Occurrence of Foot-and-Mouth Disease Virus Serotypes in Uganda and Tanzania (2003 to 2015): A Review and Implications for Prospective Regional Disease Control

Endemic foot-and-mouth disease (FMD) presents a global economic challenge to the livestock industry. The progressive control pathway for FMD (PCP-FMD) specifies successive steps through which a country/region can reduce FMD virus circulation and impact. These steps are reliant on understanding and obtaining knowledge on FMD epidemiology, to inform development of appropriate disease interventions like vaccination and quarantine programs. Currently, Uganda and Tanzania are in the early stages of the PCP-FMD. This review was undertaken to determine FMDV serotype distribution in Uganda and Tanzania between 2003 and 2015. The paper also presents the vaccine strains used in both countries for the same period viz avis the circulating topotypes. The review highlights four (O, A, SAT 1 and SAT 2) and five (O, A, SAT 1, SAT 2 and SAT 3) serotypes that occurred in Uganda and Tanzania respectively in the thirteen year period. Observations revealed that reported circulating serotypes O and A in the two countries belonged to similar topotypes, East African 2 (EA-2) and AFRICA respectively. The SAT 1 viruses in Tanzania belonged to topotype I and differed from the Ugandan SAT 1s that belonged to topotype IV. Similarly, the SAT 2s in both countries belonged to different topotypes: IV in Tanzania and I in Uganda. This review additionally, underscores the spatial distribution of FMDV serotypes in Uganda and Tanzania and highlights regions in both countries that had high serotype diversity. The paper recommends definitive disease diagnoses, molecular serotype characterisation and matched vaccination deployment for improved disease control.

disease control (Rweyamamu et al., 2008;Ayebazibwe et al., 2010b;Fèvre, Bronsvoort, Hamilton, & Cleaveland, 2006;Namatovu et al., 2013). Although Europe has managed to control FMD, serotypes O, A, C and Asia 1 have been responsible for most of the incursions that have occurred (Valarcher et al., 2008). Four serotypes (O, A, C, Asia 1) have been documented in circulation on the Asian continent, with brief reported cases of SAT 1 and SAT 2 in the Middle East (Ahmed et al., 2012;Rweyemamu et al., 2008) while, South America has had three serotypes that have been reported in circulation (O, A, and C) . Six out of the seven existing FMDV serotypes have occurred in Africa (O, A, C, SAT 1, SAT 2, SAT 3) and particularly, all six of the seven serotypes have been detected in East Africa (Kivaria, 2003;Sangula et al., 2011;Kasanga et al., 2015, Casey-Bryars et al., 2018. This makes East Africa one of the regions which has had the highest diversity of FMDV serotypes which could consequently complicate FMD epidemiology and control. This is because FMDV vaccines do not offer cross protection (Kitching, 2002;Sahle, Dwarka, Venter, & Vosloo, 2007). Generally, serotype O is most widespread globally, followed by serotype A. The SATs 1-3 are limited to Africa and are highly diverse. Although FMD is endemic in East African countries, serotype C has not been detected since 2004 (Sangula et al., 2011), may be because the serotype is in the process of becoming extinct or its detection has been overlooked. Serotype SAT 3 was previously isolated from the African Buffalo in 2007, but was last reported in 2015 having been isolated from a Ugandan Ankole calf .

Control of FMD in Uganda and Tanzania
In most of the continents, mainly in sub-Saharan Africa, eradication of FMD may be considered as a long term objective (Kitching et al., 2007). Ideally, FMD control and eradiation in endemic settings should involve vaccination of livestock twice a year, control of livestock movement and assessment of the risks associated with introduction of FMD in disease free areas (Thomson et al., 2003;Bruckner et al., 2002;Jori et al., 2009. The use of FMD vaccines has been commonly used in Uganda to prevent spread of FMD outbreaks (Ministry of Livestock Animal Industry and Fisheries [MAAIF], 2009) with annual expenses of FMD vaccines estimated at US $58 000 to $1 088 820 (Muleme et al., 2012). However, lack of routine vaccinations and reports of delayed and inadequate vaccination campaigns have been cited in causing disease spread to other areas (MAAIF, 2009). Issues with FMD laboratory diagnosis and poor reporting systems have increased the risk of disease spread and led to poor disease control options (MAAIF, 2009). In Tanzania, very low levels of vaccinations have been reported as most farmers do not vaccinate their animals (Railey, Lembo, Palmer, Shirima, & Marsh, 2018;Hasler et al., 2017). The Animal Health Strategy and Vision for Tanzania emphasizes improved epidemiological surveillance, better response to outbreaks and improved laboratory diagnostic networks in order to eradicate FMD (Maziku, Mruttu, & Gebru, 2016). The trivalent FMD vaccine used then in Uganda contained strains for serotypes O, SAT1 and SAT 2 , while in Tanzania the multivalent vaccine commonly used consists of serotype O, A, SAT-1 and SAT-2 Sallu et al., 2015). Vaccination against FMD especially in endemic areas remains complex because of multiplicity of antigenic types and subtypes given that there is limited cross protection between serotypes or topotypes within the given serotypes (Kitching et al., 2007). Issues such as poor knowledge of circulating strains, vaccine expense and the short time of protection elicited by the vaccine (often as little as 6 months) hinder the success of vaccination schemes in endemic countries (Kitching et al., 2007;Maree et al., 2014). Although the restriction of livestock and product movement is a control strategy emphasized by both countries, especially in Uganda (Kasambula et al., 2012), it still lacks proper enforcement (East African Community [EAC], 2004;Balinda et al., 2010b), making disease control difficult.

Progressive Control Pathway-FMD
The progressive control pathway for FMD was introduced by the FAO/OIE/EU-FMD to enable countries that are still endemic with FMD to develop strategies centred around existing information on the FMD status in the countries. The initial stages of the pathway, in which Uganda and Tanzania are in, are critical because comprehensive information on FMD is collected so as to move from stage 0 to 1 (FAO, 2011;FAO, 2018). The pathway has five steps through which a region/country should go until they reach a status of free without vaccination. For countries in stages 1-4 vaccination is one of the major control strategies and thus basic information on serotypes and even topotypes in circulation in that region are important (FAO, 2011).
With the recent start of the PCP-FMD implementation in East Africa in 2012, one of the vital requirements in the initial stages is the need for better understanding and knowledge on FMD epidemiology in a given country/region. Since Uganda and Tanzania are still in the initial stages of the PCP-FMD, where vaccination plays an important role in FMD control, this review aims to provide important information that can be used to develop effective vaccination strategies based on the spatial and temporal distribution of FMD serotypes in both countries between 2003 and 2013.

Data Base Search
The methodology used was adapted from Khan, Regina Kunz, Kleijnen, and Antes (2003). Relevant questions were framed for the review, after which relevant work was identified, the quality of the selected studies was assessed, evidence from the reviewed articles summarised and finally information from the compiled evidence interpreted.
The PubMed Central, Google scholar, Science Direct databases were searched using key words "foot-and-mouth disease" OR "foot and mouth disease" AND "Tanzania" and "foot-and-mouth disease" OR "foot and mouth disease" AND "Uganda". The search included theses, books, conference proceedings as well as project reports. A total of 59 items were found relevant to the topic, the years (2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015) and the defined geographical study area. Abstracts were exported to Mendeley and were later viewed to determine if they met the selection criteria on whether to be included in the study or not. Additionally, Government reports and other relevant documents were retrieved in both hard and soft copy.

Selection Criteria
Studies in which analysed samples were collected before 2003 and after 2015 were not included in this paper. This was done in order to have a more comprehensive picture of what had transpired in eleven years before the PCP-FMD was implemented in East Africa and shortly after its implementation. The review only considered studies where FMD diagnosis was either performed using cell culture, loop-mediated isothermal amplification (LAMP), antigen-enzyme Linked Immuno-sorbent Assays (Ag-ELISA) or polymerase Chain reaction (PCR). This is because the above mentioned laboratory methods are able to detect for the presence of FMDV in an appropriate sample and are serotype specific (OIE, 2012). The criteria included articles on buffalo but excluded articles on other wild animals. It also included articles on other livestock species including cattle. In the end, a total of 39 articles were selected for consideration.

Data Retrieval
Articles whose abstracts had been included for the study were then retrieved in full text format and reviewed by the first author as to whether they met the inclusion criteria. All articles reviewed in the study were in the English language. Qualitative data were compiled on country, year, serotypes, topotype, accession number, location from which samples were collected and referenced (Table 1).   (Figure 2). This review suggests that serotype O was accountable for most of the outbreaks in Uganda stretching from south western Uganda to Kaabong in far eastern part of Uganda (Figures 1 and 3). While, in Tanzania (Figure 3). We additionally observed that the Ugandan and Tanzanian FMD serotype A all belonged to genotype 1 (G-1) within the AFRICA topotype, to which the vaccine strain A/K5/1980 that was being used belonged.

Serotype SAT 1
The SAT 1 viruses in Uganda was detected in 2011 and 2015 in cattle from Kasese and Kirihura, districts respectively. The serotype was also detected in buffalo in 2007 from Lake Mburo National Park in Mbarara district. From these findings, the review shows that in Uganda, the SAT 1 viruses were found only in the south western region of the country and there was no detection of SAT 1 viruses in the central, northern and eastern regions. In Tanzania, SAT 1 viruses were detected every year from 2003 up to 2014. The SAT 1 viruses were distributed only along the eastern part of Tanzania and were found in the regions of Mara, Manyara, Morogoro and Dar es Salaam. We further observed in this review that all the Tanzanian SAT 1s belonged to the topotypeI also called North-Western Zimbabwe (NWZ) whereas the SAT 1s of Uganda belonged to topotype IV also known as East Africa-1.

Serotype SAT 2
Serotype SAT 2 was detected in samples from

Spatial Distribution
From this review, it was observed that between 2003 and 2015, the western and central regions of Uganda registered more outbreaks than the northern and eastern regions. The serotype diversity was highest in western Uganda with five serotypes (O, A, SAT 1, SAT 2 and SAT 3) documented. The central region followed with three serotypes namely, O, A, and SAT 2 and, the eastern region and northern regions had only serotype O in circulation. The outbreaks in the northern region were observed in the districts around the Lake Kyoga basin. The west Nile region had no outbreaks recorded for the entire study period (Figure 3).   Vol. 12, No. 6; argued that wide antigenic differences between the vaccines and circulating strains may result in a failure of the vaccine to generate protective antibodies. Thus the need for constant vaccine evaluation studies.

Serotype A
This paper demonstrated the rare detection of serotype A in both countries for the period covered by the review. Namatovu et al. (2015) highlighted the infrequent detection of this serotype in Uganda. Before its detection in 2013, the serotype has been last documented in 2002 . While in Tanzania, serotype A was regularly detected in the northern parts of Tanzania from 1954 up to 1971, but disappeared for over 30 years until its detection in 2008 in Iringa (Kasanga et al., 2012). The reason for the rare appearance of serotype A is still not well understood but its appearance and disappearance has been highlighted in the Asian continent as well (Kitching, 2005). Casey-Bryars et al. (2018) observed that serotypes were spreading over landscapes by waves and the distribution patterns were not random. The authors argued that certain serotypes were dominant during outbreaks for a specific time, after which they did not appear immediately (Casey-Bryars et al., 2018). This could be the same case for serotype A virus outbreak patterns in both Uganda and Tanzania. Never the less, more research is required in understanding the serotype A patterns in both countries in order to inform better control strategies. The spatial distribution of serotype A in Uganda demonstrates its confinement to the western and central parts of Uganda, suggesting certain epidemiological factors may be playing a role. The similarity in serotype A, lineages between the two countries may be indicative of FMDV serotype A spread between the two countries given the proximity between the south western region of Uganda and the northern regions of Tanzania.
Despite the few years in which serotype A was detected in Tanzania, its distribution in the country was roughly even throughout. This could mean that the serotype easily spread across the different regions in the country probably due to factors such as uncontrolled animal movement, poor uptake of vaccination programs, and lack of biosecurity measures in place among others Previous studies have shown that older Ugandan serotype A viruses isolated in the 60s belonged to genotype G-VII of AFRICA topotype to which the vaccine strain K35/1980 belongs . Although, the recent viruses belong to same topotype as the vaccine strain, Namatovu et al. (2015) highlighted antigenic differences between them that could affect the efficacy of the vaccine. Additionally, previous research has also shown that serotype A is a highly diverse serotype that is constantly evolving. This constant evolution is able give rise to new variants that may not be so closely antigenically related (Kitching, 2005;Kasanga et al., 2015;Sallu et al., 2014). Consequently these variants may complicate the control of FMD by vaccination.

Serotype SAT 1
The observed trend of the high detection of SAT 1 in Tanzania compared to that in Uganda, may highlight differences in surveillance and reporting systems, SAT 1 incursions or/and submission of samples for diagnosis. One of the challenges in comprehending the epidemiology and risk of FMD in endemic settings, is the low submission of samples both to the regional and world reference laboratories (Namatovu et al., 2013). Other reasons for the observed disparity in the frequency of reported SAT 1 occurrence in Uganda and Tanzania may be due to underreporting or limited appropriate diagnostic tools in Uganda (Dhikusooka et al., 2016;Kerfua et al., 2018).
Previously conducted studies have also shown that SAT 1 Ugandan isolates were significantly different from other SAT 1s from Tanzania and Kenya (Dhikusooka et al., 2016), suggesting that the SAT 1 virus strains from Uganda strains are most likely geographically confined to Uganda (Sangula et al., 2010). This could explain the observed differences in topotypes presented. All the Tanzanian SAT 1s belonged to the toptype I (NWZ) (Kasanga et al., 2015), whereas those of Uganda belonged to topotype IV (East Africa-1). According to Sangula et al. (2010), SAT 1 viruses in East Africa have evolved from two independent lineages from South Africa. Ugandan viruses were found to be related to West African and Sudanese viruses while Tanzanian and Kenyan SAT 1 viruses were found to be closer to each other (Kasanga et al., 2015). These observations could aid in explaining the spatial distribution of Tanzanian SAT 1s which were found in the eastern part of the country that is close to Kenya (Figure 3).
The TAN/155/71 vaccine strain that belongs to the toptype I (NWZ) has been consistently used in controlling FMD spread in the region. Since the vaccine strain belongs to a different topotype from the Ugandan strain, there may be no cross protection between these strains thus a lot of implications for FMD control. This indicates the need for more studies in Uganda on vaccine evaluation in order to ascertain the protection elicited by the TAN/155/71 vaccine strain. Dhikusooka et al. (2016) reported silent manifestation of clinical signs in the cattle from which the SAT 1 viruses were isolated. This kind of scenario presents challenges to the clinical based reporting systems (Namatovu et al., 2013), where most reported FMD cases in Uganda and Tanzania were shown to rely on clinical manifestations of the disease instead of laboratory diagnosis (Muleme et al., 2012;Kerfua et al., 2018).

Serotype SAT 2
The occasional detection of SAT 2 in Uganda may be attributed to limited outbreak surveillance and may not reflect the true representation of the situation in the country. It may also be that SAT 2 viruses were rare in Uganda during the reviewed period. However, in Tanzania, the observed widespread and frequent occurrence of SAT 2 could be attributed to regular sample diagnosis or/and the presence of high numbers of wildlife reservoirs. In this review, it was observed that the distribution of SAT 2 in Tanzania was roughly correlated to the distribution of wild life reservoirs (Figure 2). In Uganda, though, the distribution of SAT 2 viruses was mostly confined to western region of Uganda. However, the eastern part of the country that has three major national parks harbouring buffalos (Mt Elgon, Kidepo and Pian Upe), did not have SAT 2s detected. The association of SAT 2 outbreaks with wildlife-livestock interphases were observed in studies carried out in the southern and eastern Africa, and the African Buffalo has been connected with the SAT carrier status, and implicated in virus exchange with cattle (Woodsbury, 1995;Vosloo et al., 2005;Kasanga et al., 2012;Ayebazibwe et al., 2010a;Jori et al., 2016). However, other studies have shown that it's not always the case that buffalos are responsible for SAT 2 incursions (Brito et al., 2016). Nevertheless, reasons for this distribution pattern of SAT 2 in Tanzania and Uganda still remain poorly understood (Kasanga et al., 2012;Sallu et al., 2014).
According to Balinda et al. (2010b), the 2010 SAT 2 Ugandan viruses were different from viruses that were isolated in years of 1975, 1995, 1998 and 2002. We suggest that the viruses could have evolved along different lines and raises questions on whether they could have emerged from different buffalo populations. However, the limited number of SAT 2 FMD virus sequences from Uganda limits the comprehensive phylogenetic analysis of FMDVs SAT 2 from both cattle and buffalo. Furthermore, disparity in topotypes of the SAT 2 FMDV from Uganda and Tanzania either implies that SAT 2 viruses are not spreading between country or that there may be under reporting of SAT 2 incursions in Uganda. According to Balinda et al. (2010b) and Namatovu et al. (2013), Uganda has reported two virus lineages of SAT 2 FMDVs. However, the most recent isolates have been observed to belong to a similar lineage as the vaccine strain (K52/84). Namatovu et al. (2015) further demonstrated that vaccine strain had significant genetic diversity with the most recent circulating SAT 2 virus, underscoring a need for vaccine matching studies to be carried out in both Uganda and Tanzania.

Serotype SAT 3
Previous SAT 3 isolates from Uganda were only from buffalo (Kalema-Zikusoka et al., 2005;Ayebazibwe et al., 2010a). The SAT 3 isolate was retrieved from a calf that had been grazing in the Queen Elizabeth National Park (QENP). The findings from the study by Dhikusooka et al. (2015), presents a new challenge for continued surveillance in the livestock wildlife interphase areas both in Uganda and Tanzania.
The SAT 3 isolate had a 19% a nucleotide difference with buffalo FMD virus UGA/2/97 that was grouped in topotype IV. However, Dhikusooka et al. (2015), recommended that the isolated SAT 3 virus from Uganda should be categorised within a single topotype V. Currently, there is no vaccine strain that incorporates the SAT 3 serotype. This calls for further vaccine development and other strategic controls.

Conclusion
This review presents similarities and differences in the distribution of FMD serotypes in Uganda and Tanzania. It also highlights regions with the highest serotype diversity and those with the least diversity. In addition, this review has shown that there are differences between the circulating virus serotypes/topotypes and the vaccine strains currently in use in both countries. Although, underreporting of outbreaks and limited resources for diagnosis hinder the true picture of the serotype distribution in endemic countries, these findings still provide information important for strategizing control of FMD. For example, efforts for control can be specifically targeted to an area based on the prevalent serotypes and the serotype diversity.
This review recommends routine characterisation of circulating viruses, further research on FMD risk factors for certain serotypes, vaccine evaluation studies and implementation of harmonised regional FMD control programmes.