Missed immunological opportunities in pregnancy

The eminent obstetrician–gynecologist, autoimmune researcher, writer and editor at the University of Leeds (UK), JS Scott, already pointed out in the mid-1960s what a potentially wonderful experimental model, for the investigation of the human immune system, pregnancy could represent [1]. Disappointingly, in over four decades, almost nobody has taken him up on the idea. What we have learned since, however, only further strengthens his concept.

Ever since Billingham (after epochal transplantation experiments in Medawar’s laboratory [2]), referring to the pregnancy model, pointed towards ‘nature’s own solution’ in overcoming allograft rejection [3], we seem to have forgotten that the fetal–placental unit, as with any organ transplant, represents an allograft in dire need of immunological tolerance. In pregnancy, highly complex allografts, encompassing all tissue types, are tolerated every day without pharmaceutical interventions by competent maternal immune systems, while attempts at organ transplantation still offer, often insurmountable, clinical challenges.

Scott was not a transplant biologist and, therefore, was unlikely influenced by Billingham’s comments about ‘nature’s own solution’. What led him to recognize that pregnancy could be an ideal research model were unexpected autoimmune phenomena, which he observed in pregnancy (probably best known among those, the now widely appreciated causal relationship between abnormal maternal autoimmunity and congenital neonatal heart block [4–6]), and the suspicion that preeclampsia probably has an immunological etiology [7].

When he made the above noted suggestion, he was most likely still unaware of the many common pathways abnormal allograft tolerance, autoimmunity and pregnancy often share. Here are just a few selected examples: abnormalities in allograft tolerance, such as (acute) graft-versus-host disease (GVHD), are frequently characterized by pathological autoimmune responses, also seen in certain abnormal pregnancy conditions [8]. Allograft tolerance is very likely dependent on mutual cell traffic and microchimerisms between donor (organ) and host [9], both are also characteristic findings in pregnancy [8–10]. Microchimerism has, in turn, been suggested to possibly lead to autoimmune diseases [11] and has been implicated as a cause of congenital neonatal heart block [12]. Had Scott known about these and other commonalities, shared between allograft transplantation, abnormal autoimmune function and pregnancy, he would very likely have been more outspoken in promoting pregnancy as an immunological research model.

An editorial does not offer opportunity for much detail. For such purpose the reader is referred to three recent publications by this author [8–10]. The following is only a brief summary of some of the most salient points: as an allograft, human pregnancy behaves similarly to other allografts in eliciting immune responses. Despite the widely prevailing notion that the fetus is immunologically ‘tolerated’’ by the maternal immune system, it is quite apparent that this tolerance is not always complete. For example, a Rh(D)-negative female with a Rh(D)-positive fetus may mount an anti-D immune response against her fetus. [13] A mother, genetically predisposed towards autoimmune disease (in this case rheumatoid arthritis) will exacerbate her autoimmune disease during pregnancy, depending on how close her histocompatibility (HLA) is with that of the allograft. As Nelson demonstrated over a decade ago, the more compatible the male and female are (of course, reflecting on fetal maternal HLA compatibilities), the more likely the mother will exacerbate her autoimmune disease. Incompatibility, in turn, favors immunological quiescence [14].

These observations once more raise the age-old question, how does the fetal allograft, in cases of normal pregnancy, succeed in achieving successful maternal tolerance? A definite answer to this question is, of course, still lacking. However, the organ transplantation literature suggests an intriguing explanation, which would perfectly accommodate well-recognized findings in pregnancy. The transplant pioneer, Starzl, and his associates, were the first to suggest that cell traffic between donor organ and host (and vice versa), and the resulting microchimerism, are essential for tolerance of allograft transplants [15]. Such bilateral cell traffic and microchimerism are, of course, also classical features of human pregnancy [8,9]. Exposure to slowly increasing HLA antigen amounts (as offered via slowly increasing cell traffic), could be equated to slow, tolerance-inducing, antigenic desensitization in early mouse experiments [16] or to desensitization therapies, widely applied by allergists in clinical immunology [17].

Assuming that this model applies equally to organ transplantation and pregnancy, one could expect occasions of failure in tolerance induction in both conditions, leading to similar clinical complications and presentations. Such complications of organ transplantation, of course, include GVHD (acute and chronic) and allograft rejection. As was recently pointed out, abnormal pregnancy conditions, indeed, mimic many of the clinical presentations of transplantation complications, but simply have so far not been recognized as such. [10] Uniformly the etiologies of these complications of pregnancy are still considered unknown and include, as the most infamous, preeclampsia/eclampsia, the dermatoses of pregnancy and repeated pregnancy loss. The former two present with very similar clinical presentations to acute GVHD, while autoimmune-induced early pregnancy loss can easily be viewed as an early rejection process, involving the fetal allograft [10].

Scott would probably have been pleased to learn that autoimmune responses can play an important role in GVHD and general allograft rejection [18]. These autoimmune responses are characteristic phenomena of pregnancy complications and, for example, are quite typical in preeclampsia/eclampsia [8,10]. Abnormal autoimmune responses in pregnancy seem time specific and related to the maternal immune system’s exposure to unusual amounts of paternal HLA antigens. Two periods of autoimmune exacerbation are typical: aside of potential prior exposure to paternal antigen through semen [19], the maternal immune system sees paternal HLA antigens, for the first time, with implantation. Early pregnancy is, therefore, one period of autoimmune exacerbation, and is characterized by pregnancy loss (repeated miscarriages). A second classical period of autoimmune flares occurs peripartum (reaching up to 3 months postpartum), when fetal–maternal cell traffic peaks [8–10]. Again in full analogy to organ transplantation, autoimmune exacerbations are, therefore, classically associated with peak alloantigen exposure [10].

Products of conception thus, good and bad, demonstrate practically all characteristics of an allograft. In the good, the transplant ‘takes’, (at least on a temporary basis) for as long as normal pregnancy requires tolerance (we suggested before that the still unknown initiation process of labor also may represent a preprogrammed allograft rejection mechanism [10]). In the bad, such as in organ transplantation, if genetic (or other) conditions for allograft tolerance are inadequate, complications may occur. Such complications may take the form of acute GVHD (pregnancy probably does not offer enough time for chronic GVHD) and/or rejection. The GVHD of pregnancy is preeclampsia/eclampsia and/or dermatoses of pregnancy, whereas autoimmune pregnancy loss may represent an early rejection mechanism [10].

Building on ‘nature’,s own solution’, as suggested by Billingham [3], and following Scott’s recommendation from over 40 years ago [1], it now, more than ever, seems to make sense to remind immunologists of pregnancy as a research model for allograft tolerance and abnormal autoimmune function. As an author, one also could not be in better company!

Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.