WO1996027799A1

WO1996027799A1 - Method for the diagnosis of infections

Info

Publication number: WO1996027799A1
Application number: PCT/EP1996/000979
Authority: WO
Inventors: Klaus Hedman; Maria Söderlund
Original assignee: Klaus Hedman; Soederlund Maria
Priority date: 1995-03-08
Filing date: 1996-03-07
Publication date: 1996-09-12
Also published as: EP0813683A1; AU5103396A

Abstract

Described are two novel tests for patient diagnosis. One compares IgG reactivities towards conformational and linear epitopes of VP2. The other compares IgG reactivities towards denatured VP2 and VP1 antigens. Both tests are suitable for verification of the time point of human parvovirus infection.

Description

Method for the Diagnosis of Infections

The present invention relates to a method for diagnosing infections caused by microbes whereby antibody reactivities towards at least two proteins or antigenic determinants derived from the microbes are compared.

Human parvovirus B19 is structurally the simplest known human-pathogenic virus. Its icosahedral capsid is made up of two structural proteins, VPl (83 JD) and VP2 (58 kD) . The smaller protein VP2 is the major capsid constituent, while VPl comprises only <10% of the capsid. The amino acid sequence of VP2 overlaps with that of VPl, so that a portion of only 227 N-terminal amino acids is unique to VPl (Ozawa, K., et al., J. Virol. 61 (1987) , 2395-2406) .

Parvovirus causes erythema infectiosum (fifth disease) , a common childhood exanthem (Anderson M.J., et al., J. Hyg. Camb. 93 (1984) , 85-93; Plummer F.A., et al., N. Engl. M. Med. 313 (1985) , 74-79) which, especially in adults, is often coupled to transient polyarthritis or arthralgia affecting typically knees and small joints (Reid D.M. , et al., Lancet 1 (1985) , 422-425; White D.G., et al., Lancet 1 (1985) , 419-421) . The pathogen can be transmitted from mother to fetus causing hydrops fetalis and fetal death (Anand A. , et al., N. Engl. J. Med. 316 (1987) , 183-186; Brown T. , et al. , Lancet 2 (1984) , 1033-1034; Hall S.M., et al., BMJ 300 (1990) , 1166-1170) . In patients which shortened red cell survival compensated by accelerated erythropoiesis parvovirus may cause aplastic anemia (Pattison J.R., et al., Lancet 1 (1981) , 664-665) . In immunoco promised patients persistent parvoviral infection can lead to chronic bone marrow failure and prolonged anaemia (Kurtzman G.J., et al., N. Engl. J. Med. 317 (1987) , 287-294) . In chronically infected patients abnormal B19-IgG responses have been encountered, where the antibodies lack neutralizing ability and do not recognize viral capsid antigens in immunoblots (Kurtzman G.J., et al., J. Clin. Invest. 84 (1989) , 1114- 1123) . The chronic parvovirus anemia can often be cured by immunoglobulin infusion (Kurtzman G., et al., New Engl. J. Med. 32 (1989) , 519-523; Frickhofen N. , et al., Annal. Int. Med. 113(12) (1990) , 926-933; Schwarz T.F., et al., J. Infect. Dis. 162 (1990), 1214) . However, persistent B19 infections may occur also in i munocompetent (by conventional criteria) individuals (Faden H. , et al., Clin. Inf. Dis. 15 (1992) , 595-597) . The role of T-cell immunity in defence against human parvovirus infection is unknown (Kurtzman G.J., et al., J. Clin. Invest. 84 (1989) , 1114- 1123) .

In view of the high percentage of individuals being infected with viruses, for instance it is thought that about 50% of the adults have antibodies, it would be desirable to have a reliable tool for diagnosing an infection and especially the stage of an infection.

Thus, the technical problem underlying the present invention is to provide a novel assay and a kit for the diagnosis of infections caused by microbes, especially for determining the stage of an infection.

The solution to said technical problem is achieved by providing the embodiments characterized in the claims.

Accordingly, the present invention relates to a method for the diagnosis of infections caused by microbes which comprises comparing antibody reactivities towards antigenic determinants of at least two proteins derived from the microbes. In this context the term "microbe" includes microorganisms, e.g. bacteria or fungi, as well as viruses. Figure 1 shows the results of an indirect IgG-EIA with human parvovirus capsid proteins. Capsids composed of VPl and VP2 (VPl/2) or VP2 alone (VP2) on polystyrene were treated with serum pools, collected on the indicated days, and bound IgG was measured.

Figure 2 shows the results of an IgG-EIA with chemically modified proteins: A, biotin-VP2 capsids on streptavidin; and B, VP2 capsids, C, VP1/2 capsids or D, βVPl protein on polystyrene. The serum pools used are as in figure 1. The protein was treated prior to IgG binding.

Figure 3 shows immunoblots wherein the following proteins were used: A, βVPl protein; B VP1/2 capsids; C, VP2 capsids. Lanes 1 to 6 contain the same serum pools as in the previous figures, and lane 7 the control pools with no B19 antibodies. Numbers refer to molecular weights (kD) of marker proteins (M) . Arrows indicate migration positions of βVPl (140 kD) , VPl (83 kD) and VP2 (58 kD) .

Figure 4 shows an evaluation of the new diagnostic assays by serum pools.

Figure 5 shows an evaluation of the new diagnostic assay by individual follow-up samples. The 2 circles indicate a serum pair with exceptionally high relative binding to linear-VP2 determinants in the first sample, leveling off in 14 days (sample interval) . "Past imm." denotes past immunity.

In a first embodiment the present invention relates to a method wherein one of the at least two proteins originates from a modification in the primary structure of the other(s) . In a second embodiment the present invention relates to a method wherein the at least two proteins have the same primary structure but different secondary or tertiary structure.

In a preferred embodiment of the present invention the microbes are viruses.

In another preferred embodiment the proteins are capsid proteins in isolated form or in assembled capsids.

In a specific embodiment of the present invention, the two proteins are the denatured VP2 and VPl proteins of the human parvovirus B19 whereby the VPl protein can be the entire protein or the portion unique to the VPl protein.

The method of the present invention is based on an immunoassay which employs proteins derived from microbes and antibodies contained in sera or obtained according to conventional methods (see Ex. 1) . Such immunoassays are well-known in the art (see, for example, "Immunoassay: A Practical guide", D.W. Chan and M.T. Perlstein eds. 1987) and include, for example, radioimmunoassay (RIA) , enzyme immunoadsorbent assay (EIA, ELISA) , Immunoblotting, complement fixation, immunodiffusion, immunoelectrophoretic or immunofluorescent assay and the like.

Suitable assays include both solid phase (heterogeneous) and non-solid phase (homogeneous) protocols. The assays may be run using competitive or non-competitive formats, and using a wide variety of labels, such as radioisotopes, enzymes, fluorescers, chemiluminescers, spin labels, and the like. A majority of suitable assays rely on heterogeneous protocols where the ligand is immobilized on a variety of solid phases, such as dipsticks, particulates, microspheres, magnetic particles, test tubes, microtiter wells, and the like. In a preferred embodiment, the immunoassay is EIA or Immunoblotting.

The antigens used in the assays, i.e. the capsids of parvovirus B19, consisting either of VPl and VP2 together in the approximate ratio 1:11 (VPl/2 capsids) , or of VP2 alone (VP2 capsids) , can, for example, be produced in insect cells by the baculovirus expression system as described in Brown, C.S., et al., Virus Res. 15 (1990) , 197-212 or Brown C.S., et al., J. Virol. 65(5) (1991), 2702-2706.

In order to determine whether antibodies show different reactivities to conformational versus linear epitopes human IgGs to parvovirus capsids consisting either of protein VPl and VP2 in the native proportion (VP1/2) or of VP2 alone (VP2) , were directly coated onto plastic, e.g. polystyrene, and the responses were assessed using pools of sera. The sera originated from patients at successive stages of acute phase or convalescence after primary infection or from controls with immunity of several years duration (see example 1) .

An IgG-EIA (see example 3) showed strong reactivity during the acute phase (<three months of the onset) using either VP1/2 capsids or VP2 capsids as antigen (Figure 1) . After a slight subsequent decline, the VPl/2-IgG levels remained elevated. In contrast, the VP2-IgG reactivity collapsed during late convalescence and remained at non-immune background levels thereafter (Figure 1) .

In another specific embodiment of the present invention the two proteins are the native and the denatured VP2 protein. Contrary to the directly immobilized antigen, the streptavidin-attached biotinylated VP2 capsids in native form showed very high immunoreactivity at all time points after infection (Figure 2A) . In order to assess whether the difference in immunoreactivity was due to protein conformation, we exposed these capsids to protein denaturants followed by reducing and alkylating agents. Brief exposure to 8 M urea, with or without reduction/alkylation, selectively abolished the reactivity of the biotin-VP2 capsids to the antibodies of past immunity (>150 days after onset) , yet maintained strong reactivity to the antibodies of acute infection (Figure 2A) , exactly as direct immobilization (Figure 1) had done. Chemical denaturation, reduction and alkylation did not further alter the antigenicity of the directly immobilized capsids (Figure 2B) .

On the other hand, identically immobilized urea-treated, reduced and alkylated VPl/2 capsids maintained strong immunoreactivity irrespective of the time point of sampling (Figure 2C) . In fact, beyond the first week of illness the chemically modified VPl/2 antigens bound antibodies more efficiently than did the buffer-treated antigen, with no change in the shape of the temporal IgG binding curve. The unique portion of VPl in the recombinant βVPl antigen exhibited IgG binding kinetics similar to the VPl/2 capsid (Figure 2D) .

In immunoblots under denaturing conditions both βVPl (Figure 3A) and isolated VPl (Figure 3B) exhibited strong immunoreactivity at all time points after infection. In contrast, isolated VP2 (Figure 3C) was selectively recognized by acute-stage IgG, thereby verifying the capsid- EIA results at the protein level. The ratios of IgG-EIA absorbances obtained with the urea- treated capsids (VP1/2 versus VP2) were calculated and plotted against time after onset of symptoms. Employing the same serum pools as above, the ratio remained below 2 for the first 100 days, but rose dramatically thereafter exceeding 10 beyond the first year of illness (Figure 4) .

A ratio was similarly calculated for IgG reactivity with the native versus the urea-denatured VP2 capsids attached to streptavidin. The resulting curve (Figure 4) was very similar to that of the previous test.

The tests of both types were evaluated with individual samples collected at successive time points from 61 patients with symptomatic serologically verified B19 infection. The test measuring IgG binding to the denatured VPl/2 versus VP2 capsids separated the acute-phase samples very well from the samples representing past immunity (Figure 5A) . All samples taken within the first 3 months of illness yielded absorbance ratios <2.5 (mean, 1.2; SD, 0.25) . The threshold value for recent infection (mean + 3 SD) was 2.0. All the follow-up samples taken >4 months of onset gave absorbance ratios above this cut-off, as did all but two of the past- immunity controls (mean, 6.0; SD, 3.5) .

The other test comparing antibody binding to the native VP2 capsid and the denatured VP2 capsid (absorbance ratio threshold 3.6; mean + 2 SD) distinguished the recently infected patients (mean, 1.8; SD, 0.9) from the past- immunity controls (mean, 11.5; SD, 4.1) approximately as efficiently as did the previous test. The scatter of values during the acute phase and convalescence was, however, slightly wider (Figure 5B) . It was further assessed whether other virus infections induce IgG antibodies binding - to the linear B19-VP₂ epitopes. The sera of 38 patients with B19 immunity and acute primary rubella ^' virus, CMV, EBV or hantavirus infections were studied for conformational and linear VP2- IgG activity. The VP2:VP2 absorbance ratios stayed above the threshold of 3.6 for 36/38 patients, whereas lower ratios (2.6 and 2.2) were obtained for 2 patients of 10 with acute EBV infection. The mean ratio of the 38 acutely infected control patients was 11.7, with an SD of 5.4.

Another object of the present invention is to provide kits for the diagnosis of an infection caused by microbes by comparing antibody reactivity towards antigenic determinants of at least two proteins derived from the microbes.

In a preferred embodiment the kit contains denatured VP2 and VPl proteins of the human parvovirus B19.

In another preferred embodiment the kit contains the native and the denatured VP2 protein of the human parvovirus B19.

In a more preferred embodiment the VP2 protein is biotinylated.

The above-described assays show for the first time the temporal development of epitope type (linear/conformational) specificity in human parvovirus antibody responses. The new approach gives rise to novel diagnostic tests. One compares the IgG reactivities in two EIAs where the two capsid antigens have been pretreated with a protein denaturant. Another test compares the IgG reactivities with conformational and linear VP2 epitopes. Both tests accurately define the time points of primary infection for the vast majority of the B19 patients versus the controls with past immunity. The IgG antibodies of patients with acute parvovirus infection recognized both conformational and linear antigenic determinants of VP2. The IgG to the linear VP2 epitopes disappeared abruptly during convalescence, whereas the IgG to the conformational epitopes persisted. On the other hand, VPl antigenicity was invariably resistant to denaturation, reduction and alkylation. It is assumed (in the absence of direct structural verification) that predominantly linear epitopes of VPl were recognized at all time points after infection.

The presence of conformational VP2 antigenicity has been noted before (Salimans M.M.M, et al., J. Virol. Meth. 39 (1992) , 247-258) . Long-lasting VP2 IgG responses have been the rule with apparently non-denatured VP2 capsid antigens in capture immunoassays or biotin-streptavidin EIAs (Erdman D.D., et al., J. Med. Virol. 35 (1991) , 110-115; Kajigaya S., et al., Proc. Natl. Acad. Sci. USA 88 (1991) , 4646-4650; Salimans M.M.M, et al., J. Virol. Meth. 39 (1992) , 247-258), resembling the immune responses towards natural virus capsids (Cohen B.J., et al. , J. Hyg. (Camb) , 91 (1983) , 113- 130; Anderson L.J. , et al. , J. Clin. Microbiol. 24 (1986) , 522-526) . By indirect EIA or immunoblot, predominant or exclusive IgG responses for VP2 have been encountered during very early acute phase, whereas VPl reactivities have predominated months to years later (Kurtzman G.J., et al., J. Clin. Invest., 84 (1989) , 1114-1123; Schwarz T.F., et al., J. Virol. Meth. 20 (1988) , 155-168; Matsunaga Y., et al., IXth International Congress of Virology, Glasgow, UK (1993) : 175 P18-7) . Certain chronically infected or immunodeficient patients have shown impaired immunoblot reactivities towards VPl. The minor capsid protein VPl stabilizes the virus particle and, in general, is an important immunogen for neutralizing B19 immunity (Kurtzman G.J., et al., J. Clin. Invest. 84 (1989) , 1114-1123; Bansal G.P., et al. , J. Inf. Dis. 167 (1993) , 1034-1044) . The observation that urea destroyed the conformational VP2 determinants irreversibly was unexpected , since its effect on protein antigenicity is usually reversible (Hedman K. , et al., J. Clin. Immunol. 8 (1988) , 214-221; Hedman K. , et al. , Lancet 338 (1991) , 1353-1356; Hall R.A., et al., J. Virol. Meth. 32 (1991) , 11-20) . That the VP2 antigens were not simply washed out by the denaturant, was proven by the nearly undiminished capacity of the treated capsids to bind in EIA acute-stage antibodies, and was confirmed by our immunoblot data. The immunomodulatory effect of urea could not be prevented by prior cross-linking of the streptavidin- biotin-attached VP2 capsids by various concentrations of glutaraldehyde, followed by NaBH₄. All these observations suggest that the confomational antigenicity of the VP2 capsids, without the support of VPl, is extremely susceptible to mechanical or chemical damage.

The following examples illustrate the invention.

EXAMPLES

Example 1: Antibody sources

1. Sera : Acute-phase serum samples were obtained from 61 symptomatic patients with B19 infection fulfilling strict diagnostic criteria: seroconversion or -4-fold rise in titer of B19-IgG and detectable B19 IgM antibodies (Soderlund M. , et al., J. Clin. Microbiol. 30(2) (1992) , 305-311) . All the infections were primary as verified by low avidity of IgG in the acute phase, with subsequent avidity maturation (Soderlund M. , et al., J. Inf. Dis. 171 (1995) , in press) . Characteristic symptoms were rash or arthralgia/arthritis, usually with a self-limiting course. Follow-up samples >100 day after onset were obtained from 34 of these patients. Samples representing past immunity were collected from 54 members of laboratory staff with no B19-related symptoms; all had B19 IgG, none had IgM in recombinant- VPl EIA and whole-virus RIA, -respectively (Cohen B.J. , et al., J. Hyg. (Camb) 91 (1983) , 113-130; Soderlund M. , et al., J. Clin. Microbiol. 30(2) (1992) , 305-311) .

Pools of sera were created by combining equal volumes of 7-17 of the samples above at the following time points after onset: 0-7 days (a), 13-15 days (b) , 40- 100 days (c) , 150-250 days (d) and 300-400 days (e) . Pool (f) comprised the past-immunity sera, whereas pool (g) was made up of the samples of 12 non-immune controls.

2. To assess whether IgG antibodies binding to the linear VP2 epitopes are formed in other infections and to challenge the diagnostic specificity of our new tests, acute-phase sera were collected from 38 patients who had pre-existing B19 immunity and had acute primary infections by other viruses. The latter pathogens, numbers of patients and diagnostic criteria, respectively, were: rubella virus, 8 patients, IgM and seroconversion or -4-fold rise in titer of low-avidity IgG; cytomegalovirus (CMV) , 5 patients, IgM⁺ and seroconversion of low-avidity IgG; Epstein-Barr virus (EBV) , 10 mononucleosis patients, IgM⁺ for viral capsid antigen (VCA) and low-avidity VCA-IgG; hantavirus (Puumala strain) , 15 patients, low-avidity IgG (Hedman, K. , et al., J. Clin. Immunol. 8 (1983) , 214-221; Hedman, K. , et al., Lancet 338 (1991) , 1353-1356; Hedman, K. , et al., Rev. Med. Microbiol. 4 (1993) , 123- 129) . Example 2 : Assays

1. A commercial IgG-EIA (IDEIA, DAKO, Glostrup, Denmark) , utilizing VP2 capsids biotin-labeled (biotin-VP2) on streptavidin coated plates, was used in studies requiring VP2 capsids in plastic-non-adsorbed form. Antibodies for the unique portion of VPl were measured by the βVPl recombinant protein using indirect EIA

(Soderlund M. , et al., J. Clin. Microbiol. 30(2)

(1992) , 305-311) .

2. EIAs with chemically modified antigens: The VPl/2 capsids or the βVPl antigen in PBS were applied onto polystyrene microtiter wells (Biohit, Helsinki, Finland) overnight, and the biotin-VP2 capsid on streptavidin-coated plates, for 30 minutes at 22°C. The immobilized antigens underwent one of the following 30- minutes treatments at 37°C: 1) 0,5 M Tris-HCl pH 8,0

(TB) ; 2) 8 M urea in TB; 3) urea-TB with 20 mM dithiothreitol (DTT) ; 4) urea-TB-DTT followed by iodacetamide (IAA) at 5 mM, 100 mM or 0.5 M in urea-TB, for 30 min. at 22°C. After three 10-min. washes in PBS containing 0.05 % Tween 20 (PBST) , duplicate samples of serum (diluted 1:50) wre applied on the VPl/2, VP2 or the βVPl antigens for 1 hour at 37°C, or in dilution 1:200 on the biotin-VP2-antigen for 30 min. Alkaline phosphatase- (for VPl/2, VP2 and βVPl) (Orion Diagnostica, Espoo, Finland) or peroxidase- (for biotin-VP2) anti-human IgG conjugates (DAKO; 1 h) were followed by the respective substrates paranitrophenyl phosphate (30 min.) or tetra ethylbenzidine plus H₂0₂ (10 min. ) at 22°C. Example 3: VPl/2 versus VP2 EIA

The VP1/2 capsids (30 ng/well) or the VP2 capsids (20 ng/well) in PBS were separately adsorbed onto polystyrene microwell plates (Labsystems, Helsinki, Finland) overnight at 22°C. The sensitized wells were washed 2 times, 5 min each, with 8 M urea in PBS and then 3 times 10 min each with PBST. The serum samples (1:100 in PBST) were kept on the immobilized antigens for one hour at 37°C. The PBST-washed wells were treated with alkaline phosphatase-conjugated anti-human IgG and substrate, 30 min. each. The ratio of the VP1/2 and VP2 absorbances (^A4os) ^was plotted against time after onset of symptoms.

Example 4: Native versus denatured VP2 EIA

Biotin-VP2 capsids (DAKO) were first treated with 8 M urea in PBS for 2 hours at 4°C, followed by dialysis

(36 h, +4°C) against PBST, to remove the urea. For comparison, the biotin-VP2 capsids were suspended directly in PBST. Both antigens were separately mixed with sera (1:200) and applied onto streptavidin-coated microwells in a shaking incubator for 30 min at 22°C.

(As an alternative the biotin-VP2 capsids are first immobilized in native form and then denatured with 8 M urea and washed) . Peroxidase-conjugated anti-human IgG and substrate (H₂0₂ plus tetramethylbenzidine) were applied according to the manufacturer's instruction, followed by measurement of A₄5_Q. The ratios of the absorbances of the buffer-treated and the urea-treated wells were plotted against time after onset of symptoms. Example 5: Immunoblot

The antigens in reducing sample buffer (2% SDS, 5% β- mercaptoethanol, 10% glycerol, 62.5 mM Tris-HCl, pH 6.8) were heated 5 min at 95-100°C and run in sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred electrophoretically to nitrocellulose filters (Schleicher and Schuell, Germany) according to standard techniques. The filters were blocked with TEN-Tween (5 mM EDTA, 150 mM NaCl, 0.05% Tween-20 in 50 mM Tris-HCl, pH 7.4) for 45 min, and sera (1:100 in TEN-Tween) were applied for 1 hour. Peroxidase-conjugated anti-human IgG was allowed to react 1 hour followed by the substrates (dia inobenzidine plus H₂0₂) 5 min. (Soderlund M., et al., J. Clin. Microbiol. 30(2) (1992) , 305-311) .

Claims

CLAIMS :

1. A method for the diagnosis of infections caused by microbes which comprises comparing antibody reactivities towards at least two proteins or antigenic determinants derived from the microbes.

2. Method according to claim 1 wherein one of the at least two proteins or antigenic determinants originates from a modification in the primary structure of the other(s) .

3. Method according to claim 1 wherein the at least two proteins or antigenic determinants have the same primary structure but different secondary or tertiary structure.

4. Method according to any one of claims 1 to 3 wherein the microbe is a virus.

5. Method according to any one of claims 1 to 4 wherein the proteins are capsid proteins in isolated form or in assembled capsids.

6. Method according to any one of claims 1 to 5 wherein the proteins are the capsid proteins VPl and/or VP2 of the human parvovirus B19.

7. Method according to any one of claims 1, 2 and 4 to 6 wherein the two proteins are the denatured VP2 and VPl proteins.

8. Method according to claim 7 wherein the VPl protein used is the entire protein or the portion unique to the VPl protein.

9. Method according to any one of claims 1 and 3 to 6 wherein the two proteins are the native and the denatured VP2 protein.

10. Method according to any one of claims 7 to 9 wherein the protein(s) has (have) been immobilized on a plastic surface or treated with chemical or physical means for denaturation.

11. Method according to claim 10 wherein in addition the protein(s) has (have) been treated with a reducing and/or alkylating agent.

12. Method according to any one of the preceding claims wherein the assay used for comparison is an EIA.

13. Method according to any one of claims 1 to 11 wherein the IgG reactivity is measured by an immunoblot.

14. Kit for the diagnosis of an infection caused by microbes, comprising at least two proteins derived from the microbes.

15. Kit according to claim 14 containing denatured VP2 and VPl proteins of the human parvovirus B19.

16. Kit according to claim 14 containing the native and the denatured VP2 protein of the human parvovirus B19.

17. Kit according to claim 16 wherein the VP2 protein is biotinylated.