rational designs to address the limitations of adoptive NK cell therapy as well as known resistance mechanisms to NK cell-based treatments.

Conclusion

Engineering strategies to introduce CARs to immune effector cells have resulted in potent additions to the immunotherapeutic arsenal. CAR-T cell therapy has led to remarkable outcomes in patients with relapsed or refractory hematological malignances and has begun to make inroads in the treatment of solid tumors. Unfortunately, there are a number of limitations to CAR-T therapy that are a barrier to clinical use for the majority of patients. CAR-NK cells are rapidly becoming a next-generation option for cellular immunotherapy. Combining the innate ability of natural killer cells to kill malignant cells with the targeted and enhanced cytotoxicity conveyed by CAR constructs has resulted in a technically feasible, effective, and safe therapy for a variety of cancers and antigen targets. While CAR-NK therapy still faces many challenges, pre-clinical studies of CAR-NK cells have provided a roadmap for the future development of engineering strategies to overcome these issues and increase clinical efficacy (Table 2). Tailoring CAR structures to NK cell biology will maximize their cytotoxic potential. In addition, optimization of protocols for the expansion and activation of NK cells for transfusion in patients will allow for the use of CAR-NK cells as an “off-the-shelf” treatment. Finally, additional alterations of NK cells beyond CAR transduction to improve trafficking and overcome the immunosuppressive TME will overcome current challenges. The growing number of CAR-NK cell clinical trials indicates that this promising therapy has only begun to have its transformative potential realized.

Table 2.

Current Challenges to CAR-NK Therapy and Potential Solutions

Challenge	Potential Solution
Current CAR constructs not optimal for NK cells	Design CAR constructs that specifically enhance NK cell activation and cytotoxicity. Incorporate activating receptors / signaling molecules such as DAP10, Dap12, CD3ζ, and 2B4 into CAR constructs65
High NK cell sensitivity to freeze-thaw process	Identification of optimal cryopreservation and recovery protocols66
Lack of persistence in vivo without cytokine support	CAR-NK modification for ectopic expression of cytokines such as IL-1567–69
Poor migration to and infiltration of solid tumors	CAR-NK modification to express chemokine receptors such as CCR7, CXCR1, CXCR2, CXCR470–74
Tumor cell-based resistance to NK cell-mediated lysis	Knock out genes that code for NK inhibitory receptors / proteins that play a role in inhibitory pathways such as CISH and Shp-278–79
The immunosuppressive nature of the TME	Mitigate the effects of immunosuppressive cytokines such as TGF-β through receptor deletion88–89

Highlights.

• Natural Killer cells play a major role anti-tumor immune responses

• Chimeric antigen receptor engineering has led to effective cellular immunotherapies

• Natural Killer cells have several advantages over T cells for CAR-based therapy

• Challenges to CAR-NK cells include trafficking and an immunosuppressive environment

• CAR-NK cells are rapidly becoming a next-generation option for cellular therapy