Microchimerism in cord blood: Mother as anticancer drug

Maternal microchimerism (MMc) in cord blood (CB) caused by the exchange of cells at the maternal–fetal interface was first described in 1995 by Hall et al. (1). To the many known immunologic consequences of maternal–fetal cell exchange and particularly, CB MMc must now be added yet another consequence: the antitumor immunotherapeutic effect of CB transplants. Surprisingly, this effect seems to be due at least in part, to rare maternal T cells contained therein.

T Cell Priming at the Maternal–Fetal interface

Because HLA loci are highly polymorphic and mothers and fathers are generally HLA heterozygous, the offspring are nearly always HLA mismatched with the mother for at least one of the antigens encoded by inherited paternal haplotypes at the HLA loci—HLA-A,B,C (class I) and HLA-DR,DQ,DP (class II). Because of homozygosity at certain HLA loci—for example, if the offspring is HLA-A2,A2—the inherited paternal antigen (IPA; in this case, A2) will not present a target HLA-A alloantigen to the mother's T and B lymphocytes. (Note that the analysis in this study did not evaluate IPAs that are minor histocompatibility antigens that may also be bound as peptides to HLA-A2 and presented to mother's minor H antigen-specific T cells; however, typing for these loci is still in its infancy, whereas HLA typing is relatively advanced.) Similarly, from the neonate's point of view, maternal homozygosity at a given locus (for example, if the mother is B62, B62) means that the B locus noninherited maternal antigen (NIMA; in this case B62) will not present a target for its T cells. However, if one knows the complete HLA-A,B,DR typing of the mother and offspring, one can usually identify heterozygous HLA loci encoding IPA and NIMA target antigens that could potentially stimulate an alloreactive T-cell response. Thus, the maternal–fetal interface at birth is an arena with diverse and sometimes opposing players; it is occupied, on one hand, by maternal T cells specific for IPA present in the newborn, and it is occupied, on the other hand, by fetal T cells, both Tregs and Teffectors, specific for the NIMA that are found in the mother.

Oncologists and transplanters, attempting to tap the rich source of hematopoeitic stem cells (HSCs) present in umbilical CB for therapy of leukemia, did not mind the admixture of T lymphocytes also present and therefore, unwittingly, have transferred the T cells (the battlefield generals, if you will) to a third party, namely the leukemia patient. Because the mother who donates her child's CB is not herself routinely HLA-typed, the field of CB transplantation has lived in blissful ignorance of the role played by either fetal antimaternal or maternal antifetal T cells in the success of CB transplantation. Then,

The findings of van Rood et al. suggest that anti-IPA immunity is of value in cancer immunotherapy.

Jon van Rood of the Leiden University in The Netherlands found collaborators Andromachi Scaradavou and Cladd Stevens at the New York Blood Center; they routinely HLA-typed the mothers as well as the newborn CB donor. The initial result of this collaboration, published in PNAS in 2009 (2), uncovered a mechanism of cancer immunotherapy attributable to the response of the newborn's T cells to NIMA shared by the leukemia patient who receives the CB transplant. In HLA-mismatched CB transplants, the patients with mismatched HLA that included a HLA-A, B, or DR antigen identical to one that had not been inherited from the mother had a more rapid engraftment, less GVHD, and a lower rate of leukemia relapse compared with patients with HLA mismatch that was a non-NIMA. This beneficial effect was seen only in recipients with acute myeloid leukemia and not acute lymphoblastic leukemia. Now comes another bombshell from retrospective analysis of the same database: the sharing of IPA between the CB donor and recipient is advantageous to both acute myeloid leukemia and acute lymphoblastic leukemia recipients, significantly limiting relapse rates in the first 3 y posttransplant compared with patients where no IPA sharing is present (3). The benefit of IPA sharing is achieved with only a slight increase in graft-versus-host disease (GVHD) incidence (hazard ratio = 1.4) that did not reach statistical significance, whereas the benefit in terms of reduced relapse risk is substantial and highly significant (hazard ratio = 0.35, P < 0.001) (3).

Maternal T cells in CB are relatively rare, ∼0.1–0.5% of total T cells, but enough to visualize by flow cytometry (4). Unlike the neonatal T cells, which comprise the other 99.5%, it is likely that many of the maternal T cells that have crossed the placenta are T memory cells that have already been selected for anti-IPA alloreactivity. T memory cell bias could explain the resistance of the IPA effect to cyclosporine and antithymocyte globulin (ATG) immunosuppression, which is used routinely in CB transplants. When one considers that a typical CB transplant involves transfer of ∼3 × 10⁷ nucleated cells per kg recipient body weight and assumes that 1% are CD34+ HSC and 33% are T cells, this finding means that as many as 5 × 10⁴ maternal T cells/kg are transferred with each CB treatment.

In a previous paper on bone marrow transplantation (BMT) between haploidentical family members, van Rood et al. (5) found evidence of a deleterious effect of IPA reactivity on transplant-related mortality and acute and chronic GVHD risk in recipients of maternal allografts, compared with a one haplotype mismatched graft from a sibling sharing the same paternal HLA haplotype, but mismatched for NIMA. Importantly, the patients in this earlier study all received non-T cell-depleted BMT. Thus, when IPA-specific T cells were present in abundance, any antileukemia effects were overridden by the increased risk of mortality because of GVHD. Only when T cell-depleted BMT was analyzed in a later study by Stern et al. (6) did a strong antileukemia effect in mother to child transplants emerge. If one considers that T-depleted marrow transplants still have some residual T cells present, the parallel with the current finding of antileukemia effects of rare maternal T cells in CB transplantation becomes quite striking.

Wider Implications of the Discovery

The findings of van Rood et al. (3) suggest that anti-IPA immunity is of value in cancer immunotherapy. Beyond CB transplantation, this type of stealth immune function is inherent in all outbred individuals because of MMc. How it might apply to immunity to solid tumors (7), autoimmunity (8, 9), and tolerance vs. rejection of solid organ transplants (10, 11) are currently areas of intense investigation. One example of how anti-IPA immunity may be relevant to solid organ transplant success is found in the recent paper by Jankowska-Gan et al. (10). These investigators found that transplant success in one HLA haplotype mismatched kidney allografts was remarkably correlated with pretransplant bidirectional regulation. All individuals were tested by a trans vivo delayed type hypersensitvity (DTH)/bystander suppression assay, including both healthy controls and patients with end-stage renal disease. Patients and controls were found to have a strong regulatory response to maternal antigens present in a cell sonicate prepared from maternal PBMC (i.e., to the NIMA present therein). In some cases, the mother, who was about to become the transplant donor, reciprocated [i.e., her peripheral blood mononuclear cells (PBMCs) were able to mediate bystander suppression in the tv-DTH assay when mixed with IPA present in a cell sonicate of son's or daughter's PBMCs]. In such cases, the donor–recipient pair exhibited pretransplant bidirectional regulation. In contrast to this situation, maternal kidney donors with PBMCs that lacked regulatory activity to their son's or daughter's IPA were also identified. In this situation, the donor–recipient pair would be classified as having one-way regulation. Although bidirectional regulators had excellent 3-y graft survival (9/9), nearly one-half of the unidirectional or nonregulators (4/9) rejected their allografts, and nearly all (8/9) developed donor-specific antibodies by year 3. These striking results indicate a previously unsuspected negative impact of antirecipient (including anti-IPA) immunity of the kidney donor, undermining regulation in the recipient. In other words, anti-IPA immunity may be a good news/bad news situation: good for generating a graft vs. leukemia effect in CB transplantation but bad for the living donor to have in solid organ transplantation. Either way, the awareness of anti-IPA immunity will doubtless lead to additional characterization of the rare maternal T cells in the CB grafts and kidney allografts, and therefore, it will benefit patient outcomes.

If correct, the IPA hypothesis could account for the remarkable success of CB transplantation as a modality for treatment of leukemia, because the vast majority (>90%) of CB transplants are of the IPA shared variety. By transferring the fetal–maternal battlefront to the leukemia patient's own battle with cancer, oncologists have unwittingly tapped into a rich source of graft vs. leukemia activity. Now that the secret is out, thanks to the work by van Rood et al. (3), it would be wise to type mothers routinely, determine the IPAs, and avoid transplants in which the recipient does not share these IPAs with the CB donor.