Immune regulatory cells in umbilical cord blood: T regulatory cells and mesenchymal stromal cells


  • Jakub Tolar,

    1. Blood and Marrow Transplant Program, Department of Pediatrics and Center for Translational Medicine, University of Minnesota, Minneapolis, MN, USA
    Search for more papers by this author
  • Keli L. Hippen,

    1. Blood and Marrow Transplant Program, Department of Pediatrics and Center for Translational Medicine, University of Minnesota, Minneapolis, MN, USA
    Search for more papers by this author
  • Bruce R. Blazar

    1. Blood and Marrow Transplant Program, Department of Pediatrics and Center for Translational Medicine, University of Minnesota, Minneapolis, MN, USA
    Search for more papers by this author

Dr Jakub Tolar or Dr Bruce R. Blazar, University of Minnesota, MMC 366, 420 Delaware St SE, Minneapolis, MN 55455, USA. E-mail:;


A major goal in haematopoietic cell transplantation (HCT) is to retain the lymphohaematopoietic potential of the cell transfer without its side effects. In addition to the physical injury caused by the conditioning regimen, donor T cells can react to alloantigens of the recipient and cause graft-versus-host disease (GVHD), which accounts for the largest share of morbidity and mortality after HCT. Immune modulator cells, such as regulatory T cells (Tregs) and mesenchymal stromal cells (MSCs) have shown promise in their ability to control GVHD and yet, in preclinical models, preserve the graft-versus-malignancy effect. Initially, MSCs and Tregs have been isolated from adult sources, such as bone marrow or peripheral blood, respectively. More recent studies have indicated that umbilical cord blood (UCB) is a rich source of both cell types. We will review the current data on UCB-derived Tregs and MSCs and their therapeutic implications.

T regulatory cells

To guard against foreign pathogens and tumour cell growth, as well as to prevent aberrant responses to self-antigens, the immune system has evolved to encompass several non-redundant regulatory mechanisms. The focus of this section will be on the biology and clinical applications of UCB CD4+25+ T regulatory cells (Tregs). Sakaguchi et al (1995) first reported that CD25+-depleted CD4 T cells transferred into nude mice resulted in autoimmune disease, which could be reversed by adding Tregs. For more than 10 years, the critical immune modulatory properties of Tregs have been well-described in mice (Asano et al, 1996; Suri-Payer et al, 1998; Takahashi et al, 1998; Thornton & Shevach, 1998). In solid organ transplant settings, tolerance induced by the combined administration of donor-specific transfusions and costimulatory pathway blockade in vivo was dependent upon donor Tregs (Thornton & Shevach, 1998), while Tregs present in the recipient at the time of skin or cardiac allografting were critical to achieving and maintaining allogeneic cell tolerance (reviewed in Wood & Sakaguchi, 2003). Extending the studies by Sakaguchi et al, other investigators have shown that the adoptive transfer of Tregs suppressed pancreatic islet allograft rejection (Davies et al, 1999; Hara et al, 2001; Sanchez-Fueyo et al, 2002).

On the flip side, removal of Tregs from the donor allograft accelerated graft-versus-host disease (GVHD) and solid organ graft rejection (Cohen et al, 2002; Hoffmann et al, 2002; Taylor et al, 2002; Anderson et al, 2004), while freshly isolated or ex vivo expanded donor Treg infusion potently inhibited acute or chronic GVHD (Taylor et al, 2002; Zhao et al, 2008). In addition, donor or host Tregs were able to ameliorate ongoing chronic GVHD (Anderson et al, 2004; Zhao et al, 2008). In sublethally irradiated recipients of T cell-depleted allogeneic bone marrow (BM), host anti-donor alloreactive T cells were able to mediate donor BM graft rejection, which was inhibited by donor Treg infusion (Joffre et al, 2004; Taylor et al, 2004; Hanash & Levy, 2005). Donor Tregs also prevented GVHD-induced thymic atrophy and, in the process thereof, accelerated the time course of T cell immune recovery (Trenado et al, 2003). In contrast to other approaches for regulating adverse T cell alloresponses in hematopoietic stem and progenitor cell transplant recipients, Tregs have been uniformly effective in various rodent models in independent laboratories and thus represent one of the most important advances in immune modulation in the past several decades. Given the striking results in rodent allogeneic transplantation models and the high degree of morbidity and mortality associated with allogeneic cell transplantation, hematopoietic stem and progenitor cell transplant proved to be a reasonable first venue for donor Treg infusion.

Despite compelling preclinical data, clinical applications of donor Treg infusions have not come easily. Treg infusional trials in humans have lagged behind rodents due to the fact that CD4+25++ cells are present in a relatively low frequency population (∼1–2%) in adult peripheral blood (PB). In rodents, a clearly distinguishable population of CD4+ cells with a high antigen density of CD25 (CD4+25++) enables readily achievable separation from the CD4+CD25 population. In contrast, human PB contains an additional population of CD4+ cells that express an intermediate density of CD25 (CD4+25+), do not express Foxp3, and function as previously activated effector cells (Fig 1A). Clinical testing of adult PB Tregs also has been hampered by the availability of good manufacturing production (GMP) reagents for rigorous Treg purification. Foxp3, a helix-loop-helix transcription factor, is expressed in Tregs with potent suppressor cell activity, as well as some activated conventional T cells in humans. Moreover, isolation of FoxP3+ cells requires permeabilization of T cells and hence is incompatible with their viability. Other markers, such as CD27 and CD127 (IL7Ra chain), may denote fractions with the CD4+25++ population enriched for suppressor cell function, though GMP reagents are not yet available for widespread testing.

Figure 1.

 Purification and expansion of regulatory T cells from umbilical cord blood (UCB) and peripheral blood (PB). (A) Tregs were purified from UCB or PB using GMP grade to isolate CD4+25++ Treg cells. Example shown is representative of flow cytometric phenotyping before and after purification, focusing on the increased abundance of CD4+25+ cells present in the initial and purified PB samples. (B) Tregs purified from UCB or PB were expanded in vitro with anti-CD3/28 beads in the presence of high-dose interleukin-2 (300 U/ml) for 17 d (+Rapamycin for PB Treg cultures).

In contrast to PB, UCB T cells are largely naïve and, as such, there is a distinct CD4+CD25bright subset that exists in relatively high frequency compared to PB. Therefore, the likelihood for co-purification of activated or memory CD4+25+ T cells in UCB units is reduced when compared to PB, perhaps as a result of the relatively low exposure of the foetus to environmental pathogens and vaccines as compared to the adult. The skewing of CD4+25+ T cells toward a Treg suppressor cell phenotype versus conventional T cells may explain, at least in part, the relatively low incidence of acute GVHD using UCB compared to similar human leucocyte antigen (HLA)-matched BM grafts. As such, CD4+25++ Tregs can be isolated with a less cumbersome isolation procedure that relies upon the high density of CD25 antigen predominately on Tregs. The final UCB Treg isolation products contain ∼50% CD4/CD25/FoxP3 flow cytometry, consistent with a Treg phenotype.

As with conventional T cells, Tregs require T cell receptor (TCR) ligation and costimulation. These signals can be provided by anti-CD3 and anti-CD28 antibodies, attached to microbeads produced under GMP conditions that can cross-link the TCR and CD28 molecules (Fig 1B). Exogenous interleukin (IL)-2 is critical, as Tregs do not produce a sufficient amount of IL-2 for their own expansion. Using these antibody-coated beads, UCB Tregs can be expanded under GMP conditions by ∼200- to 1000-fold in <3 weeks (Godfrey et al, 2005). In contrast, PB Tregs expand only 50- to 100-fold because Rapamycin must be added to suppress the outgrowth of contaminating activated or memory T cells (Fig 1B). Such polyclonally expanded UCB and PB Tregs are routinely ≥50% CD4+25++Foxp3+ and can potently suppress host and third-party alloresponses in vitro, currently at ratios of 1:16–1:64 of Tregs to conventional T cells when manufactured under GMP conditions. In vivo, UCB Tregs can suppress unrelated donor PB mononuclear cell responses at ratios of 1:1–1:3 in a model of xenogenic GVHD. In our clinical trials at the University of Minnesota, non-myeloablated or myeloablated recipients of two unrelated UCB units have been given standard GVHD prophylaxis (cyclosporine A; mycophenolate mofetil); they also received third-party, HLA partially-matched Tregs (data not shown).

Although CD25bright Tregs are more readily purified from UCB units than PB (Godfrey et al, 2004, 2005), only ∼5–7·5 × 106 Tregs can be isolated from a single frozen UCB unit. Moreover, the numbers of Tregs available for infusion are limited by those that can be isolated from that unit alone, so multiple infusions or use of Tregs for both GVHD prevention and subsequent therapy, if needed, would be problematic. Therefore, alternative expansion procedures have been explored. Toward that end, we have explored the use of a cell-based universal artificial antigen-presenting cell (aAPC) system to provide costimulatory signals for Treg expansion and survival. K562 erythromyeloid leukaemia cells were engineered to stably express CD32 (low-affinity FcgRII). UCB Tregs expanded equally well on these aAPCs loaded with anti-CD3/28 monoclonal antibodies (mAbs) as with anti-CD3/28 mAb-loaded microbeads (Hippen et al, 2008). However, in vivo, xenogeneic GVHD suppression was favoured by cell-based aAPCs, probably due to the provision of ligands present on the former and not the latter. When the human tumour necrosis factor/tumour necrosis factor receptor family costimulatory ligands OX40L or 4-1BBL were co-expressed, such anti-CD3/28 mAb-loaded cell-based aAPCs were more effective than microspheric beads in favouring the expansion of Tregs. These genetically modified aAPCs permitted > 1250-fold expansion of UCB Tregs in <3 weeks without loss of in vitro suppressor cell potency. In vivo adoptive transfer of expanded UCB Tregs indicated that cell-based aAPC expansion cultures were comparable to beads in suppressing xenogeneic GVHD despite very high Treg cell yields. Thus, these studies suggest a novel and more effective strategy for UCB Treg expansion than the currently widely-used bead-based expansion approach. Indeed, aAPC cell-based expansion techniques for Tregs are likely to move into clinical trial expansion practises in the near future.

In preclinical xenogeneic GVHD studies, UCB Tregs that were expanded using anti-CD3/28 microbeads or cell-based aAPCs persisted in the circulation for about 7–11 d. Depending on the conditions used for Treg expansion, there was a correlation between the duration of Treg persistence in the blood and the potency of GVHD inhibition. Therefore, it may be important to ensure that adequate numbers of infused Tregs are present soon after bone marrow transplantation. Improved expansion rates and avoidance of GVHD prophylactics, such as lympholytic (steroids, anti-thymocyte globulin, or CAMPATH-IH) or anti-proliferative agents (e.g., calcineurin inhibitors) that inhibit the IL-2 production needed to drive Treg expansion, will increase the likelihood for successful Treg suppression. Attention as to whether expanded Tregs have retained the capacity to home to secondary lymphoid organs where GVHD is initiated should be taken into account when considering the final product for infusion.

In summary, UCB is a readily accessible source of highly suppressive Tregs. Improvements in isolation approaches, using new discriminatory cell surface antigens that are continually being identified, to enrich for Tregs with more potent suppressor cell capacity expansion procedures by magnetic beads or high speed cell sorting and expansion procedures, including those described above using aAPCs and multiple restimulations with these cells, will facilitate clinical applications in hematopoietic stem and progenitor cell transplant. Only with randomized trials, particularly those conducted with minimal or no use of calcineurin inhibitors, such as cyclosporine A, or, preferably, with the complete absence of immune suppression, will the exact value of UCB Tregs be uncovered.

Mesenchymal stromal cells

Initially, MSCs were reported to be absent from UCB (Wexler et al, 2003), but later investigations showed that UCB MSCs can be isolated and expanded in vitro (Erices et al, 2000; Tondreau et al, 2005; Bieback & Kluter, 2007; Flynn et al, 2007). In a fashion analogous to BM-derived MSCs, spindle-shaped UCB cells attach to tissue culture plastic and differentiate to cells capable of expressing markers of bone, cartilage, and fat. MSCs can be derived from multiple haematopoietic (e.g., BM, UCB, PB) and non-haematopoietic tissues (e.g., fat, liver, muscle) (Zuk et al, 2002; Phinney & Prockop, 2007). These rare cells (<0·01% of total cellular content) are plastic-adherent and have a remarkable capacity to expand rapidly in vitro and still, within several early cell culture passages, maintain the ability to differentiate into a number of mesenchymal cellular phenotypes at a clonal level (Pittenger et al, 1999). The ability of MSCs to form single-cell-derived clones led to their initial definition as colony-forming units-fibroblastic (CFU-F) (Friedenstein et al, 1970). MSCs typically express surface proteins integrin beta 1 (CD29), hyaluronate receptor (CD44), SH-3/SH-4 (CD73), Thy-1 (CD90), endoglin (CD105), and vascular cell adhesion molecule-1 (CD106), while they lack expression of haematopoietic markers, such as monocyte surface protein CD14, hematopoietic stem cell antigen sialomucin CD34, and leucocyte common antigen CD45 (Pittenger et al, 1999; Deans & Moseley, 2000).

The identity and function of MSCs in vivo remains an enigma, even though multiple possibilities have been proposed, such as MSCs being resident cells in the vascular wall (termed pericytes) (Caplan, 2009), or MSCs being parenchymal cells responsible for replenishing and physiological turning-over of adult mesenchymal tissues (Cetrulo, 2006; Weiss & Troyer, 2006).

In a study comparing human MSCs derived from BM, fat, or UCB, UCB MSCs had the lowest frequency of CFU-F but their proliferation rate was the highest of the three (Kern et al, 2006). Comparison of gene-expression signatures of BM MSCs and UCB MSCs revealed dominant osteogenic phenotype in BM MSC while UCB MSC expression was characterized by activation of IL-1 and TNF alpha angiogenic pathways (Panepucci et al, 2004; Flynn et al, 2007). When compared to BM-derived MSCs, UCB MSCs differentiate equally well to osteocytes and adipocytes, but demonstrate less adipogenic potential (Bieback et al, 2004; Kern et al, 2006). Importantly, the expression profiles of proteins in general, and cytokines in particular, in BM and UCB MSCs are very similar (Feldmann et al, 2005; Liu & Hwang, 2005). Taken together with the functional data, it appears that BM and UCB MSCs are more similar than different, and thus their major properties and functional domains will be considered together.

The totality of evidence suggests that MSC cultures are heterogeneous cell populations of uncertain composition, as is evidenced by the multiple terms describing these adherent cell cultures: MSCs, mesenchymal stem cells, marrow stromal cells, and multipotent mesenchymal stromal cells (Phinney & Prockop, 2007). Synthesis of the available evidence also suggests that MSC progeny does not differentiate across germinal boundaries (i.e., into ectodermal and endodermal tissues) as do more immature embryonal stem cells and induced pluripotent stem cells. In addition, there are no definitive human marker panels to date that would allow prospective isolation of MSCs; even MSCs derived from the same tissues are not functionally equivalent. Multiple isolation and expansion protocols exist and even slight differences among them (or even within them, as MSCs derived using the same isolation and expansion technique may differ) result in gene expression and phenotypic changes that make direct comparison of data difficult (Prockop, 2009).

Equally challenging has been extrapolation of the MSCs' in vitro multidifferentiation potential to the in vivo behaviour of MSCs and definition of MSC function in live tissues (da Silva Meirelles et al, 2008). Significant clinical expectations have been associated with three functional aspects of MSCs.

  • 1Tissue repair: In preclinical models of bone, skin, myocardium, kidney, pancreas, and lung injury, MSCs function as reparative cells with a remarkable ability to home to sites of tissue injury and to aid in tissue regeneration (Kunter et al, 2006). Contrary to initial expectations that MSCs would function as a reservoir to replace damaged cells, however, donor MSCs do not usually replace the damaged cells of the recipient (Phinney & Prockop, 2007). Rather, they appear to exert their healing effects by secreting large quantities of tissue mediators in response to injury, by limiting apoptosis, and by recruiting the cells of the recipient to productive repair (Prockop, 2007, 2009). This mechanistically not-yet-understood paracrine effect, whereby regenerative microenvironment is established, is frequently associated with enhanced angiogenesis (Prockop & Olson, 2007; Caplan, 2009). Clinical trials are underway for treatment of the damaged myocardium following an acute myocardial infarction and for the treatment of chronic obstructive pulmonary disease.
  • 2Haematopoietic engraftment support: MSCs are closely physically and functionally associated with blood-forming cells in the BM haematopoietic stem cell niche (Calvi et al, 2003). MSCs are a rich source of growth factors, adhesion molecules, and homing cytokines. The MSCs’ trophic effects (e.g., via stromal derived factor-1 production), their capacity to provide angiogenic support (e.g., via secretion of vascular endothelial growth factor, platelet-derived growth factor, and basic fibroblast growth factor) and perhaps even neurogenic support to the highly vascularized and innervated BM—together with experimental evidence that MSCs and other stromal cells are capable of mediating a modest expansion of hematopoietic cells in co-culture in vitro (Reese et al, 1999; Wang et al, 2004; Robinson et al, 2006)—led to the intriguing possibility that co-infusion of MSCs and hematopoietic cells can shorten time to engraftment and reduce graft failure after haematopoietic cell transplantation (Caplan, 2009). While murine experimentation and small clinical series seemed to confirm this possibility, larger well-controlled clinical trials showed beneficial effects of MSCs in the engraftment of some hematopoietic grafts (e.g., haploidentical transplants) (Ball et al, 2007; Le Blanc et al, 2007) with less clear evidence at this time as to whether MSCs are equally supportive of other grafts (e.g., UCB transplants). It is likely, though, that different trial designs, cell doses, or MSC sources need to be used to optimise or uncover the full potential of MSCs in hematopoietic cell transplantation.
  • 3Immune modulation: MSCs are immune modulatory cells that do not elicit alloreactive lymphocyte proliferative response (Aggarwal & Pittenger, 2005). In fact, MSCs inhibit proliferation of T cells and B cells, and suppress dendritic cells (Beyth et al, 2005; Rasmusson et al, 2005). This function of MSCs has been elegantly demonstrated in clinical trials using MSCs to treat GVHD that is resistant to standard therapy with steroids, calcineurin inhibitors, and anti-thymocyte globulin. Standard-therapy-resistant GVHD is an almost-always lethal complication of HCT. Thus, MSCs have been quickly applied for this purpose in GVHD clinical trials, and an initial phase III trial has completed enrollment (Le Blanc et al, 2004; Ringden et al, 2006). Trials of MSC infusion also are underway for autoimmunity disorder, such as Crohn’s disease and type I diabetes. Almost all the clinical information available, however, is based on the use of BM MSCs, and whether UCB MSCs have similar beneficial effects is unknown.

It is increasingly accepted that the functional complexity of MSCs may be a reflection of the fact that the multiple effects of MSCs are specific to cellular subpopulations concealed in the bulk MSC cultures (Prockop, 2009). Dissection of these specialized MSC populations could illuminate ways to target the MSC therapy to specific injured organs or disease processes. In addition, this could lead to expansion of the MSC uses beyond current indications. For example, systemic MSC infusion may enable cross-correction of soluble protein (e.g., enzyme iduronidase in mucopolysaccharidossis type I, Hurler syndrome) (Koc et al, 2000) or structural protein (e.g., extra cellular matrix collagen 7 in epidermolysis bullosa) (Tolar et al, 2009) deficiency, cross-correction that is at present possible only by using myeloablative haematopoietic cell transplantation. In addition, MSCs can be genetically modified to enhance their tissue repair and cross-correction abilities (Bartholomew et al, 2001; Gnecchi et al, 2005; Egermann et al, 2006). Furthermore, MSCs can be useful in targeting tissues in body sites resistant to correction by haematopoietic cell transfer (such as brain, bone, and heart valves in haematopoietic cell transplantation recipients with Hurler syndrome). Lastly, pluripotency of MSC can be extended beyond mesenchyme either by reprogramming into embryonic stem cell-like, induced pluripotent stem cells (Park et al, 2008) or by isolation of subpopulations of UCB MSC with superior “stemness”: such as unrestricted somatic stem cells (Kogler et al, 2004, 2006), embryonic-like stem cells (McGuckin et al, 2005), and very small embryonic-like cells (Kucia et al, 2006, 2007).

There is tremendous enthusiasm for the application of MSCs to clinical cell therapy and tissue engineering, especially when envisioned as a simple-to-isolate, third-party, versatile, off-the-shelf therapy for diverse congenital, immune, and ischemic medical conditions (Phinney & Prockop, 2007; Burt et al, 2008). Conducting parallel clinical trials with well-defined end points and controls, and gaining insights from mechanistic laboratory research are necessary to make this optimistic vision a reality. Furthermore, MSC therapy can, in theory, lead to significant adverse outcomes, such as immunosuppression and higher risk of infections, and tumourigenesis either in the form of teratomas or sarcomas derived from the infused MSCs or in the form of donor MSC-mediated stimulation of tumour cells in the recipient, or both (Djouad et al, 2003; Rubio et al, 2005; Tolar et al, 2007; Ning et al, 2008). Nevertheless, therapeutic advances in MSC therapy that take advantage of their trophic and immunoregulatory functions can fulfil major unmet needs in tissue regeneration.

Thus, we know with great certainty that both Tregs and MSCs are active in key HCT pathophysiological processes, such as GVHD and tissue injury, resolution of which is likely to favourably impact the overall outcome of HCT. A number of prophylactic and therapeutic approaches are currently under investigation that have the potential to enhance their healing properties separately or, perhaps, in combination. Ultimately, advances in Treg and MSC biology offer great promise in the safer and more effective treatment of a variety of malignant and non-malignant diseases treatable by HCT.