Development and characterization of candidate rotavirus vaccine strains derived from children with diarrhoea in vietnam

In order to be partially self-sufficient in vaccine production, Vietnam has pursued the development of several rotavirus strains as candidate vaccines using isolates obtained from Vietna

Contents lists available atScienceDirect

Vaccine

j o u r n a l h o m e p a g e :w w w e l s e v i e r c o m / l o c a t e / v a c c i n e

Development and characterization of candidate rotavirus vaccine strains derived from children with diarrhoea in Vietnam

Luan T Lea,∗, Trang V Nguyenb, Phuong M Nguyenb, Nguyen T Huonga, Ngo T Huonga,

Nguyen T.M Huonga, Tran B Hanha, Dang N Haa, Dang D Anhb, Jon R Gentschc, Yuhuan Wangc, Mathew D Esonac, Roger I Glassd, A Duncan Steelee, Paul E Kilgoref, Nguyen V Mana,

Baoming Jiangc,∗∗, Nguyen D Hiena

aCenter for Research and Production of Vaccines and Biologicals (POLYVAC), 135 Loduc Street, Hai Ba Trung, Hanoi, Viet Nam

bNational Institute of Hygiene and Epidemiology, Hanoi, Viet Nam

cDivision of Viral Diseases, U.S Centers for Disease Control and Prevention, 1600 Clifton Road, Atlanta, GA 30333, USA

dFogarty International Center, National Institutes of Health, Bethesda, MD, USA

eVaccines and Immunization, PATH, Seattle, WA, USA

fInternational Vaccine Institute, Seoul, Republic of Korea

a r t i c l e i n f o

Keywords:

Rotavirus vaccine

Diarrhoea

Vietnam

a b s t r a c t

In Vietnam, rotavirus infection accounts for more than one-half of all hospitalizations for diarrhoea among children less than 5 years of age While new vaccines to prevent rotavirus diarrhoea have been developed and introduced into some countries by multinational manufacturers, the ability for developing countries such as Vietnam to introduce several new and important vaccines into the routine infant immunization schedule may be challenging In order to be partially self-sufficient in vaccine production, Vietnam has pursued the development of several rotavirus strains as candidate vaccines using isolates obtained from Vietnamese children with diarrhoea This paper describes the origin, isolation and characterization of

3 human rotavirus strains being considered for further vaccine development in Vietnam The goal is

to prepare a monovalent G1P[8]rotavirus vaccine using one of these strains obtained in Vietnam and naturally attenuated by multiple passages in cell culture While this is an ambitious project that will require several years’ work, we are using the lessons learned to improve the overall quality of vaccine production including the use of Vero cell techniques for the manufacture of other vaccines in Vietnam

© 2009 Elsevier Ltd All rights reserved

1 Introduction

From 1998 to 2003, we conducted sentinel hospital

surveil-lance of rotavirus diarrhoea among children less than 5 years of

age seen in four cities of Vietnam Rotavirus was identified in

more than 50% of children hospitalized for diarrhoea[1,2],

indi-cating a strong need for a vaccine that could substantially reduce

Abbreviations: US CDC, United States Centers for Disease Control and

Preven-tion; EIA, enzyme immunoassay; GCP, good clinical practice; GLP, good laboratory

practice; GMP, good manufacturing practice; MS, master seed; PMKC, primary

monkey kidney cells; PRKC, primary rabbit kidney cells; MA104, African green

monkey kidney cells; PAGE, polyacrylamide gel electrophoresis; PL, passage level;

POLYVAC, Center for Research and Production of Vaccines and Biologicals, Hanoi,

Vietnam; RT-PCR, reverse transcription-polymerase chain reaction; WHO, World

Health Organisation; WS, working seed.

∗ Corresponding author Tel.: +84 48213820; fax: +84 48213203.

∗∗ Corresponding author Tel.: +1 4046392861; fax +1 4046393645.

E-mail addresses:luanpolyvac@gmail.com (L.T Le), bxj4@cdc.gov (B Jiang).

severe rotavirus hospital admissions, outpatient visits, and asso-ciated deaths In an analysis of rotavirus strains identified from prospective surveillance, G1P[8] was the most commonly identified strain (45%) The other common rotavirus strains—G2P[4], G4P[8], G4P[6] and G9P[8], were also present, and a few rare strains were found (e.g., G1P[4] and G9P[4])

Vietnam has a strong history of self sufficiency in local manufacturing of vaccines While many new vaccines are becom-ing available for immunization of children, the government is concerned about the affordability of these products and has encour-aged the domestic manufacture of many new vaccines In recent years, the government, with international partners, has devel-oped new manufacturing capacity for vaccines against Hepatitis

B, Japanese B encephalitis, and measles[3,4], and based on the high burden of rotavirus disease, has recently provided seed funding to develop rotavirus vaccine

The demonstration that a live, oral rotavirus vaccine could be produced from a single human strain, attenuated by multiple pas-sages, and manufactured in Vero cells, encouraged us to apply a 0264-410X/$ – see front matter © 2009 Elsevier Ltd All rights reserved.

Table 1

Adaptation and passage history of the 3 candidate human rotavirus vaccines for use in Vietnam.

Origin 6-month-old girl in Khanh Hoa

Hospital, Nha Trang, 2003

11-month-old boy in Saint Paul Hospital, Hanoi, 2000

9-month-old boy in Saint Paul Hospital, Hanoi, 2001

Isolation in MA104 cells (3×) Isolation in MA104 cells (3×) Isolation in MA104 cells (3×)

Plaque purification in MA104 cells (1×) Plaque purification (3×) in MA104 cells Plaque purification (2×) in MA104 cells

6 Passages in MA104 cells 9 Passages in MA104 cells 6 Passages in MA104 cells

18 Passages in Vero cells 10 Passages in Vero cells 12 Passages in Vero cells

Cloning by limiting dilution 2 times in Vero cells Cloning by limiting dilution 2 times in Vero cells Cloning by limiting dilution 3 times in Vero cells

Master virus seed (PL5) * Master virus seed (PL5) Master virus seed (PL5)

Working virus seed (PL6) Working virus seed (PL6) Working virus seed (PL6)

Note:* PMKC: primary monkey kidney cells; PL: passage level.

similar strategy in Vietnam since these methods are already used in

our existing vaccine production activities[5] In addition, in the

pro-cess of developing a rotavirus vaccine, we have been able to update

some methods and procedures used in the manufacturing of other

vaccines Finally, by adapting a local rotavirus strain with several

additional strains as backup candidate rotavirus vaccine strains,

we will be better able to match vaccine viruses to local rotavirus

strains in circulation and thus, achieve an equivalent or higher level

of protection compared with currently available rotavirus vaccines

2 Materials and methods

2.1 Overall strategy

Our strategy to develop a live, oral rotavirus vaccine for

Vietnam has involved the selection of the most common

circulat-ing serotype, G1P[8] recovered from an 6-month-old child with

rotavirus diarrhoea captured in the national surveillance program

along with two other strains (G1P[4] and G4P[6]) detected in

young infants These strains were adapted and passaged in cell

culture to select the isolates that grew best and to highest titre

in Vero cells The strains were also fully characterized, and

pro-ceeded through at least 40 blind passages with the assumption that

natural attenuation would occur over time Following testing of

the candidate vaccine strains which included toxicity testing and

animal studies, our next steps include human studies: to ensure

the safety of the strains, to determine if they have been

ade-quately attenuated for infants, and to ensure that they still induce a

robust immune-response This paper describes the selection,

pas-saging, initial characterization and preliminary safety testing of

our selected strains For guidelines, we followed the WHO

pub-lished procedures and methods to develop oral rotavirus vaccines

[6] All work was conducted in a special laboratory at the Centers

for Disease Control and Prevention (CDC) in Atlanta and POLYVAC,

a facility approved by national authorities in Vietnam for the

pro-duction of vaccines for polio and measles

2.2 Origin and properties of Vietnam human rotavirus vaccine

candidates

Three faecal specimens from otherwise healthy Vietnamese

infants with acute diarrhoea, that tested positive for rotavirus

by EIA and PCR and contained representative genotypes G1P[8], G1P[4], and G4P[6] of rotavirus, were selected for further develop-ment (Table 1)[7,8] These strains were first adapted to grow in MA104 cells, a cell line derived from monkey kidney cells To avoid concerns about contamination with the agent of bovine spongi-form encephalopathy, we used MA104 cells that were originally received in January 1979 at passage 14 from MA Bioproducts which has changed to Bio Whittaker (Walkersville, MD) and had been stored at the Cell Culture Laboratory, Biologics Branch, US CDC

We have complete passage history and full documentation for these MA104 cells Rotavirus strains were purified and serially pas-saged in MA104 cells and primary monkey kidney cells (PMKC) to increase the rotavirus virus titres A final stage included passage in Vero cells qualified for vaccine production of a master and work-ing virus seed The PMKC were drawn from the production facility for OPV which is operated by POLYVAC and has previously under-gone rigorous routine procedures for safety testing as required for OPV vaccine production under Vietnamese national regulatory procedures

The human rotavirus (HRV) strain G1P[8] (KH0118) was obtained from a 6-month-old girl in Khanh Hoa General Hospital (Nha Trang, Vietnam) in 2003 This is the most commonly occur-ring strain globally and also in Vietnam and proof of principle exists with the monovalent Rotarix®vaccine[5] This virus was isolated

by passaging 3 times and plaque-purified once in MA104 cells It was further passaged 6 additional times in MA104 cells and then 12 additional times in PMKC The virus was then adapted and passaged

18 times in Vero cells that were qualified for use in vaccine produc-tion Individual clones were selected by limiting dilution twice in Vero cells After four more passages in Vero cells, a master seed (MS) was produced at passage level (PL) 5 One additional passage

of MS in Vero cells produced a working seed (WS) of PL 6 for a total passage number of 48

The HRV strain G1P[4] (SP210-2000) was derived from an 11-month-old boy in Saint Paul Hospital (Hanoi, Vietnam) in 2000 The virus was isolated, plaque-purified 3 times and passaged 9 more times in MA104 cells (Table 1) The virus was then passaged 11 times in PMKC and adapted to grow in Vero cells (10 passages) After cloning by 2 rounds of limiting dilution, the virus was amplified in Vero cells to prepare MS (PL5) and WS (PL6) for a total passage number of 44

A G4P[6] strain (SP203-2001) was identified from a

9-month-old boy in Saint Paul Hospital in 2001 The virus was isolated by

initial passage and plaque purification twice, and passaged 6 times

all in MA104 cells After 24 passages in PMKC, the virus was adapted

to grow in Vero cells (12 passages) After 3 rounds of purification

by limiting dilution, the virus was amplified in Vero cells to prepare

MS (PL5) and WS (PL6) for a total passage number of 56

2.3 Quality control of master and working rotavirus seeds

The MS and WS for each virus were tested for adventitious

agents by using an array of in vitro and in vivo assays as suggested in

the WHO guidelines for the production of an attenuated rotavirus

vaccine[6] The tests used were approved by the national

regula-tory authority in Vietnam The MS and WS were routinely checked

for identity by 10% PAGE (Fig 1) and by RT-PCR (G and P typing)

to assure the correct electropherotypes and G and P genotypes,

respectively using published methods[7,8] Additional tests were

performed on the virus seeds to confirm the stability of the viral

genome and absence of adventitious agents over a defined passage

range The passage level used to qualify the virus seeds is generally

5–10 passages beyond that used for vaccine production[6] Our

working virus seed (PL6) is just one more passage of the master

virus seed (PL5) in Vero cells

2.4 Safety tests

Each master or working seed lot was tested for toxicity, safety

and immunogenicity in animals including guinea pigs, mice,

rab-bits, and monkeys (Table 2) They were also tested by RT-PCR for

adventitious viruses, including foamy virus, hepatitis viruses and

mycoplasma, by the Research and Development Department at

the Company for Vaccine and Biological Production No1 in Hanoi

[9–11] Safety tests for adventitious viruses were also conducted

in cell cultures including PMKC, neonatal rabbit kidney cells, Vero

and Hep 2 cell lines as recommended by the WHO Guidelines[6]

2.5 Analysis of NSP4, VP6, VP7 and VP4 genes of candidate

vaccines

The MS of the 3 candidate vaccines was subjected to sequence

analysis for rotavirus genes 4, 6, 9 and 10, which encode the two

neutralization antigens, the group antigen and the putative viral

enterotoxin Partial nucleotide sequences of genes 4 and 9

(encod-ing VP4 and VP7, respectively) were obtained from amplicons

generated in the first round of G and P typing RT-PCR

Phylogenetic analysis was performed on the deduced amino acid

sequences from aa 1-175 for NSP4, 1-397 for VP6, 1-288 for VP7 and

Fig 1 RNA profiles of rotavirus vaccine candidates, G1P[8], G1P[4] and G4P[6] iso-lated in Vietnam Rotavirus dsRNA was extracted from infected Vero cell cultures and analyzed on 10% polyacrylamide gel, followed by staining with silver nitrate Lanes 1, DNA marker; 2, G1P[8] vaccine seed; 3, G1P[4] vaccine seed; 4, G4P[6] vaccine seed, and 5, reference Wa strain.

16-278 for VP4 Sequence data were first analyzed using CHROMAS software (Technelysium Pty Ltd, Tewantin, Australia) Forward and reverse sequence data of each sample were aligned using the Edit Seq and MegAlign program (DNASTAR Inc., Software, Madison, WI)

to obtain the final consensus sequence Complete alignment was performed with Cluster W (DNASTAR Inc., Software, Madison, WI)

Table 2

Qualification tests used for master and working rotavirus seeds and results.

Tissue culture safety: Vero, Hep2, PMKC, PRKC * Cytopathic effect, hemedsorption Negative

Animal safety: adult mice, rabbits, Guinea pigs Observation Increase in weight and appear healthy

Safety and immune-response in young monkeys

(6–12-month-old)

Observation Increase in weight and appear healthy Measurement of neutralizing antibody

by plaque reduction assays

3-fold increase in serum antibody titres from pre- to post-third dose vaccination

*

Fig 2 Phylogenetic tree of the NSP4 gene fragments of the 3 Vietnam candidate vaccines The Vietnamese strains used to develop vaccine candidates are indicated in the boxes The 738 bp PCR product of NSP4 gene was amplified using Jrg30 and Jrg31 primer pair and directly sequenced Abbreviations of countries of isolation in strain names:

AU (Australia), BL (Bangladesh), CN (China), ID (India), JP (Japan), TL (Thailand), TW (Taiwan) and US (the United States of America).

3 Results

3.1 Preparation of candidate rotavirus vaccines

Three rotavirus strains (rotavirus genotypes G1P[8], G1P[4] and

G4P[6]) from patients enrolled in the surveillance study were

selected as candidate strains for rotavirus vaccine development

All 3 patients had diarrhoea with different degrees of dehydration

but no apparent fever The patients were otherwise in good health

We first adapted rotavirus strains to grow in MA104 cells that were

prepared before 1980 with complete detailed documentation and

then passaged the strains in PMKC and qualified Vero cells using

certified reagents under stringent GLP conditions (Table 1) Strains were purified by plaque assay and limiting dilution MS and WS banks were prepared with Vero cells in roller bottles

3.2 Preclinical safety and immunogenicity of candidate rotavirus vaccines

The MS and WS of the 3 candidate rotavirus vaccine strains were shown to be free of extraneous agents (hepatitis A, B and C viruses, SV40, foamy virus, mycoplasma, bacteria and fungi) and also appeared to be safe in tissue culture, laboratory animals including monkeys (Table 2) All 3 vaccines elicited a 3-fold increase in serum

Fig 3 Phylogenetic tree of the VP6 gene fragments of the 3 Vietnam candidate vaccines The Vietnamese vaccine candidates are indicated in the boxes The full length of VP6 gene (1356 bp) was amplified using Jrg7 and Jrg8 primer pair and directly sequenced using PCR and internal primers Abbreviation of countries of isolation in strain name:

AU (Australia), BL (Bangladesh), CN (China), JP (Japan), ID (India), NZ (New Zealand), PLP (Philippines), TL (Thailand) and US (The United States of America).

neutralizing antibody titres following immunization with 3 vaccine

doses compared to pre-vaccination titres in baby monkeys (data

not shown) All 3 strains had long RNA electropherotypes (Fig 1)

Of note, the profile of the G1P[4] strain demonstrated a high level of

similarity to that of reference G1P[8] strain Wa The G and P types of

these candidate vaccines were confirmed by RT-PCR and sequence

analysis

3.3 Strain characterization—sequencing

3.3.1 NSP4 typing

The NSP4 sequences of 3 candidate rotavirus vaccine strains

were compared with 175 aa sequence fragments of viruses

iso-lated from Bangladesh, India, Thailand and China in the same

period and with reference strains Wa, DS-1, ST-3 and M37 NSP4

gene of the KH0118 (G1P[8]) and SP203-2001 (G4P[6]) strains

belong to genogroup B (prototypes Wa, ST-3 and M37), whereas

the SP210-2000 (G1P[4]) strain belongs to genogroup A (prototype

DS-1) (Fig 2) We observed several notable differences in the NSP4 sequence between the 3 Vietnamese strains in this study, the refer-ence G1P[8] strain, and strains from other countries In particular, there were substitutions of Thr to Met at aa45, Ile to Val at aa76, Tyr to Cys at aa85, and Ser/Lys for Iso at aa97 (data not shown) The position aa131 in the NSP4 toxic peptide (aa114-135) was Tyr for G1P[4] and His for G4P[6] and G1P[8] strains, consistent with the ability of the original strains of these vaccine candidates to cause diarrhoea in children However, all of these strains have the Ile at aa

135, a marker for attenuation[12], indicating that the strain might have been attenuated after 40 or more passages in cell culture The single strain from Vietnam (KH210-2004) whose NSP4 sequence was previously available also showed high aa sequence similarity

to our G1P[8] strain and came from the same province in Vietnam

3.3.2 VP6 typing

The 3 Vietnamese vaccine candidates belong to two major VP6 genogroups The region of aa280-350 defines VP6 genogroups[13]

Fig 4 Phylogenetic tree of the VP7 gene fragments of the Vietnam G1P[8] and G1P[4] (4a) and G4P[6] (4b) candidate vaccines Vietnamese vaccine candidates are indicated

in the box The 896 bp PCR products were amplified using 9con1L and VP7R primers and directly sequenced using PCR and internal primers Abbreviations of countries of isolation in strain names: Au (Australia), BL (Bangladesh), BRZ (Brazil), CN (China), ID (India), ISR (Israel), JP (Japan), TL (Thailand), TW (Taiwan), US (The United States of America) and VN (Vietnam).

and residues RPPN at aa 296-299 are present in all 3 Vietnam

strains, suggesting that they belong in subgroup I or II (Fig 3) The

G1P[8] and G4P[6] strains showed characteristics of genogroup II

(Asn at aa 305, Asn at aa 339, Leu at aa 342 and Ala at aa 348),

while the G1P[4] showed characteristics of genogroup I and

there-fore of subgroup I (Ala at aa 305, Asn at aa 310, Glu at aa 315,

Ser at aa 339, Met at aa 342 and Ser at aa 348) Phylogenetically,

the Vietnam G1P[4] was in the same cluster as DS-1, SA11, 1076

and AU-1, whereas the Vietnam G1P[8] and G4P[6] strains were in

the same cluster with Wa, RV3 and ST-3, confirming the classifica-tion of our vaccine strains (Fig 3) The sequence of VP6 from our G1P8 strain was similar to those from Bangladesh/2002–2003 and Thailand/2001 whereas the sequence of G1P[4] strain was closely related to those in Bangladesh/2000 and India during the same period and to DS-1 and SA11 (Fig 3) In addition, the Ala at aa172

of the Vietnam G1P[4], DS-1 and SA11 (subgroup I) was distinctly different from the Met at this position for many other subgroup II strains including the Vietnam G1P[8], G4P[6], Wa, RV3 and ST-3

Fig 5 Phylogenetic tree of the VP4 gene fragments of the 3 Vietnam candidate vaccines The Vietnamese vaccine candidates are indicated in the box The 876 bp PCR product was generated using primer pair con3 and con2 and directly sequenced using PCR and internal primers Abbreviations of countries of isolation in strain names: Au (Australia),

BL (Bangladesh), CN (China), HGR (Hungary), ID (India), JP (Japan), Korea (KOR), TL (Thailand), UK (United Kingdom), US (The United States of America), VZL (Venezuela) and

VN (Vietnam).

The Vietnam G1P[4] strain also showed amino acid substitutions of

unknown importance at positions aa330 (Val-Ile), aa369 (Glu-Asp)

and aa396 (Ile-Val) compared with other strains

3.3.3 VP7 sequences

The VP7 sequences of the Vietnam G1 candidate vaccine strains

were similar to other Vietnamese strains (G1-lineage III) whereas

the VP7 sequence of the G4 strain was similar to the corresponding

gene of other strains from Vietnam, China, Japan, and Thailand and

the porcine Gottfield strain (Fig 4) The VP7 sequence fragments

of the two G1 strains were similar to those isolated elsewhere in

Vietnam during the same period (2003), and similar to those from Bangladesh and Thailand, clustered into lineage III, but different from those isolated from Japan and China or Finland (other lineages

I, II and IV) (Fig 4a) Within the same lineage III, the Vietnam G1P[4] and G1P[8] were quite different

3.3.4 VP4 sequence

The amino acid sequence of VP4 from the Vietnamese G1P[8] vaccine candidate strain demonstrated that strain KH0118 (G1P[8]) falls in the P[8]-2 F45-like lineage, having a high degree of homol-ogy with OP354 strain P[8]-3 (Fig 5) Many of the Vietnamese

G1P[8], G4P[8] and G9P[8] strains were in the P[8]-2 lineage, and

some G3P[8] and G9P[8] are of P[8]-3 lineage[14] Due to

differ-ences in the region sequenced between the VP4 sequdiffer-ences in our

study and the previous study, these strains could not be included

in the phylogenetic tree Amino acid alignments within the same

region revealed the conserved Pro at aa224 and 225, like many other

P[8] rotavirus strains Important substitutions in VP4 sequence of

the Vietnam G1P[8] included Arg-Thr (aa144) and Arg-Gly (aa268)

(data not shown)

The VP4 sequence of the P[4] strain was similar to other

Viet-namese G2P[4] strains isolated elsewhere The G1P[4] strain had

important substitutions in the VP4 sequence compared to other

strains including changes of Thr-Ala at aa103 and Arg-Gly at aa162

The Vietnam G1P[4] strain shared 96.5–98.1% homology with

TB-chen 2004 and Kmr029-2001 strains in the same cluster, and

92.6% and 95.3% homology with DS-1 and RV5 of different clusters,

respectively (Fig 5) By contrast, the VP4 sequence of the G4P[6]

strain shows high similarity to porcine strains The VP4 sequences

of P6 strains from Vietnam are alike and clustered into a separate

lineage with those reference strains isolated in 1980s (RV3, ST-3

and M37) and strains from Taiwan and Bangladesh The G4P6

clus-ters in the same lineage as the VN904 (G9P[6]), VN846 and VN592

(both are G4P[6] strains) This lineage was identified as P6-Ic, which

has more than 95% similarity to porcine strains 221/04 and 134/04

[14] The Vietnam G4P[6] did not have the same conserved residues

as other P6 strains at aa 90 and 91, Pro-Ile instead of Qln-Val, aa98

Arg instead of Lys, aa101 Val instead of Ile, aa140 Tyr instead of Phe,

aa170-172 Gly-Gly-Arg instead of Tyr-Asn-Ser, aa202 Thr instead of

Val, aa232 Val instead of Ile, aa242 Ala instead of Val, and aa255 Val

instead of Ile The Vietnam G4P[6] strain exhibited many

substitu-tions at the VP4 gene compared to other reference and field strains

in the Asian region Thus, our G4P[6] strain might have porcine

origins

4 Discussion

This is the first report to describe the development of candidate

rotavirus vaccines from human strains G1P[8], G1P[4] and G4P[6]

isolated in Vietnam These strains were adapted for growth, cloned,

and prepared as MS and WS banks in Vero cells The G1P[8] strain

could potentially be used as a monovalent vaccine with a much

lower production cost and a viral sequence that fits better with the

circulating P[8] lineage in Vietnam The other two strains (G1P[4]

and G4P[6]) may be included in a vaccine formulation, if needed,

to improve cross-protection and enhance overall vaccine efficacy

Further studies will be needed to determine whether the G1P[4]

vaccine strain induces immunologic cross-protection against both

G1 and G2 viruses, derived from different rotavirus genogroups

Additional evaluation of the candidate rotavirus vaccine strains is

needed to assess whether or not the multiple passages in cell

cul-ture led to attenuation of the virus strain, something that can now

only be determined in human studies

We have characterized the 3 HRV strains that have been used

for the development of candidate rotavirus MS G1P[8], G1P[4] and

G4P[6] These strains represented the major P and G genotypes

(G1) that have been circulating in Vietnam[2] Results of

surveil-lance in Vietnam indicate that a P[8] strain can cover over 70% of

the serotypes in circulation[2] We also developed P[4] and P[6]

vaccine candidates to cover other major P types (14.4% and 7.6%,

respectively) of rotavirus strains circulating in Vietnam The 3 virus

seeds were biologically cloned by limiting dilution in Vero cells to

enhance genetic homogeneity as well as to remove potential

adven-titious agents that may have been present in the initial virus inocula

or derived from the attenuation process[5,6] The virus seeds were

examined by performing a variety of tests and were qualified to be

free of adventitious agents[6] The vaccine candidates were also demonstrated to be free of adventitious agents in cell culture and

in experimental animal models All 3 vaccine candidates induced neutralizing antibody response in young monkeys

Additional sequence analysis of the NSP4 and VP6 genes of the strains indicates that our vaccine candidates covered the 2 major NSP4 genotypes and the 2 subgroups of VP6 Sequence of the NSP4 toxic peptide identified one of the markers for attenuation (Ile at aa135), probably acquired during repeated passages in the cells, yet still showed Tyr/His at aa131, characteristic of the original strains causing diarrhoea[12] The VP7 and VP4 sequences of the G1P[8] strain shared similarity to some strains from Vietnam, Bangladesh, Thailand, whose sequences are available for analysis[14] Analysis

of the VP4 sequence indicates that the Vietnamese P8 strain belongs

to cluster F45-like (P[8]-2), whereas Rotarix®, one of the currently licensed and available rotavirus vaccines, is P[8]-1[14] The find-ings that most strains in Vietnam fall into P[8]-2 and P[8]-3 suggest that it may be important to develop a rotavirus vaccine from local strains

Several lines of rationale have led us to consider developing a new vaccine for Vietnam First, the long-term costs associated with introduction of several new vaccines may be unsustainable for a developing country such as Vietnam and the tradition of being self-sufficient in vaccine production has led the government to invest heavily in vaccine production and to support this development activity A single strain of rotavirus prepared as a vaccine through multiple passaging and preparation in Vero cells requires a level

of technology similar to the production of OPV This production technology has been developed at institutions within Vietnam and has been adapted for use in development of the rotavirus vaccine candidates reported here

Of note, the UK vaccine prepared from bovine reassortant strains has been licensed by US NIH to emerging manufacturers in eight countries and could be a direct and less expensive competitor vac-cine to the Merck vacvac-cine, RotaTeq® This vaccine is much more complicated to prepare as it contains multiple strains that need to

be formulated together The epidemiology of rotavirus in Vietnam

is more similar to the epidemiology in a middle income country rather than a poor developing country Therefore, we hope that our candidate G1 strains might perform as well in Vietnam as the mono-valent GSK vaccine performs in Latin America and South Africa[5] Our vaccine strains have a sequence more homologous to other strains in Vietnam than strains elsewhere, including the G1P[8] strain in the GSK vaccine Therefore, these strains might well pro-vide a slightly better level of protection than those included in currently available vaccines, a speculation that remains to be exam-ined independently and after a long period of clinical development The steps to make a new vaccine are complicated, costly, time consuming and fraught with many challenges This project has resulted in 3 vaccine candidates that can move into human trials—a key step of the developmental sequence for a rotavirus vaccine that will take several years to accomplish To be compliant with WHO and international standards, we need to consider and develop processes that ensure safe, consistent, robust vaccine production according to GMP practices In addition, safety, immunogenicity and efficacy in human subjects need to be performed according to GCP Finally, this process needs to take into consideration the over-all costs and ability to produce this vaccine on a manufacturing scale All of these various activities will be critical for the ultimate success of the project

The level of international collaboration and technical support from international partners including CDC and WHO, as we have initiated this rotavirus vaccine project, has enabled us to upgrade many of these aspects including GMP and GLP in the preparation

of appropriate cell substrates and techniques, safety and toxicity testing of the oral polio vaccines currently prepared at POLYVAC

Clearly there are many hurdles ahead in the science, regulatory

issues, and clinical testing of these candidates, but the

govern-ment’s commitment to this internationally collaborative project

has helped us think more deeply about how to approach a problem

that affects infants throughout the world

Acknowledgments

We thank Dang Mai Dung, Nguyen Thi Quy, Nguyen Cong Long,

and Bui Huy Phuong for technical assistance and Wendi Kuhnert for

critical reading of the manuscript Technical support was provided

by the Gastroenteritis and Respiratory Viruses Laboratory Branch,

Division of Viral Diseases and Biologics Branch, Division of Scientific

Resources, the Centers for Disease Control and Prevention, Atlanta,

Georgia, USA

Conflict of interest

None

Funding sources

Supported by grants form the Ministry of Health and the

Min-istry of Science and Technology, Vietnam, and the World Health

Organisation

The findings and conclusions in this report are those of the

authors and do not necessarily represent the views of U.S CDC

References

[1] Nguyen VM, Nguyen VT, Huynh PL, Dang DT, Nguyen TH, Phan VT, et al The

epidemiology and disease burden of rotavirus in Vietnam: sentinel surveillance

at 6 hospitals J Infect Dis 2001;183(12):1707–12.

[2] Van Man N, Luan le T, Trach DD, Thanh NT, Van Tu P, Long NT, et al Epi-demiological profile and burden of rotavirus diarrhea in Vietnam: 5 years

of sentinel hospital surveillance, 1998–2003 J Infect Dis 2005;192(Suppl 1): S127–32.

[3] Barat LM Four malaria success stories: how malaria burden was success-fully reduced in Brazil, Eritrea, India, and Vietnam Am J Trop Med Hyg 2006;74(1):12–6.

[4] Hipgrave DB, Trung NT, Vu MH, Do TD, Nguyen TN, Hoang TL, et al Immuno-genicity of a locally produced hepatitis B vaccine with the birth dose stored outside the cold chain in rural Vietnam Am J Trop Med Hyg 2006;74(2):255– 60.

[5] O’Ryan M Rotarix (RIX4414): an oral human rotavirus vaccine Expert Rev Vacc 2007;6(1):11–9.

[6] WHO guidelines to assure the quality, safety and efficacy of live attenuated rotavirus vaccine (oral) WHO Technical Report Series, World Health Organiza-tion;2007.

[7] Gentsch JR, Glass RI, Woods P, Gouvea V, Gorziglia M, Flores J, et al Identification

of group A rotavirus gene 4 types by polymerase chain reaction J Clin Microbiol 1992;30(6):1365–73.

[8] Gouvea V, Ho MS, Glass R, Woods P, Forrester B, Robinson C, et al Serotypes and electropherotypes of human rotavirus in the USA: 1987–1989 J Infect Dis 1990;162(2):362–7.

[9] Schweizer M, Neumann-Haefelin D Phylogenetic analysis of primate foamy viruses by comparison of pol sequences Virology 1995;207(2):577–82 [10] Nainan OV, Cromeans TL, Margolis HS Sequence-specific, single-primer ampli-fication and detection of PCR products for identiampli-fication of hepatitis viruses J Virol Methods 1996;61:127–34.

[11] Pruckler JM, Pruckler JM, Ades EW Detection by polymerase chain reac-tion of all common Mycoplasma in a cell culture facility Pathobiology 1995;63(1):9–11.

[12] Cunliffe NA, Woods PA, Leite JP, Das BK, Ramachandran M, Bhan MK, et al Sequence analysis of NSP4 gene of human rotavirus allows classification into two main genetic groups J Med Virol 1997;53(1):41–50.

[13] Iturriza Gomara M, Wong C, Blome S, Desselberger U, Gray J Molecular characterization of VP6 genes of human rotavirus isolates: correlation of genogroups with subgroups and evidence of independent segregation J Virol 2002;76(13):6596–601.

[14] Nguyen TA, Khamrin P, Trinh QD, Phan TG, Pham LD, Hoang LP, et al Sequence analysis of Vietnamese P[6] rotavirus strains suggests evidence of interspecies transmission J Med Virol 2007;79(12):1959–65.