Natural killer (NK) cells are lymphocytes of the innate immune system that exhibit cytotoxicity toward cancer cells and virus-infected cells and have the capacity to produce cytokines such as interferon-γ (IFN-γ) and tumor necrosis factor-α (TNF-α) in response to stimuli. NK cells are defined as CD56⁺CD3⁻ cells and can be divided into two main subsets according to their expression of CD56 and CD16. CD56^dimCD16⁺ NK cells (CD56^dim NK cells) are cytotoxic NK cells capable to mediate direct killing of target cells via exocytosis of granules containing granzyme B and perforin, activation of cell death pathways such as TRAIL or FAS/FAS-L or via antibody-dependent cellular cytotoxicity. CD56^brightCD16^−/low NK cells (CD56^bright NK cells) are the main cytokine-producing NK cells (1). In peripheral blood (PB), up to 90% of NK cells are CD56^dim NK cells while most NK cells are CD56^bright NK cells in lymph nodes.

Natural killer cell functions are regulated by signals delivered through activating and inhibitory receptors. As opposite to T cells, NK cells are “ready to go” and can eliminate target cells without prior stimulation. However, stimulation of NK cells by cytokines leads to NK cell activation and enhanced functions, in particular enhanced cytolytic activity and proliferation. NK cells have long been considered potential candidates for cancer immunotherapy and their versatility makes them attractive cells to explore. Phase I clinical trials showed autologous NK cell therapies to be feasible and safe without adverse effects in patients with breast cancer or non-Hodgkin’s lymphoma; however, these therapies had no or little impact on relapse rates (2). The potential impact of NK cell alloreactivity in hematopoietic stem cell transplantation (HSCT) was suggested by Valiante and Parham (3). The first evidence that allogeneic NK cells could exert strong anti-leukemic activity and impact on the outcome of haploidentical transplantation stems from the study of Ruggeri et al. (4) who reported NK cell alloreactivity against leukemic cells while reducing the risk of graft-versus-host disease (GvHD) in the context of human leukocyte antigen (HLA) mismatch settings. Other trials have showed that allogeneic NK cells alone can target different types of cancers such as acute myeloid leukemia (AML), melanoma, renal cell carcinoma, Hodgkin lymphoma (5), breast and ovarian cancer (6), or refractory lymphoma (7). The same group has shown the importance of NK cell expansion in vivo, which can be accomplished by infusion of interleukin (IL)-2. However, regulatory T cells were also found to compete for this cytokine and beneficial effects on NK cell expansion were observed when regulatory T cells could be depleted (8). Interestingly, other studies also have indicated that NK cell therapy could also be of interest to treat glioma (9) or neuroblastoma (10).

Umbilical cord blood (UCB) has become an established source of hematopoietic stem cells (11) for transplantation. Advantages for the use of UCB include low risk of viral transmission from donor to recipient, rapid availability of UCB units serving as an immediate “off-the-shelf” product, less stringent requirements for HLA matching, and lower risk of GvHD. However, UCB contains between 10- and 100-fold fewer nucleated cells than other sources of HSC, limiting how many cells of interest can be retrieved from one UCB unit. Interestingly, NK cells are the first lymphocytes to recover after HSCT including after umbilical cord blood transplantation (UCBT) (12). In addition, NK cells are key effectors of the graft-versus-leukemia (GvL) effect. Especially after UCBT, as T cell immune reconstitution is delayed and there is no increased incidence of relapse, it is likely that NK cells are actually the main effectors of the GvL effect in the first year post-UCBT. However, UCB also contains different types of immune cells including NK cells and as we learn more about their specific characteristics, we will identify the conditions which might benefit of an UCB NK cell therapy. This review focuses on providing an overview of the characteristics of UCB NK cells compared to NK cells from PB and explain how they could be used as a cell therapy to cancer.

Natural killer cells constitute up to 10% of lymphocytes in PB and up to 30% in UCB (13, 14), and both CD56^dim NK cells and CD56^bright NK cells can be found in PB and UCB with some groups reporting similar proportions of both subsets or higher frequency of CD56^bright NK cells in UCB (14–16). Regarding the phenotype and functions of UCB NK cells, some groups have identified differences when compared to PB NK cells while others found them to be similar to PB NK cells (17) (Figure 1).

A number of studies have recently explored different platforms to expand UCB NK cells. Increased NK cell numbers can be achieved either by large-scale expansion techniques using artificial antigen-presenting cell (aAPC) or cytokines including IL-2, IL-15, and/or FLT-3 ligand. One such strategy employed to expand purified UCB-derived NK cells on a large scale has been reported using good manufacturing practice (GMP)-grade K562-based aAPCs expressing membrane-bound IL-21 (30). Shah and colleagues have shown that following 14 days of culture in a gas permeable culture system, a 2,389-mean fold expansion of NK cells derived from frozen UCB was achieved. The expanded NK cells presented >95% purity of CD56⁺CD3⁻ NK cells and displayed efficient killing capacity against multiple myeloma in vitro and in vivo, highlighting the use of aAPCs as an attractive approach to generate large numbers of functionally competent UCB NK cells. A further strategy to evaluate the potential use of expanded NK cells was reported by using aAPCs in the form of genetically modified K562 cells expressing membrane-bound IL-15 and 41BBL (31). The aAPCs were cultured with CB mononuclear cells for 7 days, which led to the generation of expanded UCB NK cells that displayed increased expression of NK cell activating receptors, increased perforin and granzyme expression, and increased cytotoxicity against B-cell non-Hodgkin lymphoma in vitro and in vivo. The study merits the use of expanded NK cells for adoptive cellular therapy specifically to target relapse or refractory disease after UCBT. Finally, the use of irradiated Epstein–Barr virus-transformed lymphoblastoid cell lines and IL-2 was also recently reported to generate large numbers of CD56⁺ NK cells derived frozen UCB (32). The generated NK cells exhibited higher levels of cytotoxicity against K562 leukemic cells than expanded PB-derived NK cells (32). The unique advantage of this platform is that only 1 ml of the UCB unit is selectively used to generate expanded NK cells for adoptive therapy and the remaining UCB from the same unit can be cryopreserved and used for future transplantation. It would be interesting to assess whether the use of the same UCB for early NK cell adoptive therapy and transplantation can help to prevent relapse and augment GvL post-UCBT.

Natural killer cells can be directly isolated from PB and UCB but an alternative to these cell sources is the differentiation of NK cells from HSC as a way to generate high numbers of cells (33). NK cells can be differentiated from CD34⁺ cells from the bone marrow, from embryonic stem cells, mobilized PB, or UCB CD34⁺ cells. The expansion of NK cells derived from both fresh and frozen UCB CD34⁺ cells using a cocktail of cytokines in a culture system has also been described as an efficient system to generate large numbers of NK cells. We and others have reported the characteristics of NK cells produced in vitro from UCB CD34⁺ cells (34–36). These cells are mostly similar to PB NK cells with the exception that they express low levels of inhibitory receptors. However, NK cells produced in such a way have been shown to be functional, able to kill leukemic cell lines and patient cells in vitro and in vivo and produce cytokines in response to diverse stimuli (34, 36–38). Interestingly, NK cells produced in vitro have been shown to expand to high numbers while preserving their phenotype and functions after cryopreservation (39). Thus, frozen UCB CD34⁺ cells were found to be the best source of NK cells when compared to fresh UCB-derived CD34⁺ cells and frozen PB CD34⁺ cells and could therefore be a readily available off-the-shelf product for NK cell immunotherapy.

Natural killer cells can be isolated from UCB based on CD56 purification methods. One step isolation method can be used in UCB as opposite to PB where two steps are needed in order to eliminate NKT cells. This is not necessary when considering UCB as it contains a very low percentage of that cell subset. In addition, UCB has the advantage of being readily available as UCB is cryopreserved and can be obtained from accredited UCB banks. Therefore, a NK cell product derived from UCB has the potential to be off-the-shelf. Another advantage of UCB is that HLA is less stringent, although it is not clear what level of matching will be necessary to develop a third party NK cell product from UCB. However, because of the limited volume of blood collected from the umbilical cord there are only a limited number of NK cells that can be isolated from UCB. In addition, as they are immature and have lower functionality as compared to PB NK cells; taking UCB NK cells to the clinics will require a prior activation/expansion step. Several clinical trials are currently ongoing to evaluate the safety and feasibility of UCB NK cells as an “off the shelf product” in transplant and non-transplant settings (Table 1). GMP grade expansion methods for UCB NK cells are currently available as previously described. Notably, only a handful of clinical trials are currently ongoing and recruiting patients using the latest method to expand UCB NK cells to reach the cell dose required. Two clinical phase I studies aim to use expanded UCB NK cells for the treatment of patients with chronic lymphocytic leukemia (NCT01619761, NCT02280525), while another aims to evaluate NK cell therapy in the context of autologous HSCT for patients with myeloma (NCT01729091).

Only a few groups have focused on developing cell therapy approaches based on the differentiation of NK cells from HSC in vitro. However, NK cells produced in vitro have been shown to be safe and their use feasible when considered in the context of allogeneic HSCT (62). In addition, another trial, oNKord^®, is currently ongoing testing the use of NK cells produced in vitro from UCB CD34⁺ cells in patients with AML (EudraCT number 2010-018988-41).

Immunotherapy is a promising treatment for different types of cancer allowing the possibility of personalized medicine for each cancer patient. UCB provides distinct advantages and is an increasingly attractive source for HSCT and cellular therapy. Despite low NK cell numbers within a single UCB unit and their immature phenotype, strategies to expand UCB NK cells using aAPCs or cytokines and feeder cells are paving the way for NK cell adoptive immunotherapy. NK cells have shown great potential in eliminating different types of cancer cells in vitro and in animal models. A few clinical trials are currently underway to evaluate the safety and feasibility of using UCB NK cells as an “off the shelf” product for the prevention of relapse. The results from these studies will help in understanding how to maximize the beneficial potential of UCB NK cells for the treatment of hematological malignancies and solid tumors.

All the authors contributed to writing and reviewing the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.