|Home | About | Journals | Submit | Contact Us | Français|
Exposure to non-inherited maternal antigens (NIMA) in fetal and neonatal life has lifelong immunological consequences. Though there is a plethora of evidence of effects of mother on the immune responses of her offspring, there is very little knowledge available on how exposure to NIMA can result in either tolerance or sensitization. Understanding the mechanism of NIMA effects will impact different fields of immunology including transplantation, autoimmunity, and tumor immunotherapy.
Following the discoveries of beneficial effects of NIMA exposure on clinical outcomes in solid organ and bone marrow transplantation, it has now been shown that the exposure to NIMA induces various types of T regulatory cells (TR) in fetus and adult, which may partially account for tolerance to allografts bearing the NIMA. Though all offspring are exposed to the maternal antigens, they exhibit a great variability in the NIMA effects, which can be explained by the variability in the extent of maternal microchimerism (MMc).
Exposure to NIMA can have tolerogenic or sensitizing effects on the offspring, resulting in acceptance or rejection of allografts expressing the NIMA. This variability may be partly explained by the level and distribution of maternal cells persisting in the offspring.
Graft survival is best when a kidney donor and the recipient are HLA identical, with graft half-lives of ≥20 years being common between HLA-identical siblings. The reasons for this remarkable success since the 1960’s, throughout all eras of immunosuppressive drug regimens, are still not entirely clear. Absence of anti-HLA antibody risk, and lower frequencies of T cell clones specific for donor antigens, may be 2 contributing factors. However, even without HLA differences, no competent transplant physician would recommend discontinuing all immunosuppression in HLA-identical sibling transplant recipients, since immunity to non-HLA “minor antigens” is known to be sufficient to cause rejection. We have recently speculated that the hidden secret of success of HLA-identical sibling transplants is not simply the lower frequency of T cells specific for minor H antigens only as compared to minor H plus HLA, but rather, an inherent tendency for siblings to develop low-avidity CD8 T regulatory cells specific for each others’ minor H antigens, due to exposure to maternal non-inherited minor H alleles . When, instead of HLA-identical sibs, HLA haploidentical parents or siblings are considered as potential living related donors, immune-mediated graft rejection is a more frequent problem due to potential formation of anti HLA antibodies [2,3] and a high frequency of HLA allospecific T cells  that attack the donor tissue. To counteract this problem, immunosuppressive drugs are routinely administered lifelong to prolong graft survival. Unfortunately, these agents greatly increase the risk of systemic infections  and may encourage tumor growth [6,7]. So, it is the goal of transplant immunologists to find alternative ways to maintain a functional allograft without using immunosuppressive drugs that have undesirable side effects long term. Taking advantage of natural tolerance induced by HLA as well as minor H non-inherited maternal antigens (NIMA) is one of the more promising but still relatively unexplored approaches for this purpose.
More than fifty years ago, Owen et. al. showed that most Rh− women born of a Rh+ mother failed to produce antibody against Rh antigen when they gave birth to Rh+ children, whereas most Rh− women born of Rh− mothers did make anti-Rh . Class et. al.  later found that HLA broadly-sensitized patients (> 90% + for antibody to leukocytes from a random panel of blood donors) commonly failed to produce antibody against mismatched non-inherited maternal HLA class I antigens (NIMA), but were fully capable of producing anti-HLA specific for non-inherited paternal antigens (NIPA control).
One possible mechanism behind the B cell hyporesponsiveness to NIMA is clonal deletion, an idea that was originally proposed to explain the finding that offspring exposed to foreign erythrocyte antigens in utero become lifelong blood cell chimeras . Billingham et. al.  later showed when a mouse was exposed to allogeneic cells as a newborn it could accept a skin graft from the donor mouse strain as an adult. However, an alternative interpretation of these classic experiments is that the neonate is not clonally deleted for B and T cells recognizing the foreign (maternal or sibling) antigens, but rather that the acceptance of foreign tissue requires not only an initial exposure to a tolerogenic form of the antigen, but also establishment of a “tolerance network” involving host T regulatory cells and microchimerism.
As discussed previously, exposure to NIMA in early life has life-long tolerogenic effects, which has been shown in both human and animal research.
To evaluate NIMA effects in human solid organ transplantation, Burlingham et. al.  analyzed the outcome of primary renal transplants between siblings. In the study, haplotype-mismatched kidney graft survival was significantly better in NIMA- than in NIPA- mismatched donor-recipient pairs. Indeed, NIMA sibling haplotype-mismatched kidney graft survival was equal to that of HLA-identical sibling kidney transplants at 10 years . The NIMA effect in solid organ transplants was echoed by studies of patients receiving haploidentical parental or sibling hematopoetic stem cell (HSC) transplants [13,14]. Graft-versus-host disease after bone marrow transplantation was significantly less in NIMA-mismatched siblings than in other haploidentical donor-recipient combinations . Ichinohe et. al.  showed similar results in 35 patients who underwent non-T cell depleted hematopoietic stem cell transplantation from NIMA-mismatched donors. Interestingly, all 35 donors were selected on the basis of microchimerism; however, the extent and multi-lineage vs. unilineage form of microchimerism was not analyzed prior to transplantation.
With a mouse F1 backcross breeding model (B6 × BDF1) originally described by Zhang and Miller , Andrassy et. al.  performed heterotopic heart transplantation, and showed that one could reproduce essential features of the clinical NIMA-tolerance effect found in human sibling kidney transplants. B6 (H2b/b) male mice were mated with BDF1 (H2b/d) female mice (NIMAd mating) to obtain NIMAd-exposed H2b-homozygous offspring. Conversely, BDF1 males were crossed with B6 females (control mating) to obtain NIPAd control mice (also H2b/b). Offspring (~50%) that had inherited the H-2d haplotype (H2b/d) were used as skin donors, while fully allogeneic DBA/2 strain mice were used as heart donors for transplantation experiments in homozygous H-2b offspring. This study was able to reproduce the modest skin graft prolongation in NIMA vs. NIPA controls reported previously . However, to our surprise we found that 57% of NIMAd –exposed offspring exhibited allograft survival of > 180 days without conditioning, whereas the NIPA control F1 backcross offspring uniformly rejected the DBA/2 graft within 11 days. The surviving heart allografts in the NIMAd-exposed group were devoid of any vascular intimal hyperplasia, a common sign of chronic rejection, indicating that the tolerance induced by the NIMA exposure is not only resistant to acute rejection but also to chronic rejection.
Recently the impact of gender and strain differences on the NIMA effect was analyzed in mice. Only 25% of the female recipients in the NIMAd F1 backcross model were tolerant to DBA/2 heart allografts, which was significantly more than that of the female NIPAd control mice, but less than the 57% tolerance rate of NIMAd males [17*]. The weaker NIMA tolerance effect in female offspring may be due to estrogen-mediated enhancement of the CD4+ effector T cell responses , which could conceivably overcome T regulatory cell influences in cardiac graft rejection (see below). Another important variable influencing T effector: T regulatory cell balance in response to NIMA is the mouse strain from which the particular NIMAs are derived. Out of 6 different F1 backcross models in which H-2 homozygous offspring were developmentally exposed to H-2d, H-2k, or H-2b alloantigens, only those in which the offspring were developmentally exposed to the H-2d haplotype revealed a significant beneficial effect on heart allograft survival from a H-2 homozygous donor [DBA/2 or BALB/c]. NIMA exposure to H-2b or H-2k induced sensitization rather than tolerance to C3H or C57 BL/6 heart transplants [17*]. These differences may reflect differences in the maturity of the mouse immune system at birth—for example, the CBA (H-2k) does not have mature T cells until d.2 after birth, whereas C57 BL/6 mice have mature T cells at d.17–19 of gestation. In any case, the T and B cells in the offspring exposed to NIMA-H-2k or H-2b appear to be skewed toward sensitization/effector function, and away from tolerance/suppressor function.
Using the “tolerogenic” NIMAd F1 backcross model, a beneficial NIMA effect in mouse bone marrow transplantation has also been demonstrated . In this study, BDF1 mice were transplanted with B6 bone marrow along with bone marrow cells from either the NIMAd-exposed or NIPAd control mice. The first group exhibited less graft versus host disease (GVHD) than the latter; furthermore, the NIMA effect was abolished by depleting TR cells (CD4+CD25+) from the NIMA-exposed donor splenocytes, showing that TR are indeed involved in NIMA-mediated GVHD prevention . Molitor-Dart et. al.  showed that the NIMAd-exposed mice can exhibit bystander suppression of TT-mediated delayed type hypersensitivity (DTH) reaction in presence of maternal-type BDF1 antigen, while the NIPAd control mice did not. This bystander suppression could be blocked with anti-IL10 and anti-TGF-β antibody indicating a role of TR in the NIMA-mediated bystander suppression. When the splenocytes of NIMAd-exposed mice were injected i.v into a BDF1 host, they exhibited reduced lymphoproliferation and higher TGF-β expression on the CD4+ T cells than the control mice. When the NIMAd-exposed mice were transplanted with DBA/2 hearts, the regional lymph nodes and the allografts contained a high frequency of IL-10 and TGF-β-producing cells primarily in the CD4+CD25+ T cell population. The allografts were also enriched in LAP+ and Foxp3+ cells with relatively few CD8+ T cells.
A new study of human fetal lymph nodes has had a major impact on the field, shifting the focus of NIMA effects in humans to the T regulatory cell [21**] (For detailed review see Burlingham, WJ [22**]). This study showed suppressed proliferation of fetal T lymphocytes in response to maternal antigen presenting cells (APC), but not to unrelated APC. Fetal T cell proliferative responsiveness to NIMA was restored when CD25+ cells were depleted from the culture. A key finding from the study was that when T cells in the fetal lymph nodes proliferate in response to alloantigen, they acquire Foxp3 expression, and gain suppressive functions. The acquisition of Foxp3 and suppressive function was blocked by an agent that interferes with TGF-β signaling, indicating a role of TGF-β generated in the fetal [but not in adult] lymph node in the induction of the NIMA-specific TR.
NIMA effects are not only restricted in T lymphocytes, but also impact B lymphocytes [8,9]. Vernochet et. al.  investigated the effect of NIMA exposure on the development of B lymphocytes having different affinities for a NIMA MHC class I antigen. They used the B cell receptor 3–83 μδ transgenic mice as the male breeder with heterozygous mothers expressing H2Kd along with either H2Kk or H2Kb. The transgenic B cells recognized H2Kk and H2Kb MHC class I antigens (NIMA) with high and intermediate affinities, respectively. They found that when the B cells having higher affinity for the NIMA encountered the NIMA in fetal life, they were partially deleted in late gestation and the remaining transgenic B cells down-regulated their B cell receptors. The B cells also had impaired proliferative response and higher active caspase-3 (a central mediator of apoptosis) compared to the non-exposed controls indicating that exposure to high affinity NIMA leads to B cell apoptosis. In contrast, when the B cells encountered the NIMA to which they had lower affinity, they exhibit activated phenotypes including high proliferation, elevated CD69 expression, and preferential localization adjacent to T cell zones of the spleen. This suggests a mechanism whereby NIMA exposure could lead to sensitization or tolerance at the B cell level, depending on the strain sources of NIMA and the affinities of B cell receptors involved [17*].
What is the basis of the NIMA tolerance effects on T and B cells, leading to allograft survival? One possible answer is that many of us have already managed to accept a ‘mini-graft’ from our mothers in the form of widely disseminated maternal cells. Though fetal and maternal circulations are completely separated, fetal tissue is bathed with maternal blood in hemochorial placenta (e.g, mouse and human placenta) [24,25]. Trafficking of blood cells at the fetal maternal interface is bidirectional [26,27]. Fetal cells can cross the placenta and engraft into different organs of the mother, resulting in fetal microchimerism (FMc) [28–31]. Similarly maternal cells can also cross the placenta and engraft into different organs of the offspring resulting in maternal microchimersm (MMc) [32–35]. Hall et. al.  using in situ hybridization showed that 20% of male infants contain female (XX+) cells in their cord blood indicating the presence of maternal cells. MMc has since been detected by more sensitive PCR techniques in > 50% of offspring, in different organs including blood , spleen , brain , heart , lungs , liver , and pancreas . Besides the transplacental migration of maternal cells, offspring can also acquire maternal cells during nursing after birth, which preferentially localize in the liver of the newborn .
Microchimerism has been demonstrated in a wide variety of cell types in peripheral blood and lymphoid organs including T cells [42–44**], B cells [43*], dendritic cells [44**,45], and macrophages [44**]. Khosrotehrani et. al. [43*] crossed Rag-deficient female mice with wild type male mice and recovered functional T and B cells of fetal origin from the spleen of the female mice during pregnancy. The finding of lymphoid and myeloid cells of maternal origin in adult offspring suggests the presence of microchimeric progenitor cells that differentiate into the various cell lineages throughout the life of the animal. Stevens et. al. [46*] showed presence of maternal hepatocytes in liver, renal tubular cells in kidney, and beta cells in pancreas in infants. To account for parenchymal Mc, transfer of stem cells across the placenta has been suggested by different researchers [30,47]. Bianchi et. al. , by a nested PCR, detected CD34+ male progenitor cells in the peripheral blood of 1/8 women who were pregnant with male offspring at least 27yrs before the blood sampling. Chen et. al. [48*] injected human mesenchymal stem cells into pregnant rats and showed that the human mesenchymal stem cells crossed the placenta and differentiated into different human mesenchymal cells in the organs of the rat offspring, indicating that mesenchymal stem cells can also cross the placenta, giving rise to multilineage Mc.
Besides the reports relating Mc to autoimmune disease, Mc has also been shown to be associated with tissue regeneration and allograft tolerance.
Cells with multilineage potential transferred across the maternal- fetal interface have tissue regeneration capability after differentiating into different cell lineages . The first evidence came from a study by Starzl et. al. . They studied patients with type IV glycogen storage disease characterized by deposition of amylopectin in the heart and type 1 Gaucher’s diseases characterized by deposition of glycocerebrocides in lymph nodes. The patients displayed significantly decreased depositions of the compounds and systemic donor Mc after liver transplants indicating the role of the Mc in cells of the myeloid lineage, ameliorating the enzyme deficiencies. Another study showed that children suffering from type I diabetes had higher level of MMc in blood and higher number of insulin-producing maternal beta cells in pancreas than control children . If the maternal cells present in blood of the patients have multi-lineage potential, they can conceivably regenerate insulin-producing beta cells, as well as hematopoietic lineage cells. However this point remains unproven at present.
The proposal first put forward by Starzl et. al.  that Mc is required in inducing transplantation tolerance has been controversial [52,53]. Burlingham et. al.  showed that cytotoxic T lymphocyte (CTL) unresponsiveness was linked with donor Mc in a patient who received donor-specific transfusion (DST) pretransplant and later became tolerant to a maternal kidney allograft. His PBMC, which contained approximately one donor cell in 104–105 cells showed very little donor specific CTL activity in primary MLR culture. Removal of the donor cells from the recipient PBMC failed to restore primary CTL response to the donor; however stimulation with donor stimulator cells in presence of IL-2 fully restored the anti-donor CTL activity. The anti-donor CTL activity was suppressed after addition of either fresh recipient PBMC or enriched donor cells isolated from the recipient PBMC. No inhibitory effect was seen with donor cell-depleted PBMC, indicating that the donor Mc is responsible for the reduced anti-donor CTL activity. This finding was supported by a study , in which B6 mice were engrafted with H8 cells that constitutively expressed H2Db-restricted GP33 epitope of lymphocytic choriomeningitis virus (LCMV). The microchimeric B6 mice elicited GP33-specific CTL unresponsiveness after LCMV challenge, which could be restored by experimental depletion of the microchimeric cells by injecting GP33-specific T cells into the B6 mice. These experiments together suggest that Mc can directly impact CTL responses in the host, and that removal of the rare foreign cells restores CTL activity/proliferation. However it does not account for how the Mc remains tolerogenic rather than sensitizing, to the host. For this we turn to the current focus in the NIMA field on Treg cell development and maintenance.
Molitor-Dart et. al.  showed that exposure to NIMAd results in TGF-β1+ adaptive TR development in some but not all mice. This variability in TR induction could account for the variability in allograft tolerance, since only those NIMAd-exposed mice with strong mobilization of TGFβ1+ TR to the graft were able to accept heart allografts long-term . To investigate whether the variability in the regulation induced through the exposure to the NIMA is due to the variability in the level of MMc, H-2Dd DNA level in 8–9 organs of H-2b homozygous offspring[n=36] of BDF1 mothers was determined by qPCR. All mice had previously been immunized with tetanus toxoid (TT). Spleen cell transfers into footpads of naïve B6 mice in the presence of a)PBS, b)TT, c)BDF1 cell sonicate, or d) BDF1 sonicate plus TT were performed and footpad swelling (DTH) responses were measured at 24 hrs. We found a significant linear correlation between the distribution of MMc (% of Dd + organs) and the % bystander suppression of TT-induced DTH reaction in the presence of maternal-type (BDF1) antigens [44**]. In addition, when injected into BDF1 hosts, CD4+ T cells of NIMAd-exposed mice having high levels of MMc proliferated less, and expressed significantly higher levels of surface TGF-β than CD4+ T cells of NIMAd-exposed mice having no detectable levels of MMc. This suggests a role of MMc in inducing and maintaining NIMA-specific TR; i.e. that the presence of maternal antigens is required for NIMA-specific TR. An alternative idea proposed by Mold et. al. [21**] is that NIMA-specific TR are long lived, and hence like other long-lived memory T cells, antigen-independent. It remains to be seen which hypothesis is correct.
Several mechanisms behind NIMA effects have been proposed:
Maternal antigen presenting cells present in the fetal thymus and bone marrow can cause deletion of the NIMA-specific lymphocytes. Maternal thymic epithelial cells, which can serve as APCs in the thymus, have been detected in the NIMAd-exposed offspring [44**]. Vernochet et. al.  described partial deletion of B cells having high affinity for the NIMA; however, B cells having low affinity for the NIMA were not clonally deleted. Another study  showed central and peripheral deletion of donor specific T cells after establishing mixed chimerism in the recipient with high dose of bone marrow cells. However, low-doses of BMC induced a form of tolerance that was T-regulatory cell dependent, consistent with establishment of microchimerism[Mc] rather than mixed chimerism.. Whether deletion of antigen-specific T cells following establishment of Mc as has been reported by Bonilla et. al.  is the norm for complex antigens like the major and minor H NIMAs, or merely an artifact of the LCMV epitope- GP33 system remains to be seen. However it seems likely that anergy/deletion of T effector cells due to a form of ‘suppressive’ microchimerism in APC will be a consequence of the establishment of a dominant T regulatory cell population in the host. Alternatively, it is not the microchimeric APC themselves, but maternal antigen acquisition by host APC from these few Mc cells that drives the balance of TE and TR in favor of the latter.
As discussed before, exposure to NIMA in fetal and neonatal life induce NIMA-specific TR. There are several ways by which NIMA-specific TR can be induced.
Since oral tolerance is known to generate TGF-β-producing TR [57,58], oral exposure to maternal MHC antigens present in breast milk  may generate NIMA-specific TR cells, which may prevent deletion of maternal cells by NIMA-specific effector T cells resulting in high level of MMc (Fig. 1). Campbell et. al.  reported a poor outcome of maternal renal allografts in recipients who were not breast-fed, as compared with breast-fed recipients. Similarly, NIMAd-exposed mice foster nursed by B6 mothers not only lost tolerance to a NIMA-expressing heart allograft , but also exhibited significantly reduced levels of MMc; some even became sensitized to NIMA, as if effector T cells were now pre-dominant [44**]. Interestingly, when NIPAd control mice were nursed by BDF1 mothers, 2/7 offspring exhibited H2Dd Mc in their livers but not in other organs (Dutta et. al., unpublished data), consistent with the finding by Zhou et. al.  that maternal cells present in the breast milk mainly localize in livers of neonates after nursing. Aoyama et. al. [61**] showed that nursing alone could induce some protection against graft versus host disease (GVHD) in mother after bone marrow transplantation from donors exposed to maternal antigens only through nursing. However, for the maximum protection against GVHD, both the uterine and oral exposures to the maternal antigens were required.
We have recently shown MMc in CD11b, CD11c, and CD4-positive cells sorted from spleen and bone marrow [44**]. Presence of MMc in such MHC class II-positive cell populations may indicate the presence of direct pathway of maternal antigen presentation to the NIMA-specific TR cells. This suggests the possibility for inducing a NIMA-specific TR cell able to recognize maternal antigens via both direct and indirect pathways (hybrid presentation) (Fig. 1). TR cells having both indirect and direct allo-specificity have been shown to be especially potent in inducing tolerance to allografts [62**]. Induction of similar TR cells capable of dual recognition may help explain the long-term acceptance of a renal allografts that was HLA class I-mismatched, but HLA class II- closely matched [63*].
Low levels of MMc may not be enough for inducing NIMA-specific TR cells. There should be a way of amplifying the level of maternal antigens in the offspring. We found though lymphoid organs of the NIMAd-exposed mice, such as spleen had very low levels of MMc, they had higher levels of maternal antigen-positive cells than that of NIPAd control mice [44**] indicating the presence of indirect and ‘semi-direct’  pathways of maternal antigen presentation to the NIMA-specific TR cells (Fig. 1). Molitor-Dart et. al. [17*] showed that NIMAk and NIMAb-exposed offspring that do not exhibit the NIMA-induced tolerance do not show maternal antigen acquisition, where as NIMAd-exposed offspring that exhibit the NIMA-induced tolerance do, indicating a role of maternal antigen acquisition in the NIMA-induced tolerance. The source of the maternal antigen may be the MMc present in non-lymphoid organs, which is then acquired by the offspring lymphoid cells. Antigen acquisition by a cell from another genetically non-identical cell is called trogocytosis . Dendritic cells are one of the major cell types that perform trogocytosis  by ‘nibbling’ antigens from other cells with LFA-1  and scavenger receptors . T and B cells can also acquire antigens from other cells . Antigen acquired by antigen presenting cells in this way can be presented to T cells, which is called ‘semi-direct’ antigen presentation . Though trogocytosis is a random process, several biological significances of trogocytosis have been described , eg. clearance of CD8+ T cells which had acquired cognate MHC class I and peptide ligand by antigen-specific cytolysis, CD4+ T cells acquiring MHC class II and costimulatory molecules may act as antigen presenting cells . Interestingly NK cells , CD4+, and CD8+T cells  acquiring HLA-G by trogocytosis can also acquire regulatory functions.
Although the exact mechanism of NIMA is not known, donor selection based on NIMA in solid organ transplantation could be valuable since HLA identical donors are very rare. Bone marrow and stem cell transplantations will also be benefitted from the better understanding of NIMA effects. The latest finding of our lab that the level of MMc is correlated with the NIMA effects indicates the necessity of the prescreening of the recipients for MMc along with screening assays that can accurately detect NIMA-specific TR/TE balance. The murine data from our lab, as well as preliminary studies in healthy humans  suggest that widespread/multilineage microchimerism in APC and parenchymal cells, but not unilineage Mc (for ex., in T cells only) [75**] or Mc in short-lived cell subpopulations (e.g. granulocytes) is correlated with a favorable Treg/Te balance.
Papers of particular interest, published within the annual period of review, have been highlighted as:
* of special interest
** of outstanding interest