The 4^th batch of the Biological Reference Preparation (BRP) with an assigned titer of 11 international units per vial (IU/vial) was provided by the European Directorate for the Quality of Medicines (EDQM, Strasbourg, France). Reference vaccines were stored at -20 ℃ and reconstituted with 1 mL deionized water for injection. All reference dilutions indicated in the text were prepared with phosphate buffered saline (PBS) from these stock solutions (Table 1).

Table 1.  Vaccines for mice immunization

As shown in Table 1, test vaccines were inactivated rabies vaccines for veterinary use from 4 different manufacturers, including Nobivac^® Rabies (Batch No. A154A01, Intervet International B.V), Rabvac^TM3 (Batch No. 1215337A, Fort Dodge Animal Health), Rabisin-R (Batch No. L369409, Merial SAS. France) and Rabigen^®mono (Batch No. 33XZ, Virbac Laboratories).

Rabies challenge virus standard CVS-24, a fixed laboratory strain, was propagated in mouse brains, and vials containing CVS suspensions were stored at -80 ℃. The intracerebral (I.C.) LD₅₀/0.03mL of the virus seed was 10^4.50 in mice. Dilutions of CVS-24 were made in distilled water supplemented with 2% horse serum. CVS-11 virus was grown in BHK-21 cells and used in the neutralization test to monitor the neutralizing antibody titer. The TCID₅₀/mL of the virus seed was 10^7.25 in BHK-21 cells.

The standard anti-rabies serum was obtained from Agence Fran aise de Sécurité Sanitaire des Aliments (AFSSA, Nancy, France). A negative control serum was prepared from unvaccinated mice.

The test was performed according to the standard operating procedure (SOP) of the EP ^[18]. In brief, SPF out-bred mice, strain Kunming, at the age of 4 weeks, were selected at random and placed into groups of 10 animals. Four five-fold dilutions of reference and test vaccines were freshly prepared. On day 0, each group was injected intraperitoneally with 0.5 mL of a vaccine dilution. At the same time, the challenge dose (mouse intracerebral LD₅₀) of CVS-24 suspension was measured in control mice and calculated using the Spearman-K rber formula. On the 14th day after vaccination, each treated mouse was injected intracerebrally with 0.03 mL of a CVS-24 suspension containing approximately 50 times of mouse LD₅₀. Mice with clear rabies symptoms were euthanized to avoid unnecessary suffering. The number of mice dying of rabies between the 5^th and 14^th day after challenge was recorded. The relative potency of the test vaccines was calculated by probit analysis using CombiStats software (EDQM).

Four-week-old Kunming mice were purchased from Changchun H & N Animal Breeding Center for Medicine (China) and divided in a randomized manner into seven groups (10 mice/group); i.e., Vac 1, Vac 2, Vac 3, Vac 4, Reference 1, Reference 2 and Reference 3 (Table 1) in the serological potency assay. According to prior experience ^{[14, 15, 28]}, each mouse was injected with 0.05 mL of vaccine in the upper hind leg using a 29G * 1/2″ insulin syringe (BD Ultra-Fine^TM, batch 0116106, USA), and peripheral serum samples were collected from the retro-orbital plexus on the 14th day after treatment ^[6].

To demonstrate the vaccine potency assayed by the serological method in mice, unvaccinated beagles (3–4 months old) were divided into 7 groups (6 dogs /group); i.e. Vac 1–4 and reference 1–3 (Table 1). Each dog was inoculated with one dose of vaccine in the triceps brachii muscle of the right front leg, and serum samples were collected from the right front leg vein on the 14th day after vaccination.

All the mice and dogs used in this study were treated in accordance with DIRECTIVE 2010/63/EU OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL of 22 September 2010 on the protection of animals used for scientific purposes.

The antibody titers of serum samples were measured by the FAVN test performed according to the method described by Cliquet et al. ^[3]. Briefly, 3-fold serial dilutions of both the serum samples and the controls were prepared in microplates. Each serum dilution was tested in quadruplicate. 50 μL (100 TCID₅₀) of challenge rabies virus CVS-11 was also added to each well. After 1 h incubation at 37℃ in a 5% CO₂ humidified incubator, 50 μL of BHK-21 cell suspension containing 4×10⁵ cells/mL was added to each well and the plates were incubated for 48 h at 37℃. After fixing the cells in 80% cold acetone for 30 min at room temperature, the plates were stained by adding FITC-conjugated anti-rabies monoclonal antibody (prepared in our laboratory), incubating at 37℃ for 30 min and then washing with PBS-Tween 20. The plates were examined by fluorescence microscopy (Olympus Corp., Tokyo, Japan). Presence or absence of fluorescent foci in the cells was recorded. Endpoints were calculated as the inverse of the highest dilution showing no fluorescence, using the Spearman-K rber formula. A standard anti-rabies serum (AFSSA, Nancy, France) was included as a positive control in all assays, and data were calculated as IU/mL of rabies virus-neutralizing antibody.

The median (Md) and interquartile range (IQR) of serum antibody titers of each group were calculated using the SPSS 16.0 program (SPSS Inc., Chicago, IL, USA) for Windows by a non-parametric method ^[10]. By comparing the Md of the reference vaccines with that of the test vaccine, the potency of test vaccine can be semi-quantitatively determined within a range; e.g. > 2 IU/dose. However, when the Md of test vaccine is equal to that of the reference, the IQRs should be compared to determine the potency of the test vaccine in which, when the IQR of the test vaccine is larger than that of the reference, the potency of the test vaccine should be considered smaller than that of the reference, rather than vice versa. The conformity between the single-dose mouse protection test, the serological potency assay and immune efficacy in dogs was assessed by comparing with quality standards of test vaccines.

7 groups (10 mice / group) were respectively immunized with different vaccines, and then the antibody titers of serum samples were measured by the FAVN test. As shown in Table 2, on the 14th day after vaccination, the median (20.58 IU/mL) of rabies-neutralizing antibody titers of group Vac 1 exceeded reference 3 (15.64 IU/mL), and the medians of group Vac 2, Vac 3 and Vac 4 were semiquantitated as being between reference 2 (3.01 IU/mL) and reference 3 (15.64 IU/mL). As a result, the potencies of the four test vaccines were in accord with their required qualities.

Table 2.  Antibody responses to different vaccines in serological potency assay

To evaluate the accuracy and repeatability of the serological potency assay-based FAVN, repeat tests (3 times) and a single-dose mouse protection test were performed. As shown in Table 3, all results of repeat tests by the serological potency assay and of the single-dose mouse protection test accorded with the required qualities of test vaccines.

Table 3.  Conformity between single-dose mouse protection test and serological potency assay

The antibody titers of serum samples assayed by FAVN test on 14^th day after vaccination are shown in Table 4, and the potency of test vaccines was evaluated by comparison of the protective efficacy of the test and reference vaccines in dogs. The median titer (6.15 IU/mL) of the rabies-neutralizing antibody of the group Vac 1 was equal to that of the group reference 3; however, the IQR (7.34 IU/mL) of Vac 1 was less than that (8.20 IU/mL) of reference 3; i.e., the protective efficacy of Vac 1 in dogs was higher than that of reference 3 (6.15 IU/mL). The protective efficacy of Vac 2, Vac 3 and Vac 4 in dogs was estimated to be between reference 2 (1.32 IU/mL) and reference 3 (6.15 IU/mL); i.e., the protective efficacy of the test vaccines in dogs was in line with their required potency.

Table 4.  Humoral immune response in dogs vaccinated with different potency vaccines

As described here, the serological potency assay, based on the FAVN test, is a semi-quantitative and rapid method for testing the potency of inactivated rabies vaccines. The FAVN test is a reliable method recommended by OIE (World Organization for Animal Health), and shows good reproducibility and accuracy.

In general, the potency requirements of inactivated rabies vaccines is determined as no less than 1 IU/dose or 2 IU/dose (potency requirements). The NIH test and single-dose mouse protection test are widely used in vaccine manufacturing for the potency determination of semi-finished and finished products ^{[18, 24, 26]}, in which more than one hundred mice are challenged with rabies virus, with the whole process taking at least one month to complete. However, depending on the required potency (≥1 IU/dose or ≥2 IU/dose) of inactivated rabies vaccines, a semi-quantitative assessment (e.g. > 1 IU/dose or > 2 IU/dose) of the potency is enough for vaccine quality control.

In the serological potency assay, only 10 mice are used for immunization with each test and reference vaccine. Use of the serological assay instead of the challenge test, therefore, reduced animal numbers by almost 90%. In most cases, the titers of antibody reach a peak by the 14^th day after vaccination ^[15], on which day blood samples are collected from the retro-orbital sinus of mice for the FAVN test. Therefore, the serological assay requires shorter time than the NIH test and single-dose mouse protection test.

Currently, no locally produced veterinary inactivated vaccine against rabies in the Chinese market is available, and only 4 different brands of imported inactivated vaccines (Nobivac^® Rabies, Rabvac^TM3, Rabisin-R and Rabigen^®mono) have been approved by the Chinese government. These veterinary vaccines usually contain adjuvant (e.g., alum), whereas reference vaccines do not. In the NIH test and single-dose mouse protection test, the test vaccines are serially diluted, and so the action of adjuvants in vaccines will be reduced because of their low concentrations after dilution. Additionally, the potency of adjuvanted vaccines (antigen and adjuvant) would not be measured correctly by challenge tests. However, in our serological assay, test vaccines were not diluted before testing.

Some data have demonstrated that the routes and dosage of administration can have a significant impact on the immune efficacy of vaccines in mice ^{[15, 28]}. The EP recommends intraperitoneal injection for the challenge test, and intramuscular or subcutaneous routes for the serological assay. In our experience, the intramuscular route is easier to perform for small volume (50 µL) vaccination than subcutaneous and intraperitoneal injection. We have found that 50 µL of one dose of adjuvanted vaccine for dogs is sufficient for the adult mouse immune system and can be injected completely in the back side of the upper hind leg. Less than this amount or more than the amount would produce non-detectable or excessive amounts of antibody, which would affect the judgment for the result (data not shown).

In early studies, a correlation has been found to exist between rabies-neutralizing antibody titers and protection from intracerebral challenge in mice ^{[7, 29]}. The titer of rabies-neutralizing antibody is not the only parameter of immune response, but is the most reliable indicator of immune protection. We have observed that titers of antibody induced by vaccines in mice do not follow a normal distribution, which might be due to a different out-bred species, i.e., Kunming mice, a widely used mouse breed in medical experiments in China, and therefore the median (Md) and interquartile ranges (IQR) were used to represent the immune efficacy of reference and test vaccines instead of the mean plus standard deviation ^[10]. By comparing the Md and IQR of antibody titers induced by reference and test vaccines, we can conclude whether or not the potency of the testing vaccines reaches or exceeds the potency requirement. To confirm the accuracy and repeatability of this serological potency assay, repeated tests, the single-dose mouse protection test and dogs vaccination were performed, and the results of the single-dose mouse protection test and immunization in dogs were compared with those of the serological potency assay. Our results showed that the immune efficacy in mice and dogs had high conformity with the potency assayed by the single-dose mouse protection method.

However, the antibody titers in individual mice varied significantly from 0.29 to 30.77 IU/mL for a given vaccine. Although we adapted the median (Md) and interquartile range (IQR) of serum antibody titers of each group, these variations should still influence the actual efficacy values of the vaccine, which is most likely due to the in vivo variation of immune responses to the vaccines. Therefore, an in vitro assay which determines the antigenic content (the trimeric form of G) would be expected to develop only when the adjuvant effect on the amount of G protein is excluded.

In conclusion, the serological method based on the FAVN test on the antibodies produced by certain amount of reference vaccine and a test vaccine is rapid, efficient, low-cost, and in accordance with the immunogenicity induced in dogs, and therefore can be recommended for batch potency testing during inactivated rabies vaccine production.