Specific Cryptosporidium antigens associate with reinfection immunity and protection from cryptosporidiosis

Abstract

There is no vaccine to protect from cryptosporidiosis, a leading cause of diarrhea in infants in low- and middle-income countries. Here, we comprehensively identified parasite antigens associated with protection from reinfection. A Cryptosporidium protein microarray was constructed by in vitro transcription and translation of 1,761 C. parvum, C. hominis, or C. meleagridis antigens, including proteins with a signal peptide and/or a transmembrane domain. Plasma IgG and/or IgA from Bangladeshi children longitudinally followed for cryptosporidiosis from birth to 3 years of age allowed for identification of 233 seroreactive proteins. Seven of these were associated with protection from reinfection. These included Cp23, Cp17, Gp900, and 4 additional antigens - CpSMP1, CpMuc8, CpCorA and CpCCDC1. Infection in the first year of life, however, often resulted in no detectable antigen-specific antibody response, and antibody responses, when detected, were specific to the infecting parasite genotype and decayed in the months after infection. In conclusion, humoral immune responses against specific parasite antigens were associated with acquired immunity. While antibody decay over time and parasite genotype-specificity may limit natural immunity, this work serves as a foundation for antigen selection for vaccine design.

Trial registration: ClinicalTrials.gov NCT02764918.

Keywords: Adaptive immunity; Immunology; Infectious disease; Parasitology.

Figures

Figure 1. Humoral immunity to Cryptosporidium antigens was isotype specific.

Immune responses are shown for (A) IgA and (B) IgG antibodies. The Y-axis is signal intensity after normalization and the X-axis shows Cryptosporidium antigens ranked by median signal intensity. Bars represent the interquartile range of each antibody response and are shown as red if antibodies were present in ≥ 10% of infants (seroprevalent). (C) The Venn diagram shows seroprevalent Cryptosporidium antigens with IgA- (green) and IgG-specific (orange) and overlapping immune responses.

Figure 2. Cryptosporidium antigens recognized by IgA and IgG antibodies.

The proteomic microarray was used to measure the parasite-specific antibody response in the infants enrolled in our study cohort at 1 year in age. Previously infected children (columns) and the Cryptosporidium antigens (rows) that stimulated a strong IgG and/or IgA antibody response (present in > 10% of the children; n =232 antigens) are shown. The spot signals were normalized by first determining the specific background component by use of mixture models and setting this value to 0. Bar at the top of each heat map indicates the total number of Cryptosporidium antigens each child responds to (Antibody Breadth). The side bars indicate: (a) the seroprevalence of each antigen (% Sero+) and (b) presence of a membrane-targeting signal peptide (SP).

Figure 3. Antibody responses waned with time after a Cryptosporidium infection.

(A) The t-SNE plot identified a subset of children with a similar antibody profile. Each point corresponds to the immune profile of a child. Gray squares indicate children where no previous Cryptosporidium infections were identified by qPCR in clinical or surveillance stool samples (“qPCR–”), and orange circles represent children that had previous infections detected by qPCR (“qPCR+”), with the intensity of the overlaid color indicating the days since the last Cryptosporidium qPCR+ stool sample was identified. A group of infants had similar antibody profiles and a high density of recent infections (R1). (B) The split violin plot of antibody signals against the 100 most-reactive antigens (Y-axis) for each isotype (X-axis) shows the responses of children within the R1 region of the t-SNE plot compared with the remainder of the samples in R2. The median and quartile values are shown as horizontal lines in each split violin. (C) The split violin plot shows the same comparison as (B) using the antibody breadth (count of seropositive responses) among the 100 most-reactive antigens. P values above each split violin were calculated using linear mixed effects regression (LMER) and Wilcoxon’s rank sum tests for (B and C), respectively. (D) Antibody breadth among the 100 most-reactive antigens for each isotype is shown on the Y-axis after log10 transformation with the interval (days) between the last Cryptosporidium qPCR+ diagnostic assay and the time of antibody measurement shown on the X-axis. Linear regression P values and R2 values are shown for IgG and IgA, as well as a line and confidence intervals (colored bands; pink for IgA and green for IgG) fit to each. (E and F) PLS-DA is shown for IgA and IgG responses respectively. Each point corresponds to the immune profile of a child. The purple circles indicate the antibody response obtained from plasma that was collected from children where none of the stool samples (diarrheal or surveillance) collected during the first year of life, prior to the plasma sampling time point, were ever qPCR positive for Cryptosporidium parasites (“Yr0-1 qPCR–”). Green triangles indicate that the child had a verified Cryptosporidium subclinical or symptomatic infection (“Yr-0-1 qPCR+”). The percentage of the variation in the child’s antibody profile accounted for by each axis is indicated.

Figure 4. The breadth of the anti-Cryptosporidium immune response was not correlated with protection from infection.

(A) Split violin plot of antibody breadth in plasma among the 100 most-reactive antigens (Y-axis) for each isotype (X-axis) is shown for the comparison between children that had no stool samples (diarrheal or surveillance) qPCR+ for Cryptosporidium parasites (purple) and children who had a verified Cryptosporidium infection (green). (B) Data is shown from one year old infants who had prior qPCR-confirmed Cryptosporidium infections (“Yr0-1 qPCR+”) that were subsequently uninfected (blue) or reinfected (orange) during the next 2 years. (C) Data is shown from 1-year-old infants that included both the immunologically naive infants with no prior Cryptosporidium infections detected by qPCR in stool samples (diarrheal or surveillance) as well as those with qPCR+ stool samples during the first year of life. Medians and quartiles are indicated by horizontal lines in each split violin. Significant P values from Wilcoxon’s rank sum tests are shown above violins.

Figure 5. Children with antibodies that targeted the C. hominis peptides encoded by the gp60 gene and Cp23 protein were associated with protection from reinfection.

In the protein array data, IgA and IgG antibodies against the protein encoded by the C. hominis gp60 gene (Chro.60183) and IgG against Cp23 (Chro.40414) were associated with a delay in Cryptosporidium reinfection among children with a qPCR-verified Cryptosporidium infection during the first year of life (A, C, and E) or among all children in the study (B, D, and F). The X-axis shows days after the end of year 1 (when the assayed plasma samples were collected). The Y-axis shows the proportion of children who remained uninfected. Red lines represent children seronegative for the antigen, and blue lines represent seropositive children. The Kaplan-Meier curves show the probability of survival free of Cryptosporidium species, and the tables below the graphs indicate the number of children in the seropositive or seronegative categories at select time points. (A and B) IgG against Gp60 (Chro.60183). (C and D) IgA against Gp60 (Chro.60183). (E and F) IgG against Cp23 (Chro.40414). Hazard ratios (HR), confidence intervals, and P values were calculated using multivariable Cox proportional hazards models.

Figure 6. gp60 Genotype immune response.

(A) Cartoon illustrating the proteins encoded by the gp60 gene. (B and C) Heat maps showing the intensity and breadth of the IgA (B) and IgG (C) antibody responses to the polymorphic region of the Gp40 protein. The different alleles of the peptide encoded by the gp60 allele (columns) and the signal obtained when using the plasma with antibodies raised in response to infection of parasite with different gp60 genotypes (rows). Lines at the top of each heat map indicate the protein type and on the side the genotype of the infecting parasite. Parasite genotypes: rows 1–8: IaA18R3; 9: IaA19R3; 10–16: IaA25R3; 17–20: IbA9G3R2; 21–29: IdA15G1; 30: IfA13G1; 31–34: C. parvum IIdA15G1R1. Protein alleles: columns A: IaA27R3, B: IaA26R3, C: IaA25R3, D: IaA22R3, E: IaA19R3 F: IaA18R3, G: IbA9G3R2, H: IdA14G1, I: IdA15G1, J: IeA11G3T3, K: IfA13G1, L: IfA16G1, M: IIcA5G3a N: IIdA13G1. Side panels show the intensity scale for the amount of antibody binding to alleles expressed by IVTT and spotted on the array. Antibody binding to the purified recombinant relatively conserved Cp17 peptide was included on the array as a positive control. Its signal intensity was higher than that of the IVTT values.

Figure 7. RF analysis for selection of important antigens and analysis of risk during the first year after sampling.

(A) The scatter plot represents antigens and clinical variables ranked by VIMP scores in RF using 1,000 trees constructed per model. Models were fit to survival data during one year of follow up after sampling on seropositive and seronegative children that all previously had qPCR-confirmed Cryptosporidium infections. Models using the entire cohort of children and 2-year follow-up periods are shown in Supplemental Figure 8. Each model was repeated 100 times, and the VIMP score was averaged across all runs (Y-axis). For each antigen, the percentage of runs where VIMP was greater than 0 (i.e., important to the model) was calculated (X-axis). The red horizontal dashed lines represent the mean of all VIMP scores plus 1 SD. The vertical dashed red lines represent antigens with at least 80% positive VIMP scores. The upper right quadrant shows the antigens selected as important variables in the model. (B) The horizontal bar plot represents VIMP scores for each antigen with at least 80% positive VIMP scores. The vertical red dashed line represents the cutoff for selection of important variables (equivalent to the horizontal lines in A). HRs calculated in the survival analysis were shown as protective (HR < 1, teal) or not (HR > 1, magenta). (C) Only protective antigens with at least 80% positive VIMP scores and VIMP scores above the importance cutoff were selected for individual antigen analysis. (D–G) The Kaplan Meier plots represent the 2 most significant previously unknown antigens associated with protection in children with prior qPCR+ stool samples or all children, respectively, after feature selection using RF.

Figure 8. Cryptosporidium antigens associated with the development of a protective immune response.

(A) The survival curve illustrates the 2 subgroups of children and 2 follow-up periods after plasma was collected (end of year 1) that were analyzed for protection. The blue line follows only children who had a qPCR-verified Cryptosporidium infection (subclinical or symptomatic) during the first year of life, prior to sampling plasma. The red line follows all children in the array study and includes the immunologically naive children that remained uninfected at 1 year of age as well as those known to be previously infected. Dotted and dashed lines indicate the time points (year 2 and year 3) selected for analysis of the differences between uninfected and infected groups looking specifically at the protective candidate antigens identified in Table 1. (B and C) PLS-DA on the antibody profiles of the candidate antigens shown in Table 1 associated with either reinfection or protection. Each point represents the immune profile from 1-year-old children with prior qPCR-verified Cryptosporidium infections who were subsequently uninfected (blue circles) or reinfected (orange triangles) during the 1-year follow-up period (B) or 2-year follow-up period (C) after plasma samples were collected. (D and E) Predictor loadings derived from the PLS-DA analysis in (B and C) are shown, respectively. Antibody targets are shown on the Y-axis, and the X-axis shows the absolute value of the loading weights (or PLS-DA regression coefficients); the absolute value was used to focus attention on the importance of each antigen in maximizing the covariance between antibodies and Cryptosporidium infection outcomes. Orange bars indicate antibodies more abundant in the children who subsequently had a new Cryptosporidium infection and blue bars indicate the antibodies more abundant in the uninfected children.