Biologicals 34 (2006) 21e27 www.elsevier.

com/locate/biologicals

A simple immuno-capture ELISA to estimate rabies viral glycoprotein antigen in vaccine manufacture
T. Nagarajan a, G.S. Reddy a, B. Mohana Subramanian a, S. Rajalakshmi a, D. Thiagarajan a, N. Tordo b, C. Jallet b, V.A. Srinivasan a,*
b a Indian Immunologicals Limited, Rakshapuram, Gachibowli (PO), Hyderabad 500019, India Unit Antiviral Strategies, Institut Pasteur, 25 rue du Dr. Roux, 75724-Paris, Cedex 15, France

Received 2 June 2005; revised 20 July 2005; accepted 25 July 2005

Abstract Rabies is an endemic, fatal zoonotic disease in the developing countries. Prevention and post-exposure therapy require safe and ecacious vaccines. The vaccine potency depends on the amount of immunogenic rabies viral glycoprotein antigen in the vaccine preparation. In order to estimate the rabies viral glycoprotein antigen, a specic monoclonal antibody was developed and used in an immuno-capture ELISA (IC-ELISA). The monoclonal antibody binds a conformational epitope on the natively folded rabies viral glycoprotein as indicated by specic, membrane uorescence on unxed, rabies virus infected murine neuroblastoma (MNA) cells and glycoprotein gene encoding plasmid transfected COS cells. In addition, the monoclonal antibody competes with and blocks a glycoprotein antigen site III binding monoclonal antibody (mAb-D1, Institut Pasteur, Paris, France). The monoclonal antibody was used in an IC-ELISA using an in-house standard to quantify the rabies viral glycoprotein antigen in 12 vaccine preparations with potency values ranging from 4 to 18 IU. The results indicated a good correlation with the NIH mouse potency assay (r Z 0.83). The immuno-capture ELISA described in this study can be used to quantify the immunogenic rabies viral glycoprotein antigen in the inactivated rabies viral antigen preparation in a simple and rapid format, which enables better vaccine formulation. 2005 The International Association for Biologicals. Published by Elsevier Ltd. All rights reserved.
Keywords: Rabies; Glycoprotein; In vitro potency; Monocloned antibody

1. Introduction Rabies is a viral encephalomyelitis that is invariably fatal after clinical signs appear [5]. The disease is transmitted through the bite of an infected animal, usually from a dog, and can be prevented by the timely administration of post-bite vaccination. The tissue culture derived vaccine used for prevention and post-exposure treatment is safe, ecacious and has no side eects [13]. The rabies vaccine is tested using the NIH mouse protection test, which is expensive, of long duration,
* Corresponding author. E-mail address: srini@indimmune.com (V.A. Srinivasan).

cumbersome and involves the use of live rabies virus. ELISA based methods have been described which use specic monoclonal antibodies raised against the rabies viral glycoprotein [3,6,9]. Assuming that the monoclonal antibody recognizes a correctly folded glycoprotein, the rabies viral glycoprotein content is indicative of the vaccine potency because the rabies viral glycoprotein induces neutralizing antibodies and protection [9,16]. The determination of glycoprotein content requires the use of a standard reference vaccine of known glycoprotein content (expressed as mg or IU). The present study describes the application of an immuno-capture ELISA (IC-ELISA) to determine the rabies glycoprotein antigen content using a specic

1045-1056/06/$32.00 2005 The International Association for Biologicals. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.biologicals.2005.07.004

22

T. Nagarajan et al. / Biologicals 34 (2006) 21e27

monoclonal antibody against the rabies viral glycoprotein. The method can be correlated to the NIH mouse protection test and hence is a good indicator of the amount of natively folded glycoprotein antigen in various in-process samples and vaccine formulations.

Protein G Sepharose CL-4B (Amersham Biosciences, Sweden) and used in the immuno-capture ELISA. 2.5. Preparation of monoclonal antibodies Monoclonal antibodies were prepared using standard protocols by immunizing 6e8 weeks old female Balb/c mice with BPL inactivated and zonal puried PV antigen [9]. Polyclonal hybrids were selected based on their reactivity with PV antigen in an indirect ELISA [8] and the monoclonal hybrids were selected after two rounds of single cell cloning. Monoclones were propagated and checked for rabies specic monoclonal antibody secretion in indirect ELISA. A single monoclonal antibody M5B4 (mAb-M5B4) was selected for further study. 2.6. Characterization of the monoclonal antibody (i) Rabies virus neutralization and cross-reactivity: The mAb-M5B4 was tested for cross-reactivity using the following rabies strains in the RFFIT: PV, Challenge Virus Standard (CVS 11), Street Alabama Duerin (SAD) and Flury LEP. (ii) Isotype: The isotype of the monoclonal antibody was identied using a commercial kit (Sigma, USA). (iii) In vitro specicity of the rabies monoclonal antibody: Puried rabies glycoprotein coated ELISA strips (BioRad, France) were used for determining the specicity of the monoclonal antibody against rabies glycoprotein. (iv) Specicity for conformational epitope of rabies viral glycoprotein: The specicity of mAb-M5B4 for conformational epitope of rabies viral glycoprotein was demonstrated by IC-ELISA using puried native rabies viral antigen and sodium dodecyl sulphate and/or 2-mercaptoethanol treated rabies viral antigen (disrupted antigen). (v) Specicity for glycoprotein antigen: The specicity of the mAb-M5B4 to the natively folded rabies viral glycoprotein was demonstrated by: (a) reaction of the monoclonal antibody to the expressed rabies viral glycoprotein product in COS cells and rabies virus infected murine neuroblastoma (MNA) cells and (b) in comparison with the D1 monoclonal antibody (mAb-D1) supplied by Institut Pasteur, Paris, France [4]. (a) Reaction of mAb-M5B4 to expressed rabies glycoprotein in COS cells and rabies virus infected MNA cells: The rabies viral glycoprotein gene was amplied from the tissue culture derived PV as described previously [1]. The amplied product was cloned into the pcDNA 3.1 vector (Invitrogen, USA) between the Xba I and EcoR I sites. The recombinant plasmid was used for transfection of COS-7 cells using Lipofectamine 2000 (Invitrogen, USA). The binding of the

2. Materials and methods 2.1. Animals Six-to-eight weeks old female, Balb/c and 3e4 weeks old, Swiss albino mice used in this study were procured from National Center for Laboratory Animal Sciences, National Institute of Nutrition, Hyderabad. Rabbits were obtained from the breeding colony maintained at Indian Immunologicals Ltd, Hyderabad. The care and use of the animals were according to the Institutional Animal Ethics Committee guidelines. 2.2. Viruses and cells The rabies virus Pasteur strain, L 2061 (PV), used in this study is a vaccine strain obtained from Institut Pasteur, Paris, France. Puried Vero cell derived rabies vaccine and in-process samples were obtained from the Human Biologicals Institute, Udhagamandalam, India. Street Alabama Duerin (SAD) strain was obtained from Institute of Virology and Immunoprophylaxis, Mittelhausen, Switzerland. Flury LEP strain was obtained from CZV, Espana, Spain. The Challenge Virus Standard 11 (CVS 11) was obtained from Agence Fran` caise De Securite Sanitaire Des Aliments, Malzeville, France. COS cells and MNA cells were obtained from American Type Culture Collection, Maryland, USA. 2.3. RFFIT RFFIT was performed as described earlier [12]. Briey, test sera or antibodies were mixed with CVS 11 (40 FFD50) and incubated for 90 min in Labtek 8-chamber slides (Nunc, Denmark). MNA cells were added to the mixture and incubated for 20 h. Immunouorescent foci were counted after acetone xation and potency values were calculated based on comparison with an international reference serum (NIBSC, UK). 2.4. Preparation of polyclonal antisera Polyclonal antisera against PV were prepared after hyper immunizing rabbits with b-propiolactone (BPL) inactivated and zonal puried rabies viral antigen [7]. Briey, rabbits were immunized with puried PV and test bleeding was done on 42nd day post-immunization (DPI) and nal bleeding was done on 70th DPI. Sera with titers more than 100 IU as estimated by RFFIT were pooled, puried using anity chromatography on

T. Nagarajan et al. / Biologicals 34 (2006) 21e27

23

monoclonal antibody to native rabies glycoprotein antigen was demonstrated using anti-mouse IgG FITC conjugate (Sigma, USA). COS cells transfected with wild type plasmid were used as negative controls. In another experiment MNA cells were infected with PV and mAb-M5B4 was added to infected cells after 24 h. Positive binding was visualized as membrane uorescence on unxed cells using anti-mouse IgG FITC conjugate (Sigma, USA). (b) Competition with D1 monoclonal antibody: The D1 monoclonal antibody kit for measuring the rabies viral glycoprotein antigen (Institut Pasteur, Paris, France) was used for competitive ELISA methods. The D1 ELISA kit consists of a monoclonal antibody, which binds to native rabies viral glycoprotein [4]. The rabies viral antigen is measured in a sandwich ELISA format using a peroxidaseconjugated version of the monoclonal antibody (D1 conjugate). The ELISA was performed according to manufacturers instructions with a modication for competition with the D1 conjugate. Briey, standard rabies vaccine of known glycoprotein concentration was allowed to react with the capturing monoclonal antibody. The mAb-M5B4 was added in increasing concentrations mixed with constant volumes of D1 conjugate. The binding of D1 conjugate was detected by substrate addition and the amount of reduction in absorbance values was measured. In another set of experiments, standard rabies vaccine of known glycoprotein concentration was allowed to react with the capturing monoclonal antibody and mAb-M5B4 was incubated with bound rabies vaccine antigen for 1 h at 37  C followed by the addition of D1 conjugate. The reduction in absorbance values was measured as before. 2.7. Immuno-capture ELISA (IC-ELISA) to measure rabies viral glycoprotein antigen The natively folded glycoprotein antigen was measured in an IC-ELISA using the M5B4 monoclonal antibody (mAb-M5B4) as described earlier [6] but with few modications. Appropriately diluted and puried rabbit IgG (100 ml) was used to coat ELISA plates (Maxisorp, Nunc) and incubated at 4  C overnight. After removal of unbound rabbit IgG, the un-reacted sites were blocked using 200 ml of 1% gelatin by incubation at 37  C for 1 h. The test sample (100 ml) and control were subjected to serial two-fold dilutions with phosphate buered saline (PBS)eTween 20 as a diluent in the test plate and incubated at 37  C for 1 h. After removal of unbound virus the mAb-M5B4 was added. The antibody was allowed to react with the virus for 2 h at 37  C followed by washes

with PBSeTween 20 to remove unbound monoclonal antibody. The binding of the monoclonal antibody was detected using peroxidase-conjugated anti-mouse IgG whole molecule and a substrate chromogen mixture (tetramethylbenzidine (Sigma, USA) and hydrogen peroxide (Merck, India)). The optical densities were measured at 450 nm. Rabies vaccine with known amount of glycoprotein (Institut Pasteur, Paris, France, expressed as ng/ml) was used to quantify the glycoprotein antigen content in a single preparation of the puried rabies antigen obtained after zonal centrifugation. This preparation with the known glycoprotein content was designated as the in-house standard. The in-house standard was used in subsequent measurements of rabies viral glycoprotein antigen in vaccine formulations. 2.8. Correlation of glycoprotein content with the NIH mouse potency assay The NIH mouse potency assay was performed as described earlier [11]. Ten vaccine preparations covering a range of potency values (4e18 IU) were tested using the IC-ELISA and the potency of each vaccine preparation was re-estimated using the NIH mouse potency assay. 2.9. Statistical analysis The results of the IC-ELISA and the NIH mouse potency assay were compared using the linear regression analysis. 3. Results 3.1. Characteristics of the mAb-M5B4 The mAb-M5B4 generated in this study was isotyped as IgG2b. The mAb-M5B4 neutralized PV, SAD and Flury LEP strains of xed rabies virus in RFFIT. It is interesting that this mAb did not neutralize CVS strain
Table 1 Specicity of the monoclonal antibody mAb-M5B4 for conformational and neutralizing epitope on the glycoprotein of various xed rabies viruses using IC-ELISA, indirect ELISA and RFFIT IC-ELISA PV Flury LEP PM RFFIT PV Flury LEP SAD CVS 11 Indirect ELISA ) Puried glycoprotein ) Platelia kit, BioRad Native antigen C C C Reactivity C C C Disrupted antigen

24

T. Nagarajan et al. / Biologicals 34 (2006) 21e27

of xed rabies virus in RFFIT. Hence, the mAb-M5B4 used in this study has wide specicity for PV, SAD and Flury LEP strains (Table 1). The reactivity of mAb-M5B4 was tested in IC-ELISA using PV, Flury LEP and PM strains, which indicated that mAb-M5B4 reacted well with all the three strains. Further the monoclonal antibody did not react with disrupted antigens (Table 1). The specicity for rabies viral glycoprotein was determined by in vitro assay using Platelia kit (BioRad, France) and mAb-M5B4 showed high specicity for rabies viral glycoprotein. The rabies viral glycoprotein was expressed in COS cells by transfection using recombinant plasmid containing the glycoprotein antigen gene. The binding of mAb-M5B4 to native trimeric form of glycoprotein was demonstrated by indirect immunouorescence (Fig. 1A). COS cells transfected with wild type plasmid or use of unrelated mAb of same isotype did not show any membrane uorescence indicating the specic binding of mAb-M5B4 (Table 2). Virus neutralization test was performed in MNA cells and no

uorescence was observed, indicative of neutralization. Uninfected negative controls and mock-infected MNA cells were also included in the study (Table 2). To demonstrate the specicity of mAb-M5B4 for the natively folded glycoprotein antigen the mAb was allowed to compete with the mAbeD1eHRPO conjugate. The results indicated that the anti-glycoprotein antibody eectively competes with mAbeD1eHRPO for an epitope on the glycoprotein (Fig. 2a). Addition of 10 ml of mAb-M5B4 resulted in complete abrogation of the binding of the conjugate indicating that the mAb-M5B4 can compete with the mAbeD1eHRPO conjugate. In other experiments, the mAb-M5B4 is also able to block the binding of mAbeD1eHRPO conjugate (Fig. 2b). 3.2. Estimation of rabies glycoprotein using the mAb-M5B4 Rabies vaccine producers in India conduct potency test as per WHO recommendations [11,14,15] and Indian

Fig. 1. Demonstration of membrane uorescence using the mAb-M5B4 antibody and uorescein isothiocyanate conjugated anti-mouse IgG in rabies virus infected MNA cells (A) and in glycoprotein gene transfected COS cells (C). Uninfected MNA cells (B) and COS cells transfected with wild type plasmid (D) were used as negative controls.

T. Nagarajan et al. / Biologicals 34 (2006) 21e27 Table 2 Results of indirect immunouorescence demonstrating the specicity of mAb-M5B4 for the native form of glycoprotein of rabies virus (Pasteur strain) Membrane uorescence COS cells MNA cells

25

Glycoprotein gene Mock Rabies virus Mock transfected transfected infected infected mAb-M5B4 C Negative controla
a

a standard reference vaccine preparation of known glycoprotein content. Rabies vaccine formulations were prepared using determined quantities of glycoprotein estimated using mAb-M5B4 in IC-ELISA. Twelve formulations were subjected to potency test by NIH method. The results indicated a good correlation when subjected to linear regression analysis (r Z 0.83) (Fig. 3).

4. Discussion Rabies is endemic in many developing countries and fatalities reported are mainly due to the non-aordability of vaccines and rabies immune globulins. Procedures and processes that help to reduce the cost of the vaccine may decrease rabies related deaths. Especially simplied procedures that check the antigen content and quality during dierent stages of manufacture may minimize antigen loss and also help in designing strategies to recover maximum amount of immunogenic antigen from a vaccine lot. In the developing countries the challenge in rabies control and prevention is to make the vaccine aordable for a large segment of the population. Methods that aid in the manufacture of good quality vaccines with minimal loss of vaccine antigen will help in reducing the cost of the nal product. This study describes the characterization and use of a monoclonal antibody, mAb-M5B4, in an IC-ELISA to quantitate the rabies viral glycoprotein antigen. The rabies viral glycoprotein has been shown to be able to induce neutralizing antibodies and hence protection against the disease [10,16]. Several reports describe the ELISA as a suitable in vitro potency assay for the estimation of glycoprotein antigen in vaccine preparations [3,6,8,9]. The antibody used in these assays must recognize the highly immunogenic natively folded glycoprotein and not the poorly immunogenic soluble form of glycoprotein. The soluble form of the glycoprotein (Gs) is shed
2

Unrelated monoclonal antibody of same isotype.

Pharmacopoeia standards. Formulation of rabies vaccine depends on the antigen mass, mainly natively folded glycoprotein and it is a common practice to prepare formulation using virus infectivity titer and glycoprotein content data. The mAb-M5B4 was used in IC-ELISA, which served as an adjunct to virus titer as an in-process control test to determine the rabies viral glycoprotein content using IC-ELISA. It was possible to calibrate an in-house standard (puried rabies viral antigen) using

0.7 0.6

(a)

OD 450 nm

0.5 0.4 0.3 0.2 0.1 0.0 0 25 50 75

Volume of mAb-M5B4
0.6 0.5

(b)

OD 450 nm

0.4

0.2 0.1 0 20 40 60 80 100

Log 10 NIH IU

0.3

Volume of M5B4
Fig. 2. The mAb-M5B4 binds to the native glycoprotein of rabies virus. (a) Competition of anti-glycoprotein monoclonal antibody with D1 monoclonal antibody. The anti-glycoprotein antibody was added in increasing concentrations to compete with the D1 monoclonal antibody in a sandwich ELISA. (b) Blocking of binding of D1 conjugate to the bound rabies vaccine antigen. The mAb-M5B4 was incubated with the vaccine antigen before addition of D1 conjugate. Results shown in (a) and (b) are representatives of several experiments.
0 0 1

r = 0.83 2

Log 10 GP EU
Fig. 3. Correlation of GP content estimation using the mAb-M5B4 and the NIH mouse potency assay. Rabies vaccine preparations were tested using the immuno-capture ELISA and the NIH mouse potency assay. (EU Z ELISA units).

26

T. Nagarajan et al. / Biologicals 34 (2006) 21e27

into the medium and may interfere with in vitro potency assays [2]. The soluble glycoprotein (Gs) may contribute to over-estimation of antigen and lower vaccine potency than expected. In order to demonstrate the binding of the mAb-M5B4 to the native glycoprotein, rabies infected MNA cells and glycoprotein gene transfected COS cells were allowed to react with the mAb-M5B4 antibody. The specic, membrane uorescence indicated that the antibody binds to the native form of rabies glycoprotein. Firstly, native rabies glycoprotein is abundantly present on rabies infected cells as the virus matures and buds out of the cell [7] and secondly, the expressed rabies glycoprotein antigen on the COS cells was responsible for the specic, membrane uorescence. Use of unrelated monoclonal antibody of same isotype or transfection using a wild type plasmid did not result in membrane uorescence (Table 2). The mAb-M5B4 also bound well to the puried rabies viral glycoprotein coated strips (Table 1). Further, the mAbM5B4 competed and also blocked the binding of the D1 monoclonal antibody in a sandwich ELISA (Fig. 2). The D1 monoclonal antibody has been shown to bind the antigenic site III of native but not 2-mercaptoethanol and/or sodium dodecyl sulphate treated glycoprotein [6]. The mAb-M5B4 did not bind to 2-mercaptoethanol and/ or sodium dodecyl sulphate treated rabies viral antigen. Taken together these results indicated that the mAbM5B4 binds to the conformational and natively folded glycoprotein antigen at the level of site III and not to the soluble form of glycoprotein. This nding is also conrmed by the observation that the mAb-M5B4 antibody binds to the puried virions after zonal centrifugation and not to the euents in the immuno-capture ELISA, which may contain abundant amounts of soluble form of glycoprotein (data not shown). The mAb-M5B4 antibody is quite useful in determining the rabies viral glycoprotein antigen in vaccine preparations and the estimated antigen content correlates well with the NIH mouse potency test. Vaccine preparations when tested in the IC-ELISA and the NIH mouse potency assay demonstrate a good correlation (r Z 0.83) (Fig. 3). In this study, vaccine preparations with a range of potency values from 4 to 18 IU were used. The suitability of the IC-ELISA for predicting potency values in preparation with higher potencies needs further study. The immuno-capture ELISA described in this study can be used to quantify the rabies viral glycoprotein antigen in in-process and nal vaccine preparations. An in-house standard was prepared and the glycoprotein content was estimated using the D1 monoclonal antibody kit supplied by Institut Pasteur (Paris, France). Use of a well-calibrated in-house standard of known glycoprotein content enables reliable estimation of glycoprotein without the use of imported reference vaccine preparation. The use of an in-house standard signicantly reduces the cost per assay.

The immuno-capture ELISA described in this study can be used to quantify the rabies viral glycoprotein antigen in rabies vaccine manufacture. It is simple, rapid and provides accurate information about the natively folded glycoprotein antigen content in in-process samples and vaccine preparations. By using a standard ELISA kit with a monoclonal antibody known to react with the glycoprotein antigen site III it was possible to calibrate an in-house standard preparation. This study demonstrates the use of a calibrated in-house standard to measure glycoprotein antigen content in in-process and nal vaccine preparations. In addition to harmonization of in vitro potency assays the method described in this study can be adapted by vaccine manufacturers in the developing countries for measurement of glycoprotein antigen content with a well-calibrated in-house standard.

Acknowledgements We thank Dr. Charles E Rupprecht, Centers for Disease Control and Prevention, Atlanta, USA, Dr. Alexander I Wandeler, Center of Expertise for Rabies, CFIA/ACIA, Ottawa Laboratory Falloweld, Ottawa, Canada and Dr. Scott K Dessain, Cardeza foundation for Hematology, Thomas Jeerson University, Philadelphia, USA for critical reading of the manuscript. We also thank Sitarama Reddy, Indian Immunologicals Ltd for help with the RFFIT.

References
[1] Bahloul G, Jacob Y, Tordo N, Perrin P. DNA-based immunization for exploring the enlargement of immunological cross-reactivity against the lyssaviruses. Vaccine 1998;16:417e25. [2] Coll JM. The glycoprotein G of rhabdoviruses. Brief review. Arch Virol 1995;140:827e51. [3] Ferguson M, Heath A. Collaborative study report of a collaborative study to assess the determination of glycoprotein antigen content of rabies vaccines for human use. Biologicals 1992;20: 143e54. [4] Fournier-Caruana J, Poirier B, Haond G, Jallet C, Fuchs F, Tordo N, et al. Inactivated rabies vaccine control and release: use of an ELISA method. Biologicals 2003;31:9e16. [5] Hankins DG, Rosekrans J. Overview, prevention and treatment of rabies. Mayo Clin Proc 2004;79:671e6. [6] Lafon MP, Perrin P, Versmisse P, Sureau P. Use of a monoclonal antibody for quantitation of rabies vaccine glycoprotein by enzyme immunoassay. J Biol Stand 1984;13:295e301. [7] Miceli GS, Torroba J, Torres W, Madero JE, Diaz AM. Evaluation of standard reagents for radial immunodiusion assays. In vitro control of rabies vaccines. Rev Inst Med Trop Sao Paulo 2000;42:153e6. [8] Perrin P, Lafon M, Sureau P. Enzyme linked immunosorbent assay (ELISA) for the determination of glycoprotein content of rabies vaccines. In: Meslin FX, Kaplan MM, Koprowski H, editors. Laboratory techniques in rabies. 4th ed. Geneva: WHO; 1996. p. 383e8. [9] Perrin P, Morgeaux S, Sureau P. In vitro rabies vaccine potency appraisal by ELISA: advantages of the immunocapture method

T. Nagarajan et al. / Biologicals 34 (2006) 21e27 with a neutralizing anti-glycoprotein monoclonal antibody. Biologicals 1990;18:321e30. [10] Perrin P, Thibodeau L, Sureau P. Rabies immunosome (subunit vaccine) structure and immunogenicity. Pre- and post exposure protection studies. Vaccine 1985;3:325e32. [11] Seligman Jr EB. The NIH test for potency. In: Kaplan MM, Koprowski H, editors. Laboratory techniques in rabies. 4th ed. Geneva: WHO; 1973. p. 279e86. [12] Smith JS, Yager PA, Baer GM. A rapid uorescent focus inhibition test (RFFIT) for determining rabies virus-neutralizing antibody. In: Meslin FX, Kaplan MM, Koprowski H, editors. Laboratory techniques in rabies. 4th ed. Geneva: WHO; 1996. p. 181e91.

27

[13] Sureau P. Rabies vaccine production in animal cell cultures. In: Advances in biochemical engineering/biotechnology, vol. 34. Berlin, Heidelberg: Springer-Verlag; 1989. p. 111e25. [14] WHO Expert Committee on Biological Standardization, Thirty rst report. WHO technical report series no. 658. Geneva: World Health Organization; 1981. [15] WHO Expert Committee on Biological Standardization, Thirty seventh report. WHO technical report series no.760. Geneva: World Health Organization; 1987. [16] Wiktor TJ, Gyorgy E, Schlumberger HD, Sokol F, Koprowski H. Antigenic properties of rabies virus components. J Immunol 1973; 110:269e76.