A human IgG anti-Vi reference for Salmonella typhi with weight-based antibody units assigned

Abstract

Recent data showing the high incidence of typhoid fever in young children, the demonstration of safety and efficacy of a Vi conjugate for this age group, the safety and similar immunogenicity in infants when administrated concurrently with EPI vaccines, together with the interests of manufacturers and investigators in studying such conjugate vaccines prompted us to prepare a human IgG anti-Vi standard to facilitate this work. Volunteers were injected with an investigational Vi-recombinant Pseudomonas aeruginosa exoprotein A (Vi-rEPA) conjugate vaccine. Plasmas with the highest levels of IgG anti-Vi were pooled. The IgG anti-Vi content of this preparation, assayed by precipitin analysis with purified Vi, was 33μg/ml. Accordingly, the estimated IgG anti-Vi protective level of 3.5 ELISA unit/ml, derived from our efficacy trial of Vi-rEPA in 2-to-5 years old children, is equivalent to 4.3μg/ml. This reagent is suitable for comparison of immune response of Vi conjugate vaccines or for other purposes requiring anti-Vi titration.

1. Introduction

Typhoid fever, a serious and common disease in developing countries, has become difficult to treat because of increasing resistance of Salmonella typhi to commonly used antibiotics [1-3]. Typhoid causes considerable morbidity and mortality, especially in young children [4,5]. A unique complication of typhoid is prolonged asymptomatic gall bladder infection, which may lead to malignancy and serve as a source of infection to others [6,7].

Vi capsular polysaccharide (Vi) is a licensed vaccine in more than 90 countries; its production includes western and Asian manufacturers [8-11]. However the efficacy of Vi in children < 5 year-olds is still questionable [12, unpublished data]. A Vi protein conjugate, Vi-rEPA, was shown to be safe and confer ~90% efficacy for 4 years in 2-to-5 years old children in a high endemic region of Vietnam [13-16]. A protective level of 3.5 ELISA Units (EU) anti-Vi IgG was estimated from this study based on the geometric mean (GM). Additional studies in this age group showed the dosage of 25 μg of Vi-rEPA to be optimal [16]. A recent study of Vi-rEPA, administered to healthy infants in Vietnam concurrently with their routine EPI vaccines (DTP+ OPV+ HepB), elicited protective levels of IgG anti-Vi antibody in approximately 95% of the vaccinees. No interference with responses to the EPI vaccines was found [17]. These findings raised interest in investigation and production of Vi-conjugates [18-21].

To evaluate new Vi-based vaccines, it is necessary to quantitate accurately serum Vi antibodies. ELISA has been the most common method used to determine levels of anti-Vi [22-24]. Evaluating vaccine trials successfully depends on the availability of a standardized human reference. Here we describe the preparation and characterization of a human reference, whose level of IgG anti-Vi was determined by quantitative precipitation with a purified Vi, as was done for Haemophilus influenzae type b, Streptococcus pneumoniae and Neisseria meningitides [25-28].

2. Materials and Methods

2.1 Vaccination

The study protocol was approved by the Internal Review Board of Lanzhou Institute of Biological Products (LIBP), China. Vi conjugated to recombinant exoprotein A of Pseudomonas aeruginosa (Vi-rEPA, lot 060901) was prepared by LIBP. The vaccine contained 24μg Vi, bound to 26μg rEPA, in 0.5 ml dose and passed the required safety tests by the Chinese Sino Food Drug Administration. Eight healthy volunteers, ages 18-40, were recruited, the purpose of the study explained, informed consent signed and blood samples taken for screening. All volunteers were negative for hepatitis B surface and core antigens, had no antibodies to hepatitis C, gonorrhea, syphilis, human T-cell leukemia virus type 1 and human immunodeficiency virus 1 and 2. Before injection, a physician or a nurse gave physical examinations to the volunteers and measured their temperatures. Those, whose temperature was ≤37.5°C and were not ill, received an IM injection of Vi-rEPA and their temperatures and local reactions recorded at 30 min, 6, and 24 h post immunization.

2.2 Plasma preparation

Six weeks after immunization, a blood sample was taken from each volunteer and anti-Vi IgG measured by ELISA (vide infra). Volunteers with levels higher than 30 ELISA units (EU) were asked for a plasmapheresis within 10 days. Five plasmas with the highest IgG anti-Vi were pooled, de-identified, freeze-dried and stored at −70°C (Lot 091010). The plasma pool was screened again at the NIH and confirmed to be free of blood transmissive viral diseases by molecular testing [29].

2.3 IgG purification

The freeze-dried plasma was reconstituted in pyrogen-free water (Baxter) at 100 mg/mL, centrifuged twice at 15,200 × g (RC6 SS-34, Sorvall,) at 4°C for 2h, the precipitate discarded and the supernatant stored at −70°C. IgG was purified by octanoic acid precipitation (minimum 99%, Sigma Aldrich) followed by DEAE-cellulose chromatography (Sigma Aldrich) [30,31]. The DEAE-cellulose eluent (20 mM Na-phosphate buffer at pH7.5) was assayed for IgG anti-Vi by ELISA and the first peak, containing the antibody was pooled, sterile filtered (0.45μm low protein binding, Millipore), stored at 4°C and designated as Vi-IgG_R1.

For final container bottling, 1% sucrose was added to reduce water sorption, stabilize protein structure integrity, prevent physical collapse during freeze drying cycle and for ease of solubility upon reconstitution. Ten percent sucrose was sterile filtered (0.22μm low protein binding, Millipore) and added to Vi-IgG_R1 to final concentration 1% sucrose, sterile filtered, aliquot in 1.12 ml and freeze dried. Each vial contains 1 ml of Vi-IgG_R1 and 0.1 ml sucrose. The extra 0.02 ml was for possible adherence to the vials.

2.4 Vi preparation

To avoid the potential interference of group D lipopolysaccharide from S. typhi, Vi was purified from Citrobacter freundii strain WR7011, whose capsular structure is identical to that of the Vi from S. typhi [13]. Vi purified contained <1% protein and nucleic acids, had no detectable endotoxin (<0.0001 endotoxin unit per mg) and stored as a freeze-dried powder at −40°C [11,18]. Before use, the Vi was freeze-dried again for 3 days to remove residual moisture. A stock solution was prepared by swelling the Vi with a small amount of PBS at 4°C overnight, brought to 2.0 mg/ml in PBS, and stored at 4°C.

2.5 Quantitative Precipitation with Vi polysaccharide

Quantitative precipitin analysis followed published procedures [32]. Briefly 3 ml aliquots of plasma were delivered to conical glass vials and mixed with 100 μl of PBS containing 1 to 200 μg Vi polysaccharide. Controls were 100 μg Vi in 3.1 ml of PBS or plasma alone plus 100 μl of PBS. All samples were in duplicates. The vials were capped and placed at 37°C for 1 h, then 4°C for 5 days during which they were gently mixed twice daily. The vials were centrifuged (2,800 × g, Sorvall Legend RT) at 4°C for 1 h, the supernatants decanted and inverted to drain for 1 h. The precipitates were washed 3 times with 1 ml cold sterile saline, centrifuged at 2,800 × g for 1 h and the supernatants decanted and tested for residual Vi antibody (vide infra). The precipitates were dried at room temperature and 1.0 ml of 1% sodium dodecyl sulfated added to each vial followed by gentle mixing. The precipitates were allowed to dissolve at room temperature until no particles were visible. The protein concentration was determined by UV absorption at 280 nm using the extinction coefficient ε =13.8 for human immunoglobulin and by colorimetric reaction with bicinchoninic acid (Pierce) using bovine serum albumin (BSA) (Pierce) as a reference [25,26,33].

2.6 ELISA for anti-Vi

Vi (1 mg/ml in saline) was stored at −40°C. Plates (Nunc Maxisorp, ThermoFisher) were coated with Vi (100μl of 2μg/ml in PBS pH7.4), sealed, incubated at RT overnight, and washed (0.85% NaCl, 0.1% Brij 35, 0.02% NaN₃). The coated plates were blocked at 25° C for 2 h with 1% BSA in PBS, filtered through 0.45 μm SFCA membrane (Millipore). Serum samples were diluted in dilution buffer (1% BSA, 0.1% Brij35, PBS). An in-house Vi reference serum (Vi-IgG_NIH) from a volunteer injected with Vi was assayed in parallel on each plate [14-17]. Ten serial 2-fold dilutions of each serum were carried out on the plate. The plates were covered and left at 25°C overnight. Murine MAb specific for human IgG (HP6043, CDC), IgM (HP6083, CDC) and IgA (HP6104, CDC) were used as the second antibodies; diluted 1:4000 for IgG,1:2000 for IgM and IgA. The plates were incubated at 25°C for 5 h, 100 μl/well of alkaline phosphatase labeled rat anti-mouse IgG (Jackson Immuno Research) added, and incubated at 25°C overnight. Lastly the substrate, 4-nitrophenyl phosphate disodium salt hexahydrate (Fluka) in 1 M Tris, 3 mM MgCl₂, pH 9.8 was added.

Plates were read at 405 nm after 20 to 40 minutes or when the OD reading of the highest concentration reached approximately 2.0 to 3.0. IgG anti-Vi levels were computed with an ELISA data processing program provided by the Biostatistics and Information Management Branch of the Centers for Disease Control and Prevention [34]. The titration curves, used for computation by linear regression fit, were compared with the in-house Vi reference Vi-IgG_NIH assigned a value of 75 EU [14-17]. Antibody concentrations with a mean standard deviation < 15% were accepted.

2.7 Characterization of Vi-IgG_R1

Immunodiffusion between Vi (10 μl at 200 μg/ml) and reconstituted plasma (20 μl) was conducted in 1% agarose (Invitrogen) in PBS (pH 7.2) and stained with Coomassie blue.

The molecular weight of the Vi-IgG_R1(0.07 μg to 0.7μg/loading) was analyzed by 10% SDS-PAGE Bis-Tris gels, stained with Coomassie Blue and compared to protein molecular standards (Invitrogen). Western blots with goat anti-human IgG (Sigma Aldrich) was performed for identification of IgG.

2.8 Verification of antibody concentration of Vi-IgG_R1 in the final container

Five vials of bottled Vi-IgG_R1 were reconstituted in pyrogen free water (1 ml/vial). Samples were removed from each of the 5 vials and analyzed for IgG anti-Vi concentration by ELISA, repeated 12 times per vial on 6 separate plates on different dates. The levels of IgG anti-Vi in the vials were evaluated, mean standard deviation calculated, and the GM compared with those obtained before and after addition of 1% sucrose, before and after bottling. The excessive amount of reconstituted Vi-IgG_R1 was aliquot into micro tubes with rubber-seal caps to prevent lost of moisture and stored at −35°C.

Samples of Vi-IgG_R1 from the final container stored at 4°C for 9 months and from aliquots stored at −35°C for 6 months were removed for preliminary stability study by ELISA.

3. Results

3.1 Quantitative precipitation and ELISA of Vi antibodies

Vi-IgG_R1showed a characteristic precipitation curve (Fig.1). Maximal precipitation occurred at 1.6 μg Vi/ml measured by both protein assays. The concentration of IgG anti-Vi at the equivalence zone was 33μg/ml.

Fig. 1.

Quantitative precipitin curve of Vi polysaccharide and IgG purified from plasma of volunteers immunized with Vi-rEPA. Anti-Vi concentration was determined by absorption at 280 nm. The peak value, using percent extinction coefficient ε=13.8 for immunoglobulin, is equivalent to 33μg/ml.

Assayed by ELISA along side with our in-house standard Vi-IgG_NIH, Vi-IgG_R1 contained 26.57± 1.2EU of IgG anti-Vi. Thus 1 EU is equal to 1.24μg/ml. Vi-IgG_R1 had non-detectable IgM and IgA at the dilution used for the IgG ELISA measurements.

3.2 Characterization of Vi-IgG_R1

There was a faint line of precipitation between the plasma and Vi in immunodiffusion (not shown). The molecular weight of Vi-IgG_R1was ~150 KD in SDS-PAGE (Fig. 2a). At 10 times the loading concentration, another faint band (at > 188 KD) was visible. Both bands reacted with goat anti-human IgG by western blot.

Fig. 2.

Molecular weight and subclass antibody analysis of Vi-IgGR1in precipitation with Vi*

* 1% SDS solution of Vi-IgG_R1 and Vi precipitation at the peak of the quantitative precipitin analysis. SDS-Page of 10%Bis-Tris gel was loaded with 0.07μg (lane 2) or 0.7μg (lane 3) of the anti-Vi and molecular weight reference (lane 1); (2a) Coomassie blue stain, (2b) western blot with goat anti-human IgG, (2c) western blot with goat anti-human IgM.

3.3 Validation of the antibody concentration in the Vi-IgG_R1 final container

After reconstitution with pyrogen free water, theVi-IgG_R1 in the final container dissolved instantly and the solution appeared clear. The GM IgG anti-Vi of the 5 vials tested was 32.79±1.21 μg/mL, similar to that before bottling (32.95±1.49 μg/mL). The variation among vials were < 9% which is within the experimental error (Table 1). The GM IgG anti-Vi level before and after addition of 1% sucrose were 32.95 μg/mL and 31.91μg/mL respectively, also within the experimental error. Thus the addition of 1% sucrose or freeze-drying did not change the level of Vi antibody in Vi-IgG_R1.

Table I.

Concentration of IgG anti-Vi of the Vi human reference serum Vi-IgG_Ri in the final container compared with those in the bulk measured before, during and after the bottling process

Final container   (Vial #)	IgG Anti-Vi (μg/mL)
1	31.36*
2	31.91*
3	33.01*
4	33.31*
5	34.36*
Average±SD	32.79±1.21
Bulk (no sucrose)	32.95+±1.49
Bulk (with 1% sucrose)	31.91+
Bulk (with 1% sucrose, centrifuged)	31.40+
Vial stored for 9 months, 4°C	32.61+
Aliquot store for 6 months,−35°C	31.22^

^*Each represents GM of 12 ELISA repeat runs of the same sample

⁺Each represents GM of 4 ELISA repeat runs of the same sample

^{^}GM of 4 ELISA repeat runs of 3 aliquot samples

The preliminary stability study of Vi-IgG_R1 in the final container stored at 4° C for 9 months or in aliquots stored at −35°C for 6 months showed no obvious degradation in the level of IgG anti-Vi (Table 1)

4. DISCUSSION

In line with human anti-capsular reference sera for Hib, pneumococcal or meningococcal vaccines, we prepared for S. typhi anti-capsular human reference. This new global reference should facilitate immunogenicity comparison and evaluation for typhoid Vi conjugate vaccine. We prepared Vi human reference anti-IgG (Vi-IgG_R1) from pooled plasma of volunteers immunized with Vi-rEPA. Quantitative precipitin analysis showed that Vi-IgG_R1 contained 33μg/ml of anti-Vi. Calibrating with the NIH in-house reference serum used in our Vi conjugate clinical trials, whose IgG anti-Vi was measured by ELISA, the converting factor is 1EU=1.24 μg. Accordingly, the estimated protective level of 3.5 EU/ml IgG anti-Vi, derived from the geometric mean of the 2-3 years old group in our efficacy trial of Vi-rEPA in Vietnam, is 4.34 μg/ml [15,16]. This value is likely over estimated since there were no cases of typhoid fever in the 2-3 year-olds (the younger age group) during the last 2 years of the trial and nearly half of the children had less than this geometric mean level, yet with the efficacy was >84%. This level is higher than the proposed protective level of 0.6 to 2.0 μg/ml based on the Vi polysaccharide trial in South Africa measured by radioimmunoassay [8]. The discrepancy is probably due to different serologic assays, antibody standards, and over estimation from the efficacy trial in Vietnam [16].

Asymptomatic S. typhi carriers remain to be a source of transmission and their detection is important for typhoid control [35]. Carriers harbor S. typhi in the biliary tract for prolonged periods and have high levels of anti-Vi. Measuring serum anti-Vi level is a reliable way to screen for carriers. Most investigations used passive hemagglutination or ELISA to estimate the level of Vi antibodies [35-37]. Without referring to a common reference, analyses done in different laboratories may not be comparable and there is no consensus on the range of Vi antibodies in carriers. Quantitative analyses of serum IgG anti-Vi levels from individuals with acute and chronic S. typhi infections should provide more reliable information.

A standard serum with an assigned weight-based antibody concentration would allow inter-laboratory comparisons of the levels and duration of “natural”, disease- and vaccine-induced serum IgG anti-Vi or to correlate with other assays such as functional assay [38].

High lights.

• A S. Typhi anti-capsular human reference serum (IgG anti-Vi) established

• An estimated protective level in μg/mL against typhoid fever proposed

• The reference serum is available to regulatory agencies and global manufactures

• The human reference serum can facilitate Vi conjugate vaccine development.