Cord blood T cells are “completely different”

According to some, one of Britain’s significant contributions to the modern world is the comedy of the Monty Python Flying Circus. One of their films employed the memorable catchphrase, “And now for something completely different… .” Now, decades later, in this issue of Blood, the work of Hiwarkar et al represents another notable English export.¹  This group of investigators has begun to dissect the differences in graft-versus-tumor effects of adult donor peripheral blood (PB) and cord blood (CB) T cells, and the findings are quite remarkable.

For many years, CB transplantation (CBT) clinicians have observed and reported upon the unique nature of CBT biology based on observations on the transplant floor and in the clinic, as well as findings from single-center and registry studies of adult and pediatric CBT. The first remarkable feature of CBT is that the incidence of severe acute graft-versus-host disease (GVHD) is much lower than would be expected based on the high degree of donor-recipient human leukocyte antigen (HLA) mismatch. For example, the median 8 HLA-allele match of CB units transplanted at Memorial Sloan-Kettering Cancer Center (MSKCC) is 5/8 with grafts as mismatched as 3-4/8 frequently being administered to adult patients. And yet, the incidence of severe acute GVHD in CBT recipients is no greater than that of HLA-matched adult donor allografts in many series. Recent MSKCC and University of Minnesota analyses have shown a day 100 grade 2 to 4 acute GVHD incidence of <15% in double-unit CB recipients who were transplanted with adequately dosed mycophenolate mofetil.^2,3  The manifestations of chronic GVHD in CBT recipients are also different from that of adult donor PB allograft recipients with severe chronic GVHD being quite uncommon.⁴ 

It would be logical to assume that a less than expected GVHD risk would be associated with an increased incidence of relapse. In fact, the opposite has been observed, with multiple reports demonstrating that T-replete CBT with either single- or double-unit grafts is associated with a robust graft-versus-leukemia (GVL) effect.^5-8  However, despite this tremendously important advantage, relatively little laboratory investigation has been performed to explain this biology. The study by Hiwarkar et al thus represents a long-awaited “bedside-to-bench” investigation of CB T-cell–mediated antitumor responses.

The investigators examined CB and PB T-cell antitumor responses in a xenogeneic nonobese diabetic/severe combined immunodeficiency/interleukin-2rg^null model. They found that CB T cells mediated enhanced clearance of human Epstein-Barr virus-driven B-cell lymphomas as compared with adult PB T cells. The CB T cells mediating this effect did so based on alloreactivity, as CB T cells that were syngeneic to the tumors demonstrated little to no antitumor immunity at all. CB T cells demonstrated rapid tumor infiltration with a preponderance of CD8⁺ T cells within the tumor. This was in contrast to PB T cells that exhibited delayed tumor infiltration and a greater proportion of CD4⁺ T cells, even though both T-cell groups initially demonstrated similar CD4:CD8 ratios.

This provocative study, arguing that CB T cells are intrinsically more effective at GVL, generates a host of important questions. First, it is surprising that the PB T cells did not have greater antitumor potency. Additionally, the tumor model used is distinct from most clinical CBT settings, other tumors may have different responses to the T-cell populations being studied, and it is conceivable that the methods used for injecting the tumors could arbitrarily favor one T-cell population over the other. Furthermore, there are inherent limitations of xenograft models, particularly for studies of immunology and transplantation, so the findings must be appreciated in this context and cannot automatically be applied to the human (or a standard alloreactive) setting. Beyond the xenogeneic antigenic differences and issues related to mixing cells across species, given the lack of donor (and possibly host) antigen-presenting cells (APCs) in this system and the importance of APCs for mediating GVL,^9,10  it is possible that the findings in this model may be distinct from what is occurring in clinical or other experimental transplant settings. Further mechanistic studies are thus needed to better understand the differences in the neonatal vs adult T-cell sources that are reported here.

Nonetheless, these authors’ findings are novel and intriguing, and they should prompt further investigation, both in the laboratory and in clinical correlative studies. Of great interest would be to better understand the seeming separation of GVHD vs GVL that has been observed in many human CBT recipients. Whether differences in GVL potency could be demonstrated against different malignancies (eg, myeloid vs lymphoid), in subcutaneous vs marrow-based disease, and after transplantation of single- vs double-unit grafts, are all of great interest. In the interim, thanks to this exciting study, we can agree that when considering the biology of CB T cells, as stated by the Monty Python comedians, we are now talking of “…something completely different… .”