A T‐cell diagnostic test for cystic echinococcosis based on Antigen B peptides

Summary Cystic echinococcosis (CE) immunodiagnosis is still imperfect. We recently set‐up a whole‐blood test based on the interleukin (IL)‐4 response to the native Antigen B (AgB) of Echinococcus granulosus. However, AgB is encoded by a multigene family coding for five putative subunits. Therefore, the aims of this study were to analyse the IL‐4 response to peptides spanning the immunodominant regions of the five AgB subunits and to evaluate the accuracy of this assay for CE diagnosis. Peptides corresponding to each subunit were combined into five pools. A pool containing all peptides was also used (total pool). IL‐4 evaluated by enzyme‐linked immunosorbent assay was significantly higher in patients with CE compared to those without (NO‐CE subjects) when whole‐blood was stimulated with AgB1 and with the total pool. Moreover, IL‐4 levels in response to the total pool were significantly increased in patients with active cysts. Receiver Operator Curve analysis identified a cut‐off point of 0.59 pg/mL predicting active cysts diagnosis with 71% sensitivity and 82% specificity in serology‐positive CE patients. These data, if confirmed in a larger cohort, offer the opportunity to develop new diagnostic tools for CE based on a standardized source of AgB as the peptides.

imaging-inactive cysts are biologically viable when CE4 stage is the result of treatment. [9][10][11] Therefore, improved diagnostic systems and identification of biomarkers for CE diagnosis and follow-up are needed.
Among E. granulosus compounds, antigen B (AgB), one of the most abundant antigens of hydatid fluid, has been extensively studied. [12][13][14][15][16][17][18] AgB is a multimeric protein of 8 KDa subunits encoded by a multigene family. [19][20][21] To date, five subunit genes have been identified 22 ; however, several aspects of AgB-subunit composition and oligomeric structure are poorly characterized. Moreover, AgB subunits could be differentially expressed within individuals and/or throughout the parasite life cycle. 23 Therefore, the composition of native-AgB is uncertain and may have a high degree of variation influencing its use in diagnosis. In such scenario, it is essential to assess the immunogenicity and the diagnostics potentials of each AgB-subunit.
So far, studies on the diagnostic value of recombinant AgB subunits have been focused on AgB1 or AgB2. 24,25 The performance of all AgB subunits for serological diagnosis of CE was recently evaluated with a reported reactivity order of AgB1 > AgB4 > AgB2 > AgB5 > AgB3. 26 Several studies assessed the antigenicity of selected synthetic peptides derived from AgB1 and AgB2 subunits and among them, p65, p89-122, p176, Gu4 were promising as diagnostic reagents. [27][28][29] Synthetic peptides have been used as antigens for the diagnosis of parasitic diseases as malaria, leishmaniasis and schistosomiasis. 30 Indeed, they show several advantages over native or recombinant antigens being well characterized, easily standardized and producible in large amounts, cost-saving, highly pure and endotoxin-free.
Alternative laboratory tests for the diagnosis of chronic infections as tuberculosis or toxoplasmosis are based on cytokines detection upon stimulation of whole blood with pathogen-derived molecules, including antigenic peptides. [31][32][33][34] We recently set-up a whole-blood assay based on interleukin (IL)-4 detection in response to native-AgB which showed a good accuracy for CE diagnosis and staging. 12 Thus, we set-up a pilot study to: (i) analyse the IL-4 response to multiepitope synthetic peptides spanning the immunodominant regions of the 5 AgB isoforms and (ii) determine if the selected peptides may increase the diagnostic accuracy of the whole-blood assay for CE diagnosis and follow-up. Patients with cysts in the liver or any other location were included (Table 1). Hydatid cysts were staged according to WHO classification. 35,36 Patients were classified as having "active cysts" (CE1, CE2), "inactive cysts" (CE4, CE5) and "transitional-cysts" (CE3a, CE3b). Patients having multiple cysts were classified according to the more active stage. 9 Metacestodes viability was not performed in our Institutions as the common procedures to evaluate it are limited by practical and ethical reasons and proton magnetic resonance spectroscopy was not available. A previous correlation of the cyst stages (defined by US) with their biological viability was considered to classify patients with CE. 9 Based on this evaluation, CE1, CE2 and CE3b are considered viable cysts, CE3a has equal probability of being viable or non-viable, and inactive cysts (CE4 and CE5) are usually non-viable, even if some CE4 cysts have been described as viable. 9 "NO-CE" subjects, included as controls, were healthy volunteers or subjects enrolled with suspected CE who had a CE diagnosis excluded by US, by clinical evaluation and serology.

| AgB synthetic peptides selection
Sequences of the "E. granulosus" (G1) AgB subunits were obtained by search in Genebank database. Ninety-nine full-length and partial sequences were obtained as follows: 18 for AgB1, 32 for AgB2, 18 for AgB3, 20 for AgB4 and 11 for AgB5. Consensus sequence for each AgB-subunit was generated by multiple alignment, as described. [37][38][39] The consensus sequences of the AgB subunits shared a large de-

| Stimuli
Peptide pools were used at 1 or 10 ug/mL peptide; native-AgB at

| IL-4 response to AgB1 subunit pool and total peptide pool is significantly associated with CE
To evaluate the most appropriate experimental setting for detecting IL-4-specific response, stimulation with several concentrations of AgB peptide pools was assessed in 13 patients with CE. The highest IL-4 production was measured in whole-blood stimulated with the AgB1 pool, the AgB2 pool and with the total pool at 1ug/mL, or with the AgB4 pool at 10ug/mL (data not shown). The AgB3 and the AgB5 peptide pools did not induce IL-4 production, and therefore, they were excluded from the subsequent analysis (data not shown).
Due to the limited amount of blood recovered, the initial analysis was performed on 34 patients with CE and 22 NO-CE subjects ( Figure 1). AgB1 pool or AgB total pool induced high levels of IL-4, and a significant difference was found between the patients with CE and the NO-CE subjects (P = .003, P = .007 respectively; Figure 1A).
Therefore, we performed a ROC analysis to evaluate the potential use for CE diagnosis of the whole-blood assay based on AgB1 pool or AgB total pool. We found significant results in the area under curve for AgB1 (AUC) analysis (AUC, 0.73; 95% confidence interval (CI), 0.59-0.86, P = .004) ( Figure 1B). For scoring purposes, a cut-off value was chosen to maximize the sum of sensitivity and specificity. The AgB1 pool cut-off point of 0.27 pg/mL predicted CE with 35% sensitivity (95% CI, 19.8%-53.5%) and 100% specificity (95% CI, 84.6%-100%; We compared the accuracy for diagnosing CE of the wholeblood test based on AgB total pool to that based on the native-AgB. As shown in Figure 2, the IL-4 levels in response to native-AgB ( Figure 2A), as demonstrated, 12 and to the AgB total pool ( Figure 2C) were significantly associated with CE (P < .0001 and P = .005, respectively) in all the subjects enrolled ( (Table 3a).
Overall, these data demonstrate that the AgB1 subunit is the more immunogenic AgB protein in our whole-blood assay. Moreover, the IL-4 response induced by the AgB total pool is similar to that induced by the native-AgB.

| IL4-response to the AgB total pool is significantly associated with active cysts
We evaluated if the whole-blood test based on the AgB peptide pools is useful for assessing cyst biological viability. We stratified the patients with CE according to their cyst viability: "active-cysts" group, including CE1, CE2 and CE3b, and the "inactive-cysts" group including CE4 and CE5. The CE3a cysts group was excluded from the analysis due to its small sample size. The initial analysis was performed on 17 patients with active cysts and 16 patients with inactive cysts.
Increased IL-4 levels were found between the "active" and the "inactive-cysts" groups (P = .02) in response to AgB total pool ( Figure 3).
Based on this result, we compared the accuracy for diagnosing active cysts of the whole-blood test based on AgB total pool or native-AgB in all the patients enrolled. IL-4 levels in response to native-AgB ( Figure 4A) and to the AgB total pool ( Figure 4C) were significantly associated with active cysts (P = .01 and P = .006, respectively).

| Agreement between experimental whole-blood results and serology
We evaluated the agreement between the whole-blood assay based on native-AgB or AgB total pool and the commercial serology tests used for CE diagnosis.
The agreement between the whole-blood test based on the native-AgB and the serology in all the enrolled subjects was fair (k = 0.4; P = .002) (Table S1). These results depend on the high proportion of whole-blood negative results (62.5%), mainly in patients with inactive cysts as previously found. 12 Considering all the patients with CE and NO-CE subjects, the agreement between the experimental test based on AgB total pool and the commercial serology was moderate (k = 0.5; P < .0001; cyst stage, in patients with active cysts, the agreement between the two tests was moderate (k = 0.5; P = .05, Table 4), whereas among patients with inactive cysts, it was fair and not significant (Table 4).
Therefore, these data suggest that the results of a positive serology test coupled with a negative whole-blood test are likely indicative of inactive CE cysts stage. Indeed, the sensitivity/specificity of the  whole-blood test based on the AgB total pool was calculated, and in serology-positive patients, a negative whole-blood has a specificity of 71% for inactive-cysts diagnosis; moreover, in serology-negative patients, a positive whole-blood result has a specificity of 78% for activecysts diagnosis (Table 5).

| IL4-response to the AgB total pool identifies active cysts among CE patients with a positive serology
In our study, 26 of 43 (60.5%) patients with CE were serology positive; among them, 14 had active cysts, one had CE3a cysts, and 11 had  inactive cysts. As mentioned before, CE3a samples or samples with follow-up <4 years were excluded from the analysis. Therefore, the analysis was carried out on 25 patients with CE.
IL-4 response to the AgB total pool was significantly higher in patients with active cysts compared to patients with inactive cysts (P = .03, Figure 5A). The ROC analysis between the "active" and "inactive cysts" groups showed significant AUC results (AUC, 0.76; 95% CI, 0.56-0.96, P = .03; Figure 5B). The cut-off point of 0.59 pg/mL predicted active cysts with 71% sensitivity (95% CI, 41.9%-91.6%) and 82% specificity (95% CI, 48.2%-97.7%; Table 2). According to this cut-off, we found a significantly higher proportion of positive results among patients with active cysts (10 of 14; 71.4%) compared to those with inactive cysts (two of 11; 18.2%; P = .008). The accuracy of the serology for activecysts diagnosis was also evaluated, showing a 78% sensitivity and 45% specificity (Table 3a). Therefore, the whole-blood test based on the AgB total pool may be useful, after a positive score to the serology, to identify the patients with active and biologically viable cysts.

| DISCUSSION
CE diagnosis is based primarily on imaging examination and serological tests which, unfortunately, lack sensitivity 17,18 and do not provide information regarding the cyst biological viability. Recently, we showed that a T-cell-based test based on the detection of the IL-4-response to native-AgB may have good potential for CE diagnosis and staging.
Native or recombinant antigens are powerful reagents for serology, because they contain a large spectrum of epitopes, which whilst guarantee their recognition theoretically by all individuals, may cause cross-reactions, compromising the specificity of the system. AgB is a 8KDa multimeric protein, and five subunits have been identified. So far, the immunogenicity and immunodiagnostic accuracies of AgB1 and AgB2 have been evaluated demonstrating that the AgB2 subunit provided the highest diagnostic sensitivity and specificity. 24,25 However, other studies reported a higher immunogenicity for the AgB1 subunit. 26,42 In line with these results, we found that AgB1 pool induced high levels of IL-4 and a significant difference was found between the patients with CE and the NO-CE subjects demonstrating that AgB1 is the most immunogenic AgB protein in our wholeblood assay. Moreover, as previously shown 26 the AgB3 or the AgB5 pools were poorly immunogenic. The low antigenicity of AgB3 may be related to its biochemical properties, 23  and CE1 cases, the immune-specific response is low, there is a low likelihood that the serology or the whole-blood test score positive.
On the other hands, as previously shown also for the native-AgB, 12 a positive serology score coupled to a negative score to whole-blood is associated with inactive cysts. In conclusion, although the diagnostic accuracy of the wholeblood test based on AgB synthetic peptides for diagnosing CE is suboptimal, this proof of concept contribute to generate improved and alternative tools using a standardized source of antigen for CE diagnosis and monitoring. Moreover, we showed the basis for the combination of a diagnostic algorithm based on serology followed by the whole-blood test based on AgB synthetic peptides that is associated with active-cysts identification. Further studies will validate this finding. Responses were compared using the Mann-Whitney test; differences were considered significant at P-values of ≤ .05