Clinical Relevance of Immunobiology in Umbilical Cord Blood Transplantation

Umbilical cord blood transplantation (UCBT) has been an important donor source for allogeneic hematopoietic stem cell transplantation, especially for patients who lack suitable matched donors. UCBT provides unique practical advantages, such as lower risks of graft-versus-host-disease (GVHD), permissive HLA mismatch, and ease of procurement. However, there are clinical challenges in UCBT, including high infection rates and treatment-related mortality in selected patient groups. These clinical advantages and challenges are tightly linked with cell-type specific immune reconstitution (IR). Here, we will review IR, focusing on T and NK cells, and the impact of IR on clinical outcomes. Better understanding of the immune biology in UCBT will allow us to further advance this field with improved clinical practice.


Introduction
Since umbilical cord blood transplantation (UCBT) was first implemented for children and adults, it has been a valuable alternative donor source for allogeneic transplantation given its logistic advantages and comparable clinical outcomes to other types of hematopoietic stem cell transplantations (HCTs). Umbilical cord blood (UCB) grafts contain a unique cell composition in lymphocytes, and immune reconstitution (IR) of T and NK cells following UCBT appears to be somewhat different from other donor types. Here, we review immune cell composition in UCB, IR with focus on T and NK cells, and clinical relevance of IR in UCBT outcomes.

Lymphocyte Subsets in the UCB Graft
The immune cell composition and properties of UCB units is different from peripheral blood or bone marrow. Functional and physiological relevance of the lymphocytes in UCB graft can be assessed by comparing to adult peripheral blood (PB) [1]. The absolute numbers of T, B, and NK cells per volume and the fraction of NK and B cells are higher in UCB than PB ( Figure 1) [1]. The median T cells, NK cells, and B cells are 61%, 23%, and 16% in UCB, respectively, whereas 75%, 13%, and 12% in PB. In UCB, there are two lymphocyte populations (CD45 dim and CD45 bright ), whereas in PB, lymphocytes are all CD45 bright population [2]. The CD45 dim lymphocyte population contains higher fractions of B

NK Cells
NK cells are known to be the main effector for graft-versus-leukemia (GVL) reactions early after HCT [17,18], and are enriched comprising up to 30% in the UCB graft ( Figure 1) [1,[19][20][21]. The CD56 bright /CD56 dim ratio of UCB NK cells is similar or slightly higher than that of PB NK cells ( Figure 1) [20,22,23]. Notably, the cytotoxicity of CD56 dim NK subset is poorer than that of CD56 bright NK cells in the UCB graft, whereas CD56 dim NK cells exert stronger cytotoxic effects in PB [22]. Cytotoxicity of CD56 bright subset of UCB NK cells is comparable to that of PB NK cells, whereas cytotoxicity of CD56 dim NK cells of UCB is greatly diminished compared to the counterpart of PB. A conjugate forming assay revealed that binding of CD56 dim NK cells of UCB to leukemic targets was significantly impaired secondary to a lower expression of adhesion molecules, including CD2, CD11a, CD18, and DNAM-1 [22]. In another report, L-selectin, ICAM-1 expression was significantly lower in NK cells of UCB than PB [20,23]. The expression of chemokine receptors in UCB NK cells are different from PB NK cells. CXCR1 expression in UCB CD56 dim NK cells is significantly reduced, whereas CXCR4 expression in both UCB CD56 dim and CD56 bright NK cells is increased [23]. Hence, UCB NK cells may be less responsive to inflammatory stimulation (involved in CXCR1 expression), but better capable of homing to the bone marrow (CXCR4), perhaps accounting for the effective GVL of UCBT [24]. Compared to PB NK cells, UCB NK cells less express maturation markers such as KIR, CD16, and CD57, whereas the expression of NKG2A/CD94, a phenotypic marker of NK cell immaturity, is higher in UCB NK cells [20,21,23]. In terms of NK cell activation receptors, the expression of NKG2C/CD94 and NKp46 (in CD56 bright ) is lower, but TLR-4, GITR, 2B4, and CD48 (in CD56 bright ) expression is higher in UCB NK cells than PB NK cells [2,23].
Immature CD56 − /CD16 + NK cells are a distinct population identified in the UCB graft, but rarely found in healthy adults [18,25,26]. CD56 − /CD16 + NK cells can differentiate into CD56 + /CD16 + NK cells under IL-2 or IL-15 stimulation, acquiring enhanced cytotoxicity [26]. The expansion of UCB NK cells is less responsive to low-dose IL-2 stimulation (200 IU/mL) than PB NK cells. This is likely secondary to lower expression of CD25 (IL-2Rα) in CD56 dim NK cells in UCB [23]. While resting, UCB CD56 dim NK cells express lower levels of CD107a, IFNγ, granzyme B, perforin, and FAS-L, thereby exert a lower cytotoxicity than PB NK cells. UCB NK cells can acquire a potent cytotoxicity with an enhanced production of IFNγ and cytotoxic granules by high-dose IL-2 stimulation (1000 IU/mL) [23]. Furthermore, the response of UCB NK cells to IL-12 and IL-18 stimulation measured by IFNγ production and CD69 expression is higher than that of PB NK cells [27,28].

Immune Reconstitution in UCBT
Immune reconstitution after HCT may impact clinical outcomes such as incidence of transplant-related mortality and GVHD, the risks of infections and relapse, and, ultimately, survival. Numerous factors influence IR after HCT, including conditioning regimen (myeloablative vs. nonmyeloablative, use of antithymocyte globulin, and total body irradiation), immune suppression regimen, the graft source, cell compositions of the graft, and viral infections in HCT [29]. In this section, we will focus on T and NK cell reconstitution after UCBT, and compare with other donor types when possible.

T cells
T cell reconstitution is delayed after UCBT as compared to bone marrow transplantation (BMT) [30] and peripheral blood stem cell transplantation (PBSCT) [31]. T cell reconstitution after HCT occurs in two distinct pathways: (1) Peripheral expansion of mature T cells (thymus-independent pathway), and (2) thymopoiesis from donor hematopoietic progenitors (thymus-dependent pathway) [32,33]. Early after HCT, T cell reconstitution takes place through peripheral expansion by T cells transferred from the graft or recipient T cells which have survived conditioning therapy (thymus-independent pathway). The mature T cells compete for homeostatic cytokines, such as IL-7 or IL-15, and self-MHC molecules presented by antigen-presenting cells (APCs) [34]. In the lymphopenic condition, IL-7 and IL-15 are constantly produced by immune and non-immune cells, but little is consumed. Hence, there are high plasma levels of these cytokines early after HCT [35]. The mature T cells transferred from the graft have greater access to IL-7 or IL-15 and self-MHC on APCs, which, in the context of lymphopenia, promotes expansion with limited competition [34,36]. Naturally, the T cells undergo rapid expansion with multiple cell divisions, leading to accelerated telomere shortening in the first year post-HCT [37]. In contrast to memory T cells, naïve T cells require TCR engagement with MHC molecules presented by APC, in addition to cytokine stimulation, for survival and expansion [34,38]. As a result, peripheral expansion of memory T cells is greater than naïve T cells after HCT. Another unique aspect of early immune reconstitution is the inverted CD4/CD8 ratio, secondary to better peripheral expansion of memory CD8 T cells [29]. In certain donor/recipient pairs, seropositivity to viral pathogens like CMV can polarize T cell expansion towards the viral antigens, narrowing the T cell repertoire that predominantly proliferates towards antigen-specific memory T cells, limiting polyclonal expansion [39]. In UCBT, early T cell reconstitution is primarily dependent on the peripheral expansion and may have limited T cell repertoire due to delayed thymopoiesis (especially in adults), as compared to other donor types [40]. Furthermore, CD4 + and CD8 + T cells at day +100 after UCBT have reduced capability to produce IFNγ upon superantigen and CMV stimulation, possibly indicating impaired T cell functions early after UCBT [41]. In a pediatric study using myeloablative conditioning (MAC) with antithymocyte globulin (ATG), median time to T cell recovery (CD3 + T cells >0.5 × 10 9 /L) was 6.3 months in UCBT group vs. 3.2 months for unrelated BM group (p = 0.008) [30]. This was apparently driven by CD8 + T cell reconstitution (>0.25 × 10 9 /L) that took a median of 7.7 months after UCB vs. 2.8 months after unrelated BM. However, recovery of the CD4 + T cell (CD4 + T cells >0.5 × 10 9 /L) numbers was faster after UCBT, with median time for CD4 + T cell recovery 9.3 months vs. 12 months in unrelated BMT (p = 0.003). Hence, the inverted CD4/CD8 ratio is not observed early after UCBT because of the delayed CD8 + T cell recovery after UCBT [42].
The reconstitution of T cell repertoire diversity from donor-derived naïve T cells occurs in the thymus following peripheral expansion of mature T cells post HCT (thymus-dependent pathway). This process, termed thymopoiesis, requires a longer period of time in which donor-derived lymphoid progenitors enter the thymus and undergo maturation processes (positive and negative selection) [43]. Thymopoiesis occurs weeks after HCT and can last up to 6 years [44]. After the sequential positive and negative selections, only small fractions of T cells can survive and exit the thymus, so-called recent thymus emigrant (RTE) [45]. This de novo process can be measured by T cell receptor (TCR) excision circle (TREC), naïve T cell counts, and T cell repertoire diversity [44,46,47], and is critical for broad and self-tolerant T cell immunity [32,46]. After UCBT, TREC levels correlate with CD3 + CD4 + 45RO − naïve T cell counts (r = 0.83, p = 0.0001), and TCR repertoire diversity (r = 0.83, p = 0.0001) [48]. Long-term T cell reconstitution (CD3 >1.5 × 10 9 /L) is similar between UCBT and unrelated BMT (9.3 vs. 10 months) in the pediatric population [30]. In an age-and GVHD-matched comparison of children and young adults between UCB [median age 12.6 years (3-34.6)] and matched sibling recipients, TREC and CD4 + CD45RO − naïve T cells were significantly higher, whereas CD8 + activated and memory T cells were lower at 2 years in UCB as compared to matched sibling donor group, indicating efficient thymopoiesis in UCBT [48].

NK Cells
Natural killer cells are the first lymphocytes reconstituting after HCT. NK cell immunity plays a critical role in GVL, especially early after UCBT, because of the low absolute counts and functional immaturity of T cells transferred with the UCB graft. The time to NK cell reconstitution (>0.1 × 10 9 /L) was similar between UCBT (1 month) and unrelated BMT (1.4 months), when both groups received ATG as part of the conditioning regimen [30]. Notably, after UCB with no ATG in the conditioning regimen, NK cell count reconstitution at 1 month after UCBT was similar to healthy controls [49,50]. Moreover, a better NK cell reconstitution with higher NK cell counts was observed over a 24-month period in UCBT than PBSCT [31,51].
NK cell reconstitution 1-3 months after UCBT is polarized to CD56 bright NK cells (approximately 40% of the total NK cells), as compared to healthy donor controls [49,50]. Three months after UCBT, NK cells express high levels of NKG2A and CD62L and low levels of CD16, CD8, and CD57 [49]. Even in CD56 dim NK cells, the expression of CD94/NKG2A, an inhibitory receptor recognizing HLA-E antigen, is higher early after UCBT, but gradually returns to levels similar to that of healthy controls' by 1 year after UCBT [50]. The expression of KIR2DL2/3 and KIR3DL1 of NK cells is significantly lower in the UCB graft, but becomes comparable within 3 months after UCBT to healthy donors, indicating acquisition of NK cell education [49,50]. However, KIR2DL1 levels of CD56 dim NK cells are persistently lower than that of healthy controls during the first 6 months after UCBT, consistent with the sequential acquisition of KIR commonly observed in other types of HCT [49,50,52,53]. Interestingly, NKp30, NKp46 (natural cytotoxicity receptors involving NK cell activation), and CD69 (an activation marker) of CD56 dim NK cells are transiently higher for the first couple of months after UCBT than healthy controls [50], potentially providing advantages in GVL reactions. The HLA-DR expression of NK cells is significantly higher during the first year of UCBT than that of healthy controls and UCB grafts [49]. DNAM-1 (an activating NK cell receptor) expression of CD56 dim NK cells is significantly lower in the UCB graft, but gradually increases, and becomes similar to the level of healthy control NK cells within a year after UCBT [50].
NK cells acquire unique functional characteristics after UCBT, as evidenced by high IFNγ production in the first 1-3 months [49,50]. Direct cytotoxicity of NK cells during the first 6 months post-UCBT against K562 targets and HLA mismatched primary acute myeloid leukemia (AML) samples is robust, and similar to that of healthy controls [50]. However, antibody-dependent cellular cytotoxicity (ADCC) of NK cells within 3 months after UCBT is significantly impaired [50], consistent with low expression of CD16 early after UCBT [49].

Clinical Factors Associated with Immune Reconstitution in UCBT
As summarized in Figure 3 208 samples is robust, and similar to that of healthy controls [50]. However, antibody-dependent cellular 209 cytotoxicity (ADCC) of NK cells within 3 months after UCBT is significantly impaired [50], consistent 210 with low expression of CD16 early after UCBT [49].

Viral Infections
High incidence of HHV-6 (up to 70-80%) has been associated with delayed engraftment after UCBT in multiple reports [54,55]. HHV-6 infection can also interfere with T cell reconstitution, both in a thymus-dependent and independent pathways, resulting in dysfunctional T cell population after HCT [56]. A recent retrospective study with time-dependent analysis revealed that high HHV6 viral load (>10 5 copies/mL) was associated with impairment of both CD4 + and CD8 + T cell reconstitution after HCT, including UCBT [57]. Interestingly, HHV6 infection negatively affected reconstitution of naïve, but not effector memory CD4 + T cells. In addition, reactivation of HHV-6 early after UCBT was associated with T cells expressing CD57, NKG2A, and KIR2DL2/3, surface markers of T cell senescence, and hypofunction, and it was associated in inferior clinical outcomes [58]. Other viral infections may result in limited and skewed TCR diversity towards a specific viral antigen, as observed in individuals with Epstein-Barr virus (EBV) and cytomegalovirus (CMV) infections than those without these infections [59].

GVHD
Glucocorticoid (GC) is the cornerstone of treatment for acute and chronic GVHD. GC is known to induce in vivo Treg expansion [60,61], and inhibit the JAK-STAT signaling pathway induced by IL-2, IL-4, IL-7, and IL-15 in T cells [62]. A murine HCT model suggests CD8 + T cells as the main target of GC [63]. Taken together, CG used for treatment of GVHD profoundly affects in the T cell function. GC also prevents upregulation of MHC class II and costimulatory molecules on dendritic cells [64,65], hence, compromises T cell responses against foreign or allo-antigens. High-dose GC can also induce T cell apoptosis [65]. Furthermore, GVHD can directly damage the thymus by inducing apoptosis of thymocytes (thymic GVHD) [66,67]. Among allo-HCT recipients, absolute counts of naïve T cells were significantly lower at 12 months, with narrower and more skewed TCR repertoires in patients with aGVHD than without aGVHD [47,48]. Furthermore, both sjTREC and βTREC were significantly lower in the aGVHD group, but sj/βTREC ratio was comparable between groups with and without aGVHD, indicating that aGVHD primarily impairs early-stage thymopoiesis [47]. In addition, patients with cGVHD developed markedly lower TREC [44,48]. Interestingly, steady increases in TREC levels in both CD4 + and CD8 + T cells were observed in UCBT patients receiving immunosuppression without active GVHD, suggesting that GVHD prophylaxis with immunosuppression alone does not necessarily cease thymopoiesis [44].

Conditioning Regimen
ATG has been frequently used in the conditioning regimen in UCBT. ATG delays T cell reconstitution by depleting naïve and memory T cells transferred with the graft required for early peripheral expansion of mature T cells [68]. In contrast, ATG exposure is associated with strong recovery of B and NK cells 30 days after transplant, enabling B and NK cells to compensate T cell defects in UCBT [41,43]. In addition, the timing, dose of ATG administration, and serum level at the time of the allograft infusion may influence incidence and grade of GVHD and T cell subset reconstitution [68,69]. In the absence of exposure to ATG, better T cell reconstitution after UCBT is observed, which may contribute, at least in part, to better leukemia control and lower all-cause mortality [68,70,71].
Reduced intensity regimen (RIC, Fludarabine 30 mg/m 2 for 5 days, Cyclophosphamide 50 mg/kg, and TBI 200cGy) without ATG in adults employed at the University of Minnesota was associated with comparable lymphoid reconstitution at day 180 post-HCT and significantly lower chronic GVHD at 1 year post-HCT in UCBT compared to matched sibling donor HCT. In a pediatric UCBT study where all patients received ATG in the conditioning [72], NK cell counts after transplant were higher in the MAC group. Multivariate analysis revealed that the MAC group had a higher risk of developing acute GVHD (HR 6.1, p = 0.002), increased treatment-related mortality (TRM) (OR 26.8, p = 0.008), and overall mortality (HR = 4.1, p = 0.0001). In another adult study [73], CD3 + T cell recovery was observed at 6-12 months after UCBT with higher number of CD45RA + T cells, more diverse T cell repertoire in patients treated with nonmyeloablative regimen (NMA) (fludarabine 30 mg/m 2 , cyclophosphamide 500 mg/m 2 , and ATG 30 mg/kg) when compared to a historical MA group [40]. Furthermore, TREC was detected at 12 months in NMA group, whereas at 18-24 months in the MA group [40,73]. Taken together, RIC may provide better T cell reconstitution, whereas MAC may favor NK cell reconstitution.

Age
Age is one of the most significant host factors to influence T cell reconstitution and thymopoiesis following UCBT. In adult patients with TBI-and ATG-based conditioning regimen, the number of CD8 + T cells reached normal ranges a year after UCBT, but total T cell counts remained below normal for 2 years from UCBT, and memory T cells remained 70% of the total T cell population until 12 months post-UCBT [40,41]. These data suggest that T cell reconstitution and thymopoiesis are delayed in adult UCBT. In contrast, the long-term immune recovery, including after UCBT, was similar to those who underwent adult unrelated donor [30,74] and haploientical donor HCT [75] in children, suggesting faster thymus-dependent T cell reconstitution. In a pediatric population, TREC numbers recovered to the pre-UCBT levels by 6 months [75], and reached within normal limits by 1 year [40,45], whereas in adults, sjTREC recovery took a median of 3 years after UCBT [40,41]. Furthermore, diverse T cell repertoires were observed at 1-2 years post-UCBT in children, whereas it took 3-4 years in adult patients [40]. Hence, in part, driven by reduced thymopoiesis, the delay of T cell numbers and repertoire reconstitution is a challenge in adults undergoing UCBT.

Cell Dose: CD34 + Progenitor Counts in Grafts and Single vs. Double Unit(s) of UCBT
CD34 + count (>10 7 /kg vs. <10 7 /kg) in the allograft correlates with the TREC levels after HCT in children, suggesting that CD34 + cell dose plays a role in thymopoiesis [76]. However, despite a higher combined CD34 cell dose in children receiving double (dUCBT) UCB grafts, reconstitution of lymphocyte subsets was similar to that of single UCBT (sUCBT) in children (and young adults) at 1-2 years after UCBT [77,78]. While data directly comparing single with double UCBT in adults are not available, in adult dUCBT without ATG, the TCR diversity measured by TCR deep sequencing at 6 months after HCT approached that observed in a healthy control group [59]. Others reported that, despite administration of ATG, adult patients undergoing dUCBT had a steep rise in TREC numbers between 6 and 12 months after transplant [79], while the recovery of T numbers between 6 and 12 months approached that of recipients of sibling and unrelated donors grafts in the absence of ATG [80,81]. In summary, dUCBT seems to provide better T cell reconstitution in adults. At least in part, this may be explained by the "threshold effect" of CD34 + cell doses in thymopoiesis [76] (i.e., a single UCB graft may contain sufficient CD34 + progenitors to reach the "threshold" of thymopoiesis for pediatric but not for adult recipients). This effect, however, has to be considered in the context of conditioning regimen intensity and administration of ATG.

Clinical Impacts of Immune Reconstitution on Outcomes of UCBT
Reconstitution of lymphocyte subsets after UCBT is influenced by multiple clinical factors, as discussed above. Conversely, immunological and genetic characteristics of T and NK cells are critical determinants of clinical outcomes in HCT. Figure 4 summarizes the potential impact of lymphocyte reconstitution on UCBT outcomes. 324 months in UCBT were due to infection in children [82]. A better T cell reconstitution at day 50 post-325 UCBT, including higher absolute CD4 + T cell counts, was observed in the group without subsequent 326 opportunistic infections (OI) than with OI at day 100 post-UCBT. This finding indicates that T cells 327 play a critical role to prevent opportunistic infection. Additionally, both CD34 + and CD3 + cell doses 328 were associated with lower death from OI at 6 months. In a retrospective study where UCBT 329 recipients constituted more than a half of the entire cohort, delayed CD4 + T cell reconstitution was 330 predictive of adenovirus, EBV, and HHV6 infections [83]. Moreover, CD4 + T cell reconstitution (≥50 331 × 10 6 /L) within 100 days was associated with a shorter duration of adenovirus infection. CMV 332 infection is a major life-threatening complication in HCT. In a large cohort study (n = 332) at the 333 University of Minnesota, 51% of recipients with hematological malignancies undergoing UCBT 334 developed CMV reactivation [84]. CMV-specific CD8 + T cells transferred from the UCB graft alone 335 could not eradicate CMV viremia, but clearance of CMV viremia occurred later and seemed to 336 depend on CD4 + CD45RA + T cells by thymopoiesis [79]. However, the Seattle group identified intact reactivation is likely due to insufficient numbers of these T cells in vivo [85]. Nevertheless, successful 339 T cell reconstitution by thymopoiesis is required for an optimal control of CMV reactivation.

Infections
Infection is a major cause of death in UCBT. Szabolcs et al. reported that 58% of deaths within 6 months in UCBT were due to infection in children [82]. A better T cell reconstitution at day 50 post-UCBT, including higher absolute CD4 + T cell counts, was observed in the group without subsequent opportunistic infections (OI) than with OI at day 100 post-UCBT. This finding indicates that T cells play a critical role to prevent opportunistic infection. Additionally, both CD34 + and CD3 + cell doses were associated with lower death from OI at 6 months. In a retrospective study where UCBT recipients constituted more than a half of the entire cohort, delayed CD4 + T cell reconstitution was predictive of adenovirus, EBV, and HHV6 infections [83]. Moreover, CD4 + T cell reconstitution (≥50 × 10 6 /L) within 100 days was associated with a shorter duration of adenovirus infection. CMV infection is a major life-threatening complication in HCT. In a large cohort study (n = 332) at the University of Minnesota, 51% of recipients with hematological malignancies undergoing UCBT developed CMV reactivation [84]. CMV-specific CD8 + T cells transferred from the UCB graft alone could not eradicate CMV viremia, but clearance of CMV viremia occurred later and seemed to depend on CD4 + CD45RA + T cells by thymopoiesis [79]. However, the Seattle group identified intact CMV-specific T cell priming early after UCBT (at day +42), suggesting that failure to control CMV reactivation is likely due to insufficient numbers of these T cells in vivo [85]. Nevertheless, successful T cell reconstitution by thymopoiesis is required for an optimal control of CMV reactivation. Additionally, NK cells may play an important role in CMV control. A recent study demonstrated that low NKG2C copy number of NK cells in the UCB graft was independently associated with increased risk of developing CMV reactivation (HR = 2.72, p < 0.0001) [86].

Major Clinical Outcomes: Relapse, Mortality, and Survival
Poor T cell reconstitution is associated with increased risks of relapse in HCT. Clave et al. reported that relapsed pediatric patients with hematological malignancy had lower β-TREC levels at 6 months before and after HCT (including UCBT), suggesting an association between impairment of early intra-thymic T cell development and increased relapse risks [75]. Regardless of graft types, both low CD4 + and naïve T cell reconstitution are significantly associated with increased risks of treatment-related mortality at day 100 post-HCT [87]. Successful CD4 + T cell reconstitution at day 100 after UCBT by less exposure to ATG is associated with lower non-relapse mortality (NRM), lower relapse-related mortality (particularly for AML), better event-free survival, and better overall survival (OS) [68]. Furthermore, early CD4 + T cell reconstitution (defined by CD4 + T cell >50 × 10 6 /L within 100 days after UCBT) resulted in better leukemia-free survival (LFS) (HR = 0.24, p = 0.003), improved OS (HR = 0.16, p = 0.0014) with lower NRM (HR = 0.20, p = 0.0072). Again, lower RI (HR = 0.31, p = 0.041) in association with improved T cell reconstitution was observed in AML [88]. Successful CD8 + T cell reconstitution and high TREC levels, as well as CD4 + T cell reconstitution, are associated with improved OS [79]. Moreover, CMV-specific T cell response and high NK cell counts are independently associated with better progression-free survival (PFS). As T cells in the UCB graft are unlikely to have encountered antigens of herpes viruses, herpes antigen-specific T cell response represents T cell reconstitution occurring in vivo. Hence, Parkman et al. measured herpes antigen-specific proliferative T cell responses as measurement of successful T cell IR [89]. In pediatric patients with acute leukemias undergoing ATG-based MA sUCBT, the earliest herpes-antigen specific T cell response was observed within the first month. Notably, negative antigen-specific T cell response was independently associated with higher leukemia relapse (HR = 3.7, p = 0.003) and lower relapse-free survival (HR = 3.6, p = 0.0002), indicating that successful T cell reconstitution plays a critical role in relapse prevention.
Transplant outcomes in UCBT are also tightly associated with NK cell IR. In RIC UCBT for AML, the low CD16 and high HLA-DR expression on NK cells are significantly associated with increased risks of TRM [49]. In addition, KIR-HLA typing is associated with overall survival (OS) [49]. HLA C2 homozygous recipients have much poorer event-free survival (EFS) (HR = 6.19, p = 0.002), OS (HR = 6.12, p = 0.001), and higher TRM (HR = 9.44, p = 0.026) than HLA C1/x recipients. Furthermore, poor direct cytotoxicity of NK cells against K562 measured by CD107a expression was significantly associated with poor overall survival as well.

Closing Remarks
UCBT has been a valuable alternative donor for transplantation for the past several decades. It has clinical advantages, including readily available grafts, relatively lower incidence of GVHD, and lower disease relapse [4,24,77,90]. While there have been some challenges in UCBT, advances have been made by modifying conditioning regimens [87,91], double UCB grafts [77]. More sophisticated utilization of ATG has substantially improved clinical outcomes [68][69][70]. Emergence of more robust ex vivo expansion techniques have enabled to meet the adequate cell doses for larger patients [92,93]. In addition, a recent report described the powerful GVL effect of UCB in those with acute leukemia and minimal residual disease [24]. Better understanding of the immune biology in UCBT will lead to improved graft engineering in the future.