Introduction

Hematopoietic stem cell transplantation (HSCT) is currently used to treat hematological malignancies as well as blood and bone marrow disorders. Although HSCT has become an established life-saving treatment for many patients, only about half of transplanted patients will survive the procedure. This high mortality is due to major complications such as graft failure, slow immune recovery, opportunistic infections and graft-versus-host disease (GvHD). Therefore, much effort is currently being dedicated to improving the outcome of HSCT and patient quality of life. In order to do so, different types of immunotherapy are currently being developed to tackle the main complications encountered after transplantation.

Umbilical cord blood (UCB) has been increasingly used as a source of hematopoietic stem cells (HSC) for transplantation since it was first used in 1988 to transplant a child with Fanconi anemia. UCB transplantation (UCBT) has some advantages such as less stringent HLA-matching requirements, fast availability of the graft and reduced incidence and severity of GvHD. However, UCBT is also associated with a higher incidence of infection, especially during the first month post-transplant, graft failure, slow engraftment and delayed immune reconstitution. UCB is mainly used as a source of HSC; however, non-HSC present in UCB, such as T cells, regulatory T (Treg) cells, natural killer (NK) cells and mesenchymal stem cells, could be used to treat some of the main complications post-HSCT including UCBT, as well as other diseases, thus optimizing the use of UCB for immunotherapy. For any cell type to be used as a cell therapy, it is a key to identify the best conditions for cell purification and activation as well as cell expansion while getting a better understanding of their actual properties. In this context, we aim to describe the therapies currently developed that used UCB as a cell source, focusing in particular on Treg and NK cells.

Treg cell therapy to treat GvHD

Phenotype and function of Treg cells

Treg cells are involved in the maintenance of immunological self-tolerance and immune homeostasis. They represent 5–10 % of CD4+ T cells in humans [1] and are currently characterized as CD4+ CD25highCD127lowFoxp3high. Treg cells have been shown to inhibit the functions of various immune cells such as CD4+ and CD8+ T cells [2, 3] and NK cells [4]. There are currently many different mechanisms of suppression by which Treg cells can directly inhibit the functions of target cells. These include the release of suppressive cytokines such as interleukin (IL)-10, transforming growth factor β (TGF-β) and IL-35 or by IL-2 deprivation [5–8]; alternatively, Treg cells can also suppress indirectly by interaction with antigen-presenting cells (APC) [9].

Use of Treg cells to control GvHD: clinical studies

In view of their ability to induce tolerance and suppress the functions of effector T cells, Treg cells have been proposed as an adoptive therapy to prevent or modulate GvHD post-HSCT. A few clinical studies have already demonstrated the feasibility and safety of this therapy in patients that received HSCT in comparison with historical controls using mainly peripheral blood (PB) as a cell source for Treg cells (Table 1). Trzonkowski et al. [10] showed for the first time that the infusion of expanded Treg cells could control GvHD allowing withdrawal of steroid treatment. Di Ianni et al. [11] demonstrated that freshly isolated donor Treg cells were able to counteract the potential GvHD that is induced by the infusion of a high number of effector T cells in haploidentical transplanted patients. Interestingly, Martelli et al. [12] reported reduced relapse rates after transfer of donor Treg cells using the same haploidentical HSCT protocol. Edinger and Hoffmann [13] demonstrated no adverse effects of the administration of expanded Treg cells in a cohort of patients at high risk of relapse. Finally, a recent trial showed that the use of expanded Treg cells to modulate chronic GvHD led to reduced immunosuppression in transplanted patients; however, tumors were detected in two patients [14]. In light of these clinical results, although the use of Treg cells to modulate GvHD is very promising, it is clear that more studies are needed in order to really understand the potential impact of a Treg cell therapy on tumor and viral immunity in the context of HSCT.

Table 1 List of clinical trials reporting the use of Treg cells to modulate GvHD in transplanted patients
Full size table

UCB Treg cells as cell therapy for GvHD

UCB is an attractive source of Treg cells as it contains the same frequency of Treg cells as PB [1], and due to the ready availability of UCB from accredited UCB banks, there is the possibility to develop an off-the-shelf Treg cell therapy. The requirement for HLA matching is less stringent when considering UCB; therefore, when developing UCB-derived therapies, matching between 4/6 and 6/6 alleles for CBT can be considered, although which degree of matching is necessary will need to be further evaluated in clinical studies. Another interesting feature of UCB is that it is possible to isolate Treg cells with high purity in just one single step as opposed to a two steps process when considering PB [15]. In addition, UCB Treg cells are naive CD45RA+ cells that exhibit a better capacity to maintain Foxp3 expression, allowing them to maintain robust suppressive capacity and stability after expansion [16]. Interestingly, it has also been reported that UCB Treg cells are more resistant to apoptosis than PB Treg cells, suggesting a potential survival advantage of UCB Treg cells as compared to their PB counterparts [17]. In terms of function, we and others have demonstrated that UCB Treg cells can suppress effector cells in vitro and in vivo [15, 18, 19], while other groups have shown that they display low suppressive capacity [20, 21]. Regardless, as for PB, only very few cells can be isolated from UCB, and therefore most groups have focused on developing a strategy to expand UCB Treg cells for cell therapy [22–24].

One group has already performed two phase I clinical trials using third-party UCB Treg cells (Table 1). Brunstein et al. [22] reported the feasibility and safety of infusing expanded UCB Treg cells in patients that received double UCBT. Reduced acute GvHD was observed, although higher susceptibility to viral reactivation was reported as compared to historical controls [25]. Recently, the same group demonstrated in another trial that UCB Treg cells expanded using KT64/86 APC resulted in low risk of acute GvHD in patients that received double UCBT [23]. These results are encouraging for the use of third-party UCB Treg cells for immunotherapy. However, Treg cells could only be detected in patients for 2 weeks after infusion, suggesting reduced persistence possibly due to exhaustion following in vitro expansion.

Our group has previously shown that Treg cells could be isolated from fresh UCB units using only the marker CD25 and that the isolated cells exhibited suppressive capacity in vitro [15]. However, this method led to variable purity and yield when isolating Treg cells from cryopreserved UCB units. Therefore, within the T-Control consortium (http://www.t-control.info), we developed a new method to isolate Treg cells from UCB that could be applied to fresh or cryopreserved UCB units using the streptamer reversible technology, allowing isolation of Treg cells with good recovery and purity with potential to develop an off-the-shelf Treg cell product. This allows for the selection of UCB units of specific HLA type, permitting matching of the third-party product to the patients. To try to overcome the issue of Treg cell persistence in infused patients, we are planning to use this method in a phase I clinical trial for the production of a minimally manipulated clinical Treg cell product that will be transferred to transplanted patients to control GvHD. For this study, we are planning to select UCB units that will be a 4/6 allele match to the patients and will therefore adopt a similar strategy as Brunstein et al. [22, 23]. In parallel, we are also exploring which conditions are best to expand streptamer-isolated Treg cells from cryopreserved UCB units for cell therapy. Overall, more studies are warranted to thoroughly evaluate the characteristics of third-party UCB Treg cells that are expanded or not, in order to optimize their use as immunotherapy. Currently, we have not considered the cryopreservation of Treg cells prior to infusion as we are isolating Treg cells from already cryopreserved material. It has been previously reported that the infusion of frozen Treg cells had no impact on the incidence of GvHD, probably because the cells could not be detected in patients due to low viability post-thaw [22]. Further preclinical and clinical studies will help to identify the best conditions to activate and expand UCB Treg cells for use in patients.

NK cell therapy to improve engraftment and fight relapse

Phenotype and function of NK cells

NK cells are innate immune lymphocytes able to kill target cells such as infected or tumor cells without prior activation. Human NK cells represent 10–15 % of PB and 15–30 % of UCB lymphocytes [26, 27]. NK cells can be subdivided according to their level of CD56 expression and whether or not they express CD16. They are subdivided into CD56brightCD16− (up to 10 % of the NK cell population in PB and CB) and CD56dimCD16+ NK cells (about 90 % of NK cells in PB and CB) [27]. These subsets differ in functions whereby CD56bright NK cells are cytokine-producing cells and CD56dim NK cells are more cytotoxic [28].

Characteristics of UCB NK cells

UCB is a cell source of interest when considering a NK cell therapy because of the direct availability of the material and of the high prevalence of NK cells in UCB. As with Treg cells, NK cells can be easily isolated from UCB in one single step, purifying cells based on their CD56 expression as UCB contains very few NKT cells. Results from studies characterizing UCB NK cells are conflicting. It has been reported that UCB NK cells were less cytolytic than PB NK cells although UCB NK cells could respond to cytokines such as IL-2 and IL-12 [29] and had a similar phenotype to PB NK cells [30, 31]. However, it was suggested that only a fraction of NK cells were able to react against K562 cells, because of a lower CD2, CD11a, CD18 and DNAM-1 expression in comparison with PB NK cells. Dalle et al. [31] reported that UCB NK cells do not express L-selectin and therefore are not able to kill K562 cells. Finally, other groups attributed the reduced function of UCB NK cells to higher expression of the inhibitory receptors NKG2A/CD94 along with a lower expression of granzyme B [30].

Our group showed that UCB NK cells exhibited an immature phenotype and that they expressed high levels of CD94/NKG2A and low levels of DNAM-1, NKp46, granzyme B, perforin and Fas-ligand, affecting their killing capacity [27]. UCB NK cells expressed high levels of CXCR4, suggesting a preferential homing to the BM. We also found that although UCB NK cells degranulated to a level comparable to PB NK cells in response to the K562 cell line, they were not able to kill K562 cells unless activated. UCB NK cells responded differentially to cytokines as compared to PB NK cells [32]. In particular, we found that IL-2 led to optimum activation and enhanced effector functions of PB NK cells, whereas IL-15 or IL-5 in combination with IL-2 or IL18 should be considered for producing activated and fully functional UCB NK cells.

NK cell therapy as relapse treatment

Clinical trials using autologous NK cells have shown NK cell therapy to be safe and feasible; however, none of these studies could demonstrate efficacy of the therapy [33]. An alternative is the use of allogeneic NK cells. The potential of NK cells as candidates for cancer therapy was shown for the first time in the context of haploidentical transplantation, where NK cell alloreactivity, due to a mismatch between the inhibitory receptors for self-MHC class I on NK cells and MHC class I antigens on recipient cells, was associated with reduced graft failure, reduced GvHD and relapse and improved overall survival [34]. Several groups have therefore focused on optimizing methods for ex vivo isolation, activation and expansion of NK cells, primarily for the prevention and/or treatment of relapsed disease [35, 36]. Numerous phase I/II clinical studies have shown that the use of NK cells to target and eliminate human tumors in vivo (Table 2) to be safe and feasible with some studies demonstrating clinical response for some patients. Most studies have been performed using NK cells isolated from an allogeneic donor, mainly in the context of haploidentical HSCT but also directly to target tumor cells.

Table 2 List of clinical trials reporting the use of donor NK cells to fight relapse in patients with hematological malignancies
Full size table

In terms of NK cell therapy after UCBT, one group has established a robust artificial APC mediated ex vivo expansion of UCB NK cells [37]. Two clinical trials are currently on-going using this method, one testing the feasibility and safety to infuse expanded UCB NK cells prior to UCBT in patients with chronic lymphocytic leukemia (NCT01619761) and the other one testing the infusion of expanded UCB NK cells in conjunction with chemotherapy in patients with myeloma that received autologous HSCT (NCT01729091).

We recently showed that UCB NK cells activated with IL-15 increased migration toward SDF-1α and therefore to the bone marrow as well as increasing their clonogenic capacity and engraftment of UCB HSC in a humanized animal model [38]. Our preliminary data indicate that the co-infusion of IL-15 activated autologous NK cells together with the graft could double the levels of engraftment after UCBT. We are currently testing whether the direct treatment of the UCB graft with IL-15 could improve HSC homing to the bone marrow and increase their clonogenicity in a similar manner as this therapy could lead to improved neutrophil and platelet engraftment in UCBT patients. Such an approach will need to be tested in a first instance in the context of double UCBT, infusing one non-manipulated unit and one treated with IL-15. In this context, our next step will be to test the safety, feasibility and clinical activity of one UCB unit treated with IL-15, in patients undergoing double UCBT, as a way to improve HSC, neutrophil and platelet engraftment.

Producing NK cells from HSC for cell therapy

In order to produce high numbers of NK cells to enable multiple injections, some groups have explored the production of NK cells by differentiation of HSC in vitro. This approach is promising as large number of NK cells can be generated after only a few weeks of culture. Different protocols are currently available using different sources of HSC. Notably, one group has already reported the safety and feasibility of infusing NK cells produced in vitro to transplanted patients [39]. A few groups are using UCB as a source of HSC to produce NK cells for immunotherapy. Spanholtz et al. [40] established a cytokine-based culture system for the expansion of NK cells from fresh or frozen UCB HSC using clinical-grade reagents and are currently testing the feasibility to infuse these generated NK cells in acute myeloid leukemia (AML) patients (NCT01031368). Our group modified a published protocol [41] of differentiation of NK cells in vitro using either mobilized PB, fresh or frozen UCB HSC. We showed that frozen UCB was the HSC source to consider for production of NK cells in vitro because of a higher fold expansion leading to higher NK cell numbers [42]. Notably, we showed that these cells were as functional as PB NK cells with the potential to persist for longer and in higher numbers in vivo. We are currently testing the capacity of these cells to kill AML cells from patients in vitro and in vivo. Interestingly, we also found that NK cells differentiated in vitro can be cryopreserved without any negative effects on their phenotype or function.

Conclusion

Immunotherapy is a promising option for improving the outcome of HSCT, and due to its unique characteristics and off-the-shelf availability, UCB is a source of interest when considering developing immunotherapeutic approaches. The adoptive transfer of Treg cells and NK cells from UCB in particular to control GvHD and relapse, respectively, is already showing promising results. However, more studies are warranted to gain a better insight into how best to use UCB cells.