Congenital Cytomegalovirus Infection: Improved Understanding of Maternal Immune Responses That Reduce the Risk of Transplacental Transmission

congenital cytomegalovirus, cytomegalovirus vaccine, cytomegalovirus screening, TORCH infection, placenta

(See the Major Articles by Tanimura et al on pages 1652-8 and Fornara on pages 1659-65.)

Congenital human cytomegalovirus (CMV) infection is a common cause of severe and permanent neurological injury in newborns. A meta-analysis concluded that congenital CMV infection occurs in approximately 0.65% of newborns in the United States [1]. The prevalence was similar when data from a prospective multicenter study was analyzed. This study of over 100,000 newborns identified congenital CMV infection in approximately 0.5% of all deliveries [2]. Of the 20000–30000 infants born in the United States each year with congenital CMV, at least 10% will exhibit long-term neurodevelopmental sequelae, including mental retardation, seizure disorders, cerebral palsy, microcephaly, and sensorineural hearing loss [3]. The burden on society elicited by congenital CMV is substantial, and strategies to prevent infections in women of childbearing age are urgently needed.

CMV can infect the fetus either in the context of primary maternal infection during pregnancy, or following viral reactivation or reinfection in women already known to be CMV-seropositive prior to pregnancy. Transmission of virus to the developing fetus—even in the face of preconception immunity—can result in neurological sequelae. The likelihood of congenital CMV is directly proportional to the seroprevalence rates among women of childbearing age in the population being examined [4, 5]. Overall, as many as three-quarters of congenital CMV infections occur in women with preconception immunity [6], an observation that greatly complicates both the optimal anticipatory laboratory monitoring and clinical management strategies that should be employed during pregnancies at risk for vertical viral transmission.

Irrespective of whether congenital transmission occurs in the context of primary or nonprimary maternal infection, the key (and unanswered) questions are: How do we predict which women are going to transmit CMV infection to the developing fetus; and how do we know which infants will have sequelae if transmission occurs? Unfortunately, the answers are unknown, largely because knowledge of the nature of the maternal immune response(s) required for protection against congenital CMV transmission remains incomplete. Protection clearly requires more than just immunoglobulin G (IgG) antibodies against CMV. Immunologic parameters postulated to confer protection of the placenta and fetus include both innate (natural killer cell) and adaptive (CD4⁺ and CD8⁺ T cell, virus-neutralizing, and high-avidity antibody) responses [7–9]. Virus-neutralizing antibody may not only need to target the immunodominant CMV glycoprotein B (gB), but also CMV proteins involved in epithelial and endothelial cell entry (the “pentameric complex” of glycoproteins gH/gL/UL128/130/131). Neutralizing antibodies may also need to confer cross-protection against reinfection with genetically divergent CMV strains. Some of the proposed immune responses implicated in prevention of transmission of CMV from mother to fetus are summarized in Table 1. An improved knowledge of what is required to protect the fetus against CMV infection is vital for development of an effective (but as-yet hypothetical) preconception vaccine [10].

Table 1.

Immune Correlates Proposed to Be Important in Prevention of Maternofetal Cytomegalovirus Transmission

Innate Immunity
Innate immunity: cells
• Maternal NK cells
• Decidual (uterine) NK cells
• Memory NK cells
• Phagocytic cells (neutrophils, macrophages)
• Placental macrophages (Hofbauer cells)
Innate immunity: mediators
• Chemokines
• Cytokines (IFN-γ, IL-12, TNF-α)
• β-defensins
• Micro RNA (miRNA)
Innate immunity: host genetics
• TLR gene/promoter polymorphisms
o TLR2
o TLR3
o TLR7
• Mannose-binding lectin gene polymorphisms
Adaptive Immunity
Adaptive immunity: T cells
• CD4+ T cells
o pp65 (ppUL83)-specifica responses
o IE1, IE2-specific responses
• CD8+ T cells
• Gamma delta T cells
Adaptive immunity: antibody
• Serum neutralizing antibody
o Neutralizing antibody targeting glycoprotein B
o Neutralizing antibody targeting pentameric complex of gH/gL/UL128/130/131
•High-avidity antibodyb
•Antibody-dependent cellular cytotoxicity

Source: [7, 8 ].

Abbreviations: IFN, interferon; IL, interleukin; NK, natural killer; TLR, Toll-like receptor; TNF, tumor necrosis factor.

^aSee article by Fornara et al, this issue.

^bSee article by Tanimura et al, this issue.

Against this backdrop, 2 articles in the current issue of Clinical Infectious Diseases provide us with new insights into the maternal immune correlates of protection against transmission of CMV during pregnancy. In the study reported by Fornara and colleagues from Pavia, Italy, the evolution of CMV-specific T-cell immunity was examined in a group of 44 pregnant women with evidence of primary infection during pregnancy, and results were compared to 8 CMV-seropositive pregnant women who had past (preconception) infections [11]. Primary maternal CMV infection during pregnancy is a particularly concerning situation, as vertical virus transmission occurs in up to 40% of cases [12, 13]. Indeed, in the Fornara et al study, 15 of the 44 women with a documented primary CMV infection (34%) transmitted CMV to their fetuses. This provided a unique opportunity for the investigators to compare the cell-mediated immune responses to CMV in transmitting and nontransmitting women. Previous work by the Pavia group demonstrated delayed development of the CD4⁺ lymphoproliferative response in transmitters compared to nontransmitters [14]. In the current manuscript, these studies were extended to comparisons of the specificity of the CD4⁺ response in each group. Using a cultured enzyme-linked immunospot (ELISPOT) assay, the investigators found that the magnitude of the CD4⁺ response to the CMV tegument phosphoprotein pp65 (ppUL83) was significantly higher in nontransmitting compared to transmitting mothers. They also observed a higher IgG avidity index (AI) in nontransmitting mothers. This study clearly defines, for the first time, a role for pp65-specific CD4⁺ responses in protection against congenital CMV transmission, and has important implications for subunit vaccines in clinical trials, many of which include pp65 [15]. Indeed, in a relevant congenital infection model in guinea pigs, the addition of pp65 to the gB antigen augmented subunit vaccine-mediated protection against CMV transmission, compared to the gB antigen alone [16].

The importance of the AI in predicting risk of transplacental CMV transmission is explored in a second article in this issue of Clinical Infectious Diseases. In this study, Tanimura and colleagues at the Kobe University Medical School, Japan, determined the CMV IgG AI in a cohort of pregnant women undergoing routine obstetrical care, obtained in parallel with immunoglobulin M (IgM) and IgG serologic screening [17]. Traditionally, CMV serological testing during pregnancy involves testing a woman just for IgM and IgG antibodies. However, CMV IgM assays suffer from a lack of sensitivity and specificity in this setting [18], and the finding of a negative IgM and positive IgG assay can be observed in the setting of a relatively recently acquired, primary maternal CMV infection—a (false negative) IgM assay notwithstanding. In this setting, the pregnancy is at high risk for congenital CMV transmission, but this risk may be overlooked. It is in this situation where adding the AI to the screening serological profile may be particularly helpful. To more comprehensively analyze CMV immune status in their pregnant patients, the Kobe group incorporated assessment of the AI into their maternal screening algorithm, and compared the sensitivity and predictive value of AI to IgM serology. A total of 1562 of 2193 pregnant women were seropositive for CMV IgG antibodies (71%), and these women all underwent AI and IgM testing. Women with an AI of ≤45% underwent additional testing, including CMV polymerase chain reaction (PCR) testing of serum, urine, and cervical secretions, as well as blood antigenemia testing. All pregnancies were monitored for congenital CMV infection. Among the women with an AI of ≤45%, approximately half had a very low AI of ≤35%; 2 of these women had infants with congenital CMV infection. Notably, an AI of ≤35% was found to be of more value than IgM serology vis-à-vis the assessment of the risk of congenital CMV infection. Statistical analyses indicated that an AI of ≤35% had improved sensitivity, specificity, and positive and negative predictive values, compared with IgM testing. Thus, the findings of Tanimura suggest that serological screening for CMV during pregnancy using the combination of IgG and AI may be preferable to current screening approaches.

Two other observations made in the course of the Tanimura study warrant additional comment. First, it is noteworthy that among the 631 pregnant women who were CMV seronegative in this cohort (29%), only 5 (0.8%) of these women underwent IgG seroconversion during pregnancy. This is substantially lower than the stated overall rate of 3% for CMV seroconversions during pregnancy in a recent review [19]. Indeed, annualized rates of CMV seroconversion have been found to be even higher in women with young children attending group day care, approaching 8% [20]. The remarkably low seroconversion rate reported by the Kobe group reinforces observations published by other groups regarding the impact of maternal education on reducing CMV infections during pregnancy [21]. A second important observation from this study relates to the outcomes of pregnancies in CMV-seropositive women. Among women with positive CMV IgG antibodies and a high AI (>45%), 93% also had a negative (<1.2) IgM index. These women would be predicted to be a very low risk for congenital CMV transmission—but congenital transmission nevertheless occurred in 7 infants, and 3 of these infants had symptomatic disease. Thus, a screening strategy based on tiered evaluation of IgG antibody, AI, IgM antibody, and maternal virological assays still failed to identify the majority of pregnancies that were associated with congenital CMV transmission in this study, and reminds us that much work remains in elucidating additional features of the maternal immune response that predict the risk of fetal CMV transmission, particularly in women with preconception immunity.

In summary, these 2 research studies reported in this issue of Clinical Infectious Diseases have the potential to enhance our ability to predict the outcomes in pregnancies at risk for congenital CMV infection, albeit imperfectly. In the setting of a primary maternal CMV infection, the magnitude of CD4⁺ T-cell responses to pp65 is of clear importance in reducing the risk of maternofetal transmission. The utility of clinically relevant, point-of-care assays based on detection of cellular immune responses to pp65 should be evaluated in future studies. Identification of a low IgG AI (<35%) has significant implications for risk of congenital CMV transmission, and determination of AI should probably be routinely incorporated into serological testing algorithms during pregnancy [22]. Indeed, perhaps the time has come to reconsider whether CMV serological screening should be performed as a routine standard of care in all pregnancies in the United States [23, 24]. The American Congress of Obstetrics and Gynecology (ACOG) does not currently recommend CMV serological screening during pregnancy. This organization’s ambivalence for screening is in part based on the knowledge that new strains of CMV can be acquired during pregnancy, and transmitted, in a woman already seropositive for CMV: hence, screening will not provide guidance for whether a woman will be re-infected with a different strain during her pregnancy. However, the data from these 2 manuscripts support the contention that a more nuanced and detailed screening algorithm may, in fact, be quite useful in gauging risk. ACOG should therefore continue to assess the question of routine maternal CMV screening, since these papers indicate that assays such as the cultured ELISPOT and the AI may improve the clinician’s ability to counsel in greater depth about the relative risks of transmission during pregnancy. Perhaps more importantly, incorporating universal maternal CMV serological screening into clinical practice would, by creating a dialogue and increasing knowledge among women, increase overall awareness of the problem of congenital CMV.

Congenital CMV infection produces more disability in newborns than the more widely recognized entities of fetal alcohol syndrome, Down syndrome, and neural tube defects, and yet overall knowledge about CMV is substantially less than for these other threats to newborn health [25]. Until a vaccine is available, improved prevention strategies focusing on education of women of childbearing age about strategies to avoid CMV infection during pregnancy are essential. State legislative initiatives to promote CMV education and awareness programs, facilitated by the work of organizations such as the National CMV Foundation (summarized at https://www.nationalcmv.org/), represent encouraging steps in the right direction. Even as our knowledge advances regarding the aspects of the maternal immune response that are correlated with protection against congenital CMV transmission, increased public awareness will be essential to reduce the often devastating impact exerted by this infection on neurodevelopmental outcomes in newborns.

Notes

Financial support. The author is supported by the National Institute of Child Health and Human Development (awards HD044864 and HD079918) and the March of Dimes Foundation (award number 6-FY17-849).

Potential conflicts of interest. The author has received grant funding for laboratory support at the University of Minnesota to study CMV vaccines in the guinea pig model of congenital CMV infection. The author has submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.