Skip Navigation

Serotype Diversity and Reassortment between Human and Animal Rotavirus Strains: Implications for Rotavirus Vaccine Programs

  1. Jon R. Gentsch1,
  2. Ashley R. Laird1,
  3. Brittany Bielfelt1,
  4. Dixie D. Griffin1,
  5. Krisztián Bányai2,
  6. Madhu Ramachandran1,
  7. Vivek Jain1,
  8. Nigel A. Cunliffe3,
  9. Osamu Nakagomi4,
  10. Carl D. Kirkwood5,
  11. Thea K. Fischer1,
  12. Umesh D. Parashar1,
  13. Joseph S. Bresee1,
  14. Baoming Jiang1 and
  15. Roger I. Glass1
  1. 1 Respiratory and Enteric Viruses Branch, Division of Viral and Rickettsial Diseases, Centers for Disease Control and Prevention, Atlanta, Georgia
  2. 2 Regional Laboratory of Virology, Baranya County Institute of State Public Health Service, Pécs, Hungary
  3. 3 Department of Medical Microbiology and Genito-Urinary Medicine, University of Liverpool, Liverpool, United Kingdom
  4. 4 Nagasaki University Graduate School of Biomedical Sciences Sakamoto, Nagasaki, Japan
  5. 5 Murdoch Children's Research Institute, Royal Children's Hospital, Victoria, Australia
  1. Reprints or correspondence: Dr. Jon R. Gentsch, Respiratory and Enteric Viruses Branch, MS G-04, Centers for Disease Control and Prevention, 1600 Clifton Rd. NE, Atlanta, GA 30333 (jrg4{at}cdc.gov).

Abstract

The development of rotavirus vaccines that are based on heterotypic or serotype-specific immunity has prompted many countries to establish programs to assess the disease burden associated with rotavirus infection and the distribution of rotavirus strains. Strain surveillance helps to determine whether the most prevalent local strains are likely to be covered by the serotype antigens found in current vaccines. After introduction of a vaccine, this surveillance could detect which strains might not be covered by the vaccine. Almost 2 decades ago, studies demonstrated that 4 globally common rotavirus serotypes (G1–G4) represent >90% of the rotavirus strains in circulation. Subsequently, these 4 serotypes were used in the development of reassortant vaccines predicated on serotype-specific immunity. More recently, the application of reverse-transcription polymerase chain reaction genotyping, nucleotide sequencing, and antigenic characterization methods has confirmed the importance of the 4 globally common types, but a much greater strain diversity has also been identified (we now recognize strains with at least 42 P-G combinations). These studies also identified globally (G9) or regionally (G5, G8, and P2A[6]) common serotype antigens not covered by the reassortant vaccines that have undergone efficacy trials. The enormous diversity and capacity of human rotaviruses for change suggest that rotavirus vaccines must provide good heterotypic protection to be optimally effective.

Globally, rotavirus infection is the most important cause of severe diarrhea in children. Most deaths occur in less-industrialized countries [1], and health organizations worldwide are promoting the development of rotavirus vaccines to help control this disease. The current strategy is based on the use of live, attenuated rotavirus vaccine candidates designed to elicit immunity comparable to that induced by natural rotavirus infections by providing homotypic or heterotypic protection against severe diarrhea caused by the major rotavirus serotypes in circulation [2]. The earliest vaccine candidates were individual animal rotaviruses with serotypes that were usually distinct from those of common human strains. Although these candidate vaccines provided good heterotypic protection in some vaccine trials, in other trials, they either failed to provide protection or provided the best protection only when the serotype of the animal strain matched that of the infecting human strain [3]. These results suggested that serotype-specific immunity might play an important role in protection against rotavirus gastroenteritis. This notion prompted the development of vaccine candidates based on human strains and human-animal reassortants containing the common serotype antigens of human strains. Over the past 2 decades, studies of the burden of rotavirus disease and strain surveillance in many countries have assessed the need for rotavirus vaccines and have determined the most important human rotavirus (HRV) serotypes that should be targeted by candidate vaccines.

In this report, we review the results of strain surveillance and characterization studies published through early 2004 and discuss new insights gained from these studies on the potential mechanisms of the evolution and spread of new rotavirus strains. Although other recently published articles have reviewed the strain diversity and genetic and antigenic variation of rotavirus, our aim is to provide a profile of strains that could potentially affect rotavirus vaccine programs [46].

Rotavirus Structure and Serotypes

Rotaviruses, which comprise a genus of the Reoviridae family, have a capsid with 3 protein layers that encase a genome with 11 segments of double-stranded RNA (dsRNA) (figure 1) [7]. Each segment usually codes for a single structural or nonstructural protein. The inner layer encasing dsRNA is composed of the VP2 protein and small numbers of the VP1 and VP3 proteins, which are associated with the genomic RNA. This core is surrounded by the middle protein layer, which is composed entirely of VP6, the antigen that defines group and subgroup (SG) specificities. The outer capsid layer consists of the VP7 glycoprotein layer, in which VP4 spikes are embedded. The 2 outer capsid proteins carry rotavirus serotype (neutralization)-specific antigens and are encoded by segments 4 (VP4 protease-sensitive protein and P serotype antigen) and by segments 7, 8, or 9 (VP7 glycoprotein and G serotype antigen). Because the VP4 and VP7 proteins are encoded by separate gene segments, rotaviruses can generate new P-G serotype antigen combinations through reassortment after dual infection of single cells. Because both serotype antigens are believed to be key in the development of protective immunity, it is necessary to assess their prevalence and to study genetic and antigenic variation for both G and P serotypes. Although there is good evidence from animal experiments that passively transferred VP7- and VP4-specific antibodies protect separately, it is less clear which component of the immune response is most important for protection after natural infection or vaccination [8].

Figure 1

Rotavirus structure showing protein coding assignments of 11 genome RNA segments (left). Schematic diagram (middle) and cryoelectron microscopic reproduction of a virion (right) show the location of major structural proteins. NSP, nonstructural protein. Reproduced with permission from Estes [7].

At least 10 G serotypes and 11 P serotypes and subtypes of HRV have been identified. Because identification of P serotypes is technically difficult, usually, only the corresponding VP4 genes are identified by genotyping methods in surveillance studies. To distinguish strains that have been identified by genotyping only from those identified by P-serotyping, a dual nomenclature is used. P genotypes are expressed as “P,” followed by a number in brackets (e.g., P[6]), whereas P serotypes are designated by “P” with a serotype number, followed by the corresponding genotype in brackets (e.g., P2A[6]) [7].

Rotaviruses can also be classified according to their VP6 SG specificity (I or II), by use of an EIA with monoclonal antibodies (MAbs) or by nucleotide sequencing of a VP6 gene fragment [9, 10], and according to their RNA profile (long or short electropherotype), by use of acrylamide gels and on the basis of the migration rate of gene 11. The short-electropherotype phenotype results from a partial duplication in gene 11, which causes it to migrate more slowly than gene segment 10, whereas the standard-sized gene 11 of long-electropherotype strains migrates faster than segment 10 [11].

The most common HRV strains belong to 2 distinct genome constellations (genogroups) that are only minimally related, as determined by high-stringency hybridization procedures performed with whole-genome probes [12, 13]. Strains of the Wa genogroup typically have long electropherotypes and carry the SG II antigen, whereas members of the DS-1 genogroup usually have short electropherotypes and carry the SG I antigen.

Serotypes are defined by neutralization studies, and the immune targets are usually located in the outer capsid. During the early 1980s, the first data on the distribution of rotavirus serotypes were obtained through cultivation of the most common isolates found in stool samples and their characterization in cross-neutralization studies by use of antisera that were hyperimmune to the individual strains [14]. These studies identified 4 common serotypes, now designated serotypes G1–G4, that corresponded to antigenic determinants located on the VP7 protein. Studies of animal models subsequently demonstrated that VP7 played an important role in protective immunity [15, 16]. On the basis of these pioneering studies, the 4 commonly identified G serotypes were chosen for incorporation into early and current reassortant vaccine candidates. The subsequent development of EIA methods for direct typing of rotaviruses in stool samples, by use of serotype-specific MAbs [17, 18], allowed for large-scale epidemiologic studies that documented the global incidence of those 4 original serotypes (reviewed elsewhere [19, 20]) (figure 2, left panel). In those studies, serotype G1 represented approximately half of the strains globally, followed by serotypes G3, G4, and G2 [19, 20]. Overall, MAb serotyping data suggested that serotypes G1–G4 represented at least 90% of strains circulating globally and needed to be targeted by candidate vaccines. Typically, ∼30% of rotavirus-positive stool specimens cannot be typed with MAbs and are, thus, designated as nontypeable (NT) strains [21]. Characterization of NT strains by cultivation and by antigenic and molecular analysis led to the identification of several new HRV serotypes (G8, G9, and G12). Thus, by 1990, there were 7 different HRV G types, although types G8, G9, and G12 were considered to be rare [2225].

Figure 2

Findings from typing studies conducted during 1987–1991 that used G serotype-specific monoclonal antibodies (left panel; data are adapted from Woods et al. [19]) or that used G-serotyping and P-genotyping or G- and P-genotyping (right panel). The “other” category in the chart of P and G types refers to strains that were nontypeable (NT) for P type, G type, or P and G type, or were mixed infections of common P or G types.

The serotypes of rotavirus needed to be reconsidered when VP4, the other outer capsid protein, was identified as an independent serotype antigen (P serotype) by sequencing of VP4 genes from strains with different G serotypes and by neutralization studies [15, 26, 27]. P serotypes were also shown to be important for protective immunity by studies of animal models in which reassortants containing HRV P serotypes and animal rotavirus G serotypes were used [28]. However, serotyping was a challenge in the absence of a collection of MAbs specific for the diversity of P serotypes. Although MAbs to several common P serotypes were identified, cross-reactivity between the types precluded their use for routine P serotyping [29, 30].

New methods that greatly facilitated strain surveillance studies, including reverse-transcription polymerase chain reaction (RT-PCR) genotyping and automated nucleotide sequencing, have been widely used for this purpose since the early 1990s. A multiplexed, seminested RT-PCR method to identify G serotypes by genotyping [31] permitted the detection of both the common serotypes (G1–G4) and the rare serotypes (e.g., G8 and G9) for which EIA-based serotyping antibodies were either not available or not in routine use. Finally, these multiplexed RT-PCR methods were extended to permit genotyping of the other major neutralization protein, the P protein (VP4) [32, 33]. As for G-genotyping, common P types (P[4] and P[8]) and newly identified rare P types (P[6], P[9], and P[10]) could be detected. Analogous hybridization methods for genotyping were also developed [34, 35].

In 8 countries, initial surveillance studies conducted with G- and P-genotyping or with G-serotyping and P-genotyping again showed the presence of 4 common types (described in depth elsewhere [36]). In agreement with earlier molecular and antigenic characterization of culture-adapted prototype G1–G4 strains isolated from children with diarrhea, most G1, G3, and G4 isolates found in stool specimens were of genotype P[8], with SG II specificity and long electropherotypes, whereas G2 strains were of genotype P[4], with SG I specificity and short electropherotypes. Surprisingly, these studies also detected 10 uncommon reassortants, including rare P types (P[3], P[6], and P[9]) and G types (G5), at a combined prevalence of ∼3%, suggesting that earlier studies in which serotyping was used alone underestimated strain diversity. This notion proved to be true as expanded surveillance studies, especially those conducted in developing countries where surveys were not previously performed, yielded many examples of unexpected G and P genotypes.

Review of Studies

We have reviewed the global distribution of rotavirus strains as documented by studies of rotavirus diarrhea conducted in 35 countries and involving >21,000 HRV strains. Studies were included in our review if they were published in the English language through January 2004 and if they reported G and P type combinations. These studies usually used RT-PCR genotyping exclusively, but some used a combination of G-serotyping and RT-PCR (figure 2) [37105]. The 4 common strains—P[8]G1, P[4]G2, P[8]G4, and P[8]G3—represent almost 72% of all strains, and >2% consist of 2 reassortants of the recently emerged serotype G9 (P[8]G9 and P[6]G9). Strains in the “other” category typically included mixed infections with globally common types and those that could not be typed for P or G genotype or for both P and G genotype. Approximately 6% of typeable strains are composed of 25 other P-G combinations types (figure 2, right panel).

A number of studies showed striking examples of strain diversity that, together, demonstrated the importance of serotypes other than G1–G4 as a cause of gastroenteritis in children. These studies included the detection of P[8]G5 in Brazil, P[6]G8 and P[4]G8 in Malawi, and P[6]G9 in India (figure 3). Although G5 and G8 do not appear to be globally important (figure 4), they are clearly of major importance in Brazil and Malawi [40, 95, 106]. Serotype G8 may also be epidemiologically important in other parts of Africa [107]. When a high prevalence of the P[6]G9 strain was detected in India (figure 3), it was still believed to be rare elsewhere and, thus, was considered to be only regionally common [41]. However, a recent compilation of surveillance studies performed since 1995, including many not reviewed here because they did not report P-G genotype combinations, indicated that G9 emerged as early as 1993 to become 1 of 5 important serotypes globally, with a prevalence of at least 5.8% through the end of 2001 [108]. When recent studies showing a countrywide prevalence as high as 40% are included, the estimate of G9 prevalence is likely to increase substantially [109].

Figure 3

Examples of regionally important strains from India (n = 133) [41, 42], Malawi (n = 100) [53], and Brazil (n = 130) [40]

Figure 4

Reassortants of human rotavirus P and G serotypes, including uncommon and regional strains isolated from children with diarrhea

To try to develop a current picture of strains that have the highest potential to affect rotavirus vaccination programs, we summarized the most important P-G combination types detected in surveillance studies conducted since 1996 that analyzed at least 50 strains (table 1). Strain types that had an overall incidence of <0.5% are listed in the note to table 1. Five strains, including the P[8]G9 strain, were considered to be globally common on the basis of overall prevalence. Although the P[6]G9 strain was detected first, as the emergence of G9 began around 1995, more recent studies suggested that the P[8]G9 strain may be predominant [90, 110]. At least 6 regionally common strains, including P[8]G5, P[6]G8, and P[6]G9, with a prevalence of at least 0.5%, are circulating among children. Of note, P[6] strains, which were previously identified as the third most prevalent P type [5], represented >5% of the circulating strains in 30 of 47 studies reviewed (data not shown). Genotyping studies conducted in 11 of 12 African countries (not reviewed here) documented a prevalence of 18%–75% for the P[6] strain, which suggests that the average global prevalence may prove to be substantially higher as more surveys are completed in Africa and Asia [107]. Finally, strains with G types other than the common ones (G1–G4 and G9), including G5, G6, G8, and G10 in various combinations with P[4], P[6], P[8], P[9], and P[14] types, made up ∼1.2% of the total.

Table 1

Summary of rotavirus P- and G-genotyping studies conducted since 1996.

When overall strain incidence was compared with the incidence in studies conducted in different continents, several differences are obvious. The most striking difference is the higher incidence of P[6] strains in Latin America, Africa, and Asia, compared with that in more-industrialized areas, such as Europe and North America. In addition, the incidence of unusual strains (e.g., P[4]G1 and P[4]G3) is often higher in developing countries. Finally, as mentioned above (figure 3), the regional importance of some strains (e.g., P[6]G8 in Malawi and P[6]G9 in India) is illustrated.

Discussion

The introduction of molecular typing methods has enhanced our understanding of the diversity of rotavirus strains that affect both the development of rotavirus vaccines and our understanding of viral evolution. This review of >21,000 strains from 35 countries provides several new insights.

Common Strains

These studies confirmed the continued importance of serotypes G1–G4. Globally, >90% of single fully typeable strains bear these VP7 antigens, and only Africa had an incidence of <90% (i.e., 86%) for these serotypes.

Increased Strain Diversity

Studies conducted before 1990 suggested that only 4 common serotypes accounted for virtually all rotavirus strains circulating among children. Since then, along with the widespread application of RT-PCR genotyping and molecular methods, strain surveillance and characterization studies have led to the identification of at least 42 distinct P-G type combinations among the 10 HRV G serotypes and 11 HRV P serotypes and subtypes, representing more than one-third of the 110 theoretically possible P-G combinations (figures 2, 3, and 4) [26, 111129]. In addition to the original 4 strains, at least 8 other globally or regionally common strains have been described that, overall, contain 4 additional serotype antigens (G5, G8, G9, and P[6]) and that are not included in the polyvalent rotavirus vaccine currently in development. This number is likely to increase as surveillance studies continue. Many of the 42 strains, including several that are epidemiologically important, appear to have been formed by the introduction of a new serotype antigen gene into a common HRV strain (e.g., the introduction of G8 and G12 VP7 genes into DS-1 genogroup strains) [111, 113]. Other strains may have arisen by interspecies transmission or by reassortment between human and animal rotaviruses [130, 131]. The finding of this enormous diversity among rotavirus strains provides insights into the evolution of rotavirus strains [4, 5, 132] and creates new challenges for rotavirus vaccine programs.

Vaccines currently in the late stages of development include one based on a monovalent serotype P1A[8]G1 HRV strain (Rotarix; GlaxoSmithKline) and another based on a pentavalent bovine-human reassortant strain containing G1–G4 and P1A[8] antigens (RotaTeq; Merck) [133, 134]. Rotarix is based on the concept that infection with 1 HRV serotype may elicit heterotypic protection as a result of cross-reactive antigens between serotypes, whereas RotaTeq is based on serotype-specific immunity and, in theory, should contain the antigens of the most common HRV strains to achieve optimal protection. Thus, one major challenge for the RotaTeq vaccine will be to protect effectively against the globally (G9) or regionally (G5 and G8) common G serotypes that have been identified by strain surveillance in recent years but that are not present in the vaccine. Although RotaTeq might be expected to provide a measure of protection against some of these strains (e.g., P[8]G5 and P[8]G9) because of shared P serotype antigens, other strains (e.g., P[6]G9 and P[6]G8) share neither serotype antigen and, thus, may offer the biggest challenge to this vaccine. On the other hand, the Rotarix vaccine may work well against strains that have different serotypes but many cross-reactive antigens (e.g., strains in the same [Wa] genogroup: P[8]G9 and P[8]G3), but the vaccine may be challenged by short-electropherotype P[4]G2 and P[6]G9 strains that not only have distinct serotype antigens but also belong to a completely unique genogroup (DS-1) from the Rotarix strain and, thus, have fewer cross-reactive antigens. Consequently, it will be important to determine from vaccine trials of these 2 candidates whether they elicit immunity against strains that may challenge their protection mechanism [4]. In this regard, available vaccine trial data for Rotarix show that it provides cross-protection against P[8]G9 strains, but it is not yet known whether it protects against P[4]G2 strains of the DS-1 genogroup [135].

Mechanisms of Rotavirus Evolution

In addition to the challenges posed for rotavirus vaccine programs, the great diversity observed in studies of rotavirus strain surveillance and characterization provides insights into the genetic variation and spread of rotavirus strains. Rotaviruses evolve by point mutations, gene rearrangements of primarily nonstructural genes, and reassortment events, all of which have long been known [5, 136]. An intermolecular recombination of rotavirus has been described once [137].

Reassortment

Reassortment between the common strains detected in surveillance studies is well documented. The globally common genotype P[8] strains with serotype G1, G3, G4, or G9 VP7 genes belong to the same genome constellation (Wa genogroup), as indicated by the ability of all genes, except VP7, to cross-hybridize with other genogroup members [13]. The fifth globally common strain, P[4]G2, belongs to a distinct genome constellation (DS-1 genogroup) that does not cross-hybridize with any gene segment from typical members of the Wa genogroup. Nucleotide sequencing studies provided strong evidence that reassortment occurs between circulating genotype P[8] strains with G1, G3, G4, or G9 specificity when it was found that phylogenetically distinct VP4 genes of such strains can segregate with VP7 genes of >1 G serotype [138, 139]. Other sequencing studies demonstrated that all 11 genes reassorted between typical long-electropherotype G1 and G4 strains. When fragments of each gene of cocirculating G1 and G4 strains were sequenced, it was found that the cognate genes from the 2 serotypes could be distinguished phylogenetically. Subsequently, it was shown that the distinct lineages of each gene could be identified in both serotypes, demonstrating that reassortment occurs in all 11 gene segments of such strains [140]. Together, these studies show that reassortment is a major evolutionary mechanism in common circulating rotavirus strains.

In contrast to the high level of multigenic reassortment that is believed to occur between members of the same genogroup (designated “intragenogroup reassortment”), independent segregation of genes between the major HRV genogroups (designated “intergenogroup reassortment”) is uncommon overall. The evidence, in part, comes from early studies showing that the G types of commonly circulating strains are usually associated with a single P type, SG, and electropherotype [36]. These data confirmed earlier investigations conducted by use of PAGE analysis, subgrouping, and whole-genome hybridization of partially restricted exchange of gene segments between strains in the 2 main rotavirus genome constellations (Wa and DS-1 genogroups) [13]. However, more-complete surveillance studies have documented numerous examples of the reassortment of serotype antigen genes across genogroups, and single G types have been detected in association with a wide variety of P types. The global incidence of these reassortants is low (<10%), but it is clear that they can be very common in some settings; P[6]G9 is common among both Wa genogroup and DS-1 genogroup strains [108]. Numerous examples of intergenogroup reassortment in other genes are also well documented [141144].

The generation of new P-G genotype combinations by the introduction of genes from novel serotypes represents another mechanism for the generation of rotavirus diversity. Sequencing studies of the VP7 gene of emerging serotype G9 strains detected around 1995 demonstrate that the VP7 gene is distinct from the cognate gene of G9 strains first isolated in the United States a decade previously [23]. This result indicates that the modern lineage is not directly descended from the original lineage, and may, instead, be the result of a recent introduction into humans through reassortment [145, 146].

In children with diarrhea, the modern G9 VP7 gene lineage was first detected among long-electropherotype P[8]G9 strains during the 1994–1995 rotavirus season in Japan and among P[4]G9, P[6]G9, and P[8]G9 strains over the next several years in other countries [62, 145149]. Hybridization studies indicate that these strains represent 2 distinct genotypes, P[8]G9 and P[6]G9, which were single-segment reassortants of typical members of the Wa and DS-1 genogroups, respectively [108]. The same G9 lineage was subsequently shown to be present as early as 1993 in Indian neonates, who excreted a long-electropherotype P[6]G9 strain in the absence of symptoms of diarrhea [108]. Thus, it is plausible that these unusual neonatal infections could have been an early source for the spread of this gene to children with diarrhea, from which it subsequently spread globally and reassorted to produce a variety of new genotypes.

Interspecies transmission

Another major source of HRV diversity involves the introduction of animal rotavirus genes either through transmission of whole viruses or through reassortment. Evidence for the first of these mechanisms came from hybridization studies that used whole-genome probes made from HRV strains by in vitro transcription. For example, all 11 segments of several HRVs (e.g., AU-1, P3[9]G3, and HCR-3 P5A[3]G3 strains) are virtually indistinguishable from feline and canine strains with the same serotype, suggesting that these uncommon strains with novel P serotypes were derived through interspecies transmission to humans [130, 131]. However, only a few strains have high homology to all 11 genes of animal rotaviruses, suggesting that interspecies transmissions that result in gastroenteritis are rare.

Animal-HRV reassortment

A more common mechanism for the introduction of animal rotavirus genes into HRVs is through reassortment. Examples of both rare (e.g., G6, G8, P3[9], and P5A[3]) and common (e.g., G3, G4, and P1A[8]) P and G HRV serotypes (figure 4) that have very close genetic and antigenic relationships with the same rotavirus serotype in animals are well known. Especially intriguing are recent findings that, when some common HRV G serotypes (G3 and G4) and P serotypes (P1A[8]) are sequenced, they are almost indistinguishable from the same genes in porcine or canine rotavirus strains, which suggests that even common HRV serotypes may have recent animal origins [150]. Whole-genome hybridization experiments and nucleotide sequencing show that animal-human reassortants may contain only a serotype antigen gene (e.g., strain 116E) or several genes related to animal rotaviruses [151, 152]. In the case of strain 116E, the VP4 gene is highly related to the bovine serotype P8[11]G10 VP4 gene, whereas the remaining 10 genes are related to typical HRV strains of the Wa genogroup (e.g., P1A[8]G1, SG II, long electropherotype) [121]; strain I321 (P[11]G10) contains 2 genes from HRVs of the Wa genogroup, and its remaining genes are from a bovine rotavirus [152]. Some of the G5 strains isolated in Brazil, such as Br1054 (P[8]G5), contain several genes each from human Wa genogroup strains and porcine rotaviruses related to the serotype G5 isolate, OSU [153]. Other uncommon strains, such as PA151 (P3[9]G6) and PCP5 (P3[9]G3), which were isolated in Italy, have the same P serotype as feline rotaviruses and the felinelike HRV strain AU-1. Unlike AU-1, these strains are apparent reassortants, deriving several genes each from bovine rotaviruses and AU-1-like rotaviruses [154]. Strains that have the same numbers of bovine and HRV genes or the same P and G types as PA151 and PCP5 have been detected in the United States and Hungary [113, 155]. The finding of such close sequence similarity in these geographically and temporally diverse strains suggests that they possibly share a common origin. These strains may have entered the human population at least once or more and subsequently may have acquired a G3 VP7 gene through reassortment with a human G3 rotavirus [113, 126].

A variety of other HRVs with strong associations with animal strains have been reported. The serotype G8 strain 69M (P4[10]G8), which has a supershort electropherotype pattern, contains genes related to HRVs of the DS-1 genogroup and others related to typical bovine rotaviruses [156]. Long-electropherotype G8 strains (e.g., Hal1166 and other P[14]G8 isolates) from a variety of settings share similar relatedness to the human DS-1 genogroup and bovine rotaviruses, but they also share a high homology to the typically lapine VP4 gene (P[14]), which suggests that such strains could have been derived through reassortment events in 2 different animal species [125, 156158].

Mixed infections

The great degree of strain diversity among rotaviruses, particularly in some developing countries, suggests that, for children, coinfections with 2 different rotavirus serotypes may be a relatively frequent occurrence. In fact, mixed infections with rotavirus strains appear to be quite common in some settings in developing countries, on the basis of studies showing 2 RNA profiles in the same patient specimens and on the basis of results of RT-PCR genotyping and EIA serotyping studies that demonstrated the presence of >1 genotype or serotype in the same stool sample [36, 132, 159]. In strain prevalence surveys, the highest levels of detection of mixed infections are often in developing countries, where higher numbers of different genotype combinations have been detected (figures 3 and 5) [41, 51, 99, 160]. In India, for example, multicenter surveys identified as many as 9 P-G genotype combinations among rotavirus specimens collected in a single city [41, 64]. In contrast, in studies conducted in developed countries, fewer mixed infection were seen, and fewer genotype combinations were detected per city surveyed (figure 5) [36]. Thus, the high levels of detection of mixed infection in children with diarrhea, especially in developing countries, may play a major role in generating strain diversity.

Figure 5

Differences in the incidence of mixed infections in developed versus developing countries. Data are adapted from the following references: [36, 40, 41, 9799, 101103, 105, 159]. BAN, Bangladesh; BRA, Brazil; IND, India; JPN, Japan; Malay, Malaysia; S. Afr, South Africa; USA, United States of America.

Role of neonates

Another intriguing possibility is that mixed infections in neonates could be an important source of novel rotavirus strains. It has been known since the 1970s that strains with common G serotypes and genogroups and novel P serotypes (P2A[6]) circulated in hospital nurseries, often without producing symptoms of diarrhea [161]. These strains sometimes circulated continually in the same nursery for years and thus served as an uninterrupted reservoir where mixed infections could potentially occur any time another strain was introduced by staff or visitors. Although, at first, they were thought to be confined to neonates, P[6] strains are relatively common in children with diarrhea, suggesting that reassortment in neonates could be one possible source for new strains. Strains undergoing reassortment in neonates could explain the origin of other HRVs as well. As noted, the VP7 gene of novel P[6]G9 strains was first detected in infected neonates by sequence analysis. The same VP7 gene lineage is now common in children with gastroenteritis.

The novel P2A[6]G8 strains that are common in Malawi were detected in neonates and children with diarrhea in Malawi at approximately the same time [53, 162]. Two distinct P8[11] strains with G9 or G10 specificity were first detected in neonates [163, 164]. The P[11]G10 strains are now common in children with diarrhea in some parts of India, whereas a P[11]G4 reassortant is detected infrequently in sick Indian children [42, 165].

Acknowledgment

We thank Claudia Chesley for her help in editing the paper.

Footnotes

  • Financial support: Centers for Disease Control and Prevention.

  • Potential conflicts of interest: none reported.

  • Addendum. Since this manuscript was accepted for publication, another review of the global distribution of rotavirus strains and its possible effect on rotavirus vaccines has been published [166].

References

| Table of Contents