Discovery of potential diagnostic and vaccine antigens in herpes simplex virus 1 and 2 by proteome-wide antibody profiling

Abstract

Routine serodiagnosis of herpes simplex virus (HSV) infections is currently performed using recombinant glycoprotein G (gG) antigens from herpes simplex virus 1 (HSV-1) and HSV-2. This is a single-antigen test and has only one diagnostic application. Relatively little is known about HSV antigenicity at the proteome-wide level, and the full potential of mining the antibody repertoire to identify antigens with other useful diagnostic properties and candidate vaccine antigens is yet to be realized. To this end we produced HSV-1 and -2 proteome microarrays in Escherichia coli and probed them against a panel of sera from patients serotyped using commercial gG-1 and gG-2 (gGs for HSV-1 and -2, respectively) enzyme-linked immunosorbent assays. We identified many reactive antigens in both HSV-1 and -2, some of which were type specific (i.e., recognized by HSV-1- or HSV-2-positive donors only) and others of which were nonspecific or cross-reactive (i.e., recognized by both HSV-1- and HSV-2-positive donors). Both membrane and nonmembrane virion proteins were antigenic, although type-specific antigens were enriched for membrane proteins, despite being expressed in E. coli.

Fig 1

Construction of expressible HSV ORFeome by PCR and recombination cloning. (A) Gel showing PCR amplicon library arranged by predicted size. (B) Gel showing corresponding DNA minipreparations after recombination cloning of PCR amplicons into expression plasmid pXi. All plasmids are circular/nonlinearized. C-pXi, control (nonrecombinant) pXi plasmid. Results for HSV-2 are shown. Results for HSV-1 were similar (data not shown).

Fig 2

Heat map overview of the HSV-1 and HSV-2 antibody profiles of human sera. Columns correspond to sera used to probe the array, and rows are arrayed antigens. Sera were serotyped using HerpeSelect-1 and -2 IgG ELISAs (Focus Diagnostics), as shown at the top, and used as the reference for sample categorization. The patient sera were thus classified into the groups seronegative (neg; n = 47), HSV-1 seropositive only (1; n = 32), HSV-2 seropositive only (2; n = 6), and HSV-1 and -2 seropositive (1 + 2; n = 5). For comparison, sera from a healthy population (healthy; n = 21) were probed. Only those antigens that were reactive against sera from the HSV-1- or HSV-2-seropositive populations are shown. An antigen was defined to be reactive when the average signal intensity for a donor population was greater than the mean plus 2 SDs of the control spots consisting of IVTT reaction mixtures lacking DNA template (C + 2 SDs). The HSV-1 antigens are ranked by descending average signal of the HSV-1-seropositive population, and the HSV-2 antigens are similarly ranked for the HSV-2-positive population. The sera are also ranked from left to right within each group by increasing sum of the signals. The heat map was generated from log-normalized data that were retransformed to approximate raw values, and the signal was converted into a color as shown in the legend. For the ELISA data, a positive signal is shown in red, borderline in black, and negative in green.

Fig 3

Comparisons between HSV-1- or HSV-2-seropositive and -seronegative donors. Histograms show average array signals ± SEMs of seronegative, HSV-1-seropositive, and HSV-2-seropositive donors. Only the seroreactive antigens (>C + 2 SDs) are shown. The responses to each antigen by HSV-1-seropositive donors (A) and HSV-2-seropositive donors (B) were compared to those by the seronegative donors by t tests, and the Benjamini-Hochberg-corrected P (pBH) values (shown overlaid onto the histograms) were used to classify the antigens into significant and nonsignificant responses (pBH values, <0.05 and ≥0.05, respectively).

Fig 4

Comparisons between HSV-1- and HSV-2-seropositive donors. Data are as described in the legend to Fig. 3, except that HSV-1- and HSV-2-seropositive donors were compared.

Fig 5

ROC plots for discriminatory HSV-1 and HSV-2 antigens. For HSV-1 antigens, only those that gave stronger signals with HSV-1-seropositive donors are shown. For HSV-2 antigens, only the top 10 are shown, although 29 antigens were discriminatory (Fig. 4). Antigens are ranked by AUCs, shown in parentheses. TPR, true-positive rate; FPR, false-positive rate.

Fig 6

ELISA data using two purified HSV antigens. (A and B) HSV-1 US8/gE; (C and D) HSV-2 UL44 gC. Sera are as described in the legend to Fig. 2, with serological status shown in panels A and C. Diagnoses for individuals marked with an asterisk were considered equivocal. Horizontal dashed line, average plus 2 SDs of the seronegative population minus the 6 equivocal samples. Scatter plots in panels B and D show ELISA ODs and corresponding microarray signal intensity (SI) for the seronegative population only. The six equivocal seronegatives are indicated by the open circles.