4. DISCUSSION

4.1 Phylogenetic classification of Central European DOBV-Aa and DOBV-Af strains

↓69

Our sequence comparisons and phylogenetic analyses showed that the Central European DOBV strains cluster within two separate clades according to their natural host (DOBV-Aa vs. DOBV-Af). In the S segment these strains are related to the A. agrarius-derived DOBV strains from Russia (Kurkino-Aa) but quite less to the A. flavicollis-derived virus strains (Slo/Af, Esl/400Af, AP/Af) and the Saaremaa-Aa strains.

In the M segment, the analysed Central European DOBV strains (Esl/862Aa, SK/Aa) were found to be more related to the strain Saa/160V than to the A. flavicollis DOBV strains. This demonstrates a strong phylogenetic relationship between the A. agrarius DOBV strains (DOBV-Aa) on the one hand and the A. flavicollis DOBV strains (DOBV-Af) on the other. However, both DOBV lineages could be clearly distinguished from the most closely related virus species, i.e., HTNV or SEOV.

4.2 Genetic reassortment of genome segments in Saa/160V

Our phylogenetic analyses of DOBV complete S and M segment sequences showed that Saa/160V strain possess a DOBV-Af like S segment, whereas its M segment groups together with the Central European strain Esl/862Aa as well as with the new DOBV-Aa isolate SK/Aa. These results gave evidence that Saa/160V was involved in one or more reassortment processes. Identification of the partial L segment sequence of the new isolate SK/Aa allowed additional insights into the interesting topic of reassortment during DOBV evolution. The ML analysis showed that not only in S but also in L segment, the Saa/160V strain (although found in A. agrarius rodents) is more related to DOBV-Af strains and only its M segment resembles that of DOBV-Aa strains (Figure 20). This suggests that the M segment encoding the viral glycoproteins is crucial for the hantavirus host specificity. Although at the current stage of knowledge other scenarios cannot be ruled out, it might be concluded that Saa/160V was originally an A. flavicollis-associated virus which has switched to A. agrarius during its evolution. To overcome the virus-host specificity barrier and to establish a persistent infection in the new host, it could have acquired an A. agrarius-specific M segment from the DOBV-Aa lineage. Further studies on biological consequences of reassortment in hantaviruses are needed. The availability of “non-reassorted” DOBV-Af (Slo/Af) and DOBV-Aa (SK/Aa) isolates as well as the apparently reassorted Saa/160V will now allow to study the importance of this process in vitro.

↓70

Figure 20: Schematic illustration of the proposed scenario of genetic reassortment during DOBV evolution. S and L segments of Saaremaa virus are phylogenetically related to their counterparts in DOBV-Af, whereas the M segment (determining the antigenicity of the virus envelope) resembles that of DOBV-Aa.

In its cross-neutralisation behaviour, Saa/160V has been found to differ from Slo/Af (Brus Sjolander et al., 2002). This can be explained by the fact that there are 72 aa differences betweenthe glycoproteins of the two virus isolates, Saa/160V and Slo/Af. Only 38 of them are potentially host-specific (see below). It seems to be reasonable to conclude that Saa/160V is not a good virus prototype of the Central European DOBV-Aa strains because of its DOBV-Af-like S and L segments and the large number of aa exchanges in the M-segment encoded glycoproteins which are not typical for the Central European DOBV-Aa strains.

There is one other example of natural genetic reassortment in hantavirus evolution; reassortment of genomic RNA segments has been also found between SNV strains (Henderson et al., 1995; Li et al., 1995). Moreover, genetic reassortants involving S and M segments were detected after mixed infections in tissue culture by using two distinct strains of SNV. One virus reassortant was observed also after mixed infections between SNV and genetically more distant Black Creek Canal virus (Rodriguez et al., 1998). This together with our own results indicates that genetic reassortment is not an uncommon process in hantavirus evolution.

↓71

For viruses with segmented genomes, reassortment is an additional way of generating variation and antigenic alteration. A prominent example is the influenza virus. The most abrupt changes in antigenic specificity occur through hemagglutinin and neuraminidase genes reassortment. The mixing of segments from avian and mammalian influenza viruses has in certain cases yielded a very high fitness, as exemplified by the Asian (1957) and Hong Kong (1968) pandemic flu viruses (see Hungnes et al., 2000; Hilleman, 2002). It has to be evaluated whether the genetic reassortment in hantaviruses might have similar consequences concerning altered host specificity or pathogenicity towards humans.

4.3 Genetic recombination between DOBV-Af and DOBV-Aa lineages

Natural co-infection of the same host animal and the same host cell with different virus strains is the precondition also for the occurrence of genetically recombined viruses during their evolution. It was recently shown that recombination between different members of the hantavirus species Tula had occurred in their evolution; this was the first example of natural homologous recombination in negative-stranded RNA viruses (Sibold et al., 1999a).

Now, we have found two parts in the M segment of strain Esl/862Aa and one in the S segment of Esl/8xxAa that show signs for an evolutionary history differing from the rest of the sequence. While two of the areas (S 400-600 nt and M 1972-2211 nt) show only moderate puzzle support values between 62% and 64% in the phylogenetic trees, the area from 810 – 1059 nt of M segment with a support of 81% gives some evidence for an event changing the evolutionary history in respect to the flanking sequences.

↓72

Two events could explain the creation of this significant nucleotide sequence exchange at M segment position 810 - 1059. First, directed selection could have led to such differences in parts of the sequence, but an area of 250 nt should be most likely too long to be generated by directed selection. The other possible scenario consists in the occurrence of genetic recombination. As usually in case of recombination, one would expect one bootscanning curve dropping from high to very low values, while another curve comes up, indicating an exchange of genetic material leading to a recombinant query sequence from parental sequences related to the ones belonging to the changing curves. The pattern found here (Figure 10), shows that in the region between nt positions 810 - 1059 the alignment of the query sequence (Esl/862Aa) with Saaremaa-Aa is interrupted and no curve comes up reaching the threshold value. Two recombination scenarios are most likely: First, recombination between a Saaremaa-like ancestor of Esl/862Aa with a member of a (so far) unknown DOBV lineage older than the sequences we used, creating the recombinant virus Esl/862Aa. Or, secondly, an Esl/862Aa-like ancestor of Saaremaa has recombined with an DOBV-Af like lineage forming a recombinant Saaremaa virus.

However, the phylogenetic analysis of this recombination event might be influenced by the short extent (249 bp) of the region. Hence, this issue needs to be further evaluated, taking into account biological properties of M segment gene products which are currently not very well characterised.

The availability of more sequenced virus strains would be helpful to further elucidate the scenario of recombination processes in the evolution and intraspecies diversity of Dobrava virus. Nevertheless, the recombination event in DOBV evolution, after similar findings for TULV in its natural evolution (Sibold et al., 1999a) and under in vitro conditions (Plyusnin et al., 2002), offers another example of natural homologous recombination in the evolution of hantaviruses. In addition, there are indications for recombination events within the PUUV species, too (Sironen et al., 2001).

4.4 Role of genetic differences for host adaptation

↓73

The sympatric occurrence of DOBV strains in two different host species, A. agrarius and A. flavicollis, opens the chance of a comparative genetic analysis which could contribute to an understanding of adaptation to different hosts leading to genetic differences between the virus lineages. This might allow conclusions also for suggested pathogenicity differences towards humans.

Altogether, three aa differences between DOBV-Aa and DOBV-Af strains in N protein and 38 aa exchanges in G1/G2 have been identified. This list of candidate host-specific determinants can be probably reduced when more complete sequences will be elucidated and included in the analysis.

However, there are certain limits in possible conclusions from aa comparisons even in the case that an enhanced number of DOBV-Aa and DOBV-Af sequences will be available in the future: (i) Secondary (”suppressor”) aa exchanges could restore the host-dependent functionality of a protein and could mask true aa differences, (ii) the role of the nucleotide sequence independent of their coding properties for host-adaptation of a virus (nucleic acid itself as subject of genetic selection) is unknown yet. Probably important are also the sequences of NCR regions because of their role in replication and transcription regulation (Jonsson and Schmaljohn, 2001) and binding of N protein (Gott et al., 1993). Lundkvist et al. (1997b) have shown that single mutations in the NCR of the PUUV (strain Kazan) S segment could be associated with adaptation of this virus to bank vole versus cell culture. However, in our case it is difficult to define the importance of particular nucleotide exchanges in the NCRs because of the low degree of homology between the DOBV strains in these genomic regions (especially in the M segment 3’NCR).

↓74

An ultimate proof of the role of distinct amino acid exchanges in the DOBV proteins for host adaptation (A. flavicollis versus A. agrarius) would only be possible by using reverse genetics systems to generate infectious clones. The development of this field of research is still in its first steps for bunyaviruses in general and hantaviruses in particular (Jonsson and Schmaljohn, 2001; Flick et al., 2003). Moreover, one would need sophisticated Apodemus animal models that allow cross-infection experiments in the laboratory. Moreover, a “read-out” systems to compare the antiviral responses of endothelial cells, as recently used for HTNV and TULV (Kraus et al., 2004), trigered by various reverse genetics-generated DOBV clones would allow to study the role of these amino acid exchanges for the pathogenicity differences towards humans.

There are some indications that viruses of the DOBV-Aa and the DOBV-Af lineages could not only differ in their host reservoir but also in their pathogenicity towards humans (Sibold et al., 2001). So, the detailed comparison of these virus lineages could also improve the knowledge on pathogenicity-specifying genetic determinants. Given the present stage of knowledge we may be correct to state that the S segment should be less relevant for the different pathogenicity; despite possessing an S segment related to the highly virulent DOBV-Af, Saa/160V is presumed to be less pathogenic as it seems to be the case for the Central and East European DOBV-Aa strains (Plyusnin and Morzunov, 2001). It seems reasonable that the glycoproteins are the main pathogenicity determinants since they are clearly different for DOBV-Aa and DOBV-Af strains. It is known from previous work of other authors that the exchange of as less as one amino acid in the glycoprotein of HTNV can significantly alter the virus pathogenicity (Ebihara et al., 2000; Isegawa et al., 1994). Practically nothing is known about the role of the viral RNA polymerase (product of the L segment) in host specificity and human pathogenicity.

However, at the current stage of knowledge it cannot be excluded that not only genetic differences between virus lineages in A. agrarius and A. flavicollis contribute to the different clinical outcome of DOBV-caused HFRS in Central/North-East/East Europe and South-East Europe, respectively. One of other possibilities to modulate the clinical outcome are certain genetic differences between human populations in Europe (see below).

4.5 Evolution of DOBV and the problem of hantavirus species definition

↓75

Our investigations gave new insights into the phylogeny of the DOBV virus group. It is surprising that related virus lineages (DOBV-Af and DOBV-Aa) are hosted by different rodent species (A. flavicollis and A. agrarius, respectively). Moreover, it has been known for a long time that A. agrarius in Asia (probably forming other subspecies different from the European subspecies) harbours a distinct virus species, HTNV.

To explain why two distantly related virus species, HTNV and DOBV, have been found in the same host species, A. agrarius, while two closely related DOBV lineages, DOBV-Af and DOBV-Aa, are hosted by two different rodent species, A. flavicollis and A. agrarius, ahost switch has been suggested before (Plyusnin and Morzunov, 2001). Wang et al. (2000) and Nemirov et al. (2002) recently forwarded the idea that, probably, HTNV is the original hantavirus in the (Asian) A. agrarius and DOBV entered later this host species (rather in the European region) by host switch from A. flavicollis to A. agrarius

In contrast to this, our results show the higher diversity of DOBV-Aa virus strains compared to the lower degree of divergency of the (so far determined) DOBV-Af strains, which might suggest an older evolutionary history of DOBV-Aa as compared to DOBV-Af. This would then lead to a scenario that DOBV-Af could have developed after a host switch from A. agrarius to A. flavicollis. The present amount of available data does not allow making a final decision between these possible scenarios. It should be mentioned that first experimental infections have shown that Saaremaa virus is able to infect both A. agrarius and A. flavicollis animals demonstrating the close relationship of the two virus-host systems (Klingstrom et al., 2002).

↓76

In addition, the question of host switch events playing a role in evolution of Murinae-associated hantaviruses wasraised after therecent isolation of the HTNV-related hantavirus strain NC167 from Niviventer confucianus(Wang et al., 2000). Examples of potential host switch events have been also reported for hantaviruses associated with Arvicolinae(Vapalahti et al., 1999) and Sigmodontinae hosts (Nichol, 1999; Sanchez et al., 2001).

Since DOBV-Af and DOBV-Aa are stably associated with two different host species it has been proposed to consider them as different virus species and to take at that time the only available virus isolate of the DOBV-Aa group, Saa/160V, as the prototype of the new virus species, called Saaremaa (SAAV) (Plyusnin and Morzunov, 2001). According to the species demarcation criteria in the genus Hantavirus, as defined by the International Committee on Taxonomy of Viruses (Elliott et al., 1999), hantavirus species have to meet the following essentials:

(i) Species are found in a unique ecological niche, i.e., in a different primary rodent reservoir species or subspecies. This is the case for DOBV-Af versus DOBV-Aa. Our results presented here demonstrate that the investigated DOBV strains from A. flavicollis and A. agrarius in Slovakia are clearly different despite their co-existence at the same geographical place. On the other hand, the reassortment origin of Saaremaa virus indicates the possibility of genetic exchanges between members of the DOBV-Af and DOBV-Aa lineages. Obviously, the precondition of such events was the co-infection of the same host animal.

↓77

(ii) Species exhibit at least a 7 % difference in amino acid identity on comparison of the complete GPC and N protein sequences. We found aa differences in the deduced N protein sequences within the DOBV-Af and DOBV-Aa groups between 1.9 % (SK/Aa vs. AP/Af) and 3.3 % (Kur/44Aa vs. Esl/400Af and Slo/Af). Saa/160V shows a N protein sequence difference of 2.6 % when compared to the two DOBV-Af strains investigated. Since the Saa/160V carries a DOBV-Af-like S segment, it is not surprising that it exhibits lower similarities to the other DOBV-Aa strains than to the DOBV-Af strains. In the putative GPC the aa differences between the DOBV-Af and DOBV-Aa groups have been found in the range from 5.5% (Saa/160V vs. Slo/Af) to 6.6 % (SK/Aa vs. Slo/Af).

(iii) Species show at least a fourfold difference in two-way cross-neutralisation tests. By analysis of human serum samples from the Balkans (South-East Europe) and Estonia (North-East Europe) (Brus Sjolander et al., 2002) have shown that a majority of the sera exhibited this typical titer differences with preference for the local virus (DOBV-Slo/Af and Saa/160V, respectively). The authors conclude that DOBV-Af and Saaremaa virus define unique hantavirus serotypes. Obviously, the typical amino acid differences in the glycoproteins of the viruses (see above) are sufficient to result in the distinct serological behaviour.

(iv) Species do not naturally form reassortants with other species. However, this was highly probably the case for the Saa/160V virus (see above). The occurrence of such reassortment events between DOBV-Af and DOBV-Aa should be taken as one counter-argument against the idea that DOBV-Aa strains form a separate species. Moreover, even if one would like to consider the DOBV-Aa lineage as a unique virus species, the Saaremaa virus seems not to be the best representative of this suggested species because of its hybrid genome.

↓78

These observations highlight the importance of investigating carefully the relationship between proposed new viral species and/or lineages in determining which viral isolates represent true prototypes. In some cases, it will be possible to understand the specific and subspecific relationships among isolates only after examining a variety of members of a clade, including, whenever possible, those that occur sympatrically but in different host species.

Based on our phylogenetic analysis, we like to suggest a distinction between at least three DOBV lineages. DOBV-Aa consists of strains originating from A. agrarius and is currently represented by strains from Central Europe and Russia including the Slovakia isolate. A. flavicollis-associated strains from Slovenia, Greece and Slovakia (Esl/400Af) represent the DOBV-Af lineage. The Saa/160V strain with its hybrid genome does not clearly belong to any of this two lineages and could be classified as an intermediate lineage (DOBV-Saa) (Figure 20). We believe that this classification should be kept within the DOBV species and no postulate of additional species is necessary. Particularly, grouping of in S and L segments non-monophyletic DOBV-Aa and DOBV-Saa lineages into one species and classifying DOBV-Af lineage as a second, distinct species as recently forwarded by Plyusnin et al. (2002, 2003), might not resemble the evolutionary history of these viruses.

4.6 Presence of distinct DOBV-Aa strains on single geographical locus

Our sequence analysis of six DOBV-Aa strains, detected in one trapping locality Rozhanovce, East Slovakia, revealed unusual results. Although detected in rodents captured in single locality, the partial S and M sequences, expected to be nearly identical, could be divided into two distinct groups showing the sequence diversity of up to 5.7%. In addition, phylogenetic analysis showed that the Rozhanovce-derived sequences do not cluster together in one monophyletic group.

↓79

Our hypothesis was that Rozhanovce might be situated on the border of the territories of two distinct A. agrarius subpopulations harbouring their own specific DOBV-Aa strains. In the molecular phylogenetic analysis of Apodemus mice based on 12S rRNA and D-loop (mitochondrial control region) mitochondrial markers, the obtained A. agrarius sequences appeared to be practically identical. However, an important question still remained to be answered. Do these results indicate that all individuals are members of an identical population or the selected markers are not suitable for a fine differentiation on population level?

Michaux et al. (2002) used 12S rRNA gene for studying the phylogeny of the genus Apodemus, but only on subgenus and species level. The suitability of 12S rRNA marker for these purposes was confirmed by our results showing that the sequences of A. flavicollis and A. agrarius were clearly distinct.

The mitochondrial control region was shown to be attractive to evolutionary biologists for fine scale comparative studies because it is believed to be one of the fastest evolving segments in the animal mitochondrial genome. For instance, the D-loop sequence was used to differentiate the subspecies A. agrarius coreaeand A. agrarius chejuensisThe divergence in this marker within the subspecies was determined to be 2.98% and 1.86%, respectively (Koh et al., 2000). Nemirov et al. (2002) using D-loop, showed the distinctness of A. agrarius from Korea harbouring HTNV and A. agrarius from Europe. But also within the European A. agrarius, the sequences from Slovenia, Poland and Estonia were clearly distinguishable. On the other hand, Dekonenko et al. (2003) also used D-loop marker in the sequence analysis of the Clethrionomys voles from Finland, Sweden, and Western Siberia and, in line with our study, found extremely high degree of similarity within the species. In addition, similar results were obtained also with the mitochondrial cytochrome b gene, nuclear breast cancer susceptibility gene (BRCAI), and sex-determining region of chromosome Y (Sry-HMG) (Dekonenko et al., 2003).

↓80

From other phylogenetic markers, mitochondrial gene for cytochrome b (Martin et al., 2000; Serizawa et al., 2000; Barome et al., 1998), and nuclear genes IRBP (interphotoreceptor retinoid binding protein) (Michaux et al., 2002; Serizawa et al., 2000) and tspy (testis specific protein, Y chromosome-encoded) (Schubert et al., 2000) have been used in phylogenetic analyses of Murinae rodents, but resolved the relationships only on the species, subgenus or genus level.

Altogether, our study of A. agrarius in Slovakia did not confirm the presence of two distinct subpopulations, distinguishable at least in D-loop and 12S rRNA phylogenetic markers. The occurrence of two distinct strains of DOBV in a single locality might therefore have to be explained by another scenario than the presence of two distinct subpopulations of rodent hosts. Kratochvil (1962) concluded that the current distribution of A. agrarius in Central Europe is a result of multiple invasions and regressions and that this process is still ongoing. Therefore, we like to speculate that together with its host, DOBV strains of different origin could be rather recently “introduced” at least twice in the Eastern Slovakia region. Those distinct virus strains could now circulate in the current local A. agrariuspopulation what would explain the presence of two distinct DOBV-Aa strains in phylogenetically uniform local rodent population.

It remained to be answered whether the data from Rozhanovce represent a rather exceptional or common picture of hantavirus biodiversity in natural foci and what consequences could that have for hantavirus epidemiology and ecology. It might be interesting to note that the prevalence of DOBV in rodents captured in Rozhanovce in 2001 was unusually high; 11 out of 42 (26.2%) mice were found sero- and RT PCR-positive.

4.7 A new DOBV strain isolated from A. agrarius

↓81

DOBV-Aa is an important HFRS pathogen in Central Europe; by serotyping and direct molecular proof dozens of patients with renal failure have been found to be infected with strains of this virus lineage (Klempa et al., 2004; Sibold et al., 2001; chapters 3.4, 3.5). However, all current knowledge about the genetics and molecular phylogeny of DOBV-Aa strains was generated on the basis of nucleic acid isolation from A. agrarius- or human-derived specimens and subsequent PCR amplification and nucleotide sequence analysis. Here we describe the isolation of an indigenous DOBV-Aa virus strain which can be taken as the representative of the DOBV-Aa lineage within the DOBV species.

Our sequence and phylogenetic analysis showed that the DOBV-Aa virus isolate, named Slovakia or SK/Aa in brief, is genetically closely related to the other Central European DOBV-Aa sequences. When considering the S and L segment phylogenetic trees of DOBV strains, SK/Aa is the only viable virus strain within the whole DOBV-Aa genetic lineage.

Hantavirus isolation is a tedious, time-consuming process, which is only rarely successful. When original DOBV strain from Slovenia (Slo/Af) was isolated, only one out of 13 isolation attempts was successful (Avsic-Zupanc et al., 1992). However, an isolation protocol used for isolation of Saa/160V, including three weeks passaging intervals and adding of fresh uninfected cells to passaged cells, seems to be very efficient. Nemirov at al. (1999) could isolate virus from three out of three tissue samples. We also reached 100% efficiency, when both our isolation attempts were positive. Besides the protocol, critical points might be the tissue homogenisation step and the viral load in naturally infected rodent tissues. We selected samples very strongly reacting in screening tests (ELISA, RT-PCR) and the tissues were triturated very thoroughly using FastPrep Instrument (BIO 101, USA). Usually the lung tissues of rodents are used for isolation attempts. Our results suggest that liver tissues can be used as well. This could be an advantage because the amount of lung tissue is very limited and, moreover, lung tissues are usually used also for initial RT-PCR screening of animals. Vapalahti et al. (1996) used lung tissues and a pool of liver, kidney and spleen tissues in parallel during their TULV isolation attempts, though only lung tissues revealed a positive result.

↓82

The SK/Aa isolate was grown in Vero E6 cells. Adaptation of a field virus to cell culture may be accompanied by the accumulation of mutations in the viral genome. However, when we compared the original sequence from the A. agrarius specimen used for virus isolation (Esl/34Aa) with the nucleotide sequence of the complete S segments of the virus isolate SK/Aa, we found no differences in the non-coding regions and only one putative amino acid exchange in the N protein occurred. This exchange of Ala to Thr on aa position 43 can be considered as conservative according to the criteria of Dayhoff and co-workers (1978). Both aa residues can be also alternatively found in the N protein sequences of other hantavirus species, no matter whether the respective strains were cell-adapted or „wild“ strains (data not shown). This let us conclude that our virus isolate at least in the S segment represents the natural genetic make-up of the „wild“ virus.

Similar analyses were recently undertaken with PUUV strain Kazan (Lundkvist et al., 1997b; Nemirov et al., 2003a); the only amino acid substitution was found in the L protein, Ser versus Phe at position 2053. In the entire S segment sequences, single mutations in both 5’ and 3’ NCRs have been observed which correlated with a different infectivity of the viruses to bank voles, but no differences could be found in S segment coding region and entire M segment (Lundkvist et al., 1997b).

In other studies, the complete S segment sequence recovered directly from the lung tissue sample of A. flavicollis was found to be identical to that of the subsequent Vero E6 cell culture isolate Dobrava/Ano-Poroia (Nemirov et al., 2003b). Chizhikov et al. (1995) compared the non-translated regions of the S and M segments of SNV RNAs amplified from the tissues of the original trapped and experimentally infected Peromyscus maniculatus rodents with virus harvested from the fifth passage in Vero E6 cells and found no nucleotide sequence differences among these samples suggesting that no genetic selection or adaptation is taking place during growth of the virus in the experimentally infected P. maniculatus as well as during the five passages in Vero E6 cells.

↓83

Since the amplification of hantaviral nucleic acid from patient's material is difficult and rarely successful, FRNT is the only useful method for fine typing of human hantavirus infections in HFRS diagnostics (Krüger et al., 2001). The validity of the approach to serotype neutralising antibodies in the patient’s serum by FRNT mainly depends on the availability of that virus in the assay which is nearly related to and therefore representative for the naturally infecting virus strain. Accordingly, the SK/Aa strain is the virus of choice to test sera from Central European HFRS patients in FRNT. Here we have investigated convalescent sera from 9 HFRS patients from Germany and Slovakia originally diagnosed as DOBV-positive on the basis of the neutralising ability of their sera towards the Slo/Af prototype strain (Sibold et al., 2001; chapter 3.6). When comparing the neutralisation endpoint titers against SK/Aa versus Slo/Af of these patients’ sera, we found an at least fourfold higher reciprocal end-point titer towards SK/Aa in 5/9 sera. Two out of nine sera exhibited equal neutralisation activities towards SK/Aa and Slo/Af, and the remaining 2/9 sera neutralised the Slo/Af virus strain significantly better than the SK/Aa strain (Table 18).

Since in our work (using sera of HFRS patients) as well as in seroprevalence studies in Estonia (Brus Sjölander et al., 2002; Golovljova et al., 2002), Latvia (Lundkvist et al., 2002a) and most of all in Lithuania (S. Sandmann, H. Meisel, A. Razanskiene, A. Wolbert, B. Pohl, D.H. Krüger, K. Sasnauskas, and R. Ulrich, submitted for publication) a noticeable number of sera reacted equally well with Slo/Af on the one hand and SK/Aa (or Saa/160V) on the other, it is reasonable to speculate that the envelope glycoproteins (encoded by the viral M segments and responsible for the reaction with the neutralising antibodies) of DOBV-Af and DOBV-Aa are not diverse enough to account for clear differences in the patterns of neutralising antibodies. Alternatively, when one takes an at least fourfold higher end-point titer as evidence for the infection of the patient by the respective virus strain, our FRNT data could be interpreted in the following way. The majority of patients (5/9) was infected by DOBV strains nearly related to our SK/Aa isolate, for 2/9 patients no conclusion can be made, and further 2/9 patients could have been infected by DOBV-Af-like virus strains. Interestingly, both sera exhibiting better neutralisation of the Slo/Af strain were taken from patients from Slovakia, a country in Central Europe not far from South-East Europe where DOBV-Aa and DOBV-Af were demonstrated to be sympatrically present (Sibold et al., 2001). Our data provide a first hint that in this geographical region not only DOBV-Aa but also DOBV-Af strains could be etiological agents of DOBV-associated HFRS cases (see below).

In summary, the first Central European DOBV strain has been isolated from an A. agrarius rodent. The availability of this virus strain will allow additional studies to answer such interesting questions as about the antigenic properties of DOBV lineages, their differential virulence potential, or the role of genetic reassortment and host-dependent virus evolution.

4.8 HFRS in West Slovakia associated with DOBV-Af infection

↓84

We have described two DOBV-associated HFRS cases from West Slovakia. A few DOBV cases can be detected in this region every year. Interestingly, only A. flavicollis is present in this part of the country suggesting that DOBV-Af is an etiologic agent of DOBV cases in this region. Human infections caused by DOBV-Af have been yet reported only from South-East Europe, where the HFRS cases caused by DOBV infection were found to be rather clinically moderate or severe. Nevertheless, the severity of diseases and fatality rate in West Slovakia are not similar to those in South-East Europe but rather resemble the situation in rest of Central Europe. The only two HFRS fatal cases from western part of the country were documented in 1958 and 1959 (Dornetzhuber et al., 1960). It was therefore interesting to clinically describe these infections and to investigate their association with DOBV-Af.

The association with DOBV-Af was confirmed when sera of both patients showed fourfold higher reciprocal titer against Slo/Af than against SK/Aa. This finding represents in fact the first, although only indirect evidence that DOBV-Af is an etiologic agent of DOBV cases in Central Europe.

The estimation of HFRS severity according to criteria of Lee and van der Groen (1989) showed interesting results. In both of the cases most of the parameters suggested only mild diseases. In contrast, two criteria (max. BUN and days of proteinuria) of patient A corresponded to severe disease. Both parameters are related with the acute renal failure of patient A. It is questionable, how relevant these criteria, originating from clinical data of HTNV infections in Korea, for DOBV infections are. Nevertheless, two described cases showed similar clinical course as reported previously for DOBV-associated HFRS cases from Central Europe (Sibold et al., 2001).

4.9 First direct evidence that DOBV causes HFRS in Central Europe

↓85

For the first time, DOBV genetic material could be directly detected in a specimen from a patient in Central Europe. Generally, most of the hantavirus sequences available in GenBank were obtained from rodent tissue samples or virus isolates and sequences obtained from patient material are rather exceptional. It is mostly due to a fact that patient blood samples are very often taken too late. Hantaviral RNA is usually detectable only within the first days after onset of disease and not even in all patients (Vapalahti et al., 2001; Horling et al., 1995; Plyusnin et al., 1997a; Plyusnin et al., 1999a). The second critical point might be the handling and storing of clinical material. Our first positive outcome suggests that using of RNA-preserving buffer is necessary to keep detectable level of viral RNA, if the samples could not be immediately stored at –70°C. In our case, storing the blood sample in AVL Buffer (Qiagene Viral RNA Kit) at –20°C allowed us to detect and amplify hantaviral RNA and a partial S segment nucleotide sequence, designated H169, could be obtained.

Sequence analysis of H169 clearly revealed that this strain from a HFRS patient in Central Europe belongs to the DOBV species. The putative amino acid sequence encoded by the analysed genome fragment showed a 99.4 % identity with the DOBV-Aa virus sequences from Slovakia, Central Europe. The molecular phylogenetic analysis exhibited clustering of the H169 strain with sequences of the DOBV-Aa lineage suggesting that a DOBV strain originating from A. agrarius was responsible for the described HFRS case. Interestingly, DOBV-Aa strains from East Slovakia have been shown to have the highest sequence identity to H169, although in phylogenetic tree they occupy the most distal and ancestral positions, respectively.

Interestingly, the clinical course of this infection (patient H169) was quite similar to the more severe DOBV-Af cases reported from West Slovakia. Despite of acute renal failure, according to Lee and van der Groen (1989) many parameters suggested only mild disease. It seems to be characteristic for DOBV infections in Central Europe that only laboratory parameters related to renal failure are comparable to severe HTNV infections from Asia (Lee and van der Groen, 1989). Other markers, particularly those corresponding to febrile phase, usually indicate only mild disease, when haemorrhages, fever over 39°C or leucocytosis are rather exceptional. It might be concluded that DOBV in Central Europe causes infections with acute renal failure, however, without such severe hemorrhagic complications as observed for HTNV or eventually for DOBV in Balkans.

↓86

Altogether, for the first time a direct evidence that DOBV causes HFRS in Central Europe was obtained. The sequence originating from blood sample of an HFRS patient from North-East Germany is clearly distinct from patient-associated DOBV sequences from Greece and is closely related to DOBV strains found in A. agrarius mice.

4.10 Questioning the hypothesis about different pathogenicity of DOBV-Af and DOBV-Aa towards humans

Recently, several authors forwarded the hypothesis that DOBV-Af and DOBV-Aa harbour different pathogenicity towards humans. This idea is mostly based on fact that many severe HFRS cases have been reported from the Balkans, where DOBV-Af is believed to be dominant (Avsic-Zupanc et al., 1999; Papa et al ., 1998). On the other hand, no fatalities and only mild DOBV-associated diseases occur in Central and North-East Europe, where DOBV-Aa is supposed to prevail (Sibold et al., 2001; Brus Sjolander et al., 2002; Plyusnin, 2002). However, the comparison of our three DOBV clinical cases (chapters 3.5 and 3.6) suggests that there is no obvious difference in severity of DOBV infections in Central Europe.

The species range of A. flavicollis covers almost the whole Europe (Montgomery, 1999) and it seems to be unlikely that the distribution of DOBV-Af is restricted only to Balkans. The detection of DOBV-Af in Slovakia (Sibold et al., 2001; Klempa et al., 2003b; chapter 3.1) and the putative DOBV-Af-associated clinical cases from West Slovakia (chapter 3.6) are in line with this assumption. Nevertheless, a high severity and fatality rate of HFRS is reported only from Balkans. It might be concluded that not DOBV-Af generally, but only local strains, circulating in Balkans harbour higher pathogenicity towards humans.

↓87

Moreover, these are not necessarily the properties of the virus which are responsible for the different clinical outcome of the disease. Genetic variation in immune related genes of human populations is well known to be associated with interesting immunological phenotypes including susceptibility to diseases. Four such polymorphic regions are well-established candidate regions for disease susceptibility; T cell receptor loci, the killer Ig-like receptor loci, the immunoglobulin heavy chain region, and the major histocompatibility complex (Geraghty, 2002). The recent epidemic of severe acute respiratory syndrome (SARS) in Taiwan is a recent example how the genetic differences among human population can influence the outcome of infectious diseases. Up until recently, no probable SARS patients were reported among Taiwan indigenous people who are genetically distinct from the Taiwanese general population, do not show the human leukocyte antigen (HLA) B*4601 allele and have high frequency of HLA-B*1301 allele (Lin et al., 2003).

For other examples, one even does not have to go far from DOBV. It was shown that HLA haplotypes can influence the clinical course of hantavirus infection, as shown for PUUV infection. Mäkelä et al. (2002) showed that the HLA-B8-DR3 haplotype is an important contributor to the course of Nephropathia epidemica. In addition, patients with the most severe course of the disease had a very high frequency of HLA B8, C4A*Q0, and DRB1*0301 alleles (Mustonen et al., 1996). It has also been shown that non-carriage of IL-1RA allele 2 and IL-1Beta (-511) allele 2 may contribute to susceptibility to PUUV infection. In contrast, HLA B27 seems to be associated with a benign clinical course of PUUV infections. It is also of interest that there are apparently similar associations with HIV disease; the fast progression is associated with B8 DRB1*0301 whereas B27 is associated with milder forms (see Mustonen et al., 1998).

Therefore, certain genetic differences between human populations in Europe could contribute to the different clinical outcome of DOBV-caused HFRS in Balkans in contrast to the rest of Europe. In recent population genetic studies, some differences between Balkans and other South and Middle European populations and even within the Balkan population have been found in serum proteins (Scheil et al., 2001), DNA-Short Tandem Repeat (STR) analysis (Huckenbeck et al., 2001), and allele frequencies of alpha-1-antitrypsin gene (Scheil et al., 2002). It cannot be excluded that Balkans human population differs also in some markers associated with hantavirus infection susceptibility.

↓88

Altogether, the presence of severe HFRS cases in the Balkans and only mild DOBV-associated diseases in Central and North-East Europe might be explained by different pathogenicity of DOBV-Af and DOBV-Aa towards humans, but also by higher virulence of local Balkan DOBV-Af strains or some immunogenetic differences in human populations in Europe. With the current state of knowledge, any of these scenarios might be valid. The characterisation of reasonable numbers of DOBV-Af cases outside Balkans and eventually also DOBV-Aa cases from Balkans, as well as population genetic studies are needed to clarify this important subject.

4.11  Does pathogenic TULV circulate in North-East Germany?

So far, TULV was considered to be probably non-pathogenic for humans. One singular case of (anamnestic) human infection could be found by serological evidence in a healthy blood donor indicating that TULV (or a TULV-like virus) might be able to infect humans (Vapalahti et al., 1996). Recently the first case of an acute TULV infection of a patient with fever, paronychia and exanthema but without renal or pulmonary affection has been reported. However, the disease occurred after the patient was bitten by a wild rodent, indicating an unusual route of hantavirus transmission. Similarly to our case, several bouts of fever occurred during the course of infection. Interestingly, the rodent involved in this case was later described by the patient (12-year-old boy), his mother and sister, but the description was not compatible with the characteristics of M. arvalisthe natural carrier of TULV (Schultze et al., 2002).

In our case some typical signs of HFRS (e.g. proteinuria, polyuria) could be observed. Moreover, no unusual route of infection could be suggested. On the other hand, two phases of disease and pneumonia were observed, signs not common for HFRS. In contrast to the previous report (Schultze et al., 2002), the hantavirus specific IgM antibodies could not be detected by IgM-specific ELISA. However, such disappearance of detectable IgM within four weeks after onset of disease has also been described for other cases of hantavirus infections (Elgh et al., 1998; Kallio-Kokko et al., 1998). An acute hantavirus infection was shown by the increase of hantavirus antibody titers against PUUV antigen determined by IFA from 1:128 (day 2 after onset), through 1:256 (day 12) to 1:3,200 (day 38). The FRNT is the only choice for precise identification of hantavirus infection if the viral genetic material can not be detected. The at least fourfold higher reciprocal titer against TULV when compared with the other viruses, indicated that TULV was the virus responsible for the infection of the patient. The FRNT titer against TULV was relatively low, but the same titer (160) was obtained also in the previous TULV case (Schultze et al., 2002). Nevertheless, it cannot be completely excluded that a TULV-related, albeit unidentified virus not present in the FRNT virus collection caused the infection.

↓89

It was thus important to show that TULV is endemic in this region of Germany. TULV strains were detected in M. arvalis animals trapped at places only a few kilometres distant from the home village of the patient. On ML phylogenetic tree, these strains clustered with strains from Poland and represent a new, well-supported third sublineage within the TULV species. It might be possible that the pathogenicity towards humans is unique for this newly characterised sublineage.

Up to now, two branches (sublineages) of TULV had been defined represented by strains from Russia on one hand and strains from Czech Republic and west Slovakia on the other hand . The sequences of TULV strains from East Slovakia (Y13979 and Y13980) have been considered to be recombinants between these two major sublineages (Sibold et al., 1999a). Therefore we excluded them from the phylogenetic analysis of complete S segment sequences. Recently, TULV was detected in Belgium and these strains together with sequences from Switzerland and south Germany were postulated to represent an additional sublineage (Heyman et al., 2002). However, too short partial S segment sequences (304 nt) were used for phylogenetic analysis and the overall bootstrap support for calculated tree was not plausible.

Altogether, our data let us conclude that TULV is circulating in North-East Germany and that the virus is a causative agent of renal and pulmonary affection in humans. In addition to PUUV and DOBV, TULV should be considered as the third Central European hantavirus being pathogenic for humans.


© Die inhaltliche Zusammenstellung und Aufmachung dieser Publikation sowie die elektronische Verarbeitung sind urheberrechtlich geschützt. Jede Verwertung, die nicht ausdrücklich vom Urheberrechtsgesetz zugelassen ist, bedarf der vorherigen Zustimmung. Das gilt insbesondere für die Vervielfältigung, die Bearbeitung und Einspeicherung und Verarbeitung in elektronische Systeme.
DiML DTD Version 4.0Zertifizierter Dokumentenserver
der Humboldt-Universität zu Berlin
HTML generated:
20.10.2005