b Ospea ale Ciuile di Gemona, Gemona, Italy

Transcription

1 ELSEVIER FEMS Immunology and Medical Microbiology 14 (1996) /f~d~blf~ogy MICROBIOLOGY AND Abstract IgM and IgG significant reactivity to Borrelia burgdorferi sense stricto, Borrelia garinii and Borrelia afzelii among Italian patients affected by Lyme arthritis or neuroborreliosis Marina Cinco a, *, Rossella Murgia a, Maurizio Ruscio b, Barbara Andriolo a a Istituto di Microbiologia, Uniuersith degli Studi, via Fleming 22, Trieste, Italy b Ospea ale Ciuile di Gemona, Gemona, Italy Received 26 February 1996; accepted 12 March 1996 This survey evaluates the specificity of band patterns in immunoblot of sera taken from clinically defined cases of Lyme arthritis and neuroborreliosis, towards three locally isolated strains of Borreliu burgdorj2ti, belonging to the three species: Borrelia sensu stricto, Borrelia garinii and Borrelia afzelii. To assess specificity, patient sera were statistically (x2, P ) compared with blood donors sera samples. Both IgG and IgM antibodies were considered. The overall reactivity of the three Borrelia strains in IgG immunoblots indicated that ten protein bands were significant, with a different prevalence of some of them in the two groups of patient sera: bands at 60-58, 30-33, and kda were markers for neuroborreliosis sera; proteins at , and kda behaved like markers for Lyme arthritis. The JgM Immunoblots revealed significant bands at , 72-70, 51, and kda only with neuroborreliosis sera. Though there were variable band reactivities in each strain, a correlation emerged between the three genospecies and the clinical symptoms: in fart B. ufzelii and B. garinii were prevalent in Lyme arthritis sera, (IgG Immunoblots); B. garinii was associated to neurobotreliosis (IgG and IgM Immunoblots); B. sensu strict0 was strongly reactive with neuroborreliosis in IgM immunoblots. These data indicate that the three locally isolated strains of Borrelia representing the three genospecies should be used together in immunoblot to detect antibodies elicited in neuroborreliosis and Lyme arthritis. Keywords: Borrelia burgdorferi; Immunoblot; Significant bands; Arthritis: 1. Introduction The diagnosis of ILyme disease (LD), a chronic multisystem disorder caused by the spirochete Borr&a burgdmjkri, is primarily made by clinical criteria, but since multiple differential diagnoses have to * Corresponding author. Tel: +39 (40) ; Fax: +39 (40) ; cinoa@univ.trieste.it Neuroborreliosis be considered, laboratory confirmation is required in ambiguous situations. For this purpose antibody detection is the most common procedure. Immunoblotting (IB) is unanimously reported to be a confirmatory test for Lyme borreliosis, even if there are still variable factors affecting the performance of this assay. One of these factors is the polymorphism of Borrelia antigens which is evident among the species and intraspecies. In fact several reports are concerned with remarkable differences in seroreactivity 0928_8244/Federation of European Microbiological Societies PII SO (96)00027-Z

2 160 Marin Cinco et al./fems Immunology and Medical Microbiology 14 (1996) patterns of American and European Lyme borreliosis patients with the North American strain B31 of B. sensu stricto [l]. More recently sera from neuroborreliosis [2,3], arthritis and ACA patients were found to react preferentially against B. garinii, B. sensu strict0 and B. ufzelii respectively. Further some authors [4] demonstrated that the reliability of serological investigations (both Enzymatic Immunoassay and IB) increases when antibodies were measured against the antigens of Borrelia spp. occurring in a particular geographic area. Accordingly we have reported [5] that we assayed sera of Lyme borreliosis cases at different stages and manifestations, collected in a restricted area of Italy, with locally isolated strains each representing a B. burgdorferi species [6]. We found an overall preferential band IgG reactivity to the species B. ufielii and B. garinii, though only a limited number of sera have been examined. In order to improve the performance and interpretation of the IB assessed in our laboratory, in this investigation we analyzed further the band patterns obtained with IB, focusing on sera of arthritis and neuroborreliosis patients, which occur more commonly in our region as late Lyme disease manifestations. We also carried out the following: (a> we compared the band patterns of sera from definite cases of neuroborreliosis and arthritis with sera of healthy people, to identify the more specific bands and the band marker prevalent in each of the two clinical manifestations; (b) we evaluated the reactivity of each strain-antigen used alone or in association; and (c) we studied the involvement of antibody IgG or IgM. The bands indicated by Dressler [7] and Zoller [8] were taken into account as being specific for B. burgdoderi infection. 2. Materials and methods 2.1. Patient sera Sera were obtained from patients resident in the Friuli Venezia Giulia region: 36 were affected by neuroborreliosis, and 32 by Lyme arthritis. The sera had previously tested positive for LB by IFA or ELISA. As control 24 sera were collected from blood donors, which were negative in serological tests Borrelia strains and antigen preparation In IB we used locally isolated borreliae, such as strains BITS (B. garinii) BL3, Gualtieri, Nancy (B. ufielii) and Alcaide (B. sensu stricto). After growth in Barbour-Stoenner-Kelly medium to lo8 cells/ml, spirochetes were harvested by centrifugation, washed twice, resuspended in phosphate buffered saline (PBS) ph 7.4 and the pellet obtained was sonicated as previously described [5]. Protein determination was carried out by the Coomassie brilliant blue binding assay (Biorad) Western blot Western blot was done using a mini blot system as described [6]. Briefly, electrophoresis in 12.5% polyacrylamide gels was performed according to the method of Laemmli using standard mass markers and sonicated Borrelia preparations (20 mg); the gels were stained with Coomassie brilliant blue or transferred onto nitrocellulose with semi-dry apparatus (Biorad) for 30 min, and cut into 2 mm strips. The strips were blocked with Tris-buffered saline (TBS), containing 50 mm Tris and 150 mm NaCl (ph 7.51, 0.5% Tween 20 and 2% bovine serum albumin. The strips were incubated with patient sera (1:lOO) diluted in TBS, 0.05% Tween 20 and 0.1% bovine serum albumin and then with alkaline phosphatase-conjugate goat antibodies to human IgG (DAKO), all for 60 min. Finally the substrate 5- bromo-4-chloro-indolyl phosphate and nitroblue tetrazolium chloride mixed in TBS 100 mm, NaCl 0.01 M and MgCl, 5 mm was added for about 5-10 min. A positive and a negative sample were added for each assay Statistics Frequencies of antibody reactivity to each protein band observed in subjects with Lyme borreliosis and in normal controls were compared by x2 analysis, using a GraphPad software.

3 Marin Cinco et al. / FEMS Immunology and Medical Microbiology 14 (1996) Results 3.1. Band patterns in patient sera with IB Fig. 1 shows the band patterns of IgG and IgM reactivity of sera from neuroborreliosis and arthritis with strain BL3, repreisenting genospecies B. afzelii, BITS representing gerrospecies B. garinii and strain Alcaide, of genospecies B.b. sensu stricto. About 22 protein bands were detected in the human sera, which, in both groups, were from late cases of Lyme infections. The band positions were referred to the molecular mass standard and to monoclonal antibodies H9724 and H5332., which recognized the 41 and 30 kda proteins (data not shown). Different band patterns were visible among the strains, group sera and immunoglobulin classes. More bands were revealed by IgG than IgM both in neuroborreliosis and arthritis sera. Some bands, when present, were consistently visible in all the three strains like the 24-21, 25, 30-33, 41, 45, and, mostly in arthritis sera, the kda. The 18-17,28-27,51,72-80 kda and others showed variations. In general strains BITS and BL3 seemed more reactive than strain Alcaide in immunoblots with IgG, whereas in IgM immunoblots with neuroborreliosis sera, a slight response was visible with strain BL3 and several bands were detectable with strains Alcaide and BITS Comparison of frequency of antibodies reactivities of B.b. proteins between patient and healthy donor sera for each strain The frequencies at which the various B. burgdorferi proteins were reactive in a given B. burgdorferi strain in the patient sera, were compared with those of normal blood donors by x2 analysis as reported in Table 1, Tables 2 and 3. It can be observed that among the 22 protein bands regularly detected in IgG IB, only some were significant if compared to Table 1 Comparison of IgG reactivities of Lyme arthritis sera (n = 32) to proteins of the three Borrelia strains, versus blood donor sera (n = 24) Protein (ma) 1 l Number of arthritis sera positive for strains Level of significance a BL3 BITS Alcaide BL3 BITS Alcaide X2 P X2 P X2 P O(0 b, 0 (0) l(o) NS NS NS l(o) 2 (0) l(o) NS NS NS 10 (0) 4 (0) 7 (0) NS (0) 2 (1) 2 (0) NS NS NS 10 (5) ll(2) 10 (3) NS NS 7 (0) 5 (0) 8 (1) NS NS 4 (2) 2 (4) 4 (3) NS NS NS 10 (8) 13 (4) 9 (2) NS NS NS 2 (3) 1 (0) 2 (0) NS NS NS 5 (1) 4 (0) 5 (0) NS NS NS 3 (4) 4 (0) 4 (0) NS NS NS 3 (5) 0 (2) 5 (0) NS NS NS 7 (5) ll(4) 3 (4) NS NS NS 13 (5) (4) NS NS l(o) 3 (1) 6 (6) NS NS NS 4 (4) 4 (0) 4 (1) NS NS NS 6 (4) 6 (4) 8 (6) NS NS NS 8 (5) 8 (6) 8 (4) NS NS NS 10 (0) 4 (6) 3 (2) NS NS 3 (0) 2 (0) 0 (0) NS NS NS 0 (0) 1 (0) l(o) NS NS NS 19 (1) 12 (8) 9 (3) NS NS a P and X2 values calculated with the hi-square analysis. NS, P > b Number of positive sera in the control group.

4 62 Marin Cinco et al./ FEMS Immunology and Medical Microbiology 14 (1996) aw3iv s1ia s -IQ 3awmv SllQ s 10 3amiv Sllg s -IQ 3aw3iv SUB s 19 3awmv 3anmv SUE E: -Ia atvmv sue s ia 3aw3l-v SlM s 1Q 3aiv3iv sua s 1g v) I 3aw3iv + 2 SLIM s la 2 d 3aiwv 4 sue E 1B 3amiv 5118 E 1B

5 Marin Cinco et al./fems Immunology and Medical Microbiology 14 (1996) the control group ser:a at the significance level of P I The degree of significance was also indicated by the x2 value: the greater the x, the larger the differences between two comparative band fre- :z-* d 8- NJmRoBcmREL.IIms Kdo I BL3 Ei BITS Akkk BL3 BITS Abakk BW BITS AusUnI Fig. 2. Level of significance ( x2) of IgG and IgM IB referred to the three Borrelia strains and the two groups of sera. IgG Arthritis IgG Neuro IgM Neuro Fig. 3. Preferential reactivity of bands of each strain/genospecies, recognized by sera of neuroborreliosis and arthritis. The number of bands recognized in a given strain was expressed as % of total number of significant bands detected in the three strains. Neuro = Neuroborreliosis. quencies. Referring to the three strains, proteins of significance in neuroborreliosis sera included the following, in decreasing order of significance: 30-33, 24-21, 25, 36-37, 60-58, 38-37, kda; in arthritis sera , 72-70, 18-17, 25, 45 and Among these significative bands, the 25, and the 45 kda were regularly detected in both the patient group sera, whereas the 33-30, 38-37, and kda proteins seem to be significant for neuroborreliosis and the , and kda protein behaves like markers for Lyme arthritis. The level of protein band significance varied among the three B. burgdorjki strains employed in IB, as summarized in Fig. 2. In fact, in neuroborreliosis sera, the and kda bands were constantly reactive in all the strains; band 25, 45 and kda were significant only in strains BL3 and BITS, band in BITS and Alcaide, band kda only in strain BL3. Protein bands of kda were preferentially revealed by strains BL3 and Alcaide in arthritis sera; the 25, and kda proteins by strain BL3; the and 45 by strain BITS. Reactivities of the Borreliu proteins towards IgM of patients were less significant than IgG: only in neuroborreliosis sera, proteins 18-17, 24-21, 5 1, LD

6 164 Marin Cinco et al. / FEMS Immunology and Medical Microbiology 14 (1996) I and kda were detected at P I 0.05 (Table 31, when compared with controls. Proteins and kda were detected with strain BITS, 51 and kda by Alcaide, and protein by both strains, BITS and Al&de. The few bands at 8-l 1 kda probably corresponded to lipopolysaccharides, which we have described previously [ Preferential reactivity of proteins of the three strains of B. burgdo$eri The prevalences of total number of significant bands for each strain, deduced from Table 1, Tables 2 and 3, are summarized in Fig. 3. Sera of patients with neuroborreliosis and arthritis showed higher frequency of reactivity with strain BL3 (B. ufzelii) followed by strain BITS (B. gurinii) in IgG immunoblots. When IgM immunoblots were evaluated, a prevalent reactivity was due to strains Alcaide and BITS. 4. Discussion The current investigation is an extension of a previous study [5] in which we evaluated the occurrence of the IB bands in the course of LD, and suggested that in the late stages of infection, less specific IgG antibodies were detected. Therefore, to further improve the performance of our IB, we have studied here the specificity of IgG and IgM in late Lyme sera taken from definite cases of neuroborreliosis and arthritis, comparing them with IB reactivities performed on blood donor sera. The significance analysis was carried out in relation to each of the three Borrelia strains employed in the IB and representing the three genospecies commonly occurring in Table 2 Comparison of IgG reactivities of Lyme neuroborreliosis sera (n = 36) to proteins of the three Borreliu strains, versus blood donor sera (n=24) Protein Number of neuroborreliosis sera positive for strains Level of significance a &Da) BL3 BITS Alcaide BL3 BITS Alcaide X2 P X2 P X2 P 1 l-8 0 (0) b 0 (0) 0 (0) NS NS NS 14 3 (0) 3 (0) l(o) NS NS NS (0) 5 (0) 3 (0) NS NS NS (0) 4 (1) l(o) NS NS NS (5) 16 (2) 18 (3) (0) 13 (0) 5 (1) NS (2) 10 (4) 9 (3) NS NS (8) 18 (1) 13 (2) (1) 2 (0) 3 (0) NS NS NS 35 4 (2) 2 (0) 3 (0) NS NS NS (41 10 (0) 9 (0) (5) 5 (2) 3 (0) NS NS NS (51 11 (4) 12 (4) NS NS NS (5) 10 (1) 4 (4) NS (9) 3 (1) 6 (6) NS NS NS (4) 6 (0) 4 (1) NS NS NS (5) 15 (4) 15 (6) NS (5) 3 (6) 7 (4) NS NS NS (0) 2 (6) 4 12) NS NS NS 74 2 (0) 0 (0) 0 (0) NS NS NS 78 1 (0) 0 (0) 2 (0) NS NS NS (1) 5 (8) 5 (3) NS NS NS P and X2 values calculated with the cm-square analysis. NS, P > b Number of positive sera in the control group.

7 Marin Cinco et al./ FEMS Immunology and Medical Microbiology 14 (1996) our territory [lo]. The overall data seen in immunoblots (Fig. 11, indicated a greater number of bands reacting with IgG antibodies than IgM; this datum was expected because IgM response occurs early during Lyme infection [8]. Among the 22 proteins recognized by IgG antibodies in the three strains, ten were found significant (P I 0.05) for LD infection with different prevalence rates in arthritis and neuroborreliosis sera. Of note is the dominant IgG antibody response to the proteins corresponding to OspCs: it is reported that OspCs are the major proteins expressed by 50% of the European Borrelia isolates [ill and that the IgG response is strong in early LD [ 1;!,13]. According to our data the OspCs are also significatively dominant in the IgG late immune response; moreover, they are well recognized on all the three strains employed. Another immunodominant protein we found, is at 25 kda: this antigen was fist described as an OspCs component by Wilske [ll], and was subsequently judged a strong marker in IgM blots during early LD [ 131. Our studies suggest that this protein is an important marker also in late LD infection recognized by IgG. Concerning the bands which we found prevalent in each of the two clinical manifestations, it is noteworthy that protein at kda (OspA) behaved like a marker in neuroborreliosis and not in arthritis: this is in contrast with the findings reported by others [14] who have shown that the sera of US patients affected by chronic arthritis contain strong IgG reactivity to OspA. These controversial findings might be due to the more restricted antibody response in European Lyme borreliosis, different stages of Lyme arthritis and lower expression of OspA in the European strains, we employed in our IB, than in the US. The IgM late immune response was found significant only for few bands: of interest is the strong dominance of band , in fact this core protein, also termed ~94, is generally recognized as the best marker for late LD and usually recognized by IgG E31. Concerning the correlation between the reactivity of strains belonging to the three genospecies of B. burgdoferi and the clinical symptoms, our data (Fig. 3) clearly indicates that BL3 of species B. ufielii gives the higher number of significant bands in Table 3 Comparison of IgM reactivities of Lyme neuroborreliosis sera (n = 34) to proteins of the three Borrelia strains, versus blood donor sera (n = 24) Protein Number of neuroborreliosis sera positive for strains Level of significance a &Da) BL3 BITS Alcaide BL3 BITS Alcaide X2 P X2 P X2 P (0) b 0 (0) 0 (0) NS NS NS 14 0 (0) 2 (0) 4 (0) NS NS NS (0) 7 (0) 5 (0) NS NS (0) 7 (0) l(1) NS NS (0) 2 (0) 5 (0) NS NS NS (4) 7 (1) 8 (2) NS NS NS 34 2 (0) 2 (0) 0 (0) NS NS NS 35 l(o) 3 (0) l(o) NS NS NS (5) 3 (0) 4 (0) NS NS NS ) 6 (1) 4 (2) NS NS NS (2) 3 (1) 9 (2) NS NS NS 45 5 (4) 2 (1) 3 (3) NS NS NS (7) 10 (2) 6 (6) NS NS NS 51 2 (1) 5 (0) 7 (0) NS NS (2) 5 (4) 4 (4) NS NS NS (0) 2 (2) 7 (0) NS NS (0) 3 (0) 4 (0) NS NS NS (0) 7 (0) 11 (0) NS a P and X2 values calculated with the hi-square analysis. NS, P > b Number of positive sera in the control group.

8 166 Marin Cinco et al. / FEMS Immunology and Medical Microbiology 14 (1996) Immunoblots with arthritis sera, followed by strain BITS. With sera of neuroborreliosis the prevalence of BITS (B. garinii) is also relevant, suggesting a wider involvement of this species in neurological manifestations than in arthritis. Moreover this strain, and not BL3, was able to evidentiate significant Borreliu polypeptides at 45 kda, OspCs and kda. Consequently these strains should be used in association to cover the whole band patterns of each serum. The present findings show that strain Alcaide of genospecies B. sensu strict0 is scarcely reactive in IgG IB, but it strongly reacts with , and 51 kda proteins in neuroborreliosis sera carried out with IgM. Therefore strain Alcaide should be added to the other two, testing patients with neuroborreliosis and the IB should carried on with both IgG and IgM conjugates. In conclusion the results obtained in this investigation add further indications on the specificity of the strain used to detect antibodies elicited during late Lyme infection. Moreover our data suggest that there is no evident relation between Lyme arthritis and reactivity to B. sensu stricto, but a preferential serum reactivity of neuroboreliosis sera, both in IgM and IgG, to B. garinii. This latter observation is in agreement with similar results on sera collected in other parts of Europe [2,3,15]. References [l] Barbour, A.G., Tessier, S.L. and Hayes, S.F. (1984) Variation in a major surface protein of Lyme disease spirochetes. Infect. Immun. 45, [2] Assous, M.V., Postic, D., Paul, G., Nevot, P. and Baranton, G. (1993) Western Blot analysis of sera from Lyme borreliosis patients according to the genomic species of the Borrelia strains used as antigens. Eur. J. Clin. Microbial. Infect. Dis. 12, [3] Dressler, F., Ackermann, R. and Steere, A.C. (1994) Antibody responses to the three genomic groups of Borrelia burgdorferi in European Lyme Borreliosis. J. Infect. Dis. 169, [4] Bunikis, J., Olsen, B., Westman, G. and Bergstroom, S. (1995) Variable serum immunoglobulin responses against different Borrelia burgdorferi sensu lato species in a population at risk for patients with Lyme disease. J. Clin. Microbiol. 33, [5] Cinco, M., Murgia, R. and Costantini, C. (1995) Prevalence of IgG reactivity in Lyme borreliosis patients versus Borrelia garinii and Borrelia afielii in a restricted area of Northern Italy. FEMS Immunol. Med. Microbial. 12, [6] Cinco, M., De Giovannini, R., Fattorini, P., Florian, F. and Graziosi, G. (1993) Genotypic analysis of Borrelia burgdorferi strains identifies three genomic groups. Microbiologica 16, [71 Dressler, F., Whalen, J.A., Reinhardt, B.N. and Steere, A.C. (1993) Western blotting in the serodiagnosis of Lyme Disease. J. Infect. Dis. 167, Zoller, L., Cramer, J. and Faulde, M. (1993) Western blot as a tool in the diagnosis of Lyme borreliosis. Electrophoresis 14, t91 Cinco, M., BanIi, E., Godeas, C., Balanzin, D. and Panfili, E. (1991) Evidence for (1ipo)oligosaccharides in B. burgdorferi and their serological specificity. FEMS Microbial. Immunol. 76, [loi Cinco, M. and De Giovannini, R. (1993) Protein and antigenie analysis of Borrelia burgdorferi isolated in northern Italy: computerized analysis of phenotypic characteristics. J. Clin. Microbial. 31, illi Wilske, B., Preac-Mursic, V., Schierz, G. and Busch, K.V. (1986) Immunochemical and immunological analysis of European Borrelia burgdorferi strains. Zentralbl. Bakteriol. Mikrobiol. Hyg. Ser. 263, [121 Ma, B., Christen, B., Leung, D. and Vigo-Pelfrey, C. (1992) Serodiagnosis of Lyme borreliosis by Western Blot reactivity of various significant antibodies against Borrelia burgdorfen. J. Clin. Microbial. 30, [13] Aguero-Rosenfeld, M.E., Nowakowski J., McKenna, D.F., Carbonaro, C. and Wormser, G.P. (1993) Serodiagnosis in early Lyme Disease. J. Clin. Microbial., 31, [14] Kalish, R.A., Leong, J.M. and Steere, A.C. (1993) Association of treatment resistant chronic Lyme arthritis with HLA- DR4 and antibody reactivity to Osp A and Osp B of Borrelia burgdorferi. Infect. Immun. 61, [15] Anthonissen, F.M., De Kesel, M., Hoet, P.P. and Bigaignon, G.H. (1994) Evidence of the involvement of different genospecies of Borreliu in the clinical outcome of Lyme disease in Belgium. Res. Microbial. 145,