Application of synthetic peptides to the diagnosis of neurocysticercosis

Introduction

Neurocysticercosis (NC) is the most common worldwide parasitic disease of the central nervous system and its prevalence is highest in developing countries (Del Brutto et al. 1998). In recent years, human cases of NC have emerged in developed countries, such as the USA and Europe, mostly because of migrant workers (Shandera et al. 1994; Rousseau et al. 1999; Carangelo et al. 2001; Terraza et al. 2001). Accurate diagnosis of NC still requires cranial imaging technology such as computerized tomography (CT) and/or magnetic resonance imaging (MRI), techniques that are expensive and frequently inaccessible to people of endemic areas (Sciutto et al. 2000). The diagnosis of NC without imaging studies is difficult because of the highly variable clinical symptoms, ranging from asymptomatic to a very severe neurological syndrome with intracranial hypertension or dementia, with a variety of non-specific mild clinical symptoms in between.

For immunological diagnosis, the enzyme-linked immunosorbent assay (ELISA) and the enzyme-linked immunotransfer blot (EITB) are the most widely used techniques (Rosas et al. 1986; García et al. 1995). Although the initial data suggested that the EITB using parasite glycoproteins isolated by lectin-affinity chromatography was highly specific and sensitive (Tsang et al. 1989), results from subsequent studies were less encouraging (Ordoñez et al. 1996). One of the major disadvantages of almost all reported immunodiagnostic methods is the poor reproducibility between different laboratories, presumably because of differences in the clinical status of patients examined and variability in the nature and purity of the parasite antigen extracts employed. There is therefore an inherent advantage of reproducibility in diagnostic tests based on recombinant proteins or synthetic peptides (Gevorkian et al. 1996; Hernández et al. 2000; Ferrer et al. 2003). Such a permanent source of diagnostic reagents would also be useful in traditionally non-endemic countries at risk of sporadic incursions of the parasites, or where the collection of T. solium cysticerci is difficult.

Recently, synthetic peptides based on antigenic sequences identified in immunogentic oncosphere proteins of Taenia saginata have been shown to function well in the diagnosis of bovine cysticercosis (Ferrer et al. 2003), and hence we evaluated them for the diagnosis of human NC using samples from Mexico and Brazil. The five peptides consist of: one peptide (HP6-3) derived from the sequence of the 18 kDa T. saginata surface/secreted oncospheral adhesion antigen identified by McAb-HP6, two peptides (Ts45W-1 and Ts45W-5) derived from the sequence of the T. saginata homologue of the T. ovis 45W protective gene family, one peptide (TS45S-10) derived from a T. saginata sequence with significant similarity to the T. ovis 45S protective antigen, and one peptide (TEG-1) derived from the sequence of the T. saginata homologue of Echinococcus spp. main surface protein. The corresponding proteins, which may also be potential protective antigens, consist of: the 18 kDa T. saginata surface/secreted oncospheral adhesion antigen identified by McAb-HP6 (Benítez et al. 1996; Bonay et al. 2002; Ferrer et al. 2003), the T. saginata homologue of the T. ovis 45W protective gene family (Johnson et al. 1989; Waterkeyn et al. 1995; Ferrer et al. 2003), the T. saginata homologue of T. ovis 45S protective antigen (Waterkeyn et al. 1995; Ferrer et al. 2003) and the T. saginata homologue of Echinococcus spp. surface protein (Benítez et al. 1998).

Material and methods

Cerebrospinal fluid

Cerebrospinal fluid (CSF) samples were collected in Mexico and Brazil. The Mexican samples formed three clinically distinct groups: 27 CSF from patients (12 men and 15 women, mean age 40 ± 13 years) with radiologically diagnosed NC taken before any treatment, 16 CSF from patients with clinically diagnosed NC taken 1–12 months after specific drug treatment (albendazole +prednisona) and 31 CSF from non-NC control patients. We also collected and tested 11 CSF of the 27 NC patients 1–8 months after cysticidal treatment (albendazole + prednisona).

The CSF samples from Brazil were obtained by sub-occipital puncture at the Neurological Department of the Medicine, University of Botucatu and consisted of 24 CSF samples from control, non-NC neurological patients and 32 from diagnosed NC patients (24 men and eight women, mean age 38.6 ± 9.4 years), 1–360 days after drug treatment (albendazole).

Clinical details and characteristics of each CSF sample, including the number of cells, were obtained from the medical records. The diagnosis of NC, as determined by neuroimaging, was classified as ‘active’ when at least one viable cysticercus or arachnoiditis was present, and cases were classified as either single- or multiple-viable cysts. Alternatively, when neuroimaging studies only detected calcifications or meningeal fibrosis, which are the permanent sequelae of previous infections (Sotelo et al. 1985), the disease was classified as ‘inactive’. Active cysticercosis patients were further classified as ‘inflammatory’ or ‘non-inflammatory’ by counting mononuclear cells in their CSF; inflammatory being defined as having more than six mononuclear cells per cubic millimetre. Controls were patients for whom the diagnosis of NC had been excluded by neuroimaging, and who were seen in neurological institutes for inflammatory, vascular, infectious and degenerative disease.

Peptides

The peptides (Ferrer et al. 2003) examined in this study and described in the introduction are summarized in Table 1. Peptides were synthesized by stepwise solid-phase synthesis with N^α-tert-butyloxycarbonyl (BOC) derivatives of l-amino acids on phenyl-acetamidomethyl (PAM) resin (Sigma Chemical Co, St Louis, MO). The peptides were 95% pure as judged by high-performance liquid chromatography (HPLC) on analytical C₁₈ reversed-phase column (3.9 × 150 mm; Delta Pak, Waters). The correct amino acid sequence of each peptide was confirmed by protein sequencing on a pulsed liquid-phase protein sequencer (Applied Biosystems) at the National Institute of Cardiology, Mexico City. Peptides were coupled to maleimide activated ovalbumin (OVA) and keyhole limpet haemocyanin (KLH) (Pierce Warriner Ltd.) following the manufacturer's instructions.

Antibody detection by ELISA

Cerebrospinal fluid antibody titres were assessed by indirect ELISA using either T. solium cyst fluid or the peptides as antigens as previously described (Ramos-Kuri et al. 1992). Plates were set up with duplicate sample determinations. The assays were routinely repeated to check for reproducibility. A sample was considered positive if the specific optical density (OD) values were greater than the mean of the values obtained with negative CSF samples plus two standard deviations (SD). Five to 10 negative CSF from confirmed non-NC neurological patients different from those included in the study to determine the specificity were used in each plate to determine the cut-off value.

Statistical analysis

Sensitivity was calculated as the number of proven NC cases that were ELISA positive, and specificity as the number of negative control samples that were ELISA negative. Dichotomous data were compared using the chi-square Pearson test or the Yates corrected chi-square test. Differences in mean values between groups were analysed by Student's t-test or by a non-parametric test (Mann–Whitney).

Results

The five peptides were first tested in a free unconjugated form and the most obvious conclusion is that the peptides function best for antibody detection in cases of active inflammatory NC (Table 3), especially in patients aged 30–50 years (Table 2).

As different peptides do not necessarily bind with equal efficiencies to ELISA plates, as an attempt to standardize conditions, the five peptides were coupled to protein carriers, OVA and KLH. These preliminary experiments indicated that there was no significant difference in the results obtained using free simple peptides or peptides coupled to the carrier proteins (Table 4).

These peptide carrier protein conjugates were used as probes for the Mexican and Brazilian CSF samples. Employing Mexican active inflammatory NC, the sensitivity using Ts45W-5 and HP6-3 coupled to KLH or Ts45W-1, HP6-3, TEG-1 and Ts45S coupled to OVA was as good as that obtained using cyst fluid antigen. Similarly, with the Mexican active non-inflammatory NC, peptide HP6-3 OVA gave the same sensitivity as cyst fluid antigen (Table 5). Finally, using Ts45W-1, Ts45W-5 and HP6-3 coupled to OVA with the Brazilian active inflammatory NC, the same sensitivity was obtained as with cyst fluid antigen (Table 5).

As immune responses to parasites are characteristically highly heterogeneous, the effect of combining the results from certain individual peptide assays was examined. With the Mexican samples, a combination of HP6-3 OVA and Ts45W-1 OVA gave a sensitivity and specificity quite similar to those obtained with cyst fluid (Tables 5 and 6). Similarly, with the Brazilian samples, combinations of Ts45W-5 OVA and Ts45S OVA or TEG-1 OVA also indicate an improved sensitivity (Table 5) with a good specificity (Table 6), which were between 93% and 100%. Interestingly, cysticidal and corticosteroid treatment did not significantly modify CSF levels of antibodies or sensitivity of the assay employing the different peptides as a source of antigen (Table 7).

Discussion

Diagnosis of human, bovine and porcine cysticercosis still relies on the detection of antibodies using parasite extracts or multi-component glycoprotein fractions. The use of sufficiently sensitive and specific recombinant proteins or synthetic peptides as antigens in such assays would provide the much needed reproducibility and circumvent reliance on the less reproducible parasite extracts. Five peptides of high antigenic indices from the sequences of four potentially protective T. saginata proteins already shown to have utility in the diagnosis of T. saginata cysticercosis (Ferrer et al. 2003) were tested as possible reagents for the diagnosis of human neurocysticercosis. It is worth mentioning that the HP6-3 peptides were taken from the sequence of a oncospheral T. saginata antigen with proven protective capacity against bovine T. saginata cysticercosis (LJS Harrison and RME Parkhouse, unpublished data), and will be of interest to determine the relevance of homologous anti-HP6-3 antibodies in the human disease.

Most of the data presented was obtained with peptide–protein conjugates. As seen in Table 4 these peptides gave comparable results in soluble form or coupled to proteins (OVA or KLH). Peptides HP6-3 and Ts45W-1 coupled to OVA detected the majority of the active inflammatory NC patients [83.3% (Figure 1) and 88.9% (Figure 2) of the Mexican, respectively, and 85% of the Brazilian]. Notably, as active NC was mainly found in patients between 30 and 50 years of age, the test could be more effective in this age group. The fact that the results obtained with the Mexican samples mirror those of the Brazilian samples suggests that similar antigenic epitopes are shared by T. solium from these two countries. More than 45% of these patients also recognized the other four peptides that were tested, a fact that points to a low variability of cysticercal antigens and an important cross-immunity between cestodes. A combination of two peptides [HP6-3 (OVA) and Ts45W-1 (OVA)] reached a sensitivity of 100% and a specificity as good as that obtained using cyst fluid and thus merits serious consideration for the design of a diagnostic kit.

In summary, this study has identified peptides of potential use in NC diagnosis. Simple and reproducible antibody detection assays for NC would find utility in hospitals and in epidemiological studies in endemic areas and would also be a useful resource in non-endemic countries to determine possible exposure to the parasite.