When the first anticancer therapies based on engineered T cells hit the market a few years ago, they offered the possibility of what would have once been perceived as a medical miracle: a one-shot cure for certain blood cancers. Chimeric antigen receptor (CAR) T cell therapies, as they are known, involve harnessing the patient’s own immune cells, genetically modifying them with cancer-specific receptors for maximum potency against cancerous cells, then reinjecting them into the patient. But for all that cancer-fighting ability, CAR T cells come at a cost: potentially severe side effects, massive price tags, and slow manufacture.

Now a new cell therapy for cancer is edging into the spotlight. Natural killer (NK) cells have potential as a cellular anticancer therapy that could be significantly safer, cheaper, and faster, researchers say.

So far, NK cell therapies haven’t shown any of the...

So far, NK cell therapies haven’t shown any of the significant toxicities that plague CAR T cell therapies.

While T cells are part of the adaptive immune system—they are primed to recognize a specific threat by the immune proteins (antigens) on a foreign cell surface—NK cells are part of the innate immune response, meaning that they respond to anything that appears to be non-self. This broad action makes them suitable for use not only as engineered cell therapies, but as unmodified cells administered on their own. 

Both the unmodified and the engineered forms of NK cell treatment are showing promise in early clinical trials in patients with cancer. And so far, they haven’t shown any of the significant toxicities—such as graft-versus-host disease, in which the transplanted cells attack the host as foreign, or cytokine release syndrome, in which immune cells pour out dangerous amounts of inflammatory signaling molecules—that plague CAR T cell therapies. 

“You have these cells that have an innate capability to recognize tumor cells, and they don’t cause graft-versus-host disease, so you could potentially manufacture multiple doses of these cells . . . to treat multiple patients,” says Katy Rezvani, an immunotherapist at the University of Texas MD Anderson Cancer Center. This combination of efficacy, safety, and relative ease of supply is “the holy grail of cell therapy.” 

It is still early days for NK immunotherapies, which now face many of the same challenges that have limited CAR T cell therapies’ broader application, particularly in targeting harder-to-treat cancers such as solid tumors. NK cells also have their own disadvantages compared to their adaptive immune cousins: they don’t last as long in the body, for example, and they don’t proliferate as easily.

But the excitement surrounding experimental NK-based cancer treatments is nevertheless translating into serious commercial interest. Biotechnology company Nkarta last year raised $114 million to take its NK cell therapy into clinical trials, and Celgene has paid a total of $83 million since 2017 in a partnership with Dragonfly Therapeutics for its NK cell programs. Rezvani and her colleagues’ own research, meanwhile, has led to a partnership with the Japanese pharmaceutical company Takeda to take their NK cell work into multicenter clinical trials.

Natural killer cells’ broad immune response

NK cells were first described in the 1970s, when Swedish and British researchers independently discovered a new class of immune cells that didn’t match the features of T cells or B cells, but still laid waste to cancerous cells. Ever since that discovery, scientists have tried to harness the cells’ powers to fight cancer. But it took the more recent development of immune checkpoint inhibitors, which showed that it was possible to unleash and enhance the immune response against cancer, to throw open the door for cell-based immunotherapies.  

Unlike T cells, NK cells don’t need any antigen-specific priming to provide a therapeutic anti­cancer effect.

As it happened, CAR T cell therapy was the first such therapy out of the gate—albeit with a few obstacles. Most notably, T cell treatment currently has to be autologous—only a patient’s own T cells can be used. That’s because of the way in which T cells interact with the human leukocyte antigen (HLA) complex, a group of cell surface proteins that identifies a cell as being part of the self: any change to the HLA complex on the surface of a cell signals to a T cell that the cell is foreign. While this sensitivity makes T cells effective immune defenders, it has potentially deadly consequences for T cell–based therapy, as any mismatch between introduced and host cells can lead to cytokine release syndrome or graft-versus-host disease. 

NK cells, by contrast, are much less choosy about the HLA complex, says Soyoung Oh, a cancer immunologist at the biotechnology company Genentech. Stressed cells, such as those that are malignant or infected with a virus, may have reduced expression of HLA proteins, or, in some cases, none at all. They can also produce stress-related proteins on their surface. Either of these changes triggers NK cells to release two types of proteins that perforate a target cell’s membrane, damage its vital organelles, and induce cell suicide. This broad activity means that, unlike T cells, NK cells don’t need any antigen-specific priming to provide a therapeutic anticancer effect. 

But typically, the mere presence of any HLA complex on a cell surface can be enough to signal NK cells to stand down. Consequently, using another person’s NK cells is less likely to trigger the dangerous immune reaction that an unfamiliar T cell might, says Oh, allowing researchers to envisage mass-producing NK cell therapies as an off-the-shelf product that doesn’t need to be immunologically matched to a patient. 

Unmodified NK cell therapies

Biotech companies are exploring various sources of NK cells. Nkarta, for example, harvests cells straight from the peripheral blood of a donor using a technique called leukapheresis, in which immune cells are separated out from red blood cells. Other companies are looking to umbilical cord blood, which has a more concentrated supply of NK cells and their progenitors than peripheral blood.

Still others are investigating the use of stem cells, which can in principle be differentiated into hematopoietic stem cells (HSCs) that then generate NK cells. Dan Kaufman, a hematologist at the University of California, San Diego, has been using both embryonic stem cells and induced pluripotent stem cells—stem cells derived from normal cells such as skin cells—to generate NK cells. (Kaufman is also collaborating with California-based Fate Therapeutics on its NK cell immunotherapy program.) 

Netherlands-based biotechnology company Glycostem, meanwhile, manufactures its NK cells from hematopoietic stem cells harvested from cord blood. In 2015, Glycostem released results from a Phase 1 clinical trial in which 10 patients with acute myeloid leukemia who had relapsed after chemotherapy were infused with varying doses of the company’s unmodified cord blood–derived NK cells. The study recorded no instances of graft-versus-host disease or cytokine release syndrome, and patients showed significantly better survival compared with historical controls. Two patients even showed evidence of eradication of minimal residual disease—a low but persistent level of leukemic cells after treatment that usually heralds relapse

“For these patients, . . . they have no therapeutic option actually, they just have to wait for the relapse,” says Didier Haguenauer, chief medical officer at Glycostem. After the success of that single-dose study, the company is now launching a clinical trial testing a three-dose course of treatment. “We expect repeat administration to improve the efficacy of the treatment.”

Technologies to make NK cells more effective

Results from trials such as Glycostem’s suggest that simply boosting a patient’s population of unaltered NK cells could be enough to help their immune system overcome a cancer. But many researchers are also looking into mechanisms that might enhance or better engage NK cells’ anticancer actions.

As is the case for T cells, NK cells are regulated and inhibited by immune checkpoints. They are “a natural brake, so that you potentially don’t lead to autoimmune diseases or immuno-pathologies from excessive activation,” says Nicholas Huntington, who directs the Cancer Immunotherapy Laboratory at Monash University in Melbourne and is the cofounder and CSO of oNKo-Innate, an Australia-based biotech developing NK cell therapies. Overriding those checkpoints, which Huntington helped identify for NK cells, could be useful when it comes to targeting cancer. “If we genetically delete these checkpoints, then the natural killer cells remain active and hyperfunctional, and they can eradicate cancer much quicker, can eliminate metastases much quicker than normal natural killer cells,” he says. Huntington and his colleagues are currently researching these molecular switches, and mechanisms that might turn them on or off.


Natural killer (NK) cells can be extracted directly from umbilical cord blood or from the peripheral blood of a donor (1), or generated using stem cells from these or other sources (2). Some biotechs are investigating the use of unmodified NK cells as cancer therapies (3); others are genetically engineering cells to carry chimeric antigen receptors (CARs) and other modifications (4) to make them more effective at targeting and killing cancer cells while sparing healthy tissue.


Other groups are interested in boosting NK cell activation. One receptor in particular, CD16, appears to be the equivalent of a gas pedal for NK cells. “It’s expressed on the NK cell and then once you engage it, it basically will [trigger] killing by the NK cell,” says Genentech’s Oh. Plans to target this receptor with drugs were the subject of a $96 million deal between Genentech and biopharmaceutical company Affimed in 2018. Oh notes that Genentech is additionally investigating whether the cancer-targeting precision of CD16-activated cells might be enhanced through CAR-style engineering involving other protein receptors. 

Rezvani and colleagues are also combining NK cells with CAR technology. They have been engineering cord blood–derived NK cells with a CAR that targets the CD19 antigen—a well-studied molecule characteristic of certain B cell lymphomas. “We wanted to target CD19-positive
malignancies, because that’s where the best results have been published with CAR T cells,” Rezvani says.

To overcome NK cells’ relatively short half-life and poor proliferation in the body, the team also introduced a gene coding for a cell signaling protein, interleukin-15. This molecule is known to encourage NK cells to increase in number and hang around longer than normal. Lastly, they engineered a safety mechanism into the NK cell genome: a so-called suicide switch gene. Already used in some CAR T cell therapies, the switch can be activated with a drug in the event that the cell therapy shows signs of serious toxicity in patients.

The researchers administered these triple-engineered NK cells to 11 patients with relapsed or treatment-resistant lymphomas or leukemias that were positive for the CD19 antigen. According to results published a few months ago, seven patients showed complete remission of disease and one showed partial remission. Most importantly, there were no signs of toxic effects such as graft-versus-host disease or cytokine release syndrome. “The study has been so safe that the [US Food & Drug Administration] has allowed us to move into an outpatient setting,” Rezvani says. The team is now initiating an international multicenter study in partnership with Tokyo-based Takeda.

Gaps in the science on NK function

Despite recent progress toward NK cell therapies, there are still many un-answered questions about how the immune cells function. For example, why does the normal complement of NK cells present in the body fail to automatically eliminate all cancer cells in the first place? One possibility is that cancer cells find ways to evade NK cells’ detection. “There have been some studies that [show that] tumor cells can lose [the] ligands or proteins on the surface that NK cells normally recognize,” says Kaufman. 

In addition to helping the cancer hide from NK cells, this loss of recognizable ligands could potentially play a role in the development of resistance to cell therapies, he adds. “Typically with a one-time dosing of NK cells, the disease does come back eventually; it might be months or years later. Whether that’s due to resistance to the natural killer cells or other mechanisms, I don’t think we really know.”

It’s also still not clear whether NK cells will have better luck than T cells getting into solid tumors—a harder-to-reach environment than blood cancers. Huntington says he thinks it’s possible, particularly if researchers can find the right antigen to target. “I think that is certainly feasible,” agrees Kaufman, “but it will take a little bit more development.” 

Even if researchers do overcome these challenges, Oh says it’s unlikely that NK cell immunotherapy will entirely supplant CAR T therapy. Instead, the technologies might be effective in combination, particularly as there’s emerging evidence that “NK cells could produce other factors to recruit other immune cells that may actually then further potentiate the anti-tumor response,” she says. “I could envision where there may be benefits to using both together.” 

Bianca Nogrady is a freelance science writer based in Sydney, Australia.

Interested in reading more?

Magaizne Cover

Become a Member of

Receive full access to digital editions of The Scientist, as well as TS Digest, feature stories, more than 35 years of archives, and much more!
Already a member?