Intended for healthcare professionals

Editorials

Pregnancy and immunity

BMJ 1994; 308 doi: https://doi.org/10.1136/bmj.308.6941.1385 (Published 28 May 1994) Cite this as: BMJ 1994;308:1385
  1. G M Stirrat

    For mammalian pregnancy to succeed, large physiological adjustments are required in the mother: these changes result from signals passing between the conceptus (especially the trophoblast) and the mother throughout pregnancy. Every system in the body is affected, including the immune system, which is part of a complex signalling system between cells that has developed the ability to recognise self and non-self.

    Immune adaptation is not required for the mother to cope with the fetus as an allograft. The lack of HLA antigens on the syncytiotrophoblast and the presence of only the non-classic HLA G antigen on the cytototrophoblast cells precludes the fetal trophoblast from playing any part in currently recognised types of allogeneic immune reactions.1 All these reactions depend on the cellular recognition processes associated with the major histocompatibility complex classes I and II. Not only, therefore, will the maternal immune system fail to be stimulated by allogeneic trophoblast, but allogeneic trophoblast cannot be the target for otherwise armed maternal cytotoxic T cells. Furthermore, according to current understanding of the phenomenon of “major histocompatibility complex restriction,” the absence of classic major histocompatibility complex antigens on the trophoblast will prevent the corecognition of any other form of cell surface antigens that it might express. The mother is not “immunodeficient”: in terms of classic mechanisms dependent on T cells she remains immunocompetent throughout pregnancy.2

    The maternal immune response does, however, undergo some changes. Data from inbred mice and less good evidence from human pregnancy (reviewed by Wegmann et al3) suggest that the maternal immune response may be modulated away from cellular responses towards humoral immunity, not all of which depends on recognition of major histocompatibility complex antigens. For example, antibody production during pregnancy tends towards the IgG 1 isotype and away from the complement fixing IgG 2a isotype, particularly for antibodiesagainst fetal alloantigens. In addition, cytotoxicity mediated by non-specific natural killer cells, seems to be dampened by inhibition by the TH1 helper cells, which produce cytokines such as interleukin-2 and interferon gamma. Activation of natural killer cells in pregnant mice is known to result in fetal resorption. This suggests that the bias towards the production of helper cells in pregnancy may be a protective mechanism tending to promote fetal survival. Dudley et al have observed that the cytokines produced by activated lymphocytes during murine pregnancy tend to favour antibody production over cytotoxic responses.4 This effect is most prominent within the uterine decidua, but it also has systemic effects. Dudley et al also contend that the regulation of cytokine production during normal pregnancy changes as a result of the pregnancy itself and not in response to specific fetopaternal antigens.

    Another possible immune modification might be some adaptation to inhibit non -specific complement mediated damage to trophoblast. Holmes and Simpson have shown that this is achieved at the level of the trophoblast itself without requiring any systemic dampening of the complement cascade, which presumably needs to remain intact if the mother is to combat infections adequately.5 The human trophoblast expresses three membrane bound complement regulatory proteins, which protect it from maternal complement mediated damage arising not only from activation of the classic or alternate pathways but also after systemic complement activation in response to microbial infection.

    Do the changes in the immune system during pregnancy have any untoward clinical effects? There is little firm scientific evidence that pregnant women are more susceptible to infectious diseases.6 In general, neither viral nor other infections (such as tuberculosis) seem to occur more commonly in pregnancy, nor are localised infections more likely to become generalised, as occurs in immunosuppressed patients. Clinical experience suggests that varicella infections tend to be more severe during pregnancy - in particular, varicella pneumonia seems more common in pregnant women, but this might be related to the mechanical effects on the diaphragm of the enlarged uterus.7 Falciparum malaria may be more common and severe during pregnancy, particularly during a first rather than subsequent pregnancies.8,9

    The potential effects of pregnancy on HIV infection and AIDS are of growing concern. Current evidence does not suggest that pregnancy makes any difference to the natural course of HIV infection in women.10 Access to medical care, particularly for drug using women, seems to be a more important determinant of outcome than pregnancy. This is, however, only an interim judgment.

    The misconception that pregnancy is an immunodeficent state may well be a consequence of the outdated paradigm that some degree of immune paralysis of the mother was required to protect the fetal allograft against immune attack. Biased observers therefore tended to find what they expected to find when looking for clinical sequelae of this presumed immunodeficiency. Our current concepts recognise that fetal trophoblast is not susceptible to attack by T cells but suggest that other aspects of the immune system may be modulated during pregnancy as part of a series of maternal adaptations necessary for successful fetal development. Current evidence does not suggest any clinical consequences, but - as always - we must recognise that our knowledge remains partial and in need of continual review.

    References