Parasite Immunology, 2010, 32, 20–28
DOI: 10.1111/j.1365-3024.2009.01152.x
Detection of the Sm31 antigen in sera of Schistosoma mansoni – infected patients from a low endemic area G. S. SULBARN,1 D. E. BALLEN,1 H. BERMDEZ,2 M. LORENZO,2 O. NOYA2 & I. M.CESARI1 1
Unidad de Trematodiasis, Centro de Microbiologa y Biologa Celular, Instituto Venezolano de Investigaciones Cientficas (IVIC), Caracas, Venezuela, 2Seccin de Biohelmintiasis, Instituto de Medicina Tropical, Caracas, Universidad Central de Venezuela (UCV), Caracas, Venezuela
SUMMARY
INTRODUCTION
Schistosoma mansoni cathepsin B (Sm31) is a major antigen from adult worms that circulates in the blood of infected patients (Li et al., Parasitol Res 1996; 82: 14–18). An analysis of the Sm31 sequence (Klinkert et al., Mol Biochem Parasitol 1989; 33: 113–122) allowed the prediction of seven hydrophilic regions that were confirmed to be exposed on the surface of a 3D model of Sm31; the species specificity of these regions was checked using BLAST analysis. The corresponding peptides were chemically synthesized in polymerazible forms using the t-Boc technique. Rabbits developed a high humoral response against these peptides as tested by a multiple antigen blot assay; it recognized native Sm31 in crude S. mansoni extracts and as circulating antigen in sera of S. mansoni-infected patients by western blot. Relevant antigenic determinants were located at the N- and C-terminus sequences. Antibodies against these regions recognized the native enzyme in an ELISA-like assay called cysteine protease immuno assay in which the immunocaptured enzyme was revealed by the intrinsic cathepsin B hydrolytic activity of Sm31. The method successfully and specifically detected Sm31 in sera of infected individuals, most of them (83Æ3%) with light infections, offering a rationale for the development of parasite enzyme capture assays using anti-synthetic peptide antibodies for possible use in the diagnosis of schistosomiasis.
Schistosomiasis is a parasitic disease caused by trematode worms of the genus Schistosoma affecting over 200 million people in tropical countries (1). The demonstration of schistosome eggs in faeces or urine of patients is so far the only unequivocal diagnosis of an active infection. This strategy may fail if the egg output is low like in low transmission areas or in old chronic infections. Many serological tests, mostly based on antibody detection, are available (2). Although providing in some cases high sensitivity and specificity, they cannot discriminate between active and past infections. This has led to a limited identification of active cases, necessary for drug treatment decisions and post-treatment following up (3). Recently, PCR has been proposed as an alternative method for the diagnosis of schistosomiasis; it has been tested to detect the infection in a murine model (4), mammalian hosts (5) and used with human sera, urine and stool samples (6–8). This strategy offered a diagnostic sensitivity of 94Æ4% and a specificity of 99Æ9% (6); however, some difficulties with this technology have arisen in areas of low transmission where the sensitivity of the technique seems to be diminished (7). Detection of circulating antigens by immunological means may be also an effective strategy to discriminate active infections. A variety of schistosome specific molecules are released into the host circulation by the adult parasite’s regular regurgitation of the gut content. Presence of adult worm antigens (AWAs) in serum and urine of mice and hamsters infected with S. mansoni was first reported by Berggren and Weller (9). Two major intestinal polysaccharide antigens have also since long been described: a negatively charged circulating anodic antigen and a positively charged circulating cathodic antigen (CCA) (10,11). Sensitive and specific immunoassays using monoclonal antibodies have been developed for the detection and quantification of these two antigens in serum and urine of patients (12–15). Although these methods work well in
Keywords circulating antigen, immunodiagnosis, Schistosoma mansoni, Sm31, synthetic peptides
Correspondence: Dr I. M. Cesari, Laboratoire de Parasitologie, Facult de Mdecine, Universit Libre de Bruxelles, CP 616, 808 route de Lennik, B-1070 Bruxelles, Belgium (e-mail:
[email protected]). Disclosures: None Received: 10 March 2009 Accepted for publication: 30 June 2009
20
2010 Blackwell Publishing Ltd
Volume 32, Number 1, January 2010
endemic areas with moderate and high prevalence of the disease, they are less sensitive in low infection areas (13,14). Among the gut antigens that enter host circulation through the adult parasite’s regurgitation is cathepsin B, a host-haemoglobin digesting cysteine protease (CP) antigenically known as Sm31 because of its mobility at 31kDa in SDS-PAGE (16,17). This prominent antigen whose sequence is fully known (18) has been employed for the diagnosis of schistosomiasis in antibody-detecting ELISA assays in endemic areas (19–21). The use of synthetic peptides has opened new perspectives for the diagnosis of diseases like malaria, Chagas disease, leishmaniasis and schistosomiasis (22–24). Previous work has shown that some linear and polymeric synthetic peptides from the Sm31 sequence are highly immunogenic in rabbits and are specifically recognized by antibodies of infected human sera (25). In the present study, we attempted to develop an antigen detection assay by immunocapturing Sm31 from the biological fluids of the infected hosts, using rabbit polyclonal antibodies raised against polymerized synthetic peptides selected from predicted hydrophilic regions of Sm31. The immunocapture event was revealed by taking advantage of the intrinsic enzymatic (CP SmCB1) activity of Sm31, a technical advantage already used in other enzyme antigen tests (26,27). The antibody-captured CP activity present in patients from low transmission areas (having parasitic loads lower than 100 eggs per gram (epg) of faeces) was detected by using the chromogenic substrate N-a-carbobenzoxy-phenylalanyl-arginylp-nitroanilide (CBZ-Phe-Arg-pNA) at pH 6Æ9.
Detection of circulating Sm31 antigen in sera
were homogenized in 50 mM Tris–maleate pH 6Æ9 containing 100 mM NaCl and 0Æ5 mM HgCl2 for 2 min over ice and centrifuged for 2 h at 100 000 g ⁄ 4C. The soluble phase supernatant (AWA) was distributed in aliquots and stored at )20C until used. Protein concentration was determined by the Bradford (31) method using bovine serum albumin as protein standard.
Predictive studies and selection of Sm31 peptide sequences The hydrophilic (epitopic) regions of the Sm31 sequence (18) were determined using the Hopp and Wood algorithm (32) (Figure 1). These regions were localized and visualized on a 3D-protein model of the Sm31 cathepsin B molecule and their surface accessibility confirmed. A BLASTp analysis (http://www.ncbi.nlm.nih.gov/blast/Blast.cgi) for each sequence was also performed to refine selection, avoiding as much as possible homology to peptide sequences of other proteins to diminish possibilities for cross-reactions. An additional criterion for selection was to select those Sm31 segments that are distant from regions probably involved with the active site. The selected linear segments were synthesized by a manual t-Boc strategy developed by Merrifield (33) and modified by Houghten (23) and Tam et al. (34). Extra N-terminal Cys–Gly and C-terminal Gly–Cys dipeptides were added to favour polymerization under oxidative (air) conditions. The selected peptides and their physico-chemical properties were calculated as in http://www.innovagen.se.
MATERIALS AND METHODS Rabbit anti-Sm31 peptide sera Sera
Adult S. mansoni worms (Venezuelan JL strain from IVIC) obtained from hamsters 6–7 weeks after infection
Two groups of New Zealand rabbits were used. In the first group, animals were immunized each with different homopolymers of single peptide sequences. As the obtained anti-sera were limited in their peptide recognition (see ‘Results’ section), a second group of rabbits was immunized with heteropolymeric mixtures of two or three different peptides (Table 1), the peptides of these mixtures were chosen according to the results of an immunoprecipitation assay (see below). Three immunizing doses were given to both groups at 0, 15 and 30 days. Each animal was injected subcutaneously with 0Æ5 mg ⁄ mL of peptide dissolved in 0Æ85% NaCl and mixed with Complete Freund Adjuvant (first injection), or Incomplete Freund Adjuvant (subsequent two doses). Negative (adjuvant alone) or positive (adjuvant plus AWA) immunized rabbits were included as controls. Ear bleedings were made before the first immunization and 10 days after the third immunization. Rabbits were bled under anaesthesia (ketamine chlorhydrate: 10 mg ⁄ kg) and their maintenance and manipula-
2010 Blackwell Publishing Ltd, Parasite Immunology, 32, 20–28
21
Sera were randomly selected from a group of sera taken from people living in the low transmission area of Beln (Carabobo State, Venezuela). All patients had stool examination by Kato–Katz (KK) (28–30) and triplicate slides were made from each stool specimen (n = 65, all age groups). Sera were divided into three groups: (i) negative to KK, ELISA and the alkaline phosphatase immunoassay (n = 20); (ii) negative to KK and the above serological tests, but positive for other parasitosis (n = 21); (iii) positive for S. mansoni in the KK test and by the above serological tests and negative for other parasitosis (n = 24); 83Æ3% of these samples corresponded to individuals with light infections ( 23 ⁄ 40 (492) > 13 ⁄ 40 (491) > 9 ⁄ 40
Figure 2 Recognition (IgG) of Sm31 synthetic peptides by S. mansoni-infected patient sera in multiple antigen blot assay. 1–24: a representative selection of infected patient sera from Beln (Venezuelan low endemic area); 25: conjugate control. 2010 Blackwell Publishing Ltd, Parasite Immunology, 32, 20–28
23
G. S. Sulbarn et al.
(487); peptides 488, 489 and 493 were weakly recognized or unreactive. Noninfected control sera were unreactive with the synthetic peptides and also with AWA (data not shown).
Inmunogenicity of the selected synthetic peptides in MABA Polyclonal rabbit anti-Sm31 synthetic peptide antibodies strongly and specifically recognized in MABA most of their respective homopolymeric peptide immunogens (Figure 3a). On the other hand, rabbit anti IMT-491 did not recognize peptide IMT-490 (comprising the sequences of IMT-491 and IMT-492 together), whereas rabbit anti IMT-492 did it. When mixtures of peptides (Table 1) were used, some epitopes appeared more dominant than others. Thus, the anti-487 (N-terminal) ⁄ 493 (C-terminal) recognized 487 strongly (dominant N-terminal region), but weakly the peptide 493; anti-488 ⁄ 493 recognized 488 strongly, but peptide 493 was not recognized. Anti IMT492 and anti IMT-493 reacted with AWA (containing native Sm31) only after repeated boosting, whereas none of the rabbits immunized with the peptide mixtures did react with this preparation (Figure 3b). Rabbit preimmune sera did not recognize these peptides (data not shown).
Parasite Immunology
Recognition of original Sm31 in AWA by rabbit anti-peptide antibodies in western blots Contrary to results obtained in MABA (Figure 3a, b), heteropeptide-immunized rabbit sera clearly recognized Sm31 in AWA when tested by western blot (WB) (Figure 3c). However, a 31-kDa band was recognized in WB both of infected and uninfected patient sera (data not shown). Serum from rabbit injected only with adjuvant did not recognize this band.
Effect of anti-peptide antibodies on the Sm31 cathepsin B activity Pre-immunized (without specific anti-Sm31 peptide antibodies) and post-immunized (with anti-Sm31 peptide antibodies) rabbit sera did not inhibit more than 21% the CP activity measured in AWA, indicating low presence of inhibitory antibodies, as should be expected from the selection of peptides from nonactive site regions (data not shown).
Immunocapture of CP by rabbit anti-Sm31 sera Only rabbits immunized with the IMT-487 ⁄ 493 mixture (corresponding to the N- and the C-terminal domains,
Figure 3 Recognition (IgG) of Sm31 synthetic peptides by anti homo- and anti heteropeptide rabbit sera in multiple antigen blot assay. (a) Anti-homopeptide rabbit sera: (1, 2) anti-493; (3, 4) anti-adult worm antigens (AWA); (5, 6) anti-489; (7, 8) anti-488; (9, 10) anti-487; (11) anti-491; (12) anti-492; (13, 14) serum of rabbit injected only with adjuvant. (b) Anti-heteropeptide rabbit sera: (1) anti-487 ⁄ 488 ⁄ 489; (2) anti-487 ⁄ 493; (3) anti-488 ⁄ 493; (4) anti-487 ⁄ 488 ⁄ 493; (5) anti-489 ⁄ 491 ⁄ 493; (6) serum of a rabbit injected only with adjuvant. (c) Recognition (IgG) of Sm31 in AWA by peptide-immunized rabbit sera as assessed by western blot: (1) anti-487 ⁄ 488 ⁄ 489; (2) anti-487 ⁄ 493; (3) anti488 ⁄ 493; (4) anti-487 ⁄ 488 ⁄ 493; (5) anti-489 ⁄ 491 ⁄ 493; (6) serum from rabbit injected only with adjuvant; (7) conjugate control.
24
2010 Blackwell Publishing Ltd, Parasite Immunology, 32, 20–28
Volume 32, Number 1, January 2010
Detection of circulating Sm31 antigen in sera
Figure 4 Cathepsin B immunocapture assay with rabbit antiIMT487 ⁄ 493 antibodies. The cut-off was obtained from the average captured activity of sera from healthy donors € 2 SD. Values of captured cathepsin B in human sera: (1) negative by Kato– Katz (KK) and serological tests (n = 20); (2) negative by KK and serological tests but positive for other parasitosis (n = 21); (3) positive only for S. mansoni by the KK and serological tests and negative for other parasitosis (n = 24).
useful tool for diagnosis (26,27,41). Cathepsin B, like other CP, has been highly conserved through evolution and the human and S. mansoni cathepsins B share 58% homology (18); nonetheless, it has been reported that host responses to this parasite enzyme are quite specific (19). On this basis, several criteria were used to identify and select epitopic linear segments of Sm31 to be chemically synthesized. A predictive analysis of the hydrophilic (epitopic) regions according to Hopp and Wood (32) was first performed on the sequence published by Klinkert et al. (18). This algorithm produced an initial series of possible theoretical sequence regions to be synthesized. A BLASTp analysis was then performed to restrain to more specific S. mansoni domains. These domains were then visualized in a 3D-model of Sm31; those loops that appeared as more exposed to solvent and far from the active site region were targeted. Seven linear sequences were thus finally selected and the corresponding peptides synthesized (Figure 1). To favour polymerization, Cys–Gly was added to their N-terminus and Gly–Cys to the C-terminus thus disulfide bonds could easily form in aqueous media under oxidative conditions. The resulting peptide solutions were sterilized and injected into rabbits to produce different anti-Sm31 peptide sera. Antibodies raised against the homopolymeric peptide preparations were able to recognize the original Sm31 molecule in WB, but not to capture it on a flat solid phase surface as that of an ELISA plate (data not shown), indicating either conformational modifications upon adsorption to the plastic surface or possibly that several binding sites are required to fix Sm31; it might be also more probable that the amount of Sm31 fixed in the plates was too low. Absence of catalytic activity in negative results with the homopeptide antisera was probably not because of the blocking of the enzymatic activity by allosteric inhibition of the catalytic site because the Sm31 segments for synthesis were purposely chosen distant from the active site. This was otherwise confirmed by the lack of inhibition of the positive reactions with the corresponding anti-peptides sera (data not shown). On the other hand, heteropolymeric peptide mixtures induced stronger antibody responses in rabbits, some mixtures appearing better antibody inducers than others (Figure 3a, b); these mixtures (Table 1) were formulated according to their higher immunoprecipitating properties towards Sm31 molecule. Anti IMT-487, -488, -489, and 493 sera showed good recognition of the corresponding peptides in MABA (Figure 3a, b) and the anti-peptides mixtures recognized the original Sm31 present in AWA by MABA and WB (Figure 3c) confirming that the selected Sm31 surface-exposed epitopes were accessible to the B cells receptors; however, they recognized also a 31-kDa band (presumably a cathepsin B-like polypeptide) in both
2010 Blackwell Publishing Ltd, Parasite Immunology, 32, 20–28
25
respectively) produced antibodies that were able to capture, in a solid phase matrix, CP from S. mansoni – infected patient sera (Group 3, mean € SEM = 0Æ19 € 0Æ13; n = 24). Captured CP activity was significantly lower in healthy patient sera (Group 1, mean € SEM = 0Æ05 € 0Æ03; n = 20) and only a few patients with others parasitic diseases (Group 2, mean € SEM = 0Æ08 € 0Æ05; n = 21) showed values above the cut-off (Figure 4). Differences between groups 3 and 1 were highly significant by the t (P £ 0Æ001) and the median (P £ 0Æ008) tests, but not between group 2 and 1 (paired t-test, two tails) (Figure 4). Sensitivity by cysteine protease immuno assay (CPIA) was 70Æ8% and specificity 80Æ5%. The predictive positive value was 68%, the predictive negative value 82Æ5%, and the efficiency 76Æ9%. Although some values from group 2 were above the ‘cut-off’ (Figure 4), their mean (0Æ08 € 0Æ05, n = 21) was not significantly different (P = 0Æ06) from that of healthy donors (Group 1). The kappa agreement between CPIA and KK (gold reference test) was positive (0Æ508 € 0Æ22). There was also a positive correlation between CPIA results and the presence ⁄ absence of infection as assessed by KK (Spearman rho = 0Æ69); however, no correlation was found between amount of activity and worm burden (as expressed by egg output) (data not shown).
DISCUSSION In the present work, an assay was developed to detect circulating S. mansoni cathepsin B (Sm31 antigen) in human sera as indicator of active infection by this trematode. The use of enzyme antigens has been reported previously as a
0·5
nmol/h
0·4 0·3 0·2 Cut off
0·1 0·0 1
2 Groups
3
G. S. Sulbarn et al.
Parasite Immunology
S. mansoni-infected and uninfected human sera (data not shown). Thus, in spite of all the different and complementary criteria used to select specific Sm31 peptide epitopes, we confronted specificity problems in WB analyses. This could be partially because some short sequence segments, bound to a solid matrix like NC, do not probably have the same specificity than in a 3D conformation, but may probably cross-react with similar proteins migrating at the same site. The CPIA was performed using biochemical incubation conditions favouring the parasite enzyme activity rather than the host cathepsin B activity with CBZ-Phe-Arg-pNA as substrate at pH 6Æ9. These conditions decreased the background activity caused by the captured host enzyme, as seen in WB experiments with anti-488 ⁄ 493 and infected ⁄ uninfected human sera (data not shown). To our knowledge, no anti-Sm31 antibodies have been shown to inhibit the catalytic activity of Sm31 and anti-Sm31 peptide sera tested in vitro with the soluble S. mansoni AWA extract having cathepsin B activity did not show significant inhibition of this activity. Thus, inhibitory antibodies would not seem to limit the usefulness of any CPIA assay. However, the capacity of the CPIA-negative sera (group # 3, Figure 4) to inhibit the Sm31 cathepsin B activity was not tested. It worth noticing that this study was conducted with serum samples of patients with low circulating levels of parasite components (light infections) having high antiparasite antibody titres that do not correlate to the parasite load. Presumably, CPIA should be more sensitive with sera of individuals having high parasite loads where the parasite ⁄ antigenic load ratio is higher, thus favouring immunocapture and detection of Sm31. Sm31 capture in CPIA worked only with anti IMT-487 (N-terminal region) and anti IMT-493 (C-terminal domain) sera; these antisera discriminated between infected and healthy patient sera (Figure 4). The CPIA results would seem to indicate that there is a mixed population of anti-cathepsin B antibodies in the antisera, some expressing more specificity for Sm31 than others. Interestingly, both the N- (487) and the C- (493) domains, highly immunogenic in rabbits, were poorly antigenic in humans (Figure 2). Although the CPIA sensitivity (70Æ8%) and specificity (80Æ5%) were low, the assay significantly discriminated between sera of S. mansoni-infected patients from low endemic areas (expulsing
