Usutu Virus Persistence and West Nile Virus Inactivity in the Emilia-Romagna Region (Italy) in 2011

Background The circulation of West Nile virus and Usutu virus was detected in the Emilia-Romagna region in 2008 and 2009. To evaluate the extent of circulation of both viruses, environmental surveillance, based on bird and mosquito testing, was conducted in 2008 and gradually improved over the years. Methods In February–March 2009–2011, 5,993 hibernating mosquitoes were manually sampled, out of which 80.1% were Culex pipiens; none tested positive for the viruses. From 2008 to 2011, 946,213 mosquitoes, sampled between May and October, were tested; 86.5% were Cx. pipiens. West Nile virus was detected in 32 Cx. pipiens pools, and Usutu virus was detected in 229 mosquito pools (217 Cx. pipiens, 10 Aedes albopictus, one Anopheles maculipennis s.l., and one Aedes caspius). From 2009 to 2011, of 4,546 birds collected, 42 tested positive for West Nile virus and 48 for Usutu virus. West Nile virus and Usutu virus showed different patterns of activity during the 2008–2011 surveillance period. West Nile virus was detected in 2008, 2009, and 2010, but not in 2011. Usutu virus, however, was continuously active throughout 2009, 2010, and 2011. Conclusions The data strongly suggest that both viruses overwinter in the surveyed area rather than being continually reintroduced every season. The lack of hibernating mosquitoes testing positive for the viruses and the presence of positive birds sampled early in the season support the hypothesis that the viruses overwinter in birds rather than in mosquitoes. Herd immunity in key bird species could explain the decline of West Nile virus observed in 2011, while the persistence of Usutu virus may be explained by not yet identified reservoirs. Reported results are comparable with a peri-Mediterranean circulation of the West Nile virus lineage 1 related strain, which became undetectable in the environment after two to three years of obvious circulation.


Introduction
West Nile virus (WNV) was first reported in the Emilia-Romagna region in 2008 [1], and evidence of circulation was obtained until 2010 [2]. The high identity from partial nucleotide sequencing of positive mosquitoes, humans, and wild birds suggests that the virus overwintered rather than being reintroduced several times [2], [3], [4].
Usutu virus (USUV) has been detected in the same area since 2009 [3], [5], and biomolecular data demonstrate that the virus has overwintered in Emilia-Romagna [2]. This flavivirus was first reported in Europe in Vienna in 2001 [6]; from there, the virus was reported in neighboring countries and has become endemic [7]. Both viruses belong to the Japanese encephalitis antigenic complex [8] and are phylogenetically closely related [9], but while WNV is a known pathogen for humans, the pathogenicity of USUV to humans is not fully understood [10], [11].
Similar to other arboviruses (vector-borne viruses), WNV and USUV exploit complex biotic and trophic interactions between vectors and reservoirs, to persist in a particular area. These interactions are influenced by numerous ecological and environmental factors that can confer unique characteristics on the virus cycle in a particular ecosystem, including differential ways of transmission and permanence. Due to this variability, characterizing the epidemiological picture in a specific environment, identifying all transmission ways and discovering the predominant ones, is difficult. This is especially true for WNV, which may exploit many arthropod vectors and reservoir species [12]. Certainly, wild birds are the most significant reservoir species for WNV, with different viremia duration and intensity according to the species [13], and a different role in virus transmission according to the species abundance. The USUV cycle, however, is less well known but seems to be similar to the WNV cycle.
Overwintering is a critical point for virus permanence in temperate regions. Several mechanisms have been hypothesized to explain overwintering of arboviruses. When the vector survives from one season to the next by arresting development and reducing metabolism (diapause), as in the case of the Culex pipiens mosquito, the virus can persist in the adult vector. The virus can also survive the winter in the vector eggs, or more rarely, in the juvenile stages, when they are the diapausing stages [14].
Vertebrates can also contribute to arbovirus overwintering. For example, a delayed viremia can be produced in hibernating animals at their awakening, or in some specimens a recurrent or long-lasting viremia could be produced, or the virus could persist in specific organs of animals, and be reactivated after ingestion by predators or scavengers [14], producing an infection.
Since 2008, an environmental surveillance program targeted toward WNV and USUV has been active in the Emilia-Romagna region [15], [16]. This system has been improved over the years, and this paper reports the 2011 monitoring results. To explain the permanence mechanisms of the two viruses in the region, the data obtained in previous years [2], [3], [5], [16] were also analyzed retrospectively.

Surveillance results for 2011
Mosquitoes. From February 11 to March 14, 4,109 overwintering mosquitoes were sampled manually from winter hibernacula (of which 3,524, equal to 80.1%, were Cx. pipiens) ( Table 1). A particular effort was made in the Modena province, where 2,333 mosquitoes were collected at one site on three sampling days (1985 Cx. pipiens, 348 An. maculipennis s.l.), and 699 Cx. pipiens specimens were collected at another site during one day.
At three sites, by using carbon dioxide attraction traps weekly from March 10 to May 20, 258 mosquitoes were collected, of which 252 (97.7%) belonged to the species Cx. pipiens (Table 2).

Detection and isolation of viruses
The viruses were not detected in overwintering mosquitoes or in the mosquitoes sampled early in the season. Assuming a potential USUV minimum infection rate of 0.4, obtained with the 2010 data (213,235 Cx. pipiens specimens sampled from the sampling of the first USUV-positive pool and 89 USUV-positive pools), at least one USUV-positive mosquito was expected among the 3,524 overwintering Cx. pipiens mosquitoes sampled.
In 2011, all mosquito pools tested negative for WNV, while 74 Cx. pipiens pools, six Ae. albopictus pools, one Ae. caspius pool, and one An. maculipennis s.l. pool were USUV positive (Table 3, Figure 1). Fifty-four of these pools were sampled with plan traps and 28 with Modena traps.
The first Cx. pipiens USUV-positive pool was sampled in the province of Bologna on July 14, and the last was sampled in the province of Modena on September 30 ( Figure 2). The first Ae. albopictus USUV-positive pools were sampled on July 22 and the last on October 4, both in the province of Modena. The positive An. maculipennis s.l. pool and the Ae. caspius pool were sampled on September 13 with a Modena trap.
Flavivirus sequences were detected in 22 Ae. albopictus pools and in three Ae. caspius pools; these sequences had a high identity rate for Aedes flavivirus and Ochlerotatus flavivirus, respectively, two mosquito-only flaviviruses previously reported in Italy [17].
No birds tested positive for WNV; however, 25 were USUV positive. Of these birds, 14 were collected in active surveillance, and 11 were collected in passive surveillance. Almost all positive birds collected in active surveillance were Magpies (13/14), and one was a Eurasian Jay. Of these birds, six birds were sampled in Bologna province, five were sampled in Ravenna province, and three were sampled in Ferrara province. In passive surveillance, one Magpie was positive for USUV. Of the other birds passively

USUV sequence
The 42 sequences obtained with the flavivirus-genus PCR showed an identity range of 99.1% to 100% out of 214 positions compared. Compared to the 2009 sequence, the 2011 sequence differed by one base pair (bp) and by another base pair compared to the 2010 consensus sequences (out of 264 bp compared). Interestingly, one of these mutations was present in the 2010 and 2011 partial NS5 consensus sequences, but was not present in the 2009 consensus sequence or in any other GB sequence (G.A in genomic position 9060 of EF206350). This mutation could have been acquired while the virus was expanding.
The 55 sequences obtained with USUV-specific PCR showed inter-identity ranging from 99.4% to 100%, out of 361 positions compared. The 2011 PCR consensus sequences were identical to the 2009 and 2010 consensus sequences obtained from the mosquitoes and birds sampled in the Emilia-Romagna region [2], [3]. There were no mutations in the obtained sequences; therefore, no changes were observed in the sequence of the translated proteins.

Surveillance results from 2009 to 2010
Mosquitoes. In March 2009 and 2010, overwintering mosquitoes were sampled manually from winter hibernacula. In 2010, 894 specimens were collected (of which 764, equal to 85.0%, were Cx. pipiens) at 50 sites selected within a 5 km radius from stations that had at least one positive mosquito pool in 2009. In 2009, 991 female specimens (of which 512, i.e., 51.7%, were Cx. pipiens) ( Table 1) were collected in 28 suitable sites in areas with WNV-positive horses, birds, and mosquito pools in 2008 [1], [16]. The different ratio between the two species recorded in this year was probably due to the high density of An. maculipennis in the sampled area and to the wide seasonal variations in the density of the two species, which may also affect the number of overwintering specimens. Early seasonal samples were collected at 48 sites in April 2009, and 2,587 mosquitoes were collected. A high proportion of these were Cx. pipiens (2,040 specimens, 78.9% of the total) ( Table 2).
Birds. In 2010, a total of 1,596 birds were tested: 1,385 were actively collected, and 211 were obtained passively. Of these birds, 157 (74.4%) were from WRCs, and 54 (25.6%) were found dead in the field (Table 5). In 2009, 1,464 birds were tested: 1,378 were actively collected, and 86 were passively collected. Of these birds, 60 (69.8%) were from WRCs, and 26 (30.2%) were found dead in the field (Table 5). These results are partially reported in Calzolari et al. [2] and Tamba et al. [5].     albopictus pools (two in 2009 and two in 2010) ( Table 5). WNV was detected in 40 birds in 2009 and in two birds in 2010. USUV was detected in 11 birds in 2009: seven actively collected and four passively collected (of which two Blackbirds found dead in the field). USUV was detected in 12 birds in 2010, one actively collected and 11 passively collected (of which two Blackbirds found dead in the field) ( Table 5, Table S1). The permanence time in a WRC was recovered for six of these birds. All died within two days after arrival. Positive bird species are reported in Table S1. The details of and results for the mosquitoes collected from May to October in 2008-2010 and birds collected from May to October in 2010 and from May to November in 2009 have been reported in previous publications [2], [3], [5], [16].
Estimation of reservoir importance of bird populations. An estimation of the relative reservoir importance of birds tested in active surveillance was performed using 2009 data, a year with a relevant WNV circulation. The infection rate of Magpies, Crows, Jays and Starlings was found by utilizing data from surveyed areas, and then the number of potential infected specimens were estimated according to the bird abundance (Table 6). In the 2009 scenario, among the four actively collected species, Starlings seemed to be the more important WNVreservoir hosts, followed by Magpies. Due to the low number of USUV-positive birds actively collected, this estimation was not performed for USUV.   Figure 3). Despite the limited length of the 2011 USUV sequences, obtained with the two diagnostic PCRs, the high identity with previous year sequences suggests that the virus overwintered in the region [2]. Previous data also strongly sustain the hypothesis of overwintering for WNV between 2008 and 2010 [2], [3], [4]. The geographic distribution of the viruses in the territory over several years of surveillance also supports the hypothesis of overwintering of both viruses, rather than different reintroductions.

Discussion
A possible mechanism of arbovirus overwintering in temperate climates could be through infected diapausing mosquitoes [14], [18]. USUV-positive mosquitoes in diapause have been reported in Europe [19]. Furthermore, WNV-positive overwintering mosquitoes were detected several times in the United States [20], [21], [22], [23], but always at low rates, and often after seasons with a high infection rate in mosquitoes [23]. Moreover, a reduced overwintering capacity was recorded in engorged mosquitoes [24], and various field observations reported a low parous rate in mosquitoes in diapause, which decreased during the winter [25], [26], often reaching zero at the new season start [27]. In addition to this scarce probability of virus overwintering in directly infected diapausing mosquitoes, vertical transmission could be another mechanism supporting WNV overwintering. Vertical transmission of the virus has been shown by WNV detected in mosquito larvae sampled in the field [28] and in adults reared from field-collected larvae [17], [29]. However, a low rate of vertical transmission was often found in WNV experimental infection of mosquitoes [30], [31], [32], [33], [34], [35], suggesting that this phenomenon is rare in nature. Furthermore, if virus overwintering did occur in mosquitoes, it would have to occur in enough infected mosquitoes to restart viral circulation in the following season. This number must be sufficient to ensure the passage of the virus to a competent host, since a female mosquito feeds a limited number of times, to mature the eggs. Limited presence of blood from different hosts was found in engorged Cx pipiens mosquitoes, which seems to demonstrate that multiple feeding is a rare event for this species [36], [37], [38], and therefore limits the capacity of this species to transmit the virus. During the three years of surveillance, two overwintering mosquito species were sampled more abundantly, Cx. pipiens (80.1%) and An. maculipennis s.l. (19.8%), but none tested positive, despite the large sampling effort expended in areas where positive mosquito pools had been sampled the previous summer. Furthermore, the first positive mosquito pool was recorded in July, even though mosquito sampling was performed early in the season (Table 2) and the majority of sampling started in June. The data suggest that diapausing mosquitoes were not the main source of virus overwintering in the Emilia-Romagna region; another mechanism probably exists.
Vertebrate hosts could also provide a way for viruses to overwinter. Long-term viremia was described in the House Sparrow [39]. WNV was detected in the urine of human convalescent patients [40] and in the kidneys of laboratory hamsters [41] months after infection, showing the virus persists for long periods in some organs. Ingestion of infected animals by susceptible hosts, by hunting or carrion-feeding behavior, may be an alternative virus transmission method [13], [14], [39], [42]. The mortality of birds at the end of winter could facilitate this mechanism of infection in opportunistic scavengers, such as European Magpies and Hooded Crows, and, therefore, WNV overwintering in birds.
The relevance of this mechanism in WN and USU viruses overwintering in Emilia-Romagna seems to be confirmed by finding birds that tested positive for the viruses early in the season, when the mosquito density was low, as the Spotted Eagle died on April 12, 2011, and tested USUV-positive, and the WNV-positive magpie was sampled in May 2009 [5].
The surveillance system reported an increase in WNV circulation from 2008 to 2009, with a successive decrease in 2010, and failed to detect the virus in 2011 (Table 5), despite comparable sampling effort. Herd immunity within the reservoir bird populations could explain the drastic drop in WNV circulation, as already reported [43]. In effect, durable and  Data partially published in [3], [5], [16], [2]. doi:10.1371/journal.pone.0063978.t005 protective WNV immunity has been described for various bird species [44]. This reported decline in WNV circulation is consistent with the dynamic observed in many European countries, with a progressive decrease in WNV activity in the years after the first onset of an outbreak [45], [46]. These reports and our results are comparable with a peri-Mediterranean circulation of WNV lineage 1 related strains, which disappeared from an area, or circulated into nidus at an undetectable sylvatic cycle, after about two to three years of epidemic circulation, perhaps due to the trigger of herd immunity, and are probably transported to a new territory by short-ranging migrant birds, not necessarily from Sub-Saharan Africa. Eurasian Magpies, Hooded Crows, and Eurasian Jays tested during the survey represent a relevant part of the estimated population of these birds in the surveyed area (Table 6). In addition, an relevant number of Common Starlings was tested, of which five tested WNV-positive in 2009 [5] (Table S1). Crows are more abundant in Emilia-Romagna western part and in specific environments, like poplar plantations, while Magpies are more abundant in eastern part of the region, especially in particular plain environments [47]. Although the three corvids are sedentary in Emilia-Romagna, Common Starlings have a different tendency to migrate, and large part of the regional population is sedentary [47], [48]. These birds belong to the Passeriformes order, and a primary reservoir role was recognized for passerines from experimental studies [13] and European field data [49]. Moreover, these birds are abundant in Emilia-Romagna, so they are good candidates for WNV-reservoir hosts. According to the abundance estimation and 2009 data, a more important reservoir role between these species could be assigned to Starlings and Magpies ( Table 6). These two species are more abundant in the plains and prefer arable land and urban area compared to Crows and Jays, which prefer woodland areas, and are more common at higher altitude [47], [50]. According to the habitat preference of Starling and Magpie, the number of mosquito WNV-positive pools was significantly higher in rural areas in 2009 [3], confirming the greater circulation of this virus in a rural environment.
Moreover, the observed two-to three-year decline in WNV circulation is in accordance with the 90% turnover in the population of possible bird reservoirs, such as the European Starling, which was estimated with ringing activity [48]. This is consistent with a primary role in the WNV herd immunity mechanism carried out by hatch year birds, which has been reported in the United States [51]. The high rate of immunization in hatch year birds could be due to greater exposure to mosquito bites, higher sensitivity to the virus, and the possible direct passage of the virus to the nestling, as mechanisms of passive immunity from mother to offspring exist but seem to have a short duration [52].
In contrast to WNV, the circulation of USUV did not show evidence of a decrease. After the virus in 2009 was first detected, the surveillance system constantly detected this virus in mosquitoes and birds until 2011, with similar prevalence in different years (Table 5). Possible cross immunization between USUV and WNV could influence this dynamic. However, birds tested positive for both viruses, which was reported in the Emilia-Romagna region [2], [5] seems to rule out this hypothesis. The absence of a decrease in the circulation of USUV (Table 5) could be due to a different mechanism of immunity in birds for the two viruses; however, a USUV herd immunity phenomenon was described in Austria [53]. If this is the case, the persistence of this virus could be explained by not yet identified reservoirs, in addition to wild birds, unable to develop sufficient herd immunity due to an ineffective immune response or because of a fast population turnover, as in the case of rodents [14].
A possible difference in the epidemiology of the two viruses seems also supported by the detection of USUV in Ae. albopictus pools (obtained in 2009, 2010, and 2011) and An. maculipennis and Ae. caspius pools in 2011. These findings do not necessarily confirm the USUV vectorial competence of these species, but the feeding habits of these mosquitoes may indicate the possible involvement of some mammals in the cycle of this virus. Ae. albopictus is a marked opportunistic feeder mosquito, whereas species of the An. maculipennis complex and Ae. caspius mosquito are mostly mammophilic [54], [55].
The level of USUV circulation detected by the surveillance system in 2011 (Table 5) could not be explained by the mosquito abundance recorded, lower in 2011 than in 2010, indicating a weak correlation between mosquito density between years and USUV circulation, as previously observed in Emilia-Romagna [2]. This confirms that a high density of vectors does not automatically mean higher viral circulation [2] as other factors may influence the arboviral load in the environment, such as the level of immunity in the reservoirs and several environmental parameters. The dynamics of the two viruses are similar in all years, with the circulation peaking 1-2 weeks earlier in mosquitoes than in birds (Figure 3), which seems to confirm the primary role of mosquitoes in amplifying the viruses already present in reservoirs.
Reported data highlighted differences in the epidemiology of the two viruses but leaves open several questions, and points out the arbovirus ability to exploit complex interactions between species to persist in the ecosystem, indicating that only a complete understanding of the ecology can explicate its maintenance mechanisms. This is also true in the highly anthropized environment, typical of the surveyed area. In fact, the success of an arbovirus could be linked to the simplification of the ecosystem, which causes an exponential increase in several species, with a consequent increase in their reservoir potential. This is in accord with reports that indicate a rise in the burden of vector-borne diseases as biodiversity decreases [56], [57], implying a possible buffer role of biodiversity on the impact of these diseases.

Survey area
The survey area is part of the Po valley administratively in the Emilia-Romagna region, approximately 12,000 km 2 . This area is the largest Italian floodplain, characterized by intensive agriculture and animal husbandry. The eastern part of the area is on the Adriatic Sea and is marked by large natural wetlands (Valli di Comacchio and Po River Delta). The climate is typically subcontinental, gradually becoming sub-Mediterranean toward the coastal part of the region. All of the monitored territories are densely populated, with cities, villages, and industrial areas. Industrial activity is highly developed, and the Pianura Padana is strongly influenced by human activity, characterized by intensive agriculture, with a few hedges, scattered natural protected areas, and a dense irrigation network. Non-cultivated zones are rare (e.g., riverbeds, disused quarries, newly established wetlands). The most common crops are cereals and maize; vineyards or orchards are locally abundant, and poplar cultivation is widespread in flood zones near rivers. Pinewood and Mediterranean thicket vegetation is typical of the coastal part of the region.

Mosquito collection
From February to March, hibernating mosquitoes were manually sampled in 43 suitable places, sited in a radius of 5 kilometers (km) from stations with at least a positive mosquito pool in the previous season. The locations were typical places where mosquitoes hibernate, such as damp rooms of old brick buildings, especially basements. The mosquitoes were collected when they were leaving their hibernacula with a large batterypowered backpack aspirator or a small hand-held mechanical aspirator.
Sampling sites were managed with various methods: (1) Ninety regularly sampled fixed stations, placed in a 10 km grid in the surveyed area, were activated every 14 days from the end of May to the end of September (plan traps). Mosquitoes were identified to the species level using morphological characteristics according to three classification keys [54], [55], [59], [60]. The Ochlerotatus taxon was considered an Aedes sub-genus [61]. Mosquitoes were pooled according to date, location, and species, with a maximum of 200 individuals per pool [62]. For every species, a maximum of 1,000 specimens were tested per sampling. The pooled mosquitoes were ground and centrifuged, and an aliquot of the sample was collected according to the method reported in Calzolari et al. [2].

Bird collection
Birds were sampled in the plain and hill areas of the region, and birds born in the year were preferentially examined. Programmed sampling was performed from April to October by dividing the surveyed area into 1,600 km 2 quadrants and collecting 5-10 specimens in each quadrant every month. Passive surveillance of wild birds was carried out on animals found dead in the field or deceased at WRCs.
Birds were autopsied, and organ samples (brain, spleen, heart, and kidney) were pooled, ground, and submitted to biomolecular analysis. Samples from every bird were processed individually.
Estimation of bird populations. The estimation of bird populations is difficult, due to the scarcity of data, the fluctuation in the different years and the heterogeneous abundance trough the territory. To obtain a range of breeding couple (c) abundance in surveyed area, data provided by MITO2000 (www.mito2000.it) [63] were applied to the estimates of Italian breeding population [64] (Table S2). Moreover couple abundance were assessed through faunal reports [50], [65] and experts' experience in 1-40 c/km 2 for Magpie, 1-5 c/km 2 for Crow, 0.5-10 c/km 2 for Jay, 5-40 c/km 2 for Starling. According to both estimations, density of 1.3 c/km 2 was assessed for Magpie, of 1 c/km 2 for Crow, of 0.8 c/ km 2 for Jay and of 20 c/km 2 for Starling. Non-breeding individuals were estimated in 40% of total population for Magpie and 50% for Crow and were considered not relevant in Starling and Jay populations.

Virus survey
The virus survey was conducted according to the methods used in previous season [2]. RNAs present in samples were extracted using TRIzolHLS Reagent (Invitrogen, Carlsbad, CA); cDNA synthesis was achieved using random hexamer (Roche Diagnostics, Mannheim, DE) and SuperScriptH II Reverse transcriptase (Invitrogen, Carlsbad, CA) according to the manufacturers' instructions.
Mosquito and bird samples were analyzed using three different polymerase chain reactions (PCRs): (1) traditional PCR, targeted at the NS5 gene fragment, to detect flavivirus-genus according to Scaramozzino et al. [66]; (2) traditional PCR to detect USUV [67]; and (3) real-time PCR to detect WNV, according to Tang et al.'s method [68]. Fragments obtained with PCRs were sequenced with an automated fluorescence-based technique following the manufacturer's instructions (ABI-PRISM 3130 Genetic Analyzer, Applied Biosystems, Foster City, CA). To obtain WNV sequences, a traditional PCR protocol was performed on WNV-positive samples with the primers described in Lanciotti et al. [69], targeted at C and prM genes.
The sequences were used to perform a basic local alignment search tool (BLAST) in the GenBank library to confirm the specificity of the positive reaction and to estimate the degree of identity of the detected strains. The sequences were compared with available GenBank sequences using the ClustalW alignment algorithm in the freeware program MEGA 5 [70]. All sequences are available from the authors upon request.
Virus isolation was attempted starting from the remaining part of the PCR-positive mosquito homogenates using the Vero cell line (American Type Culture Collection, Rockville, MD), incubated at 37uC, and the C6/36 cell line [71], incubated at 28uC.

Statistical analysis and point pattern analysis
All the collected data were managed by a dedicated Microsoft Access database. Due to the non-normal distribution of the dataset, the abundance of mosquitoes collected during 2009-2010 was tested using the non-parametric Kruskal-Wallis test, with p,0.05. Statistical analyses were performed using Intercooled Stata 7.0 software (Stata Corporation, College Station, TX).

Supporting Information
Table S1 Positive birds for WNV and USUV collected in active and passive surveillance (A/P) in the different years of survey.