Diversity of Rotavirus VP7 and VP4 Genotypes Associated with Severe Childhood Diarrhea in North West Cameroon: Detection of Unusual Strains G1P[6], G2P[6], G2P[8] and G3P[6]

Aims: To determine rotavirus genotypes and distribution among children 0 - 5 years old with severe acute diarrhea in the North West Region (NWR) of Cameroon. Study Design: Cross sectional. Methodology: Fifty-six Enzyme Immunoassay (EIA) and Polyacrylamide gel electrophoresis (PAGE) - positive rotavirus stool specimens were analyzed for VP7 and VP4 genotypes by reverse transcriptase polymerase chain reaction (RT-PCR). Results: A total of 51 (91.1%) samples genotyped as G and P types while the remaining five samples (8.9%) were partially characterized. Four VP7 genotypes (G1, G2, G3 and G9) and two VP4 genotypes, P[6] and P[8] were detected. The predominant G and P types were; G3 (34%) and P[8] (41.1%). Genotypes G1-G3 accounted for 85.8% of isolates while G9 represented 3.6% of isolates. Genotype P[6] represented 30.4% of all VP4 genotypes and was the most widespread strain occurring in all age groups. Twelve single G and P-type combinations were identified. Genotype followed by G3P[6] Unusual strains detected were G1P[6], G2P[6], G2P[8] and G3P[6] for 32% of cases. Mixed infections were detected from 15 (27%) isolates comprising G1/G3P[6], G2/G9P[6], G2P[6]/P[8] and G1/G3/G8 P[6]/P[8]. Twenty isolates unusual genotype / electropherotypes combinations. All G9 strains, mixed G genotypes strains children Future studies will investigate post vaccine prevalence and characterization of non-typeable strains by other methods.


INTRODUCTION
Rotavirus-associated diarrhea accounts for about 500,000 deaths yearly in children under years of age worldwide with about 85% of cases occurring in developing countries of Asia and Sub-Saharan Africa [1].
Rotavirus, a genus of the family Reoviridae, has 11 segments of double-stranded RNA as its genome which are enclosed in a triple-layered capsid. These segments encode six structural proteins (VP1-VP4, VP6 and VP7) and six nonstructural proteins (NSP1-NSP6) [2]. Two outer capsid proteins, VP7 and VP4, contain neutralizing epitopes which define serotypes of rotavirus and form the basis for a dual classification of rotaviruses into VP7 (G types) and VP4 (P types) genotypes [3]. Based on diversity of the VP7 and VP4 genes, G and P genotypes have been defined for group A rotaviruses (RVA), respectively, and at least 27 G types and 37 P types have been described [4]. In human rotaviruses, G1, G2, G3, G4, G9, and G12 usually combined with P [4], P [6], and P [8] and are frequently detected throughout the world, with G1P [8] being the most prevalent strain in humans [5].
In order to assess vaccine efficacy rotavirus strain surveillance studies are needed prior to introduction of vaccines in any community [6].
There are very few studies on rotavirus in Cameroon and the Central Africa region.
Reports of rotavirus studies in the West and South West Regions of Cameroon showed a wide diversity of strains including unusual genotypes [7]. It was therefore, necessary to determine the circulating strains in the North West Region (NWR) where no such studies have been conducted before the introduction of rotavirus vaccines. Following the World health Organization recommendations to participating nations especially those with high diarrhea related infant mortality Cameroon became the 21 st country to introduce rotavirus vaccines in the Expanded Program on Immunization in 2014. This study was supposed to provide baseline data on the diversity and distribution of rotavirus genotypes circulating in the NWR of Cameroon before the introduction of rotavirus vaccines in the country.

Study Area and Specimens
In this study we analyze rotavirus strains isolated from an earlier molecular epidemiologic study conducted in the North West Region of Cameroon between January and December 2004 [8]. A total of 543 stool samples were collected from involving children aged 0 -59 months who sought medical care for acute diarrhea at the Bamenda Regional Hospital; divisional hospitals at Bali, Batibo, Ndop and Santa; Presbyterian Health Center Bafut and the Catholic Health Center Bali in respective health districts. A case of acute diarrhea was defined as a child with  3 bowel movement per day with decrease in stool consistency (loose, watery or liquid) and presenting within ≤ 7 days of onset. A child < 60 months of age, male or female suffering from acute diarrhea was considered for inclusion in the study. Any child who met the criteria for inclusion and whose parents or guardian gave informed consent was eligible and enrolled in the study. Consenting parents or guardians completed an interviewer assisted structured questionnaire for socio-demographic and biodata of each child including outcome of the visit whether the child was admitted or treated at the outpatient department.

Polyacrylamide Gel Electrophoresis
Rotavirus dsRNA genome was extracted from 128/153 stool specimens that previously tested positive for rotaviruses by Enzyme Immunoassay. Briefly, rotavirus antigen detection was performed on 10% stool suspension in phosphate buffered saline using commercial DAKO IDEIA Rotavirus kit (DAKO Diagnostics, Sweden) following the manufacturer's instructions. Rotavirus dsRNA was extracted by the Bender method [9] with slight modification for polyacrylamide gel electrophoresis (PAGE) analysis. The extracted dsRNA was applied to separate lanes of a 10% polyacrylamide gel in a discontinuous buffer system overnight for 18-20 hours at 100V. The RNA bands were visualized by the silver nitrate staining technique over an illuminated box [10]. The presence of rotavirus in stool specimens was determined by detection of 11 RNA segments of rotavirus by PAGE as previously described [11].

Genotyping of Rotavirus Isolates
Rotavirus dsRNA genome was extracted from 56 rotavirus positive stool specimens that were determined to have sufficient genomic RNA by PAGE. The viral genome was extracted from 100 L of PAGE positive stool specimens in 10% phosphate buffered solution by sodium dodecyl sulphate and Phenol/ Chloroform following standard procedures as previously described [12] and purified with the RNaid® Kit (Bio 101, Carlsbad, USA) as described [11]. Rotavirus G and P types were determined by the standard two-step reverse transcription polymerase chain reaction (RT-PCR) as previously described [13,14].

Statistical Analysis
Odds ratios with their 95% confidence intervals (CI) were calculated to establish associations between age and rotavirus genotypes using EPI-Info version 3.

RESULTS AND DISCUSSION
We present results of rotavirus genotypes and genotype-electropherotype combinations detected in specimens collected in an epidemiologic study conducted in in North West Cameroon in 2004. Results of rotavirus prevalence, electropherotypes, clinical and sociodemographic data were previously reported [8].
Although G1 is usually the predominant G type in worldwide studies [5,15] G3 predominated in the NWR of Cameroon. The predominance of this strain in the absence of G4 and P [4] also contrasts with reports in the neighboring West and South West regions where both genotypes were detected at low levels [7]. The prevalence of rotavirus genotypes vary from season to season and between different geographic regions [16] which could account for the differences observed. Four rotavirus genotypes (G1-4 and G9) generally constitute about 80-90% of strains detected in surveillance studies but the increasing reports of unusual strains could impact their global epidemiologic significance. A typical example is the increasing prevalence of P [6] in clinical infections [5]. High prevalence of G3, P [6] and mixed P types involving P[6]P [8] are common in this region and have been reported at high levels in in Ghana, Nigeria and many parts of Africa [15]. Our study detected a higher rate of P [6] (30.4%) compared to data from other regions of the country [7] where the strain represented only 0.1% of genotypes in circulation. Continuous detection of high levels of P [6] genotypes in clinical specimens may imply strain evolution with increasing pathogenicity or simply due to increasing rotavirus surveillance especially in developing countries. The strain was initially associated with asymptomatic neonatal infections but now causes significant clinical infection in all age groups representing about 33% to 50% of rotavirus disease cases [5,17].
Infections with strains showing mixed G and P types are not uncommon in Africa. Earlier studies reported significant levels of mixed infections in Guinea Bissau [18] and other parts of the continent [15]. Apart from increasing the risk of reassortment they are also detected more in patients with more severe disease requiring hospitalization. Although it is not clear why mosaic strains are prevalent in some regions it is possible that co-infections with other disease agents and malnutrition that weaken host defense against reinfection may be the mechanism promoting multiple infection and clinical disease manifestation. Common environmental factors such as overcrowding, mass movement and poor living conditions that are prevalent in many developing countries could also facilitate disease transmission leading to mixed infections.

Age Distribution
A total of 54/56 (96.4%) cases of all G and P types occurred in children less than 24 months of age (Fig. 1). Genotype G1 was not detected from children less than 6 months old while P [6] circulated in all age groups. The odds for any age group having a specific genotype compared with other age groups were: 0 -6 months, G2  1) were more likely to occur in male children who are generally more active than females. Peak rotavirus infection in the NWR was previously reported in ages 7 -12 months old [8] and this age group also had the highest prevalence of both G and P genotypes. Most cases of rotavirus occur in children younger than 24 months [8,19] although studies in the neighboring Western and South West Regions of the country reported higher prevalence in older children [7]. Mixed genotypes were seen mainly in children less than 24 months of age defining the greater susceptibility of younger children to rotavirus diarrhea but it was not clear why ages 13-18 showed the most prevalence of mixed G types. However, this is usually the transitional age when children begin to temporally move away from supervising parents and could increase the risk of exposure to contaminated fomites and infection.
Mixed infections are frequently reported in rotavirus studies worldwide [3,12,16,19] and could be responsible for the high rate of emergence of unusual strains secondary to genetic reassortment.

Sex Distribution
A majority of the genotypes 30/56 (53.8%) were from male patients. Fig. 2 shows the sex distribution of rotavirus genotypes. All G9 strains occurred in female children who were also more likely be infected with G1 (OR = 1.46; CI: 0.40 5.32) while male children were more likely to be infected with G2 (OR = 1.28; CI: 0.3 G3 (OR = 1.81; CI: 0.51-6.57) strains but the odds were also not significant (P>0.05). 5 majority of the genotypes 30/56 (53.8%) were from male patients. Fig. 2 shows the sex distribution of rotavirus genotypes. All G9 strains occurred in female children who were also more likely be infected with G1 (OR = 1.46; CI: 0.40ildren were more likely to be infected with G2 (OR = 1.28; CI: 0.3 -5.92) and 6.57) strains but the also not significant (P>0.05). Sex analysis of P types showed that male children were more likely to be infected with P[ 1.79; CI: 0.12 -53.18) and non-typeable P types (OR = 1.22; CI: 0.37 -4.1) and these relationships were also not statistically significant (P>0.05). Reports of higher rate of infection in male children are common [19] but sex did not affect the distribution of rotavirus genotypes causing childhood diarrhea in North West Cameroon. analysis of P types showed that male children were more likely to be infected with P [8] (OR = typeable P types 4.1) and these relationships were also not statistically significant (P>0.05). Reports of higher rate of infection in male children are common [19] but sex did not distribution of rotavirus genotypes causing childhood diarrhea in North West Fig. 3 shows the monthly distribution of genotypes. Most strains circulated between December, 12/56 (21.4%), January 30/56 (53.4%) and February 9/56 (16.1%) which is characteristic of most rotavirus infections [7,15,17]. A majority of G2 strains 9/11(75%), P [6] 10/18 (55.6%) and mixed G, 3/4 (75%) and P, 8/13 (61.5%) genotypes circulated mostly in January. Many studies have shown wide geographical variation in the prevalence of G and P types across continents as well as local and global temporal changes in the frequency of dominant strains [5,15,19]. Knowledge of the seasonal distribution of rotavirus genotypes is important for strategic planning and effective control measures. However our data was collected only for a period of one year and the results might not reflect the true seasonality of strains but could nevertheless shed light on point prevalence of genotypes.
West Africa is known to show great diversity in rotavirus strains and unusual genotypes as well as mixed infections [7,15,20]. Therefore the detection of strains with dual reactivity to either G or P types was not unexpected. However, the detection of a strain with multiple reactivity showing both VP7 and VP4 genotype specificities (G1/G3/G8 P[6]/P [8]) was rather unique in that such combinations have not been previously reported to the best of our knowledge. The mixed reactivity in this case could have resulted from co-infection with more than one mosaic strains with long (possibly G1P [8]) and short electropherotypes (possibly G2P [6], G2P [8], G9P [8], G9P [6], or G2/G9P [6]/P [8]) based on the G and P genotypes detected and high prevalence G2P6 strains in the region with short profiles [17]. Five globally common G/P combinations (G1P [8], G2P [4], G3P [8], G4P [8] and G9P [8]) generally represent about 75% of rotavirus strains in circulation [21]. Only two of these (G1P [8] and G3P [8] were detected and accounted for 18/56 (32.1%) of isolates. The significance of this variation will unfold in future surveillance and vaccine efficacy studies. Genotype G1 P [8] is usually the most common strain detected in most epidemiologic studies [6,7,15,18]. Occurrence of G3P [6], 10/56 (17.9%) as the second most common strain overall could pose a problem in vaccine effectiveness if the current vaccines do not produce sufficient heterotypic antibody protection against globally less common strains such as P [6] if they become prominent in any region. G3P [6] strains have also been reported as the dominant strain in pediatric inpatients in Zimbabwe [19] and protection against the strain in addition to G1P [8] could play a role in reducing rotavirus disease hospitalization in children.
The segmented nature of the rotavirus genome and high rate of mutation [20] suggest that during a mixed infection with different rotavirus strains genetic reassortment (substitution of RNA segments between different RNA strands) could occur giving rise to mosaic strains and unusual genotypes. Mosaic strains occurred mainly in hospitalized children implying that rotavirus could be more severe in West and central African children where such strains are common [7,15,17].
Unlike with many studies in the region that report high levels of untypeable starins [7,15,18], most of our isolates could be characterized by RT-PCR genotyping methods. Failure to type rotavirus strains could result from several factors including presence of rare genotypes, insufficient viral RNA genome, destruction of viral genome during repeated freeze thawing, or due to mutations at specific binding sites in the target genome or primers [18,20,22]. Strains that do not react with RT-PCR genotyping primers could be analyzed by a modification of standard procedures, virus culture and sequencing [18,20,23] that were beyond the scope of our study.

Age (months) Distribution of G/P Combinations
The predominant G/P combinations in different age groups were: 0-3, G2P [ (Table 2). All mixed genotypes occurred in children under 24 months of age while genotype G3P [6], 8/40 (20%) was the most common strain in hospitalized children. All strains with unusual genotypes occurred in children less than 2 years of age with 13/18 (72.2%) appearing in ages 6 -18 months and could pose an increased risk in this age group. Scheduled immunization targeting peak age should protect children in this region from severe rotavirus disease.
G9 strains may also combine with P[6] [5] but strains presented with long RNA profiles and mixed VP4 specificity of P[6]P [8]. Three types of G9 strains have been reported based on electropherotype profile, VP7 and VP4 specificities in which strains either have long or short electropherotypes reacting with either P [4], P [6] or P [8] genotype specific primers [5]. This raises further concerns about the possibility of a complex genetic polymorphism among rotavirus strains that remain to be uncovered and questions about the specificity of genotyping primers for G9 that could react with G3 [20] as well as those for G8 strains that often appear in mixed infections [23]. The detection of all G9 strains, mixed G types and those with atypical G/P -electropherotype combinations in hospitalized children further suggests the possibility of these strains being more virulent although host and environmental factors could also play a role in disease severity. Mixed infections can provide an opportunity for genetic reassortment among rotaviruses [20] that could result in new genotype combinations and unusual strains. Both host and environmental factors could facilitate genetic reassortment among rotavirus strains causing emergence of unusual strain. It is common for humans to cohabit with domestic animals such as goats, sheep, pigs and chicken or cattle in the homes or backyards which could also account for the high rate of unusual strains in the region. Another possibility for unusual strains is the intrinsic nature of the virus as a segmented RNA virus which predisposes to spontaneous mutations during replication.

CONCLUSION
This report presents the diversity of rotavirus G and P types circulating in the NWR of Cameroon in which genotypes G1-3 constituted over 85% of strains with predominance of G1P [8]. G3P [6] emerged as the second most common strain overall and as the predominant strain in hospitalized children with. Unusual strains and mixed infections circulated at high levels including P [6] strains detected in all age groups. These results provide the background for future evaluation of vaccine efficacy and strain evolution in Cameroon and the Central Africa region. Continuous strain surveillance is necessary for a better understanding of rotavirus strain diversity and the epidemiologic significance of G3, G8, G9, P [6] and unusual strains. Follow up studies intend to investigate post vaccination strain prevalence. Further research is also needed to confirm the association of mixed infections and unusual strains with disease severity that may elucidate the high rotavirus disease burden in Sub-Saharan Africa [1] and to clearly identify partially characterized strains and those with atypical genotypes or genotype / electropherotype combinations.

CONSENT
Informed consent was obtained from parents or guardians of all individuals who participated in the study.

ETHICAL APPROVAL
This study was approved by the North West Regional Delegation of Public Health, Cameroon.