Alkhurma hemorrhagic fever virus (AHFV) is a variant of Kyasanur Forest disease virus (KFDV), and like KFDV, is a member of the mammalian tick-borne encephalitis group [family Flaviviridae, genus Flavivirus]. AHFV and KFDV cause similar disease syndromes in humans marked by fever, severe headache, myalgia and arthralgia, followed in a subset of cases by a hemorrhagic syndrome or diffuse neurological sequelae [1]–[4]. Recent epidemiological studies of AHFV outbreaks describe case fatality ratios close to those associated with KFDV infection (2–10%), suggesting that previous estimates of 25% might have been exaggerated due to unrecognized asymptomatic or mild cases [5]. Beyond supportive care, there is no specific treatment for either AHFV or KFDV infections.

AHFV and KFDV have distinct geographic ranges in Saudi Arabia and India, respectively (Figure 1). AHFV was isolated in 1994 in Makkah, Saudi Arabia, from the blood of a fatally infected butcher [6]. Since then, AHFV cases have been confirmed in Jeddah, Jizan, and Najran, and most recently, outside of Saudi Arabia near the Egypt-Sudan border in 2010 (Figure 1) [7]. Some early reports used an alternate spelling of Alkhumra [1], [2]. Human AHFV infections have been associated with tick bites, and AHFV has been isolated from an Ornithodoros tick in Jeddah [8], and Ornithodoros savignyi and Hyalomma dromedarii ticks in Najran [9]. However, another common risk factor for AHFV infection appears to be close contact with domestic animals, particularly sheep and camels [5], although no disease has been reported in livestock or other animals. The host range of KFDV is quite different from AHFV; KFDV was first identified as the cause of nonhuman primate die-offs (Presbytis entellus and Macaca radiata) as well as a concurrent outbreak of fatal human disease in the Kyasanur forest region of India in 1957. All known confirmed laboratory cases of KFD in people have been associated with contact with ticks (primarily Haemaphysalis spp.), activities within the Shimoga forest region, or through laboratory infections.

Despite apparent differences in their hosts and geographic ranges, AHFV and KFDV share high sequence identity [10], [11]. Their positive-sense RNA genomes are approximately 11 kb in length and encode a single 3416 amino acid polyprotein that is post-translationally cleaved into a total of 3 structural (C, M and E) and 7 nonstructural (NS1, NS2a,NS2b, NS3, NS4a, NS4b, and NS5) proteins. Given the notable genetic similarity and the later emergence of AHFV in Saudi Arabia, it has been speculated that AHFV arose following an introduction of KFDV from India. Previous phylogenetic analyses of AHFV [12] and KFDV [10] have relied on partial gene sequences, including regions of the structural envelope (E), the RNA-dependent RNA polymerase (NS5), and the viral protease/helicase (NS3). These studies indicated a recent divergence of AHFV and KFDV, occurring sometime between 1828 and 1942, presumably spawned by the introduction of KFDV into Saudi Arabia. However, given that the vector ecology, mammalian hosts and ecological niche of AHFV differ markedly from KFDV, it is possible that a longer period of divergent evolution between AHFV and KFDV could explain these significant biological differences.

A more complete understanding of the evolutionary history of these viruses would provide insight into the circumstances surrounding their emergence. This is particularly important, as AHFV and KFDV are serious public health threats, and the recent identification of AHFV in Egypt demonstrates the potential for either viral spread and range expansion, or a larger range than previously thought. To gain insight into the relationship between AHFV and KFDV, we sequenced the full-length genomes of 16 AHFV and 3 KFDV isolates and analyzed those with two existing sequences available in GenBank. Our analyses revealed a higher overall diversity amongst AHFV strains than previously thought, particularly within the tick population. Surprisingly, these analyses indicated a much earlier divergence of AHFV from KFDV approximately 700 years ago, suggesting AHFV and KFDV might have broader geographic ranges, and raises the possibility of closely related but undiscovered virus variants existing in the regions between Saudi Arabia and India.

Here we describe the first AHFV/KFDV phylogeny produced using full-length genomes. To sequence the available 16 AHFV and 3 KFDV isolates (Table 1), primer pairs were designed to amplify 3 overlapping segments of approximately 4 kb each. These RT-PCR products span the entire 10.7 kb genome excluding the terminal 26 nucleotides (nt) of the non-coding 5′ untranslated region (UTR). The terminal 33 nt of the 5′ UTR and the terminal 39 nt of the 3′ UTR of KFDV-P9605 and AHFV-809006 were identical using 3′ RACE on the vRNA and vcRNA. The full genome lengths of AHFV and KFDV differed by only 1 nt; a conserved A in position 10415 of the noncoding 3′ UTR of AHFV was absent in KFDV, resulting in genome sizes of 10775 and 10774, respectively.

Analysis of full-length AHFV and KFDV genomes (Table 3) confirmed a high level of sequence identity between the viruses, with an overall diversity of 8.4% (nt) and 3.0% (aa). The region showing the most nucleotide diversity (11.6%) within AHFV isolates was the 23 aa (69nt) C-terminus of the NS4A protein, referred to as the 2K peptide [20]. However, the 2K amino acid sequence was completely conserved amongst AHFV isolates. The ratios of nonsynonymous to synonymous substitutions were determined for each AHFV were determined and showed no strong indication of positive selection.

The mean rate of molecular evolution was estimated to be 9.2×10⁻⁵ sub/site/year (95% highest posterior density [HPD]: 1.6×10⁻⁵ sub/site/year [low] to 1.9×10⁻⁴ sub/site/year [high]). The phylogeny clearly showed AHFV and KFDV as 2 separate lineages (Figure 2). The TMRCA for all AHFV isolates was recent, 84 years before 2009 (∼1925) [HPD range: 29-165 years]. All KFDV isolates shared a recent TMRCA as well, 75 years before 2009 (∼1934) [HPD range: 56–104]. Despite their genetic similarity and individual recent TMRCAs, our analyses revealed that AHFV and KFDV diverged approximately 687 years before 2009 [HPD range: 121-1487 years].

This analysis encompassed 18 AHFVs isolated from either southern or western Saudi Arabia. Eight AHFVs were derived from ticks, including one published previously [8]. The remaining 10 were derived from human cases, including the prototypic AHFV 1176, also previously published [11]. The resulting phylogeny showed all known AHFV isolates fall into 3 sublineages (Figure 2). Sublineage I consisted entirely of 5 tick pool isolates from the Al Balad Magan camel market in Najran. All were isolated from Ornithodoros savigyni ticks and had a pairwise diversity of 0.5%. Sublineage II included two tick pool isolates taken at the same time from a nearby camel market, also in Najran, in the Al Mishaaliyia district. Despite the close proximity of the camel markets, the genetic variation between tick pool isolates of Sublineage I and II was 1.1%, slightly higher than the overall diversity seen within all human AHFV isolates (0.7%). The tick pool isolates of Sublineage II came from different tick species (one from Hyalomma dromedarii and one from O. savignyi) but differed from one another by only 3 nt. There were no defining amino acid changes found between human and tick isolates. Sublineage II includes isolates from 1994-2009, the full range of available full-length genomes, as well as the small gene fragments from the 2010 Egyptian isolate. Sublineage III encompasses isolates taken from ticks and humans in Makkah, Jeddah and Jizan. In summary, the genetic differences between AHFV isolates did not obviously correlate with geographic, temporal or host origin.

The emergence of AHFV in the mid-1990s and its remarkable genetic similarity to KFDV suggested that AHFV arose from a recent introduction of KFDV from India into Saudi Arabia. Using partial E and NS5 sequences of KFDV, the AHFV/KFDV divergence was estimated to have occurred 68 years ago [10]. Similarly, Charrel et al., found a TMRCA of 182 years ago with partial E, NS3, and NS5 sequences of AHFV [12]. Using only the concatenated E and NS5 sequences of our isolates (as in [10]), we also found a recent time of divergence (∼100 years before 2009). However, analyzing full-length genome sequences, we show 2 distinct lineages, with KFDV and AHFV diverging approximately 687 years ago. Analyzing full-length genomes allowed us to incorporate the full diversity of AHFV isolates, and therefore to provide a more accurate estimate of an older common TMRCA.

The use of full genomes also revealed a slower rate of molecular evolution (9.2×10⁻⁵ substitutions/site/year) than determined previously with partial sequences of KFDV [10]. This may be explained by the slightly higher NS5 and E diversity in KFDV relative to the rest of the genome (Table 3), however, the limited number of KFDV isolates available for this study might not be fully representative of the extant KFDV diversity. In accordance with the evolutionary rate patterns for vector-borne flaviviruses [21], the AHFV/KFDV rate is slower than mosquito-borne flaviviruses, including Dengue virus (DENV) [22] and yellow fever virus (YFV) [23] and more similar to tick-borne flaviviruses, including the closely related tick-borne encephalitis virus (TBEV Siberian subtype, 1.64×10⁻⁴ sub/site/year) [24]. The evolutionary rate is also similar to that of an unrelated tick-borne member of the Bunyaviridae family, Crimean-Congo hemorrhagic fever virus (0.58–1.52×10⁻⁴ sub/site/year)[25].

In our analyses, the sublineages of AHFV isolates did not show clustering by geography or date of isolation. Although these isolates primarily came from two distinct geographic areas, the Jeddah/Makkah region in the west and the Jizan/Najran region in the south, frequent travel of people and livestock between these regions effectively eliminates geographic boundaries. The short sequences available from the 2010 Egyptian isolate placed it within Sublineage II, the most inclusive of AHFV sublineages, possibly suggesting a relatively recent introduction to Egypt from Saudi Arabia. Alternatively, the limited sequence data available from the Egyptian isolate is not representative of extant diversity within East Africa and the full-length sequence could indicate a different evolutionary history. The narrow range of time encompassed by the viral isolates (∼15 years) and the slow evolutionary rate of AHFV may also explain the lack of detectable temporal clustering.

The centuries between the divergence of the ancestral AHFV and KFDV (∼year 1322), and the individual TMRCAs of AHFV (∼1925) and KFDV (∼1933) isolates, coupled with a slow evolutionary rate, indicate the 2 viruses evolved separately and AHFV is not a result of recent KFDV introduction in Saudi Arabia [10]. This separate evolutionary history spanning several hundred years and the vast geographic distance involved raises the possibility of an existing spectrum of thus far unknown tick-borne encephalitic/hemorrhagic viruses in the regions between Saudi Arabia and India. In support of such a viral spectrum in that region, Gould et al. [26] cite the presence of close AHFV/KFDV relatives, Karshi virus (KSIV) and Farm Royal virus (FRV), in Uzbekistan and Afghanistan, respectively. A similar dispersal corridor of TBEV has been described from Eastern Europe westwards [27], although more recent studies did not support such a pattern [28]. Dispersal of the ancestral viruses of AHFV and KFDV may have been accomplished through the movement of animals, including camels presumably carrying ticks, along the Silk Road, which by the 1300s stretched from Europe to China.

The presence of potentially competent tick vectors in these regions [29] also supports this possibility of AHFV/KFDV-like viruses existing between Saudi Arabia and India. Although the viruses are primarily associated with two tick genera, AHFV with Ornithodoros and KFDV with Haemaphysalis, both have been isolated from multiple genera. At least 16 tick species have been shown to transmit KFDV [3]. The experimental finding of transovarial and transtadial transmission of KFDV in O. cruzi [30], coupled with the ubiquitous nature of Ornithodoros spp. in India, Saudi Arabia and the intermediate regions, suggests these ticks may have played an important role in the spread of AHFV/KFDV-like viruses. Although the transmission of KFDV has been primarily associated with H. spinigera, a tick species not documented in Saudi Arabia, these ticks are present between Saudi Arabia and India [29]. Given the diversity of potential vectors and their extensive ranges, it is feasible that AHFV and KFDV, as well as other similar but thus far undiscovered viruses, are circulating more broadly throughout Saudi Arabia, India and beyond.

The genomic diversity of AHFV was most apparent within isolates from the tick population. Ticks sampled on the same day in one city (Najran) harbor a more genetically diverse AHFV population than isolates obtained from human cases over the course of 15 years throughout Saudi Arabia. Similarly, an analysis of the European subtype of tick-borne encephalitis virus (TBEV) isolated from ticks shows relatively high variation in a limited geographic area and temporal period, perhaps due to importation of TBEV on migratory birds [31], [ 32]. The notable genetic heterogeneity of AHFV in Najran may be explained similarly, as camel and other livestock traders often travel large distances to Najran to market their animals. Like all soft ticks, Ornithodoros spp. feed for only short periods, spending most of their time in burrows or nests, allowing areas such as a camel market to become a potential focus of AHFV infection and high virus genomic diversity. Presumably, like KFDV, AHFV can persist for long periods within a tick, and given the nature of RNA viruses, variants may evolve over time within the invertebrate hosts. The lower diversity within human cases may reflect limited transmissibility or reduced pathogenicity of some of these variants in people.

In our analysis, the use of full-length AHFV and KFDV genomes demonstrated a deeper evolutionary history than suggested by previous partial genome analyses. The divergence of AHFV and KFDV almost 700 years ago indicates a long period of divergent evolution, and suggests that a range of as-yet undiscovered tick-borne hemorrhagic/encephalitic viruses could exist between Saudi Arabia and India. The notably high AHFV diversity found within tick populations, coupled with the extensive geographic range of competent tick vectors, raises the possibility of broader AHFV and KFDV geographic ranges, and is supported by the recent discovery of AHFV in Egypt. As AHFV and KFDV are both associated with significant human morbidity and mortality, the potential spread of these viruses should be of serious concern and warrants further study of these significant pathogens.