Can We Prevent Congenital Infection by Cytomegalovirus?

Abstract

The prevention of congenital cytomegalovirus (CMV) infection and disease by vaccination has been studied for more than 50 years without successful application in practice. This failure has led to doubts as to whether immune prevention is possible. Indeed, some contend that “cytomegalovirus is not rubella,” meaning that induction of maternal CMV antibodies that neutralize viremia will not protect the fetus as they do against rubella. However, although prevention of congenital CMV will not be as straightforward as was the case for congenital rubella, we contend that the mechanisms of congenital CMV are becoming clear and that a vaccine could be successful in prevention. Knowledge concerning CMV and potential protection against congenital infections has grown over the years, led by a number of scientists and their associates, including Suresh Boppana in Birmingham, Alabama; Paul Griffiths in London, UK; Marianne Leruez-Ville in Paris, France; Daniele Lilleri in Pavia, Italy; Mark Schleiss in Minneapolis, Minnesota; and Sallie Permar in Durham, North Carolina, and New York, New York.

The human CMV is ubiquitous and, in lower- and middle-income countries, most pregnant women have been previously infected in childhood. In high-income countries in North America and Europe, approximately one-third to one-fourth of child-bearing women are seronegative [1, 2]. If infected by CMV in early pregnancy, those women will have about a 30% to 40% risk of transmitting the virus to their fetuses [3–5].

During pregnancy, seropositive women will also be exposed to CMV carried by others, including their prior-born children, and those women will be reinfected or may reactivate latent cell–associated CMV they carry from a previous infection. Thus, some children of CMV-seropositive women will undergo intrauterine infection, but the risk of that infection is between one-tenth and one-fourth of the risk in previously seronegative women [3, 5]. Prior infection confers strong (75%–90%) protection against congenital CMV infection of fetuses [6].

The question as to which immune responses are responsible for a reduction in risk of congenital infection has been debated over the years. Candidates include antibody to the glycoprotein B (gB) surface glycoprotein, antibody responses to the pentamer, or other proteins also present on the surface of the virus, and T-cell memory, particularly that evoked by the internal tegument pp65 protein [7]. In 1 study, the higher the quantity of maternal neutralizing antibodies in seropositive women, the lower was the risk of congenital CMV infection [6].

However, classical neutralization is not the only antibody function involved [8]. Recent studies have shown beyond doubt that functions such as antibody-dependent cellular cytotoxicity mediated by natural killer cells may be at least as important as classical neutralization. In effect, it has become clear that multiple immune responses in a seropositive mother act together to prevent infection of the fetus, including antibodies, CD4+ T cells, and CD8+ T cells [8–14].

Interestingly, neutralizing antibodies to gB also seem to be key in the prevention of CMV infection in transplant recipients [15], and antibodies to the AD2 and AD6 epitopes of gB are the specific actors in that prevention [16, 17]. The recent identification of a prefusion form of gB may permit improvement of the antigen structure and immunogenicity in primates [18].

A proposed concept that could explain protection is that, in seropositive women, little cell-free virus is circulating to expose the fetus, whereas in seronegative pregnant women, cell-free virus is present to some degree during the early stage of infection. Recent studies confirm that CMV reinfection or reactivation in seropositive pregnant women usually does not lead to infection of fetuses [19, 20]. However, cell-to-cell spread in those women has not been excluded [21]. In addition, a human challenge study showed that whereas seropositive volunteers usually were protected, high doses of CMV could overcome prior immunity. Consequently, an elevated titer of viremia after infection in a seropositive mother may lead to fetal infection [22].

In my opinion, the important concepts with regard to the induction of congenital CMV are as follows: first, the virus is primarily cell associated, except briefly after a primary infection. Second, after exposure to CMV, a child or a seronegative pregnant woman will develop a pharyngeal CMV infection that rapidly becomes intracellular. Third, once antibodies develop in the infected woman, cell-free virus is likely greatly diminished and infection of the fetus across the placenta is less likely, as has been observed in an in vitro model of human placenta [23]. Thus the risk of fetal infection and abnormality is less after CMV infection in seropositive women.

However, the antibodies operative in preventing the access of virus to the fetus are complex. Certainly classical neutralization plays a role, but the Permar group and others have shown that Fc effector functions operating on natural killer cells are at least as important as neutralization in the prevention of fetal transmission [10–12, 24]. Moreover, CD4+ T cells are critical to development of antibodies and CD8+ T cells can suppress viral replication [7,9]. Interestingly, experiments with the guinea pig CMV confirmed that the gB and pp65 analogs working together gave protection against congenital infection in that animal model [25]. Clearly, protection against fetal infection in humans depends on multiple antibody and T-cell responses.

Aside from the protective functions of antibodies and cellular immunity, a simple fact should be taken into account with regard to explaining the frequency of fetal infection, and that is the relative number of challenges to maternal immunity. In low- and middle-income countries, there is more frequent exposure of women to children and other individuals excreting CMV. Thus, pregnant women are highly exposed to CMV, in particular from young children. The size of the virus challenge to women may also be greater in those crowded environments than in rich countries.

Vaccine development against CMV started back in the 1970s, far in advance of the knowledge summarized here [26]. Prevention of infection in seronegative transplant recipients who received tissue from seropositive donors was achieved in studies conducted both in the United States and the United Kingdom using a live attenuated vaccine or a killed virus vaccine [15, 27]. Additional candidate vaccines are now being tested including an mRNA vaccine from Moderna [13, 25, 28, 29].

Given that there has been limited success in prevention of congenital CMV despite notable advances in our knowledge over the past 50 years, it is time that a concerted effort be made to identify the best vaccine antigens and to show that they do indeed prevent primary infections. The characterizations of an ideal vaccine are listed in Table 1. Although a vaccine may not provide 100% protection against CMV infection, it should be capable of giving vaccinated women of child-bearing age a greatly reduced risk of bearing affected infants in subsequent pregnancies. In addition, modeling has shown that vaccination of young children to render them seropositive could result in reduced CMV exposure of their mothers in subsequent pregnancies [29]. We now have indications as to which antigens should be in the vaccines and the responses that should give protection against transfer of virus from mother to fetus. A major effort to license a CMV vaccine in the near future should have a high priority for public health [30].

Notes

Financial support. No external funding was used.

References

Author notes