I have read the journal's policy and the authors of this manuscript have the following competing interests: BPS, IdlRO, VvH, VB, TK, TGU, RZ, JHHVC, HS, and DEM are employees of Janssen Infectious Diseases and Vaccines (Pharmaceutical companies of Johnson & Johnson). JL was employed at Janssen while contributing to the work described and has currently moved to Aecsela Biologics Little Compton, Rhode Island, USA. This does not alter our adherence to PLOS Pathogens policies on sharing data and materials.
Conceived and designed the experiments: BPS TK TGU JL VB RZ JHHVC HS DEM. Performed the experiments: BPS IdlRO VvH YS GC LEC. Analyzed the data: BPS IdlRO VB DEM JC JM RZ HS JC EW. Wrote the paper: BPS JHHVC VB HS DEM.
Current Address: Aecsela Biologics, Little Compton, Rhode Island, United States of America
The poliovirus vaccine field is moving towards novel vaccination strategies. Withdrawal of the Oral Poliovirus Vaccine and implementation of the conventional Inactivated Poliovirus Vaccine (cIPV) is imminent. Moreover, replacement of the virulent poliovirus strains currently used for cIPV with attenuated strains is preferred. We generated Cold-Adapted Viral Attenuation (CAVA) poliovirus strains by serial passage at low temperature and subsequent genetic engineering, which contain the capsid sequences of cIPV strains combined with a set of mutations identified during cold-adaptation. These viruses displayed a highly temperature sensitive phenotype with no signs of productive infection at 37°C as visualized by electron microscopy. Furthermore, decreases in infectious titers, viral RNA, and protein levels were measured during infection at 37°C, suggesting a block in the viral replication cycle at RNA replication, protein translation, or earlier. However, at 30°C, they could be propagated to high titers (9.4–9.9 Log10TCID50/ml) on the PER.C6 cell culture platform. We identified 14 mutations in the IRES and non-structural regions, which in combination induced the temperature sensitive phenotype, also when transferred to the genomes of other wild-type and attenuated polioviruses. The temperature sensitivity translated to complete absence of neurovirulence in CD155 transgenic mice. Attenuation was also confirmed after extended
The vaccines that are used to protect against poliovirus infection have been available since the 1950s and have brought the eradication of poliomyelitis to our doorstep. For the post-eradication era, an Inactivated Poliovirus Vaccine (IPV) based on attenuated Sabin strains is recommended, as these strains are currently the only option to move to safer manufacturing of IPV. Here we describe three novel poliovirus strains that cannot replicate at 37°C. Their lack of pathogenicity was confirmed by intracerebral inoculation of susceptible transgenic mice that subsequently did not develop any symptoms of poliomyelitis. The inability to replicate at 37°C is caused by multiple mutations which do not revert to virulence after passage in cells. Furthermore, when used as vaccines, these viruses were capable of inducing a potent immune response in rats. At low temperature (30°C) these viruses showed high productivity on the PER.C6 cell line, which has the potential to significantly reduce costs of goods, as previously shown for conventional poliovirus strains. Taken together, these new strains could contribute to a safe, genetically stable, efficacious and affordable IPV.
There are two vaccines that can effectively protect against poliomyelitis which have been available for more than 60 years and are still used today. The Inactivated Poliovirus Vaccine (IPV), today referred to as conventional (c)IPV, was developed in 1955 by Jonas Salk and contains three formalin inactivated, wild-type and neurovirulent poliovirus strains (Mahoney, MEF-1 and Saukett) [
Despite the efficacy of OPV, the Sabin strains have the propensity to revert to neurovirulent form [
However, even if eradication is achieved, immunization against poliomyelitis will remain necessary to maintain a polio-free world [
Currently the Sabin strains used in OPV are the preferred candidates to replace the wild-type strains. Sabin-based IPV’s (sIPV) have been recently licensed in Japan [
Our aim was to develop novel attenuated strains for IPV manufacture that can address the biosafety issues of cIPV without altering immunogenicity. Our approach for viral attenuation was to develop strains with impaired growth at physiological temperature (≥37°C) but that are still capable of replication to high infectious titer yields at lower (manufacturing) temperatures. We hypothesized that inability to replicate at 37°C would impede reversion to neurovirulent form, resulting in a non-pathogenic phenotype in the natural host. Cold-adaptation (adaptation to growth at low (<37°C) temperature by serial passage) is often associated with reduced replication (or sensitivity) at higher temperatures (37–40°C)). Historically, cold-adaptation has been frequently used to generate attenuated RNA and DNA viruses (reviewed in [
By combining empirical and rational methods of attenuation, we generated temperature sensitive poliovirus strains incapable of replication at physiological temperature, that grow to high titers at 30°C, and that have the antigenic profile of (wild-type) cIPV strains. The strains were obtained via serial passage at low temperature and genetic engineering, ultimately resulting in three Cold-Adapted Viral Attenuation (CAVA) vaccine strains, namely: CAVA-1 Mahoney, CAVA-2 MEF-1 and CAVA-3 Saukett. We characterized the CAVA strains with respect to
The highly temperature sensitive poliovirus strains were derived from, Brunenders, a Type I partially-attenuated poliovirus [
Panel A) Average and standard deviation of two (n = 2) replication kinetic curves of the Brunenders strain versus 3 selected clones (clone A, B and C) derived after passage with impaired growth at 37°C. Panel B) Average and standard deviation of three (n = 3) independent infections of Brunenders and the CAVA backbone, which contained all mutations from Clones A, B and C combined. Panel C) Average and standard deviation of three (n = 3) independent infections of the Brunenders strain versus the CAVA vaccine strains (CAVA-1 Mahoney, CAVA-2 MEF-1 and CAVA-3 Saukett).
Each clone had 18 nucleotide mutations (either shared or unique) with respect to the parental Brunenders strain. Overall, 31 distinct mutations were detected across the three different clones (see
Black vertical lines represent the synonymous CAVA mutations whilst red vertical lines represent non-synonymous CAVA mutations, dispersed over the poliovirus genome; a detailed description of the individual mutations is given in
All 31 mutations were cloned into the Brunenders cDNA plasmid and transfection of the resulting
To generate CAVA-IPV vaccine strains for poliovirus serotypes 1, 2 and 3, the capsid sequence of the CAVA backbone was replaced by the capsid sequences of each of the three cIPV strains, to mimic their antigenic profiles. This resulted in three new synthetically-derived viruses named CAVA-1 Mahoney, CAVA-2 MEF-1 and CAVA-3 Saukett. The remainder of the genome maintained 24 of the CAVA mutations spread over the 5’ Untranslated Region (5’UTR) and Non-Structural proteins (see
To visualize signs of infection, PER.C6 cells were infected with CAVA-1 Mahoney at 37°C and 30°C, at an MOI of 1 and crude harvests were taken 24–48 hours post infection for examination by electron microscopy (EM). PER.C6 cells were infected using the same conditions with either wild-type Mahoney or PBS (mock) at 37°C as positive and negative controls, respectively.
Panel A) PBS (Mock) infected cells, Panel B) Mahoney infection at 37°C, Panel C) CAVA-1 Mahoney at 37°C, Panel D) CAVA-1 Mahoney at 30°C.
We used intracerebral (i.c.) inoculation in of CD155 transgenic mice [
Virus | Virus Titer (log10 TCID50/ml) | Intra cerebral dose (log10 TCID50/mouse) | Experiment 1 | Experiment 2 | PLD50 (log10 TCID50) |
---|---|---|---|---|---|
CAVA-1 Mahoney | 9.9 | 8.4 | 0/3 |
0/5 | >8.4 |
Sabin 1 | 9.6 | 8.0 | 1/3 | 2/5 | >8.0 |
Mahoney | 10.1 | 4.0 | 2/2 | 5/5 | 2.0 |
CAVA-2 MEF-1 | 9.9 | 8.4 | 0/3 | 0/5 | >8.4 |
Sabin 2 | 9.2 | 7.7 | 0/3 | 0/5 | >7.7 |
MEF-1 | 10.0 | 6.0 | 2/2 | 5/5 | 4.5 |
CAVA-3 Saukett | 9.7 | 8.2 | 0/3 | 0/5 | >8.2 |
Sabin 3 | 9.9 | 8.4 | 0/3 | 0/5 | >8.4 |
Saukett | 9.7 | 4.0 | 3/3 | 5/5 | 2.6 |
* Proportion of mice with signs of paresis or paralysis. TCID50 = Tissue Culture Infectious Dose 50%, PLD50 = Paralytic of lethal dose (50%).
Upon i.c. inoculation of CD155 transgenic mice all three CAVA vaccine strains showed a highly attenuated phenotype. The maximum dose possible (8.2–8.4 Log10 TCID50/mouse) did not induce paresis or paralysis in any of the mice 21 days post inoculation (
To further investigate the temperature sensitive phenotype associated with the CAVA viruses and potential mechanisms of attenuation, we compared the levels of infectious titers, viral RNA and viral protein translation of the parental Brunenders, CAVA backbone, CAVA-1 Mahoney, and two intermediate viruses containing either the CAVA mutations in the internal ribosomal entry site (IRES) or the CAVA mutations in the Non-Structural proteins in the background of the Brunenders genome (see
Infections were performed in PER.C6 cells at an MOI of 1, at 30°C and 37°C. Harvests were subjected to infectious titer determination, quantitative reverse transcription PCR (RT-qPCR) and Western Blot analyses. Infection harvests were freeze-thawed and clarified prior to analysis and therefore represent the viral components in the cells and supernatant. At 30°C, all viruses showed similar replication kinetics and maximum infectious titers as compared to the parental Brunenders strain (
Viruses used were Brunenders (WT), Brunenders with the CAVA mutations in the IRES (IRES), Brunenders with the CAVA mutations in the Non-Structural proteins (NS), the CAVA-1 Mahoney (C1) vaccine strain and the CAVA backbone virus (CBB); Control (Ctrl) is an uninfected control. Data depict one representative infection (n = 1) measured once for infectivity and (n = 3) times for viral RNA and protein levels. Error bars represent standard deviation from the mean and one representative of three independent western blots is shown.
To study the changes in viral RNA levels during infection, the fold increase of genome copies from 0 to 48 hours post infection were quantified by RT-qPCR (
Detection of viral proteins was by performed by Western Blot (
Of the 24 CAVA mutations in the vaccine strains, 14 were selected for synthetic incorporation into different poliovirus backgrounds based on 1) conservation amongst a panel of polioviruses and 2) mutations causing amino acid changes or located in the IRES. The 14 mutations were incorporated in the Brunenders, MEF-1 and Sabin 3 genomes (Brunenders-14, MEF-1-14, and Sabin 3–14, respectively,
Viruses used were Brunenders, Brunenders—with 14 CAVA mutations, MEF-1, MEF-1 –with 14 CAVA mutations, Sabin 3, Sabin 3 –with 14 CAVA mutations and the CAVA backbone virus (see overview of all viruses in
To identify more exactly the molecular determinants of temperature sensitivity CAVA-1 Mahoney was serially passaged at 37°C and at low MOI (0.01 TCID50/cell), however, this always resulted in an inability to detect quantifiable virus, already after the first passage. Therefore, to select for viruses that had regained the ability to replicate at 37°C, conditions were altered by gradually increasing the infection temperature each subsequent passage. The first three passages were done at 33°C as previous experiments had shown productive infection of PER.C6 cells by the CAVA backbone virus at this temperature (
Serial passage was performed using CAVA-1 Mahoney in suspension PER.C6 cells infected at a cell density of 107 cells/ml at low MOI (0.01) and harvested at 3–4 days post infection. Temperature was gradually increased (33–37°C) or temperature was kept constant at 30°C for the control viruses. Panel A depicts the viral titers at each passage when titrated and incubated at 30°C or 37°C for the two independent experiments (n = 2). Panel B lists the reverting CAVA mutations of the viruses at passage number 6 where the nucleotide number refers to the position from the start of the viral genome.
During the first four passages the CAVA-1 Mahoney viruses passaged at 30°C and at increasing temperatures retained their temperature sensitive phenotype, as shown by the high virus titers in the crude harvests when titrated and incubated at 30°C, but there was an absence of quantifiable virus when titrated and incubated at 37°C. Capacity to replicate at 37°C was only regained between passage 4 and 5 where infection temperature was gradually increased per passage (
Full viral genome sequencing was performed at passage number 6. Reversion (partial and full) of four CAVA mutations to the nucleotide residues of the parental Brunenders genome was observed in both passaging experiments. Reversions to the Brunenders sequence were observed at nucleotide nt142 in the IRES, in the 2C (nt4428 I [101] V) and 3D (nt6210 M [74] V and nt6848 M [286] I). However, the mutations at nt142 and nt6848 were also observed in viruses that were passaged at 30°C as parallel controls and which still retained the temperature sensitive phenotype, this suggests that these mutations alone do not revert temperature sensitivity. Sequencing of the viruses passaged at increasing temperature also revealed 5 and 6 new mutations across the viral genome, for experiment 1 and 2, respectively. Of these, two mutations (nt127 in the IRES (which forms a base pair with CAVA mutation nt163) and nt918 in VP4 K [58] E) occurred in both experiments.
To evaluate stability of attenuation during vaccine manufacture, the CAVA vaccine strains were tested for
Prior to
Full genome sequencing of
To determine the immunogenic potential of the CAVA vaccine strains
Groups of ten (n = 10) rats were immunized with a full dose (FD: 100% 40:8:32 DU/dose or 150% 60:12:48 DU/dose) or a 1:2, 1:4 or 1:16 dilution of the full dose. Poliovirus type 1, 2 and 3-specific neutralizing antibody titers were determined by Sabin Virus Neutralizing Assay at day 21 post immunization. Each dot represents one individual animal; the connected line represents the geometric mean at each dose. Relative potency estimates and 95% confidence intervals of the difference between the CAVA vaccine strain and cIPV reference based on the number of seroconverting animals are depicted in the table, horizontal dotted line represents the seroconversion limit for each assay.
As eradication of poliomyelitis draws closer, the poliovirus field is moving towards novel vaccines and vaccination strategies. To serve as novel IPV strains, we generated three attenuated poliovirus strains using a combination of empirical and rational attenuation methods with specific focus on (genetic stability of) attenuation, immunogenicity, and affordability. Our approach for viral attenuation was to develop strains with impaired growth at physiological temperature (≥37°C) with high replicative capacity at (manufacturing) temperature. The CAVA strains were empirically derived by serial passage at low temperature, much like the Sabin strains; however, the subsequent synthetic combination of multiple mutations into one genome was essential to obtain a complete block in viral replication at 37°C.
The CAVA strains showed no sign of successful
The intermediate viruses with CAVA mutations in the IRES or Non-Structural proteins showed impaired, but not completely halted, growth at 37°C. Therefore the combination of CAVA mutations in these regions is required. More specifically, a combination of 14 mutations within those two regions was sufficient to cause the CAVA temperature sensitive phenotype. This was confirmed by introduction of the 14 mutations into other wild-type and attenuated polioviruses of differing serotypes, indicating that the mechanism of action is unique and independent of the parental Brunenders backbone.
The CAVA temperature sensitivity is likely exerted by multiple molecular mechanisms (as exemplified by the synergistic, combinatorial effect of the CAVA mutations) which work together to hamper replication at 37°C. However, at 30°C these mutations do not appear to obstruct virus replication, protein translation, or RNA replication. One explanation for this may be that the introduction of multiple mutations decreases the thermal stability of the viral proteins and/or RNA, resulting in folding defects, conformational changes and subsequent losses of biological functionality of the viral (precursor) proteins and/or functional RNA elements. When environmental thermal energy is lowered (for example at 30°C) the decreased thermal stability may not be sufficient to cause significant changes in protein/RNA structure and function to such an extent that virus replication is restricted. For example, the CAVA mutations in the IRES may destabilize the secondary RNA structure of this essential RNA element. Predicted secondary RNA structures of the CAVA and Brunenders IRES-domains II and VI show that the free energy (ΔG) is raised in the CAVA domains (as well as an altered domain II structure) indicating decreased thermostability (
Circled nucleotides (at positions 133, 142, 146, and 163 in domain II and at positions 597, 609 in domain VI) represent nucleotide changes between CAVA and Brunenders. The last remaining CAVA IRES mutation (nt579) lies outside of any IRES domains and in the spacer region between Domains V and VI. After serial passage at increasing temperature (
The CAVA mutations in the 2C and 3D proteins, which reverted after gradual increase of infection temperature (and which are part of the 14 selected mutations), may play particularly prominent roles in inducing the CAVA temperature sensitivity. However, the compensatory impact of the other CAVA mutations, or the additional new mutations identified after regaining replication capacity at 37°C, cannot be ruled out. The highly conserved 2C protein has multiple functions in host-cell membrane alteration, encapsidation, viral RNA binding and RNA replication [
The CAVA mutation 3D[74] is close to residue 73 in the palm of 3D, which in Sabin 1 has been implicated to play a role in temperature sensitivity via a temperature dependent decrease of VPg-uridylation compared to wild-type [
The CAVA mutation at residue 286 of the 3D is located in the middle finger domain of the polymerase close to a putative translocation domain [
Further research is required to pinpoint the exact molecular mechanisms of CAVA temperature sensitivity, and the responsible mutations. However, the vast number of mutations and combinations thereof makes full understanding of the CAVA phenotype a challenging endeavor. To illustrate, even the extensively studied Sabin strains do not have a fully understood molecular mechanism to explain their attenuation [
As expected, the inability to propagate at 37°C
Formaldehyde-inactivated versions of the CAVA vaccine strains were immunogenic and induced high neutralizing antibody titers
Although unpredicted, another explanation for the slightly lowered immunogenicity could be an incompatibility of the CAVA-backbone with the cIPV capsids and/or the lower culture temperature which may slightly alter the conformation of the virion during virus assembly causing changes in antigenicity and immunogenic potency. Indeed, biophysical characterization of the CAVA vaccine strains as compared to their cIPV counterparts did result in some unexpected differences. For example, antigenicity as measured by D-antigen unit per infectious unit (DU/TCID50) was similar for CAVA-1 Mahoney, higher for CAVA-2 MEF-1 and lower for CAVA-3 Saukett as compared to their cIPV counterpart (
Cost of IPV is a critical parameter to enable immunization of the developing world, an essential endeavor to achieve and maintain eradication. Therefore, strategies to increase IPV affordability are encouraged by the WHO[
Next generation IPV vaccine strains should ideally portray a non-infectious phenotype to reduce the risk of transmission and disease, should dissemination into the environment occur. The novel CAVA strains are characterized by an inability to replicate at 37°C and capacity to propagate to high titers at 30°C. Their unprecedented temperature sensitivity translated to a high level of
The Brunenders, MEF-1 and Saukett viruses were derived from virus seeds kindly donated by SBL (former Swedish Bacteriological Laboratories). Sabin 1, 2 and 3 were purchased at The National Institute for Biological Standards and Control (NIBSC, catalogue number: 01/528, 01/530, and 01/532, respectively). The Mahoney virus was purchased at the European Virus Archive (EVA).
All remaining viruses used were rescued via RNA transfection for which the RNA was transcribed
PER.C6 cells [
Infectious titer determination was performed in multi-well 96 plates seeded with 6.5x104 adherent PER.C6 cells per well in DMEM supplemented with 10% FBS and 10mM Magnesium Chloride. Eleven serial virus dilutions with a five-fold dilution factor were prepared and added to the cells with subsequent incubation for 13 days at 30°C for all titration assays, unless indicated differently. On day 13 each well was scored for CPE and titers were calculated by method of Spearman and Kärber [
EM was performed at Leiden University Medical Center. Infection harvests were fixed at 1 hour at room temperature in 1.5% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.4) and stained with 1% osmium tetroxide for 1 hour. Samples were pelleted in 3% agar where resulting pellets were cut and gradually dehydrated with an ethanol series. The samples were then infiltrated for 1 hour with a 1:1 mixture of propylene oxide and epoxy LX-112 resin (Ladd Research). After an additional hour in 100% epoxy LX-112, the samples were polymerized for 48 h at 60°C. Cell sections of 50 nm were cut, placed onto carbon-coated formvar grids, and counterstained with 7% uranyl acetate and lead citrate for 20 and 10 minutes, respectively. Imaging was performed with a Tecnai 12 BioTwin transmission electron microscope (FEI company) operated at 120 kV.
For detection of viral proteins 100 μg of total protein from clarified crude harvests (after 3 freeze thaw cycles) was loaded into a NuPAGE Novex Bis-Tris 4–12% protein gel (Life Technologies) and blotted onto Nitrocellulose membranes (Life Technologies). Membranes were blocked in 5% non-fat dried milk (Bio-Rad) and incubated overnight with a 1:1000 dilution of goat polyclonal antibodies against poliovirus types 1,2,3 (ProSci), followed by 2 hours with a 1:15000 dilution of donkey anti-goat IRDye 800CW secondary antibody (Westburg). Proteins were visualized using Odyssey infrared imaging system (Li-Cor, BioSciences).
Quantification of poliovirus RNA was performed by RT-qPCR using viral RNA isolated from clarified, freeze-thawed infection harvests using a QIAamp viral RNA isolation kit (Qiagen). Viral RNA was reverse transcribed to cDNA and subsequently amplified with the Power SYBR Green RNA-to-Ct 1-Step Kit (Life Technologies), using 400nM forward primer (5’ TCTCCTAGCCCAATCAGGAA 3’) and 400nM reverse primer (5’ TCTCCCATGTGACTGTTTCAA 3’) flanking an amplicon (86nt in length) in the 3D polymerase gene. Real-time PCR was performed in a 7500 Fast thermocycler (Life Technologies) starting with 30 min at 48°C for reverse transcription and 10 min at 95°C for activation of DNA polymerase, followed by 40 amplification cycles of 15 sec at 95°C for denaturation and 1 min at 63°C for annealing and extension. Purified,
Sequencing of the full viral genomes was performed by RT-PCR and Sanger sequencing as described previously [
RNA secondary structures predictions of the IRES domains II and IV of CAVA and Brunenders were executed by the MFOLD program (
Neurovirulence testing was performed at Stony Brook University using transgenic mice expressing the poliovirus receptor (CD155) [
CAVA harvests were treated with domiphen bromide to remove host cell DNA and consequently clarified of a series of filters. Prior to Cation Exchange Chromatography (CEX) the clarified harvests were acidified using 25mM sodium citrate. CEX was performed using Sartobind S cationic membranes. The CEX eluate was subjected to a Size Exclusion Chromatography step for further purification (polish) and buffer exchange. The SEC eluate was conditioned using M199 and glycine prior to inactivation. Inactivation was performed according to the EP guidelines and in line with Salk’s description of poliovirus inactivation procedure in the 1950s [
Rat Potency testing was performed at the National Institute of Biological Standardization (NIBSC).
The purified and inactivated monovalent CAVA samples were tested for D-antigen content by ELISA. This D-antigen ELISA utilizes polyclonal capture and monoclonal detection antibodies raised against active Sabin viruses. The inactivated CAVA viruses were consequently tested for monovalent
All mice used for
NIBSC’s Animal Welfare and Ethical Review Body approved the application for Procedure Project Licence Number 80/2523 which was approved by the UK Government Home Office and under which animal care and protocols for Rat potency testing were conducted. All animal care and protocols used at NIBSC adhere to UK regulations (Animals, scientific procedures, Act 1986 that regulates the use of animals for research in the UK) and to European Regulations (Directive 2010/63/EU of the European Parliament on the protection of animals used for scientific purposes). The experiments in rats shown here were carried out following protocol 1 within Home Office Procedure Project Licence Number 80/2523 referred above.
Viruses were capable of replication at 37°C.
(PPT)
(PPT)
(PPT)
(PPT)
(DOCX)
(PPT)
(DOCX)
An unpaired t-test was performed to assess if the difference in DU/TCID50 ratio between the CAVA strains and the respective cIPV strain is significant (two-tailed, α = 0.05). P-values are shown for each combination and an asterisk (*) represents a significant difference.
(DOCX)
Protocol adapted from [
(DOCX)
We would like to especially give thanks to Shraddha Dubey, Alies Brandjes, Philip Brouwer, and Masha Ivanova for their excellent work and technical contributions. Furthermore, we are very grateful for the many colleagues who contributed in the form legal support, and/or scientific discussions: Gert Scheper, Soumitra Roy, Wilfred Marissen, Beckley Kungah Nfor, Gerbrand Korten, Marloes Naarding, Maarten Santman, Annemieke Manten, Richard Verhage, and Ken Singleton. Our thanks also go to Montserrat Barcena and Aat Mulder from the LUMC for performance of the EM imaging.
We would like to give special thanks to Joanne Wolter for critical reading and reviewing of the manuscript.