The pendulum swings

Abstract

Fetal and/or perinatal exposure to noninherited maternal antigens (NIMA) has been reported to induce NIMA-specific tolerance. This tolerant state is highly beneficial in transplantation settings; enhanced graft acceptance has been observed when transplanted tissues express NIMA. Reduction in severe graft-vs-host disease has also been noted when bone marrow grafts originate from donors exposed to NIMA in early life. However, there is emerging evidence that exposure to NIMA can alternatively lead to specific priming. The processes regulating tolerance versus priming to NIMA are poorly understood and probably multifactorial. Based on studies in both humans and mice, we propose that both the quality and the quantity of NIMA exposure will be found to be key determinants of these opposing outcomes.

Recently, much attention has been paid to the beneficial, tolerizing effects of exposure of the developing offspring to non-inherited maternal antigens (NIMA). This beneficial effect is most evident in settings of solid organ transplantation, wherein grafts expressing NIMA are more readily accepted than non-NIMA-containing grafts.^1–4 An additional benefit is observed with bone marrow transplantation in which grafts from donors exposed to NIMA in early life cause a less severe graft-vs-host disease in sibling recipients expressing that NIMA.^5–7 Exposure to NIMA may be a relatively common event in fetal and neonatal life,^8–14 a developmental time period associated with susceptibility to tolerance induction. Factors that may contribute to this tolerizing NIMA effect include the development of regulatory T cell activity or clonal deletion in the offspring, the level and type of maternal microchimerism, the genetic backgrounds of mother and offspring and gender differences.^15,16

In spite of the wealth of information pointing to a positive, tolerizing outcome to NIMA exposure, there are observations indicating that priming may be a rarer but alternate outcome. CTL¹⁷ and antibody priming¹⁸ to maternal antigens as well as acute rejection of NIMA-bearing grafts¹ have been described in humans. In the mouse, poorer graft survival and/or enhanced adaptive immune responses to NIMA have been demonstrated.^4,19–21 At first blush, the observation that early life exposure can lead to priming seems to fly in the face of the prevailing dogma that fetal and neonatal animals are highly susceptible to the induction of tolerance. However, one of the most striking findings in the recent past is that full-fledged priming of adaptive immune responses can occur in both human and murine neonates and even in utero.²² Therefore, it is actually not surprising that robust priming to NIMA can occur during gestation and/or perinatally.

An important question that arises from these observations is why NIMA-specific tolerance develops in some individuals, while priming occurs in others. Although there are clearly many possible explanations, we propose that the quantity of NIMA exposure may influence whether tolerance or priming occurs. Tolerance may arise in response to relatively high doses of NIMA from persistent exposure throughout development. Breastmilk contains both immune cells and soluble HLA molecules;^23–25 therefore, high doses of NIMA could be attained through a combination of exposure to NIMA in utero and during breastfeeding after birth. Such chronic exposure may then lead to the development of tolerance-inducing processes. Several studies support this idea, demonstrating that chronic exposure to antigen can lead to the development of functional Tregs, both in vitro and in vivo.^26–28 On the other hand, priming to NIMA may occur in response to lower, more transient doses of NIMA, such as that occurring in utero only. The idea that low level NIMA exposure may lead to the absence of tolerance or even primed responses is supported by studies in both humans and mice. In humans, cord blood cells have been reported to develop NIMA-specific cytotoxic responses in vitro,^29,30 demonstrating a lack of tolerance after in utero exposure alone. Breastfeeding was found to be required, together with in utero NIMA exposure, to specifically downregulate CTL alloreactive responses to NIMA²⁹ as well as to improve the survival and function of kidney allografts from sibling³¹ and maternal³² donors. Similarly, although some level of tolerance was observed upon oral exposure only,³³ maximal tolerizing effects have been reported only with the combination of in utero and breast milk exposure to NIMA in mice.^3,33 In humans, priming to NIMA has been reported in newborns; Goulmy and colleagues¹⁷ demonstrated the pre-existence of NIMA-specific cytotoxic cells in cord blood—i.e., priming following prenatal exposure only. Moreover, a recent study from van Rood et al.³⁴ indicated that cord cells may exhibit primed responses to NIMA in vivo; cord blood transplants in NIMA-matched versus non-NIMA matched recipients showed lower leukemia relapse rates, suggesting that the NIMA-matched transplanted cells mounted an anti-leukemic response. In addition, our studies in the mouse provide compelling evidence that low level NIMA exposure may lead to priming. Using an allogeneic transfer system that mimics the natural exposure to NIMA, we have found that, while large doses of NIMA-like antigens lead to CTL non-responsiveness, small doses induce specific CTL priming.^19,20 Small numbers of maternal cells may be insufficient for stable chimerism, thereby failing to support sustained Treg function and leading to priming instead. Indeed, as recently documented by McCune and colleagues,⁸ human fetal cells can develop substantial anti-NIMA responses. Alternatively, small numbers of cells may lead to selective microchimerism, which, as discussed in the next section, may more likely induce priming rather than tolerance.

In addition to the level, we propose that the type of NIMA exposure may also be instrumental in deciding the balance of tolerance versus priming. There are several ways that the quality of NIMA exposure may be important. First, the type of cell(s) that the fetus and/or neonate is initially exposed to may influence the direction of the resulting NIMA-specific response. Indeed, there are several studies that have shown that exposure to different cell subsets can induce the development of either tolerance or priming. Matzinger and colleagues³⁵ showed that whole spleen cells tolerized while mature dendritic cells (DC) primed neonatal animals. Further, in a model system that mimics early neonatal development, exposure to inoculum containing mature donor T cells induced tolerance to skin grafts, while grafts lacking T cells did not.³⁶ Most of the available evidence suggests that maternal immune cells of many types can be found in the offspring.^8,10,33 Therefore, exposure to maternal populations that contain mature DC may induce priming, while those that lack them may generate tolerance. Likewise, maternal populations that include T cells may tolerize, while their absence may lead to priming. Alternatively, the type of maternal chimerism that is stably maintained after the initial exposure may also influence whether tolerance or priming is achieved. In support of this, Anderson and colleagues³⁷ have shown that the presence of T cells alone may not be sufficient to induce tolerance; mice exhibiting persistent unilineage donor T-cell chimerism failed to develop tolerance whereas those with long-term multilineage chimerism were tolerant.

In summary, although tolerance to NIMA is the desired outcome in transplantation settings, the importance of priming against NIMA should not be overlooked. Overall, the challenge of exploiting NIMA exposure for successful transplantation relies on understanding how tolerance versus priming is achieved. We believe that both the quantity and quality of NIMA exposure will be found to play important roles in this outcome.