Discrimination of primary and chronic cytomegalovirus infection based on humoral immune profiles in pregnancy

Abstract

BACKGROUNDMost humans have been infected with cytomegalovirus (CMV) by midlife without clinical signs of disease. However, in settings in which the immune system is undeveloped or compromised, the virus is not adequately controlled and consequently presents a major infectious cause of both congenital disease during pregnancy as well as opportunistic infection in children and adults. With clear evidence that risk to the fetus varies with gestational age at the time of primary maternal infection, further research on humoral responses to primary CMV infection during pregnancy is needed.METHODSHere, systems serology tools were applied to characterize antibody responses to CMV infection in pregnant and nonpregnant women experiencing either primary or chronic infection.RESULTSWhereas strikingly different antibody profiles were observed depending on infection status, limited differences were associated with pregnancy status. Beyond known differences in IgM responses used clinically for identification of primary infection, distinctions observed in IgA and FcγR-binding antibodies and among antigen specificities accurately predicted infection status. Machine learning was used to define the transition from primary to chronic states and predict time since infection with high accuracy. Humoral responses diverged over time in an antigen-specific manner, with IgG3 responses toward tegument decreasing over time as typical of viral infections, while those directed to pentamer and glycoprotein B were lower during acute and greatest during chronic infection.CONCLUSIONIn sum, this work provides insights into the antibody response associated with CMV infection status in the context of pregnancy, revealing aspects of humoral immunity that have the potential to improve CMV diagnostics.FUNDINGCYMAF consortium and NIH NIAID.

Keywords: Adaptive immunity; Antigen; Immunoglobulins; Immunology; Infectious disease.

Figure 1. Diagram of study cohort participants.

Figure 2. Antibody features distinguish primary from chronic CMV infection but not pregnancy status.

(A) Uniform manifold approximation (UMAP) biplot of antibody features excluding IgM. Distinct clusters of participants with primary (blue, n = 158) and chronic (green, n = 76) infection but not pregnancy status (hollow and filled symbols) are observed. (B) Volcano plot of each CMV-specific antibody feature assessed. Volcano plot represents the log₂ fold change (x-axis) against the –log₁₀ P value (Mann-Whitney test: *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001). Antibody specificities (antigen) are indicated by shape and Fc characteristics (detection) indicated by color. (C) IgM, IgA, and IgG binding to CMV antigens (further described in Supplemental Table 1). Data are the mean median fluorescent intensity (MFI) values of technical replicates. Solid red line indicates median.

Figure 3. IgG subclasses display distinct profiles in primary and chronic infection depending on antigen specificity.

(A) Levels of gB-, pentamer-, and tegument-specific IgG1, IgG2, IgG3, and IgG4 antibodies in individuals with primary (blue, n = 158) or chronic (green, n = 76) CMV infection. (B) Scatterplot and fit lines of levels of (UT) pentamer-specific IgG1 and IgG3 responses by infection status. Pearson correlation coefficient (R_P) indicated in inset. (C) Scatterplot and fit line of levels of (UT) pentamer-specific IgG1 (left) and IgG3 (right) versus (UT) pentamer-specific FcgRIIIaV-binding antibodies. Pearson correlation coefficient (R_P) indicated in inset. (D) Pearson correlation coefficients (R_P) for each pentamer antigen tested across FcgR by infection status. Solid lines denote group medians; differences between groups were assessed by Mann-Whitney test (*P < 0.05, **P < 0.01, ***P < 0.001, and ****P <0.0001); values presented are median fluorescent intensities (MFI).

Figure 4. Machine learning accurately predicts primary or chronic CMV infection status.

(A) Schematic overview of cross-validated machine learning workflow employing antibody profiling data to discriminate between primary and chronic infection status in discovery and validation cohorts. (B) Prediction accuracy for 100 repeated 5-fold cross-validation runs on actual (black) and permuted (gray) class labels. (C) Confusion matrix of predicted versus actual class labels in the median model for 5-fold cross validation. (D) Class probabilities of each sample in the discovery set when evaluated as a test sample in the cross-validation run exhibiting median performance (left) and for the cross-sectional (center) and Erasme (right) validation cohorts using the final model. (E) The identities and coefficients of the features making the largest positive (n = 5) or negative (n = 5) contributions to the final model. Solid lines denote group medians; differences between groups or conditions were assessed by Mann-Whitney test (*P < 0.05, **P < 0.01, ***P < 0.001, and ****P <0.0001).

Figure 5. Longitudinal models define a molecular clock of CMV primary infection.

(A) Analysis overview. The infection status classification model trained on the cross-sectional cohort was applied to longitudinal samples available from the validation cohort. (B) Class probabilities of each sample in the longitudinal cohort over sample collection visits for individuals with primary (blue) and chronic (green) infection. (C) Scatterplot of class probabilities for participants defined as having primary infection at visit 1 over time. (D) Scatterplots of features employed by classification model to predict infection status over time in the longitudinal cohort. (E) Analysis overview. Primary infection samples from the longitudinal cohort samples were used to train a regression model to predict time since infection (days after symptom onset) that was applied to the primary samples from the cross-sectional cohort. (F) Scatterplot of model predictions of time since infection when primary samples used for predicting days after symptom onset. The cross-sectional Pavia (left) and Erasme (right, n = 23) samples w0ere used as distinct validation cohorts. Data shows the measure of the predicted label and its closeness to the true label.