Epidemiology and Genetic Diversity of Rotavirus Strains in Children with Acute Gastroenteritis in Lahore, Pakistan

Pakistan harbors high disease burden of gastro-enteric infections with majority of these caused by rotavirus. Unfortunately, lack of proper surveillance programs and laboratory facilities have resulted in scarcity of available data on rotavirus associated disease burden and epidemiological information in the country. We investigated 1306 stool samples collected over two years (2008–2009) from hospitalized children under 5 years of age for the presence of rotavirus strains and its genotypic diversity in Lahore. The prevalence rate during 2008 and 2009 was found to be 34% (n = 447 out of 1306). No significant difference was found between different age groups positive for rotavirus (p.0.05). A subset of EIA positive samples was further screened for rotavirus RNA through RT-PCR and 44 (49.43%) samples, out of total 89 EIA positive samples, were found positive. Phylogenetic analysis revealed that the VP7 and VP4 sequences clustered closely with the previously detected strains in the country as well as Belgian rotaviruses. Antigenic characterization was performed by analyzing major epitopes in the immunodominant VP7 and VP4 gene segments. Although the neutralization conferring motifs were found variable between the Pakistani strains and the two recommended vaccines strains (Rotarix TM and RotaTeq TM), we validate the use of rotavirus vaccine in Pakistan based on the proven and recognized vaccine efficacy across the globe. Our findings constitute the first report on rotavirus' genotype diversity, their phylogenetic relatedness and epidemiology during the pre-vaccination era in Lahore, Pakistan and support the immediate introduction of rotavirus vaccine in the routine immunization program of the country. Copyright: ß 2013 Alam et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: The authors have no support or funding to report.


Introduction
Diarrheal infections are considered as a significant cause of infant and childhood morbidity and mortality in both developing as well as developed countries [1]. Most of these infections have a viral etiology principally rotaviruses that alone account for approximately 527,000 (475000-580000) deaths per annum among children less than 5 years of age [2,3]. Globally, more than 2 million children are hospitalized every year for rotavirus infection and 90% of rotavirus associated mortalities occur in Africa and Asia [4,5]. Rotaviruses have been classified into seven heterogenic groups (A to G) based on their genetic and antigenic properties. Group A rotaviruses are the most important and frequently detected viruses among the three groups A, B and C that infect humans [6]. The virus belongs to the family reoviridae with triple-coated icosahedral virion particle, enclosed within is the 11 segmented double stranded RNA genome. The triple layered capsid consists of 2 proteins in the outer shell (VP7 and VP4), the intermediate layer constitutes VP6 while VP2 forms the inner layer enclosing two proteins VP1 and VP3. Based on VP6 capsid gene, the virus has been classified into seven major genogroups while VP7 and VP4 are the basis of a binary-system for further classifying the genogroup A viruses into 27 G-and 35 P-genotypes respectively [7]. Globally, the most important types causing majority of infections are G1P [8], G2P [4], G3P [8], G4P [8] and G9P [8] [8,9]. However, significant diversity of rotavirus genotypes have emerged with several novel combinations due to accumulation of point mutations, genome re-assortments or zoonotic transmission of animal strains resulting in the introduction of new antigenic variants [10,11].
Rotavirus strain identification is considered as the key component of epidemiological surveys, disease distribution, genotype prevalence studies, vaccine administration and efficacy monitoring programs. Many past studies have highlighted the significance of continued monitoring of circulating rotavirus strains in order to maintain sufficient population immunity [12]. In Pakistan, there is no well-developed surveillance system for rotavirus strain identification although country's Ministry of Health has initiated a hospital based surveillance network to serologically test the stool samples from children presented with gastroenteritis at central district hospitals in 3 major cities; Karachi (Sindh province), Lahore and Rawalpindi (Punjab province). These sentinel sites perform ELISA for the diagnosis of rotavirus infection without any further analysis for viral genotype identification. This study is in continuation to our previous work where we identified an emerging rotavirus genotype G12 in two children admitted to a hospital in Rawalpindi [13]. To further explore the epidemiology and genotypic diversity of rotavirus in Pakistan, here we report the findings of rotavirus subtypes detected in children hospitalized, due to severe dehydrating diarrhea, at Children's Hospital Lahore.

Materials and Methods
Samples from hospitalized children suspected of rotavirus gastroenteritis were collected as per World Health Organization's standard case definitions that describe a suspected case as a child ,5 years of age, admitted to a designated sentinel hospital for treatment of gastroenteritis while a confirmed case is a suspected case in whose stool the presence of rotavirus is demonstrated by means of an enzyme immunoassay. The study concept and design was approved through the Pakistan's National Institute of Health Internal Review Board. The samples were collected after informed and written consent from the patient's parents/guardians.
A total of 1306 stool samples were collected from hospitalized patients at 'The Children's Hospital' Lahore during January 2008 to December 2009. Majority of these samples were collected from children below than 5 years of age who were hospitalized with suspected rotavirus gastroenteritis. The collected stool samples were processed for the detection and confirmation of Rotavirus antigen. ELISA test was performed using the ProSpecT TM Rotavirus Microplate Assay (Oxoid Ltd., Basingstoke Hants, UK) as per World Health Organization's recommendations (http://whqlibdoc.who.int/publications/2011/ 9789241502641_eng.pdf).
A subset (20%) of EIA positive samples were transported to Department of Virology, National Institute of Health, Islamabad for rotavirus RNA detection through RT-PCR and further genotype determination on the basis of VP7 and VP4 gene segments using the protocol as described by Gouvea et al 1990 [14] and Gentsch et al 1992 [15] respectively. Amplified products from round 1 PCR reactions were purified using QIAquick PCR purification kit (Qiagen, Germany) and were directly sequenced for VP7 and VP4 genes using the Big dye terminator sequencing kit v3.0 by automated Genetic analyzer ABI 3130xl (Applied Biosystems).
Phylogenetic analyses of VP7 and VP4 sequences were performed in comparison to the strains belonging to different geographical regions as retrieved from GenBank. Evolutionary tree and distances (number of base substitutions per site) were generated by Neighbor Joining method with Kimura-2 parameter using MEGA 4.0 (http://megasoftware.net/). The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates) is shown next to the branches. The GenBank accession numbers, country, year of sample collection and respective genotype information has been given where available. The VP7 sequences obtained in this study have been submitted to GenBank under the accession numbers KC896141-KC896157 and the VP4 sequences under accession numbers KC896127-KC896140.

Prevalence, Epidemiology and RVA-Genotypes
Screening of total 1306 samples collected during two years period (2008-2009) for the presence of rotavirus antigen yielded 447 (34.22%) samples positive for group A major inner capsid protein (VP6 antigen) common to all known rotavirus genotypes. The prevalence rates during 2008 and 2009 were 34.12% (n = 243 out of 712) and 34.34% (204 out of 594) respectively. No significant difference (P.0.05) was found between different age groups positive for rotavirus through ELISA; however, high infection rate was observed among children between 12-17 and 24-59 months during 2008 and 2009 respectively. The rotavirus positive cases appeared to occur throughout the year with peaks during January to April and July to September (Figure 1a and 1b).
A subset of EIA positive samples, based on availability, was further screened for the presence of rotavirus RNA through RT-PCR. Out of total 89 EIA positive samples available from archived stock kept at -80uC, 44 (49.43%) samples were found positive for rotavirus through PCR. G and P types were found with the prevalence as G1P

Phylogenetic Analysis Based on RVA-VP7 and -VP4 Gene Segments
The genetic sequence of 18 viruses included in this study was determined for the VP7 gene. These samples were randomly selected on the basis of sample quantity and resource limitations which showed that all viruses are highly identical (99-100%) to each other within their respective genotypes; G1, G2 and G9. The G1 strains shared close sequence similarity with the previously reported viruses from Pakistan (JN001862) as well as to those from Belgium (HQ392261, HQ392204) and USA (JN258346). The VP7 sequences of G9 strains were found 100% identical among themselves as well as highly similar level of homology was found with the viruses from South Africa (GQ338887), Russia (FJ447573) and Australia (AY307090). These viruses also hold 98% similarity to previously reported G9 strains (JN001865, JN001866) from Faisalabad city of Pakistan (unpublished data). Likewise, the G2 viruses showed 99-100% identity among themselves and a similar genetic closeness was found among the G2 strains from Bangladesh (EF690778, EF690782) as well as the already identified G2 strains from Pakistan (JN001883; unpublished data).
The genetic relationships of rotavirus strains from this study were also determined on the basis of VP4 gene. 14 viruses were grouped among three genotypes P [4], P [6] and P [8] with the highest prevalence of P [8] in combination with G1, G2 and G9. The genotype P [6] was found with G1 and G9 counterparts while P [4] was found with G2 as the most common circulating strains during the study period with 48% prevalence. In a pattern similar to the G-types, P [6] and P [8] strains detected from Lahore during 2008 and 2009 revealed closest identity to already reported viruses from Pakistan (JN001876-79 unpublished data).
Phylogenetic reconstruction revealed clustering of all rotavirus strains from Children Hospital, Lahore with viruses (JN001862 (G1P [8]); JN001865 (G9P [8]); JN001883 (G2P [6]) previously detected from the Faisalabad district, located in the South-West of Lahore at a distance of 118 kilometers (Figure 2a and 2b). The viruses classified as G1 are categorized into two separate lineages. Lineage A contains rotavirus strains from both Lahore and Faisalabad. Two of the strains (NIHPAK-210 and NIHPAK-417) were distinguished as Lineage B clustering with viruses from USA (JN258346) and Belgium (HQ392204).

Comparison of RVA-VP7 Antigenic Regions between Pakistani and Vaccine Strains
The effectiveness of rotavirus vaccines can be determined by analyzing the amino acid (aa) differences, in the neutralizing epitopes of VP7 and VP4 proteins, between vaccine and the circulating strains. We compared the VP7 aa sequences of Pakistani viruses, constituting the antigenic epitopes with those of the two available vaccines strains (RotaTeq TM and Rotarix TM ). The VP7 protein contains three antigenic epitopes defined as 7-1a, 7-1b and 7-2 which comprise of 29 amino acid residues from positions 87 to 291 based on rhesus Rotavirus strain numbering (GenBank accession No. AF295303) [16]. The comparison of all Pakistani rotavirus strains found in Lahore and vaccine strains showed that out of total 29 amino acids, only five are conserved (Figure 3a).
Analysis of Pakistani G1 strains showed that the 29 amino acid residues across the three antigenic sites are highly conserved when compared to Rotarix TM G1P [8] strain while 02 differences were found at positions 97 (D.E; Aspartic acid to Glutamic acid) and 147 (S.N; Serine to Asparagine) within the 7-1a and 7-2 regions when compared to RotaTeq TM G1P [5] strain.
Among the G9 viruses presented in this study, the VP7 protein was compared with Rotarix TM G1 strain as well as all constituent strains of RotaTeq TM vaccine (G1-4 with P [5] and G6P [8]). Thirteen aa changes were found when compared to Rotarix TM strain across all three epitopes. On comparison with the RotaTeq TM G1 strain, 14 variations were found in Pakistani viruses. In comparison to G2 RotaTeq TM strain, 19 aminoacid changes were noticed. Similarly, 11 variations were observed between G3 RotaTeq TM strain and the Pakistani G9 viruses with majority of differences found in 7-1b epitope in which all the aa residues were changed except Glutamine at position 201. For G4 RotaTeq TM strain, 10 aa changes were found in study strains Figure 1. Phylogenetic analysis of VP7 gene segment of the study viruses isolated in hospitalized children at Lahore. Each of the G1, G2 and G9 genotype viruses from this study are indicated with round circle, while the previous available sequences in GenBank from Pakistan have been given with an arrow-head. The closely matched sequences have retrieved from NCBI GenBank and included to reconstruct a phylogenetic tree (Figure 1a). Phylogenetic analysis of VP4 gene segment of the study viruses isolated in hospitalized children at Lahore. Each of the P [4], P [6] and P [8] genotype viruses from this study are indicated with round circle, while the previous available sequences in GenBank from Pakistan have been given with an arrow-head. The closely matched sequences have retrieved from NCBI GenBank and included to reconstruct a phylogenetic tree (Figure 1b). doi:10.1371/journal.pone.0067998.g001 across the three epitope regions. When compared to G6P [8] RotaTeq TM strain, 12 dissimilarities were noticed across the three antigenic regions mainly within the 7-1b region.

Comparison of RVA-VP4 Antigenic Regions between Pakistani and Vaccine Strains
The VP4 protein of rotavirus contains four antigenic regions defined as 8-1 to 8-4 spanning between aminoacid positions 88 to 196 based on rhesus Rotavirus strain (GenBank accession No. AY033150) [17]. The VP4 sequences of Pakistani viruses were compared to those of vaccine strains and found significant variations among the four regions except aminoacid residue 180 and 132 which were well conserved amongst all Pakistani and vaccine strains (Figure 3b).
When compared to Rotarix TM P [8] strain, the Pakistani P [8] strains with G1 and G2 counterparts showed higher degree of disparity than those with G9. Importantly, the Pakistani strains contained 3 to 12 residues that differed from the RotaTeq TM P [8] strain while Rotarix TM strain P [8] was more variable with 7 to 13 aminoacid differences. Interestingly, all the three aminoacid residues (87-89; NTN) of 8-4 epitope were well conserved among all P [8] Pakistani viruses.
A high degree of variation was observed for P [4] and P [6] strains as well. For P [4] strains, there were 13 aminoacid differences when compared to Rotarix TM while 22 to 23 residues were different from RotaTeq TM strains. Comparison of P [6] strains showed 17 differences with Rotarix TM whereas 20 aminoacid differences were found when compared to RotaTeq TM strains.

Discussion
Diarrhea is accountable for approx. 5% of all deaths among children below five years of age. Viral pathogens are the most common cause of gastroenteritis in both developing and developed countries [18][19][20] especially rotaviruses, noroviruses, astroviruses and adenoviruses [21]. Three rotavirus serogroups (A, B and C) are known to cause gastroenteritis in humans [22]. The most commonly associated VP7 genotypes of group-A rotavirus with different combinations of VP4 genetic counterparts are G1, G2, G3, G4 and G9 [23]. However, unusual genotypes with different G and P specifities have recently been reported from various parts of the world [1].
Seasonality of rotavirus gastroenteritis in Pakistan has never been determined although the cases have been detected round the year [24]. In our study, rotavirus was found during all seasons especially early and late calendar months. This may be attributed to the rainy and post-monsoon season in our country and support the previous findings that low air temperature and dry environment increase the rotavirus incidence [25]. The same pattern is peculiar to tropical areas where rotavirus infections appear throughout the year without any significant epidemic peaks [26]. Similar findings have been reported from Bangladesh, South Asian regions, Bahrain and Costa Rica where rotavirus is detected all year round with less obvious seasonality [27][28][29][30]. However, our findings contrast those in USA, Japan, Northern Asian region, Australia and Europe where rotavirus diarrhea peaks in winter and is rarely identified in summer [31][32][33][34][35].
The higher percentage of samples positive for rotavirus was found among the children between 12 and 17 months of age. These findings are slightly variable than the previous study conducted in Karachi city of Pakistan [36] where most of the rotavirus infections were found among children less than 12 months of age. This variation might reflect a different study design and target population i.e. our study includes the samples from patients hospitalized with severe gastroenteritis whereas Qazi et al. [36] studied the incidence of rotavirus associated infections in lowincome communities of Karachi.
During the last 15 years, very few reports about circulating genotypes of rotavirus are available from Pakistan [36][37][38] in which circulation of G1, G2, G4 and G9 in combination with P [4], P [6], P [8], P [11] have been described. The current study is focused on the detection of circulating genotypes of rotavirus causing gastroenteritis in Lahore city of Punjab province and provides relatively comprehensive epidemiological as well as genetic information. Although G1 is the most common genotype found in diarrhea or gastroenteritis cases worldwide, we found G2P [4] as the most prevalent genotype in 48% of study samples. Genotype G9 was found among 20% of our samples in combinations with P [4], P [6] and P [8]. A similar increased prevalence of G9 strains have been found in neighboring countries including India and Bangladesh [39]. This genotype has recently gained sufficient epidemiological concern worldwide due to its variable vaccine response and infectivity rate [9,40]. We found G9P [8] as the third most prevalent genotype which is consistent with the recent findings from India [41]. 20% of our samples were found positive for G1P [6] which has not been reported previously from Pakistan. We did not detect any G3 and G4 genotypes in our samples which substantiate the global scenario of their significant decline during the recent years [41]. G1P [8] was found in only 1 sample even though it has remained the most common strain detected globally and constitutes one of the currently recommended vaccine, Rotarix TM . Surprisingly, we did not find any G12 genotype in this study which has recently emerged with numerous reports of global spread [42,43]. We cannot speculate the absolute absence of this genotype from our population due to limited time period and number of samples selected for this study. Such variation and vast genetic differences of rotavirus strains evoke the need for continued surveillance to assess and consider the role of strain variability in the design of new vaccine candidates as well as to measure the impact of vaccine introduction in the community. The continued surveillance programs and characterization of rotavirus strains during the post-vaccination era is needed to monitor the emergence of 'escape' strains due to longterm pressure exerted by homotypic immunity [44].
In the present study, we compared the amino acid motifs constituting the neutralizing epitopes of VP7 and VP4 proteins between the circulating Pakistani rotaviruses and available vaccine strains. Multiple antigenic variations were noticed among both VP7 and VP4 epitopes highlighting the need for detailed antigenic mapping of prevalent rotavirus genotypes in the country. Complete genome sequencing will therefore be required to generate more comprehensive information on molecular epidemiology and evolutionary dynamics of rotaviruses on account of its segmented genome and the possible role of internal genes in immunity [45]. Although the aminoacid residues within the VP7 protein of G1 strains showed higher similarities with the Rotarix TM strains, significant differences were found amongst all four epitopes of VP4 proteins. In addition, a comparable degree of disparities was noticed among the RotaTeq TM and the Pakistani strains in both VP7 and VP4 proteins. Although such differences do not change the genotype specificity of the rotavirus strain, they may significantly influence the binding of neutralizing antibodies and hence viral fitness through selective pressure [45]. Similar findings have been reported from countries such as Belgium where rotavirus immunization program was initiated in 2006 followed by rapid massive coverage of up to 88% [46].
The vaccine introduction has significantly reduced the rotavirus disease burden in many countries like Australia, Austria, Belgium, Brazil, El Salvador, Nicaragua, Panama and the United States and both vaccines have been found to carry equivalent efficacy against G1, G3, G4 and G9 strains with P [8] specificity [47]. The Rotarix TM vaccine efficacy against G2 strains is somewhat lower as compared to other genotypes while data pertaining to G12 strains is not available as yet [47]. In addition, the vaccine introduction has modified the epidemiological patterns of circulating strains in certain regions such as the upsurge of G3 strains in the United States and Australia after introduction of RotaTeq TM [48][49][50]. Therefore, further detailed epidemiological studies must be planned to monitor the vaccine response during pre-and postvaccination periods in Pakistani Population.
The mutations in the antigenic regions of rotavirus play an important role in the outcome of vaccine response. When compared to the respective prototype strains, our data substantiates the previous findings [51,52] that the currently circulating G9 genotypes are different from their prototypes that were isolated during 1980s. There are about 15 known mutations in the antigenic regions of G9 genotype that can modify the antigenicity of the respective region [53]. Similarly, for G1 genotypes, the substitutions at positions 94, 97, 147 and 291 of VP7 protein are found to play a significant role in antigenic recognition [54]. In respect to G2 rotavirus genotypes, the substitution D96N within the antigenic region A was found responsible for failure of G2specific monoclonal antibodies to react [55], but none of these mutations were found in Pakistani strains, except for G2 strains where D96N; Aspartic acid to Asparagine, substitution was found amongst all viruses from Pakistan (data not shown). Jin et al reported that most of the currently circulating G1 strains and the Rotarix TM vaccine RRV-S1 strain differ in their antigenic properties and mutations in these antigenic sites may ultimately cause vaccine failure [56].
Despite these variable epidemiological reports, it has been now established that rotavirus vaccines are equally effective against vast diversity of rotavirus genotypes by generating heterotypic immune response. In the recent vaccine trails, both Rotarix TM and RotaTeq TM have been proven as equally effective in regions with high child and adult mortality [57]. Multiple reviews from high and middle income countries have reported a substantial reduction in disease burden within a few years of vaccine implementation through decreased magnitude of rotavirus associated diarrhea and deaths [58,59]. For instance, studies in Mexico and Brazil reported a reduction in diarrhea related deaths in infants and young children after the introduction of rotavirus vaccine [60,61]. Rotarix TM efficacy has been evaluated in a large clinical trial of more than 63,000 infants from 11 Latin American countries and Finland, and was found to be safe and highly immunogenic [62,63]. Similarly, in a randomized, double-blind, placebocontrolled study conducted in 6 European countries, Rotarix TM was observed to be highly immunogenic [64]. RotaTeq TM efficacy has also been evaluated in two phase-III trials among healthy infants, including a large clinical trial of more than 70,000 infants enrolled primarily in the United States and Finland, and found to be highly immunogenic [65,66]. As a final point, in April 2009, the World Health Organization's (WHO) Strategic Advisory Group of Experts (SAGE) on immunization recommended the inclusion of rotavirus vaccination of infants into all national immunization programs, with a stronger recommendation for countries where ''diarrheal deaths account for $10% of mortality among children aged ,5 years'' [67]. In addition, it has been emphasized that the introduction of rotavirus vaccine should be accompanied by measures to ensure high vaccination and timely administration of each vaccine dose.

Conclusion
Our present findings conclude that the rotavirus genotypes circulating in different geographical areas of Pakistan are quite variable and large scale studies must be conducted to calculate the burden of disease as well as the epidemiological understanding of contributing viral genotypes. Similar diversity of prevalent rotavirus strains within the Indian subcontinent including Pakistan has been reported by Miles et al., [41]. The annual birth rate in Pakistan is approximately five million; the right rotavirus vaccine would be greatly helpful to protect newborns from this serious disease and its associated mortality. Although our report does not describe a thorough representation of the prevailing genotypes throughout the country, but it provides significant information for the health policy makers to review and implement informed immunization policy in the country.