Umbilical Cord Blood Hematopoietic Stem and Progenitor Cell Expansion for Therapeutic Use Umbilical Cord Blood Hematopoietic Stem and Progenitor Cell Expansion for Therapeutic Use

Hematopoietic stem cell transplantation (HSCT) is a potentially curative therapy for severe hematological malignancies and other severe disorders of the blood, immune system, and bone marrow. It is the most successful regenerative therapy to date, with 2013 marking the one millionth HSCT and the 25th anniversary of the first umbilical cord blood (UCB) HSCT. UCB has most often been used for allogeneic HSCT when a matched bone marrow or peripheral blood donor is unavailable. Recently, novel genome editing technologies to correct inherited gene disorders or to modulate biomarkers/ receptors on HSC and the potential use of HSCT for a variety of other nonmalignant conditions have led to a surge of interest in autologous HSCT, with the HSC source depending on the condition to be treated. UCB HSCs may be used to generate red blood cells, granulocytes, or platelets ex vivo for transfusion into difficult-to-transfuse patients. Alternatively, UCB may be reprogrammed to induced pluripotent stem cells or used to generate cell lines, which can then be differentiated into different cell lineages for transfusion or used as diagnostic reagents. Disadvantages of UCB are its restricted cell numbers and delayed hematological engraftment. Here, UCB HSC expansion/ manipulation ex vivo and clinical applications are addressed. Adhesion Molecule-A is highly


Introduction
Hematopoietic stem cell transplantation (HSCT) is a potentially curative therapy for severe hematological malignancies and other severe disorders of the blood, immune system, and bone marrow. Its usage has increased rapidly and substantially since the first HSCT in 1957, almost 60 years ago, and with the one millionth HSCT being performed in 2013 [1]. Over 40,000 HSCT were performed in Europe in 2014 for more than 36,000 individuals for all disease indications, with 58% being autologous and the remainder receiving HSCs sourced from related and unrelated allogeneic donors of bone marrow (BM), peripheral blood after induced mobilization (mPB) or umbilical cord blood (UCB) [2]. The main indications for these HSCTs were leukemias, lymphoid neoplasias, solid tumors, and nonmalignant diseases [2].
The year 2013 also marked the 25th anniversary of the first UCB HSCT in a child with Fanconi's anemia using an human leukocyte antigen (HLA)-identical sibling UCB donation [3]. The recipient remains alive and well, having achieved full donor chimaerism. Worldwide, there are estimated to be over 730,000 unrelated UCB units banked for public use in more than 160 international cord blood banks, and more than 35,000 have been transplanted [3,4]. In 2014, UCB accounted for 2% of allogeneic HSCT in Europe, with substantially more in the USA and Japan [2]. From available worldwide data, the Center for International Blood and Bone Marrow Research (CIBMTR) reported that UCB accounted for 8% of allogeneic HSCTs [2][3][4][5][6]. In their international survey report of 2015, the CIBMTR recorded that 32% of pediatric and 10% of adult-unrelated donor HSCTs used UCB as the HSC source [6]. Indications for allogeneic UCB HSCT have included hematological malignancies, bone marrow failure, severe anemias, metabolic storage diseases, immune-deficiencies, and some cancers. Studies in December 2014 estimated that there were over 4 million UCB units banked for private/family use (see [4]) of which at least 1015 (530 autologous and 485 allogeneic) had been transplanted by December 2013 (see [4]). Indications for family UCB HSCTs have included the severe hemoglobinopathies and the treatment of brain injury (see [4]), and these and other indications will be discussed further.
UCB has several advantages over bone marrow or mobilized peripheral blood as the HSCT cell source. It is noninvasive, can be collected, tested, HLA-typed and banked ahead of use, is readily available for urgent HSCTs, and for black and minority ethnic recipients who do not have a matched bone marrow or peripheral blood donor, demonstrates less-associated graft versus host disease (GvHD) in the allogeneic setting, particularly where less rigorous HLA matching is possible and is not subject to donor attrition (reviewed in Refs. [7-10]). Its main disadvantages are its limited cell dose, delayed hematological engraftment (early neutrophil and platelet and long term immune reconstitution), lack of additional donor lymphocytes for infusion and interbank variability in viable hematopoietic stem and progenitor cell (HSPC) content on product release. It may also be subject to changing accreditation and quality control standards between the time of banking and its use, and the UCB donor may have immunological or hematopoietic disorders that are not manifested at the time of donation or during transplant follow-up (reviewed in Refs. [7-10]). UCB HSPC can also be used for other purposes. These include generating adequate supplies of mature blood cells from the HSPCs ex vivo for transfusion into difficult-to-transfuse patients or modulating resistance to specific acquired infections and correcting monogenic gene disorders, in dividing HSC, using new genome editing technologies [11][12][13][14].
A single UCB unit generally has sufficient HSCs for pediatric, but not adult HSCT [11][12][13][14]. Different approaches have been used to enhance the HSPC numbers in UCB grafts so that their use can be extended to adult HSCTs, to develop further uses for banked cord blood units in other transfusion and transplant therapies or as diagnostic or research reagents, or to improve their engraftment in the bone marrow hematopoietic stem cell niche. These include the use of double UCB units for transplant without or with ex vivo manipulation of cells to enhance their HSPC dose or improve engraftment. The readouts following the manipulation and expansion of HSPCs include assessing the levels of defined hematopoietic progenitor cell subsets using specific markers and using functional assays to demonstrate hematopoietic repopulation capacity in vivo in surrogate animal models (see Refs. [15,16]). The majority of HSPCs in human UCB are found in the CD133+ and CD34+ fractions of CD45+ cells [17][18][19][20][21][22][23][24][25][26][27] and these progenitors can be further segregated into or enriched for HSC or their immediate myeloid and lymphoid progeny with the discriminatory marker sets identified by Notta et al. [15].
In this review, we will describe these lineage hierarchies in human UCB, the procedures used to expand or manipulate the engraftment of these HSPCs, and the established and potential clinical uses of both unmanipulated and ex vivo manipulated autologous and allogeneic UCB units.

Indications for hematopoietic stem cell transplants and the use of umbilical cord blood
The indications for which allogeneic and autologous HSCTs are most often used have recently been provided as guidelines [28] from the American Society of Blood and Marrow Transplantation (ASBMT). Table 1 summarizes those for allogeneic HSCT based on the published ASBMT disease categories and recommendations with more specific details discussed in, but without reference to, HSC source. This does not necessarily exclude their use for autologous HSCT and, where this is appropriate, this is also described in Ref. [28]. Allogeneic UCB HSCT may be used to treat blood disorders, cancers including hematological malignancies, and metabolic and immune disorders. Table 2 shows a list of 95 diseases in adult and pediatric patients, which Cord:Use defines as being treatable with UCB HSCT [29]. A potential curative option for β-thalassemia is allogeneic HSCT from an HLA-matched sibling donor with reported disease-free survival of 65% in adults and 88% in children (reviewed in reference [30]). For severe sickle cell disease, similar transplants are reported to result in 85-90% disease-free survival in children [31]. While allografts are usually curative for young patients with an HLA-matched sibling donor, this is not an option for the vast majority of patients with sickle cell disease. HLA-matched sibling directed UCB HSCTs (with or without preimplantation genetic HLA-matching), therefore, can provide curative therapies for children suffering from hemoglobinopathies [32,33]. Better outcomes have also been reported in children transplanted with UCB HSCs in such metabolic disorders as Hurler syndrome, metachromatic leukodystrophy and Krabbe disease, and for congenital bone marrow failure and immunodeficiencies [4,[34][35][36][37][38]. • Inclusion-cell disease (mucolipidosis II)
Graft failure and delayed immune reconstitution in UCB HSCT with myeloablative therapy are not without risk, and only 25% of the patients have a matched sibling donor. Lower UCB HSC engraftment rates have been observed where resistance to engraftment occurs (e.g., hemoglobinopathies, chronic myeloid leukemia, and acquired aplastic anemia, see Ref. [4]). Autologous UCB HSCTs have been less common than allogeneic transplants, but the recent development of novel genome editing technologies opens the way to using this new technology to correct certain inherited or acquired gene disorders in autologous HSCs and sourced from UCB at birth or alternatively from mobilized peripheral blood and bone marrow as appropriate, and to then transplant these cells into the affected individual to correct the disease. HLA matching of these grafts and hence GvHD has not, to date, been a problem for these autologous transplants. However, the use of myeloablative conditioning creates a substantial risk.
Recently, studies in mice suggest that the risk of myeloablative conditioning can be greatly reduced by using CD45-saporin toxin conjugated antibody treatment to make space in the bone marrow for transplanted cells to treat sickle cell disease in the autologous setting without significant adverse effects on graft recovery [39], but this has not been conducted in the human. However, earlier studies using rat CD45 antibodies produced in Cambridge, UK [40,41] have demonstrated the safety and efficacy of an 111 In-labeled CD45 conjugate in bone marrow transplant patients with acute leukemia [42]. This may then provide a safer approach with gene-modified HSCTs for treating the β-globin-associated severe hemoglobinopathies, as well as congenital immunodeficiencies and HIV AIDS. While α-thalassemia affects the production of the α-globin chain in β-thalassemia and sickle cell disease, mutations in the β-globin gene result in absent or reduced β-globin and abnormal hemoglobin structure, respectively [43][44][45]. Importantly, the inherited hemoglobin disorders, the thalassemias, and sickle cell disease constitute the most common monogenic disorders worldwide [43][44][45][46][47]. Around 300,000 children are born with sickle cell disease each year [43][44][45][46][47], and there are around 36.7 million people infected with HIV [48]. These conditions result in a reduced life expectancy and quality of life [48][49][50][51][52][53].
An analysis of European trends to 2014 [2] suggests a peak of allogeneic UCB HSCTs in 2012 and a slight decline since this time paralleling an increase in haploidentical HSCTs when combined with posttransplant cyclophosphamide prophylaxis. Two parallel Phase II clinical trials using haploidentical versus double UCB HSCTs in a reduced intensity conditioning regime setting indicate that the 1-year disease-free survival is similar (see [4]). A comparative trial of these two approaches is currently recruiting. In a recent review, Kurtzberg [54] cites higher relapse rates with haploidentical HSCTs for hematological malignancies in a limited number of studies.  [54] indicate that greater allele-level HLA mismatching of UCB HSCT between donor and recipient in a significant number of patients with hematological malignancy undergoing double UCB HSCTs could protect against disease relapse without affecting engraftment, GvHD, and nonrelapse mortality. In a further recent survey of UCB HSCT in older patients (50 years) presenting principally with acute myeloid leukemia, myelodysplasia, and non-Hodgkin lymphoma in Europe and North America, Rafii et al. [5] confirmed the efficacy of UCB HSCTs with reduced intensity conditioning in these patients. Further studies on donor selection are warranted. However, leukemic relapse is a major cause of mortality in HSCT recipients, and this must be taken into account in donor selection strategies.
Although there is the potential for autologous UCB HSCTs to rise substantially with new technological developments in the treatment of inherited monogenic diseases and acquired immunodeficiencies such as HIV and AIDS, current indications for autologous UCB use are low or are in development. The main use of autologous UCB grafts (82%) has been for brain injury, and as described in Ref. [4], this includes cerebral palsy, ataxia, apraxia, traumatic brain injury, hypoxic ischemic encephalopathy, and periventricular leukomalacia [56,57]. Autologous UCB transplants (7%) have been used in clinical trials to treat type 1 diabetes but responses were transient [4,58,59].

Other uses of UCB
UCB hematopoietic stem and progenitor cells (HSPCs) may also be used to generate red blood cells, granulocytes, or platelets ex vivo for transfusion [11,60]. Alternatively, and although initial studies have used fibroblasts [61][62][63], UCB may be reprogrammed to induced pluripotent stem (iPS) cells [64], which can then be differentiated into different cell lineages, e.g., to generate red blood cells and platelets. As these end cells lack nuclei, they may allay certain safety concerns with respect to iPS cells and tumorigenicity (provided that enucleated cells can survive transplantation). Currently, however, these strategies do not replicate the production of over 3 × 10 11 blood cells that are generated in adults per day. Other cell types may be isolated, used, expanded, or manipulated from UCB or the umbilical cord (UC) to enhance engraftment, to eradicate malignancies, to prevent GvHD, or prevent infections. The strategies to improve UCB engraftment and immune reconstitution are listed in Table 3. This is updated from data presented in [65].

Increasing cell dose
Improved collection and processing of cord blood

Infusion of two cord blood units (double cord blood transplantation) in adults
Ex vivo expansion of cord blood HSC/HPC Infusion of cord blood with third-party donor cells (haploidentical graft)

Improving selection of cord blood units
Enhanced HLA-matching in some clinical settings Detection of donor-specific anti-HLA antibodies

Modifying UCB transplant regimens
Using reduced-intensity conditioning Using T-replete protocols Using CD45-toxin conjugates for autologous UCB HSCT

Assessing the quality and potency of UCB cells
UCB dosages are often 5-10% of those obtained for BM and mPB and as many as 10-20% of UCB HSCTs can result in graft failure [65,66]. The total nucleated cell (TNC) and CD34+ cell counts have most often been used in selecting UCB units for transplantation as a particular threshold dosage of these cells in a graft correlates with better engraftment and better clinical outcomes [3]. Poorer outcomes of HLA-mismatched UCB HSCTs are reported with dosages of CD34+ cells and TNCs of less than 1.7 × 10 5 cells/kg and 2.5 × 10 7 cells/kg recipient body weight, respectively, and much improved outcomes with median TNCs of 10 × 10 7 cells/kg [66][67][68][69].
Although double cord blood transplants increase cell numbers in the graft, the time for engraftment does not increase in comparison with single UCB HSCTs that are appropriately dosed [70].
Because the potency of the unmanipulated UCB unit following cryopreservation has been cited as the most important parameter in predicting engraftment [71], Kurtzberg and colleagues have developed an Apgar Score for enhancing the quality and hence the potency of UCB HSPCs at the time of collection and banking [72]. For Caucasoid babies, they predict the best quality UCBs are likely to be obtained if birth weights are >3500 g, deliveries are between 34 and 38 weeks of gestation, and the UCB units are processed within 10 h of collection. This is because both CD34+ cells and progenitor cell number decrease at and after 40 weeks gestation, even though TNCs increase, and also due to loss in cell viabilities with delays in processing and cryopreservation postcollection [72,73]. It is of note that in the USA, African-American UCB units contain 30% less TNC after processing and therefore, may not meet the criteria for TNC numbers that have been set for Caucasoid donors in some banks [72]. Although there are moves by some cord blood banks to bank UCB units with >1.75 × 10 9 TNCs, Kurtzberg's studies provide a note of caution and suggest that this would limit banking of UCB units collected to 5% of UCBs collected, adding to cost, but more importantly, potentially compromising the potency of UCBs for transplantation since as many as 25% of such units would be predicted to have insufficient progenitors for successful engraftment.
Measuring the HSC content of UCB presents some difficulty as the gold standard is in vivo transplantation over an individual's lifespan or alternatively in surrogate nonhuman primate models, which are likely to be closer to the human situation than in vivo immune-deficient murine models [3,16,74]. However, this may be particularly important in the genome-editing context. Viability of banked UCB units is generally assessed using in vitro colony forming unit (CFU) content. Although this is not a measure of the repopulating HSCs, CFU content correlates with neutrophil engraftment and posttransplant survival and can be performed in 2-3 weeks [75,76]. In human UCB, aldehyde dehydrogenase (ALDH) bright cells as assessed by flow cytometry of viable CD45+CD34+ or CD45+CD133+ cells have been strongly correlated with CFU content and may represent a more rapid surrogate potency assay for predicting at least early neutrophil engraftment [76] provided that the thaw-wash protocol is followed after thawing cryopreserved UCB units. The coexpression of MA6 on CD34+ cells has been reported recently as predictor of platelet recovery [77].  [86] have added further to our understanding of the human hematopoietic lineage and the developmental changes it undergoes from fetal liver to UCB to bone marrow hematopoiesis. Importantly, these investigators developed an in vitro single-cell assay that exclusively assesses myeloid, including erythroid and megakaryocytic, lineage potential of individual CD34+ cells by combining MS-5 stromal cultures with LDL and eight cytokines, stem cell factor, thrombopoietin, FMS-like tyrosine kinase 3 ligand, interleukin 6, interleukin 3, interleukin 11, granulocyte macrophage colony stimulating factor and erythropoietin (SCF, TPO, Flt3L, IL6, IL3, IL11, GM-CSF and Epo respectively) and found that 72% of the enriched UCB CD49f+ HSCs gave rise to such clones [86]. Of these, approximately half formed high proliferative potential (HPP)-CFU [86]. This level of clonogenic potential was not observed in semisolid methocel cultures used to assay CFU content. Results are summarized from those presented by Notta et al [86]. Oligopotent or bipotent myeloid progenitors, as the offspring of HSCs, have previously been categorized into common myeloid progenitors (CMP), granulocyte-macrophage progenitors (GMP), and megakaryocyte-erythroid progenitors (MEP) on the basis of differential expression on CD34+CD38+ cells of CD123 [15,86] or CD135 [87,88] and CD45RA [84]. Using additional biomarkers to those described above or in Ref.
[15], particularly CD110 and CD71, but not lineage markers (used in [15]) as shown in Table 4 [86], has allowed progeny of HSCs to be segregated further (e.g., into 3 MPP, 3 MEP, 3 CMP, and 2 GMP subsets), with the erythroid/ megakaryocytes now being shown to originate from the HSC compartment predominantly and with megakaryocytes emerging from the multipotent HSC and MPP compartments, rather than the CMP subset. This again redefines the lineage relationships in UCB and has implications for expanding distinct HSPC subsets ex vivo.

Improving UCB HSPC grafts with cytokine and small molecule treatments ex vivo
Different approaches to improve the efficacy of UCB units clinically have been and are being taken. Whether UCB is used in the related, unrelated, or autologous HSCT settings described above or to generate blood cells ex vivo, enhancing UCB HSC self-renewal or graft content, improving delayed hematological reconstitution, and improving UCB homing to and engraftment in patient bone marrow niches are important issues to address and some of the approaches are described below.
Clinical trials have been based on double or single UCB HSCTs and are exemplified in the following findings: i. Systemic infusion of stigaliptin, a CD26 inhibitor, aimed at enhancing the HSC homing/engraftment response to the key chemokine CXCL12 by inhibiting CXCL12 degradation. The mechanism of CD26 action has been studied and found not only to degrade CXCL12 but also to truncate a number of other growth factors such as GM-CSF, IL3, M-CSF, EPO, Flt3L, and SCF. Truncation alters their growth factor activity. For example, truncated EPO blocks the activity of the full-length EPO molecule, while truncated GM-CSF binds to its receptor with higher affinity [96,97]. Two multicentre Phase II trials have commenced where Stigaliptin is given orally for 3 days in the myeloablative conditioning, single UCB HSCT setting. Although initial outcomes indicated that oral treatment with Stigaliptin showed a median time to neutrophil engraftment of 21 days, if the UCB unit were red-cell depleted, neutrophil engraftment correlated with DPPIV suppression [98]. This approach does not require ex vivo manipulation of the UCB graft, but dosage of Stigaliptin needs further investigation for this treatment to be more effective [99].

ii.
Priming of UCB grafts with C3a. This complement pathway protein is produced by bone marrow MSCs and interacts with the C3aR on UCB HSPCs to enhance CXCL12mediated migration [100,101]. A Phase I clinical trial involving C3a priming of a one of the two UCB units in the nonmyeloablative conditioning double cord blood setting [101] did not, however, demonstrate preferable neutrophil recovery in the manipulated UCB unit in most cases, although the CD3 content of the graft correlated with engraftment. Lund et al. [102] have reviewed this trial recently and concluded that the study has added value for designing further clinical trials using manipulated and unmanipulated double UCB units.

iv.
Ex vivo treatment of one of the two UCB units using fucosyl transferases, fucT-VI, or fucT-VII, aimed at enhancing HSC entry to the bone marrow niche by fucosylation of homing receptors, e.g., the selectins (ASC101) [75,87,107].
The latter two approaches [83,[102][103][104][105][106][107] require relatively short exposures of UCB cells to fucosyl transferases or dmPGE2 prior to transplant. dmPGE2 was first identified as a potential agent for expanding HSC in a high throughput screen in zebrafish [105]. The initial clinical study on nine patients receiving nonmyeloablative conditioning and double UCB HSCTs (with 1 unit primed with dmPGE2) did not demonstrate improved engraftment [104]. In a subsequent similar clinical trial, but for which the dmPGE2 treatment of the UCB cells was first optimized, improved time to neutrophil recovery (median 17.5 days) was observed when compared with historical controls (21 days) and with the dmPGE2 primed UCB unit engrafting in over 80% of the patients. This has preceded to a Phase II clinical trial [106], and recent studies suggest that dmPGE2 may modulate Wnt signaling in UCB T cells and enhance immune reconstitution posttransplant [103].  [87,107]. However, the fucosylated UCB unit contributed to engraftment in approximately half of the patients. Biologics that specifically modulate homing and engraftment activities of UCB grafts thus warrant further investigation and optimization.

Ex vivo expansion of UCB units prior to infusion
It has been predicted that a fourfold expansion of HSCs in UCB would allow the majority of banked UCB units to be used for single UCB HSCT. Currently, the principal practice is to transplant an unmanipulated UCB unit with an ex vivo expanded UCB unit, with the former generally engrafting longer term and the latter contributing to early neutrophil engraftment. The preferred aim is to move to a single UCB unit where a portion of the UCB unit is expanded and transplanted with the unexpanded portion or to expand a single UCB unit ensuring that the manipulated cells can enhance short-term engraftment without compromising long-term engraftment. Approaches have used cytokines with or without small molecules or MSCs to expand UCB units.

Classical cytokine expansion
Current expansion protocols for UCB HSPCs are still in development and evolving continually with improvements in understanding the bone marrow niches and advances in cell and molecular technologies. However, they have evolved from many studies conducted over the past four decades or more commencing with studies in the 1960s and 1970s on in vitro cultures of murine hematopoietic stem/progenitor cells [108,109] and the identification of monoclonal antibodies [110]  removal of inhibitory factors is also beneficial in promoting human UCB HSPC expansion using a fed-batch approach [111,112]. Our own unpublished studies also indicate that expansion of HSPCs is a multistage process in which exposure to different cytokine combinations over time influences HSC self-renewal and differentiation ex vivo. In the initial clinical trials using cytokine-based expansion, in which part of the graft was unmanipulated and CD34+ cells from part of the same UCB unit were expanded in a limited number of cytokines (e.g., SCF, granulocyte-colony stimulating factor (G-CSF) and megarkaryocyte growth and differentiation factor (MGDF) for 10 days or in a perfused bioreactor with Flt3L, Epo, and GM-CSF-IL3 fusion protein for 12 days) and then both manipulated and unmanipulated cells transplanted, no improvements in neutrophil or platelet recovery were observed [66]. These studies, however, provided the impetus for the identification of new factors for HSPC expansion and for the design of further clinical trials.

Further cytokine and small molecule addition to expand UCB HSPCs
With an improving knowledge of stem cell niches and microenvironments [118,119] and technological advances, numerous factors have been identified that regulate HSPC proliferation and differentiation and some may potentially also control HSC self-renewal. Here, we will restrict our discussion to UCB expansion ex vivo and, as appropriate, discuss clinical applications, while other approaches to generate and assay (in xenograft in vivo models, e.g., in zebrafish) patient-specific HSPCs derived from ES or iPS cells have recently been reviewed and will not be discussed further [118].
Human UCB CD34+CD38-HSPCs cultured in SCF, Flt3L, TPO, IL6, and IL3 and with Fc immobilized DLL1 Notch ligand over several weeks in vitro demonstrate a tenfold increase in CD34+ cells with enhanced repopulating ability in immunodeficient mice [124,125]. Similar to Notch signaling, another key developmental signaling pathway, Wnt, also serves as a potential target for maintaining the HSPC multipotency during ex vivo expansion. Increased early engraftment and chimaerism levels in immunodeficient mice were observed when UCB CD34+ cells were expanded in cytokines supplemented with an inhibitor of glycogen synthase kinase-3β (GSK3β), BIO [141]. Another GSK3β inhibitor CHIR99021 also appears to maintain HSC functionality, at least when tested on murine HSPCs [142]. A higher level of chimaerism was observed 4 months postsecondary HSC transplantation for UCB cells expanded with CHIR99021 and the mTOR inhibitor rapamycin for a week [142]. Thus activating Wnt/β-catenin signaling and inhibiting mTOR may serve to increase UCB HSC numbers in future studies.
Wnt and TGFβ pathways also have opposing roles in regulating the balance between HSC selfrenewal, quiescence, and differentiation [143], and these are controlled by miRNAs. By regulating the balance of these two signaling pathways, the miR-99a/100∼125b tricistronic miRNAs are reported to promote human HSCs expansion and to favor megakaryocytic differentiation [143]. Additionally, miR-126 regulates HSC proliferation and differentiation by targeting PI3K/AKT/mTOR signaling [131]. Regulating these miRNAs may serve as a potential strategy to modulate HSPCs by targeting multiple functional, but opposing, signaling pathways.
p18 [144] as a specific inhibitor of cyclin-dependent kinase (CDK) is a potential direct target of cell cycle regulation. Two small-molecule compounds P18IN003 and P18IN011 were identified in this study as being able to enhance the proliferation of mouse HSC cells and increase the reconstitution to bone marrow by at least threefold. This may have a potential for human UCB HSC expansion.
Fares et al. screened 5289 small molecules for expansion of mPB CD34+CD45RA-cells, and this and subsequent synthesis of derivatives of one compound UM729 led to the identification of UM171, a pyrimidoindole derivative, which expanded human UCB CD34+ cells over 100fold with limited differentiation in 12-day fed-batch cultures supplemented with three essential cytokines, SCF, TPO, and Flt3L [122,145]. Cells expanded with UM171 show improved hematological reconstitution in NSG mice for at least 18 weeks postsecondary transplantation (13-fold higher than DMSO control). By comparing the RNAseq profiling of cells treated with DMSO or UM171 at different concentrations, UM171 suppressed erythroid/ megakaryocyte transcripts. Of note, TMEM183A and PROCR (CD201) encoding genes were found to be significantly upregulated after UM171 treatment. Both TMEM183A and PROCR are cell surface molecules, with the latter being expressed highly on murine HSCs [146]. A direct comparison [122] has been made between the effects of UM171 and the purine derivative SR 1, which was also identified by a high throughput screen on mPB CD34+ cells in the presence of SCF, TPO, Flt3L, and IL6 and which functions as an aryl hydrocarbon receptor antagonist. SR 1 was reported to expand UCB CD34+ and immunodeficient mouse in vivo repopulating cells by 670-and 17-fold, respectively, over 3 weeks of culture [123]. Fares et al. [122] have suggested, however, that SR 1 expands less durable engrafting cells than UM171.
Other small molecules used for HSPC expansion include nicotinamide and TEPA. The vitamin, nicotinamide, generates oxidized nicotinamide adenine dinucleotide (NAD) that regulates the function of the sirtuins (SIRTs). As well as generating oxidized NAD, nicotinamide acts as a specific inhibitor of SIRT-1, and when it is added to human UCB cultures containing SCF, TPO, Flt3L, and IL6 for 3 weeks, then expansion of in vivo (in NOD/SCID mice) repopulating HSCs occurs [129,132]. When UCB CD133+ cells were cultured in TEPA with SCF, TPO, Flt3L, and IL6 for 3 weeks, an 89-fold increase in CD34+ cells was observed, together with increases in in vivo NOD/SCID repopulating cells [132].
Poycomb group (PcG) genes, identified as global epigenetic transcriptional repressors, have been demonstrated to work through negatively regulating Hox genes [147]. Genes in the Hox families appear to be highly expressed in murine long-term repopulating HSCs [148]. This supports the idea that HSPC expansion can be regulated epigenetically. Histone deacetylases (HDACs) are classed as important epigenetic modifiers of the eraser type. There are 11 HDAC family members that function as zinc dependent deacetylases of histone and nonhistone proteins. These are divided into four classes. Class I comprises HDAC1-3 and HDAC8, class II consists of HDAC4-7 and HDAC9-10, class III are the sirtuins SIRT1-7 that require NAD as a cosubstrate for their activity, and class IV comprises HDAC11 [149]. Elizalde et al. [150] demonstrated HDAC3 as a potential target for regulating CD34+ cells. Hoffman and colleagues subsequently tested eight HDAC inhibitors, which inhibit class I and II HDACs and found that VPA had a robust influence in promoting the expansion of human UCB CD34 + C90 + CD184 + CD49f + CD45A − HSCs, which expressed key biomarkers such as CD90, CD49f, and CXCR4 and were ALDH high [121]. To evaluate the repopulating activity of expanded cells, sublethally irradiated NSG mice were injected with cells expanded in cytokines (SCF, IL-3, FtL-3, and TPO) and HDAC inhibitors for 7 days. Assessment of the human CD45+ cell chimaerism 13-14 weeks posttransplant demonstrated higher levels of chimaerism in grafts expanded with VPA plus cytokines (32.2±11.3%) than grafts with cytokines alone (13.2±6.4%).
The former cells showed secondary transplantation activity (measured at 15-16 weeks posttransplant). Additionally, when compared with the uncultured cells, VPA with cytokines generated 36-fold more SCID-repopulating cells (SRCs) ex vivo. Using limiting dilution analyses, VPA-expanded grafts were also found to contain significantly more SRCs (1 in 31) than control primary grafts (1 in 1115) or cytokine alone expanded grafts (1 in 9223). However, although three of eight HDAC inhibitors were more effective in improving HSC expansion, further studies are required to more accurately define the mode of action of these HDAC inhibitors in HSC expansion. As indicated above, it has been suggested that HDAC3 serves as a target for regulating HSC expansion [150], but not all HDAC inhibitors, and particularly not VPA, target HDAC3 only. Further investigations are warranted to determine the mechanism of action by which HDAC inhibitors provide improved UCB HSC expansion.

Clinical trials of cytokine and small molecule expanded UCB HSPCs
In these clinical trials, a second, unmanipulated UCB unit with adequate cell numbers and/or the unmanipulated CD34-or CD133-fraction of the expanded UCB unit are generally cotransplanted to ensure the presence of durable long-term engrafting HSCs. Outcomes have generally reported improved times to early neutrophil engraftment.
Clinical trials which have been or are being progressed include the following:

i.
Notch ligand (DLL1) enhanced expansion of HSC/HPC ex vivo of one of the two UCB units prior to transplant. Delaney et al. [151] expanded CD34+ UCB with SCF, Flt3L, TPO, IL6, IL3, and DLL1 and noted an improved median neutrophil engraftment of 16 days in a double UCB transplant setting. T cells were not expanded, whereas myeloid cells (CD33+, CD14+) from the expanded UCB unit predominated. Over 3 weeks, TNC expansion averaged 562-fold and CD34+ cell expansion averaged 164fold. Longer-term engraftment (of <1 year) was observed for two of the nine patients evaluated, but in one patient, this was not maintained and in the second, the patient died from sepsis at 6 months posttransplant [102].

ii.
Ex vivo expansion of one of the two UCB units prior to transplant using bone marrowderived MSCs from a third party haploidentical family member or purified bone marrow Stro3+ MPCs protected under patent to Mesoblast. For the Mesoblast clinical trial [140], where Stro3+ MSCs were cocultured with UCB CD34+ cells for 2-3 weeks in the presence of SCF, G-CSF, FL, TPO, median TNC, and CD34+ cell expansions were c.12-and 30-fold, respectively. Expansion of the HSPC generated more myeloid cells, while NK cells were preserved. Neutrophil engraftment was enhanced (median 15d) in patients receiving HSCT with myeloablation, and the MSCs were equally efficient if used as an off-the-shelf product or sourced as a haploidentical product. Chimaerism from both expanded and unmanipulated UCB units was 46% on days 21-30; this reduced to 13% at 6 months, and by 1 year, the unmanipulated UCB unit had engrafted. The expanded cells, therefore, contributed to early neutrophil recovery, and the unmanipulated UCB unit to longer-term repopulation.
iii. Tetraethylenepentamide (TEPA) enhanced expansion of HSPC ex vivo of part of an UCB unit prior to transplant of the treated and untreated UCB unit (StemEx, Gamida Cell). In this clinical trial [152], the smaller part of the UCB unit was expanded and transplanted with the remaining unmanipulated fraction of the UCB unit. For expansion, UCB CD133+ cells were cultured in SCF, TPO, Flt3L, IL6, and TEPA for 3 weeks and infused 24 h after the unmanipulated UCB unit. However, the median time to neutrophil and platelet engraftment did not appear to be enhanced.

iv.
Nicotinamide (pyridine-3-carboximide) enhanced expansion of HSC/HPC ex vivo of one of the two UCB units prior to transplant (NiCord, Gamida Cell). Gamida Cell used nicotinamide to expand 1 UCB CD133+ selected unit over 3 weeks. Cells that were CD133-were cryopreserved and subsequently infused with the expanded and nonexpanded UCB units into patients with myeloablative HSCTs for hematological malignancies [129]. Results show that median time to neutrophil engraftment was 13 days. T cell recovery was similar to double UCB transplants, and the Nicord unit generally engrafted as assessed with a median follow-up of 21 months posttransplant (8/11 engrafted with the Nicord unit). There was one nonengraftment and two patients engrafted with the unmanipulated UCB unit. This was the first trial to demonstrate longer-term engraftment of the expanded UCB unit. A new trial has commenced recruiting with planned HSCT of a single Nicord expanded unit. The first two grafts have reported median neutrophil recoveries by day 10/11.  [153] indicate that SR 1 promotes the production of megakaryocyte precursors from CD34+ cells with 90% reaching the proplatelet stage with TPO addition, thereby potentially contributing to the ex vivo production of platelets from normal cells for transfusion.

vi.
A Phase I/II clinical trial (NCT02668315) involving UM171-expanded UCB HSCs produced in a fed-batch culture system is recruiting patients with hematological malignancies from January 2016.
With most of these clinical studies, the common denominator is that it is possible to reduce the time to neutrophil engraftment to 15-17 days and possibly down to 10-11 days. Thus, improved early neutrophil engraftment is possible with current expansion protocols and the key question going forward relates to whether it is possible to maintain or promote long-term hematopoietic engraftment, particularly where genome editing is applied and where there is a need to expand blood cells ex vivo for difficult-to-transfuse patients.

Conclusions
The key considerations in expanding human UCB HSC include (i) cost to health providers of GMP cell products and clinical trials, (ii) the number of clinical trials that UCB recipients can be entered into and alternative protocols for HSCT, (iii) restrictions in the use of some compounds or protocols related to intellectual property rights, (iv) defining which UCB units will engraft long term and why 1 UCB unit will engraft in preference to another, (v) addressing variability in donor cell response, particularly related to platelet engraftment and long-term reconstitution of the expanded UCB unit, (vi) the availability of licensed facilities in which to expand cells, and (vii) the optimal development of the best and cheapest protocols to allow rapid engraftment of neutrophils and platelets, longer-term lymphoid reconstitution, and expansion of the HSCs without their differentiation for long-term hemopoietic reconstitution at affordable costs to the healthcare provider and for transplant recipients.
Clinical trials to date have demonstrated improved homing or retention in the bone marrow niche and improved early neutrophil engraftment. Further research and development is required to regulate the self-renewal of the HSC without significant differentiation in order to facilitate novel genome engineering studies and the development of this into a cost-effective GMP-grade process. Ensuring that the cells can effectively home to and engraft in the bone marrow after this manipulation may require the addition of small molecules (e.g., fucTVI, dmPGE2) for a short period prior to transplant after expansion with specific growth factor cocktails. In the short term, better characterization and refinement of newer and existing UCB products will result in shorter hospital stays and improved prognosis for HSC transplant recipients as to-time-to-hematological-reconstitution shortens, the incidence of graft failure reduces and treatment of residual malignant or cure for acquired or inherited disease improves. Better characterization of UCB self-renewal and differentiation should also ex-pand the treatment choices available and will be a valuable resource to patients in ethnic minority and other groups where it can be difficult to find a matched graft or where graft engineering and genome editing offer the best choice for a cure. In the longer term, a full understanding of the molecular mechanisms that govern HSC commitment and differentiation, the homing of HSCs to bone marrow, and the control, retention, and engraftment of normal HSCs within the specialized bone marrow niches will lead to achievable and cost effective translation of a great deal of research into effective clinical practice for many millions of individuals worldwide.