Improved Antigen Yield in Pandemic H1N1 (2009) Candidate Vaccine Viruses with Chimeric Hemagglutinin Molecules

Abstract

The candidate pandemic H1N1 vaccine virus NIBRG-121 was derived by reverse genetics and comprises the hemagglutinin (HA) and neuraminidase (NA) genes from A/California/7/2009 (CAL) on an A/Puerto Rico/8/34 (PR8) backbone. NIBRG-121 was found to grow poorly in eggs, compared to seasonal H1N1 candidate vaccine viruses. Based on our previous study with H5N1 candidate vaccine viruses, we generated two new viruses with chimeric PR8/CAL HA genes. Here we show that these new viruses have considerably improved growth in eggs and are therefore better candidate vaccine viruses for use in production of pandemic H1N1 (2009) vaccine.

TEXT

Wild-type pandemic A (H1N1) influenza viruses grow exceedingly poorly in eggs, in line with seasonal A (H1N1) viruses. The candidate vaccine virus (CVV) NIBRG-121 was generated at the National Institute for Biological Standards and Control by using reverse genetics (6, 8, 9, 11). It comprises the hemagglutinin (HA) and neuraminidase (NA) genes from A/California/7/2009 (CAL) on an A/Puerto Rico/8/34 (PR8) backbone. Early characterization of NIBRG-121 showed that it grew to a much higher titer in eggs than the wild-type strains; however, the virus did not grow as well as seasonal H1N1 vaccine viruses (8). Due to the potential high demand for a pandemic vaccine, a virus that grows to high titers and provides for a maximum number of doses per manufacturing batch is highly desirable.

Based upon results obtained in this laboratory in a previous study that was focused on improving candidate vaccine viruses for H5N1 vaccine manufacture (2), we created two new viruses from NIBRG-121. In these viruses, the six internal genes were derived from PR8 and the NA gene was derived from A/California/7/2009, while the HA gene was a chimera with sequences derived from both PR8 and A/California/7/2009 HAs. We demonstrate here that one of these viruses, NIBRG-119, has significantly improved growth in eggs and antigen yield.

Two chimeric HA genes were constructed in a series of overlapping PCRs (Fig. 1a shows a schematic diagram of the constructs). The modified HA genes were inserted into the pPST vector (9) by using unique SapI restriction sites and standard cloning protocols. Two new viruses with these HAs were derived by reverse genetics (6, 9). Virus NIBRG-118 has an HA in which the 3′-noncoding region (NCR), signal peptide, and ectodomain are derived from A/California/7/2009, while the transmembrane domain, cytoplasmic tail, and 5′-noncoding region are derived from PR8. Virus NIBRG-119 has an HA with the ectodomain of A/California/7/2009 and all other domains (3′- and 5′-noncoding regions, signal peptide, transmembrane domain, and cytoplasmic tail) from PR8. Alignments of the 3′-NCR and signal peptide (Fig. 1b) and transmembrane domain, cytoplasmic tail, and 5′-NCR (Fig. 1c) for PR8 and NIBRG-121 are shown in Fig. 1. The HA and NA sequences of the new viruses were confirmed by sequencing.

Fig. 1.

(a) Schematic diagram of HA chimeras. For each construct, areas of sequences which were derived from A/California/7/2009 are shown as white boxes, and areas of sequences derived from PR8 are shown shaded in gray. SP, signal peptide; TM/tail, the transmembrane and cytoplasmic tail components of the HA. Designations of the corresponding reverse genetics viruses are shown to the right of each HA construct. (b and c) Alignments of nucleotide and amino acid sequences between PR8 and NIBRG-121 HA. Nucleotides that are common between the two strains are indicated by an asterisk under the aligned nucleotide. Amino acids that are different between the two strains are indicated in bold with underlining. (b) The 3′-noncoding region and signal peptide region. The noncoding region is shown from the start of the sequence until the dashed line. The signal peptide is located between the two vertical dashed lines. The nucleotides and amino acids shown after the second dashed line are the start of the ectodomain. (c) The transmembrane domain, cytoplasmic tail, and 5′-noncoding regions of PR8 and NIBRG-121. The transmembrane domain is the sequence before the dashed line, and the cytoplasmic tail region is the sequence between the dashed line and the stop codon, which is indicated inside the boxed area. The 5′-noncoding region follows the stop codon.

Hemagglutination titers of viruses NIBRG-118 and NIBRG-119 were determined from harvested allantoic fluids, prior to the preparation of virus concentrates, by using standard methods with a 0.7% suspension of turkey red blood cells. HA assays showed that PR8 (HA titer, 4,096), NIBRG-118 (HA titer, 1,024), and NIBRG-119 (HA titer, 1,365) grew significantly better than NIBRG-121 (HA titer, 256; analysis of variance [ANOVA] P values of <0.01) (4) (Fig. 2a). Fifty percent egg infective dose (EID₅₀) assays were performed as described previously (2), and the results were calculated as described by Reed and Muench (7). NIBRG-118 and NIBRG-119 allantoic fluids had statistically significantly higher EID₅₀ titers than NIBRG-121 (10¹⁰ and 10^10.1 EID₅₀/ml, respectively, versus 10^8.7 EID₅₀/ml; P < 0.001) (Fig. 2b). Figure 2c shows the growth kinetics of the viruses in embryonated hens' eggs. NIBRG-118 and NIBRG-119 displayed improved growth compared to NIBRG-121. Changes in packaging signals of the HA segment can affect virus growth (5), and to investigate whether the above differences in growth were due to viral RNA (vRNA) packaging effects, we used a quantitative PCR (qPCR) assay to assess the amount of HA vRNA packaged compared to the NP vRNA segment, which served as a control. Briefly, vRNA was extracted from infected allantoic fluids or purified virus concentrates. The vRNA was reverse transcribed into cDNA by using a universal primer, and a probe-based multiplex qPCR assay was used to measure the amounts of HA and NP template in each cDNA sample. The results (Fig. 2d) are expressed as the HA/NP ratio. The data show no significant differences in the amounts of HA vRNA packaged between NIBRG-121, NIBRG-118, and NIBRG-119.

Fig. 2.

(a) Hemagglutination titers of allantoic fluids prior to virus concentrate preparation. (b) Infectivity titers of chimeric HA viruses. (c) Multicycle growth curve for PR8, NIBRG-121, NIBRG-118, and NIBRG-119. Each data point is the geometric mean value of 4 samples, and error bars denote the standard deviations. (d) Ratio of HA and NP RNA in virus particles, assessed using reverse transcription-qPCR. Values shown for panels a, b, and d are the averages of at least 3 independent experiments, and error bars show standard deviations. In all cases where a sample was significantly different from NIBRG-121, the P value is indicated.

Purified concentrates of the two new viruses, as well as of PR8 and NIBRG-121, were prepared as previously described (2), and their HA content, reflecting the amount of HA per virion, was analyzed using SDS-PAGE including a deglycosylation step as previously described (3) (Fig. 3a). Quantitation was carried out by scanning gels in an Imagescanner using the associated software (GE Healthcare). Calculated HA content is shown in Fig. 3b relative to NIBRG-121, which was taken as 100%. PR8 had a higher HA content (120%) than NIBRG-121, but this difference was not found to be statistically significant by ANOVA (P = 0.1) (4). NIBRG-118 and NIBRG-119 had values of 105% and 101%, respectively, and thus showed no significant differences in their HA contents compared to NIBRG-121 (P values of 0.9 and 0.99, respectively).

Fig. 3.

Yield analyses of chimeric viruses. (a) SDS-PAGE analysis of virus concentrates. All virus samples were reduced before loading on 10% SDS-PAGE gels (NuPage; Invitrogen). Viruses were compared with (PNGase F +) or without (PNGase F -) deglycosylation. (b) Quantitation of HA content. Values are expressed as the percentage of the control virus NIBRG-121. (c) TPY of purified virus concentrates are shown as total protein (in mg) per 100 eggs. (d) Total yield of HA. HA yields were calculated using SRD data (gray columns; taking into account the predilution of samples) and from data obtained from the SDS-PAGE and TPY assays (white columns). Values are shown as mg HA per 100 eggs. Values for panels b, c, and d are averages obtained from analyses of three virus concentrates; error bars denote the standard deviations. In all cases where a sample was significantly different from NIBRG-121, the P value is indicated.

Total protein yield (TPY) of purified virus preparations is often used as a good indicator for yield potential of CVVs (1, 8); therefore, we determined the TPY from each of the purified virus concentrates by using a bicinchoninic acid assay kit (Pierce); the results were read using a MultiSkan plate reader, and concentrations were calculated based on a standard curve using the associated software (Thermo Labsystems). The results, expressed as milligrams of protein per 100 eggs, are shown in Fig. 3c. Both NIBRG-118 and NIBRG-119 had reproducibly higher total protein yields than NIBRG-121. NIBRG-118 yielded 9.5 mg/100 eggs, NIBRG-119 yielded 11.1 mg/100 eggs, and NIBRG-121 yielded 5.5 mg/100 eggs; only the result for NIBRG-119 was found to be statistically significantly better than NIBRG-121 (P values of 0.02 and 0.06 for NIBRG-119 and NIBRG-118, respectively).

The final yield of HA protein is a function of the percent HA content (Fig. 3b) and the TPY (Fig. 3c), the latter being affected by virus growth (Fig. 2a to c). By using the results from TPY and SDS-PAGE analyses, the total yield of HA protein for each virus was calculated (Fig. 3d, white columns). Both viruses containing chimeric HA genes had statistically significantly improved total HA yields (4.4 and 5.0 mg HA/100 eggs for NIBRG-118 and NIBRG-119, respectively, versus 2.5 mg HA/100 eggs for NIBRG-121; P < 0.01 for both viruses).

As an alternative way to assess total HA antigen yield, we used the single radial immunodiffusion (SRD) assay (10). Samples were diluted to the same total protein concentration; A/California/7/2009 reference antigen reagent 09/146 (http://www.nibsc.ac.uk/documents/ifu/09-146.pdf), which was diluted to a concentration of 30 μg HA/ml, was included as a reference. SRD zones were measured using a Synoptics Image Analyser, and all comparisons were performed using the Combistats analysis software, v4.0 (EDQM). The results (Fig. 3d, gray columns) confirmed statistically significantly higher HA yields for both NIBRG-118 and NIBRG-119 than for NIBRG-121 (P < 0.01 for both [ANOVA]). Total HA antigen yield calculated from the SRD data and the dilution factors used in SRD mirrored the results obtained from TPY and SDS-PAGE analysis (Fig. 3d, compare the gray and white columns). The TPY and SDS-PAGE method measures the amount of HA protein physico-chemically, whereas the SRD method, which is used in vaccine manufacturing, measures an antigen-antiserum reaction and assigns a potency to the antigen relative to a standard reagent.

Finally, a hemagglutination inhibition assay (HAI) employing a panel of antisera was used to ascertain whether or not the antigenic characteristics of NIBRG-118 and NIBRG-119 were altered compared to NIBRG-121. The results (Table 1) confirmed that no change in antigenicity compared to NIBRG-121 occurred in viruses NIBRG-118 and NIBRG-119.

Table 1.

Antigenic analysis of viruses in the hemagglutination inhibition assay

Virus	Titer of ferret serum sample against indicated virusa
	F11/09	F37/09	F11/10	F12/10	F31/10	Negative control
A/Cal/7/09	1,280	2,560	1,280	2,560	1,280	<10
A/Narita/1/09	2,560	5,120	1,280	5,120	2,560	<10
A/Eng/195/09	1,280	2,560	1,280	1,280	1,280	<10
NIBRG-121	5,120	10,240	2,560	5,120	2,560	<10
NIBRG-118	2,560	5,120	2,560	5,120	2,560	<10
NIBRG-119	2,560	5,120	2,560	5,120	2,560	<10

^aFerret sera were raised against viruses as follows: F11/09, A/California/7/2009; F37/09, A/Narita/1/2009; F11/10, F12/10, and F31/10 were raised against A/England/195/2009.

Altogether, these data show that the two viruses containing chimeric HA genes are significantly improved as CVVs compared to the parental virus, NIBRG-121. Total protein yield was higher in NIBRG-118 and statistically significantly higher for NIBRG-119 compared to NIBRG-121. More importantly, due to the improved growth of the two new viruses, the HA antigen yield, whether measured by SRD assay or calculated using results from SDS-PAGE and TPY, was significantly improved for both NIBRG-118 and NIBRG-119 compared to NIBRG-121, with a yield from NIBRG-119 of >200% that of NIBRG-121. This should translate into higher antigen yields in vaccine production.

In our previous study with an H5N1 virus, the replacement of all domains of the H5 HA with the corresponding domains of PR8 except for the ectodomain led to a higher HA content (124% compared to NIBRG-14) in the virus particles; combined with a slight improvement in growth, the overall increase in HA yield was more than 100% (2). In the present study, using the same approach, we found that 2009 A (H1N1) viruses with chimeric HA genes had a significant improvement in HA yield via improved growth in eggs rather than improvement in the HA content of virus particles. In both studies, we made the modified viruses more “PR8 like” genetically. The inclusion of PR8 domains in the HA gene/protein of the H5N1 (2) and H1N1 viruses so far analyzed has resulted in improved yield characteristics, and further investigation is required to determine the underlying mechanism(s).

The results of the qPCR assay (Fig. 2d), which showed no difference in the packaging of HA vRNA, suggest that the mechanism is not based on differential packaging of vRNAs. It is more likely that the improvement observed is due to altered behavior of the chimeric HA protein itself; furthermore, we cannot rule out an effect on transcription, mediated by the altered RNA sequence. As PR8 is highly adapted to growth in eggs, it is likely that the improvement seen with these chimeric viruses is very specific to the egg substrate. Importantly, we have shown here improvement of the antigen yield of pandemic 2009 H1N1 virus, demonstrating that this chimeric approach is not limited to the improvement of a single virus or subtype but may be more widely applicable as a way to improve HA antigen yield. Ultimately, if the introduction of PR8 regions into HAs of various subtypes is shown to reproducibly improve HA yields, a new type of CVV for seasonal and pandemic influenza vaccines could be introduced for routine use.