Before introducing the story of CAR-NK cells, let’s first introduce the foundations of CAR-T cell therapy research.

Background: CAR-T Cells 

What is CAR-T? The full name of CAR here is Chimeric Antigen Receptor, its structure includes extracellular domain, transmembrane domain, and an intracellular signal domain. The extracellular domain contains antibody single-chain variable (VH-linker-VL, scFv) and hinge regions that target the recognition of tumor target antigens. The intracellular signal domain consists of a co-stimulating molecule (CD28 or 4-1BB) and a CD3ζ signal domain. CAR-T cells are T cells armed with CAR: the scFv segment of the CAR molecule can give T cells targetability, and the intracellular signal domain aims to replicate the events to activate T cells. Therefore, compared with T cells, the combat effectiveness of CAR-T cells is significantly enhanced.

As a superstar in adoptive cellular immunotherapy, CAR-T cells cannot perfectly apply to all kinds of diseases. First, CAR-T is a highly personalized, custom therapy “weapon” to fight cancer, but can only treat an individual patient; this means that it is extremely difficult to produce and apply for manufacturing at an effective scale. Second, many cytokines will be released by CAR-T cells when they are targeting and killing the cancer cells, which may over-stimulate immune activity and could cause a cytokine storm and promote inflammation. Third, one of the challenges for CAR-T cells is not only targeting solid tumors, but also avoiding targeting normal tissue and the effects of off-target toxicity. This directly relates to selecting the appropriate target antigen in CAR design to ensure the limitation area.

In order to get over these limitations, many researchers are trying different solutions, such as engineering CAR-T cells (with the CRISPR-Pro system) to generate allergenic universal CAR-T cells by removing TCR and CD52, in order to avoid graft versus host disease (GVDH). This approach can be used to broaden the application range for manufacturing and allogeneic, or universally-applicable, CAR-T cell therapy treatments. At the same time, many researchers have focused their attention on researching the potential applications of CAR on relatives of T cells’- our natural killer (NK) cells.

Natural Killer (
NK) Cells

In 1975, natural killer (NK) cells were first discovered in mice as a subpopulation of lymphocytes. NK cells were subsequently found in humans, and comprise up to 10% of cells in human peripheral blood mononuclear cells (PBMCs). NK cells play several major roles in the innate immune system, such as in microbial defense, virus infections, and control of several types of tumors. NK cells have a wider range for treatment applications, as they can be used for allogeneic treatment. Unlike T cells, NK cells themselves express a variety of activating and inhibitory receptors to recognize and respond to inflamed or infected tissues.

NK cells mainly use three methods to defend and fight against tumors:

  • By releasing cytoplasmic granules containing perforin and granzymes that lead to tumor-cell apoptosis by caspase-dependent and -independent pathways
  • By death receptor-mediated apoptosis or secreting various effector molecules, such as TNF-α and IFN-γ.
  • Through antibody-dependent cellular cytotoxicity (ADCC) by expressing CD16 to destroy tumor cells.

For these reasons, NK cells can be an important weapon to fight against tumors, especially as the NK cells are loaded with "CAR"-like molecules, it can also greatly enhance its combat effectiveness against tumors, and it is also expected to overcome many disadvantages of CAR-T cell therapy.

Similarities and Differences in CAR Molecular Design

So far, most of the CAR molecular structures used in CAR-NK cell research are originally designed for CAR-T cells. Although these CAR molecules can also apply for NK cells, many researchers are still working on further designing and optimizing CAR molecules that are more suitable for NK cells. Table 1 summarizes the most commonly used CAR molecular structures in recent CAR-NK cell therapy research. Studies have shown that compared with CAR containing CD28/4-1BB costimulatory molecules, CAR-NK cells expressing the NK cell-specific costimulatory domain molecule 2B4 have stronger proliferation ability, cytokine secretion ability, and anti-tumor activity; indicating that NK cell-specific costimulatory domain molecules play an important role in enhancing CAR-NK cell therapy performance. With more customized CAR molecular structure design for NK cells, it is believed that the anti-cancer effectiveness of CAR-NK cells will continue to improve.

                                      Table 1: The most recent commonly used CAR molecular structures in CAR-NK cell


Preparation of CAR-NK Cells

Unlike T cells, there is extremely low risk of GVHD caused by NK cells used for therapy, so CAR-NK cells are expected to become a new approach that can be used to develop allogeneic therapies. At the same time, this characteristic also expands the potential sources of viable NK cells. As shown in Figure 1, there are already multiple sources of NK cells that have achieved clinical-grade preparation, including peripheral blood (PB), cord blood (UCB), induced pluripotent stem cells (iPSC), and the NK cell line known as NK-92.

Most recently, human peripheral blood mononuclear cells (PBMCs) are the most important source of NK cells. High purity of NK cells can be easily obtained through NK cell sorting kits, which is great for clinical-grade NK cell preparation. CAR-NK cells prepared from PBMC-derived NK cells are usually dominated by more than 90% CD56dimCD16+ NK cells, which has a strong toxic effect, but weak proliferation ability. Cord blood (CB) is also an important source for NK cells, and just like NK cells derived from PBMCs, NK cells from cord blood can also be enriched by sorting.

Unlike PBMC-derived NK cells, CB-derived NK cells mostly exhibit immature phenotypes and low cytotoxicity. However, given that PBMC and CB-derived CAR-NK cells have different origins, this aspect makes it difficult to standardize. iPSC-derived NK cells are expected to address the difficulty of standardizing CB- and PBMC-derived NK cells. Featuring unlimited passage characteristics and gene editability, iPSCs can be made into CAR-iPSCs, which can be used as a stable source for continuous production of highly homogeneous CAR-NK cells. Notably, both iPSC-derived and CB-derived NK cells have a similar immune phenotype, which are dominated by immature NK cells. 

Figure 1: Source of NK cell and different CAR-NK preparation processes[1]

Research Advances in CAR-NK Cells

CAR-NK cell therapy products are not yet available, and CAR-NK cell therapy research is in the early stages compared to the number of clinical studies of CAR-T cells. However, in recent years, there have been increasing numbers of publications and reports on CAR-NK cell therapy research, and many companies have actively deployed CAR-NK cell therapy development projects. By searching for CAR-NK from, we found that there are currently 35 in-progress or discontinued CAR-NK cell clinical trials, most of them in either phase 1 or 2 of clinical trials.

In terms of target distribution, hematological tumors are mainly concentrated in CD19 (10 items, including one dual target), BCMA (3 items), CD33 (2 items), CD22 (2 items, including one double target); The main therapeutic targets of solid tumors are: NKG2D (4 items), ROBO1 (3 items), mesothelin (1 item), PSMA (1 item).

In terms of the geographical distribution of research, the most clinical trials are currently carried out in China, followed by the United States, and then other countries and regions, as shown in Figure 2.

In terms of efficacy, the therapeutic effect of CAR-NK cells has been initially verified in some research and clinical studies: in a clinical trial of CAR-NK targeting CD19, 7 of 11 patients got treated. The patients achieved complete remission for their tumor disease with no side effects, such as cytokine release syndrome (CRS) or neurotoxicity, during treatment, showing a good application prospect for CAR-NK cells. 


                 Figure 2: Distribution of CAR-NK clinical studies  


Compared with T cells, NK cells have a unique mechanism which does not require pre-activation to stimulate its immune function, and the recognition of the target is not limited by the presence of major histocompatibility (MHC) molecules. Thus, NK cells are suitable for allogeneic cell therapy, or universal patient applications from a single set of standardized cells. In addition, CAR-NK cell therapy has weaker side effects and typically does not cause severe CRS or neurotoxic effects as seen with CAR-T cells. However, CAR-NK cell therapy also faces some challenges, such as the proliferation and purity of NK cells, the choice of CAR gene delivery tools, and delivery efficiency. Since the research on CAR-NK is just beginning, these problems will likely be solved by researchers and CAR-NK cell therapy will quickly make revolutionary progress in the field of oncology and cancer treatments.


[1]Xie G, Dong H, Liang Y, Ham JD, Rizwan R, Chen J. CAR-NK cells: A promising cellular immunotherapy for cancer. EBioMedicine. 2020;59:102975.