Introduction

Arthropods and the viruses (arboviruses) they transmit are highly sensitive to seasonal changes in temperature and other climatic variables that can alter insect physiology and behavior as well as the kinetics of viral replication [1], [2], [3]. Bluetongue virus [BTV] is the cause of bluetongue, a globally emerging arboviral disease of domestic and wild ruminants that is transmitted by certain species of biting Culicoides midges [3], [4], [5]. These insects occur throughout most tropical and temperate regions of the world. It has been proposed that the recent emergence of BTV in Europe is linked to climate change, likely through the effects of temperature on the biology of the midges that serve as biological vectors of BTV and the rate of BTV replication within its insect host [3], [6], [7]. Bluetongue virus infection of animals at temperate latitudes is distinctly seasonal (typically July – November in the Northern Hemisphere) but it remains uncertain how BTV persists during winter months when cold temperatures restrict vector activity and reproduction [8], [9]. The mechanism of this so-called “overwintering” of BTV in endemic temperate regions has vexed scientists since the original published descriptions of BT in South Africa over a century ago [10]. Even the term overwintering is somewhat misleading because the interseasonal period of transmission cessation typically is greater than 6 months [8], [10], [11], [12].

Bluetongue was first identified beyond Africa and the Mediterranean Basin over 60 years ago when the disease was described in the USA [13], [14]. Ongoing surveillance since that time, including our studies, has confirmed that BTV infection of ruminant livestock in California is endemic but highly seasonal (late July – November), thus the interseasonal period (November – July) when BTV is apparently “inactive” lasts approximately 8 months [9], [15], [16], [17]. In the current study we undertook intensive surveillance of sentinel cattle and vector midges during both the seasonal and interseasonal periods of BTV activity at a dairy farm in Northern California (latitude 39.590° N). The goal of this surveillance effort was to characterize the interseasonal maintenance and dynamics of BTV activity throughout the entire calendar year, both in vector insects and sentinel cattle.

Methods

Results

The duration (July – November) and peak occurrence (late September) of seasonal BTV infection among C. sonorensis midges were similar to those described in our previous studies [9], [17], [26] (Figures 1&2). Similarly, the duration of seasonal BTV infection among sentinel cattle in cohort 1 (September-February) as determined by qRT-PCR was similar to that described in our previous studies (Figure 2). Within cohort 1, 3 cows were infected in September, 2012, 21 cows were infected in October, 2012 and 9 were infected in November, 2012. It is to be stressed that the Ct values in the cattle in cohort 1 were low (strongly positive) during the seasonal period of virus activity and became increasingly elevated (weaker) during the interseasonal period (after November, 2012) when the weakly positive results likely represented the prolonged presence of non-infectious viral RNA [27], [28]. Consistent with the literature, infectious BTV is usually associated with Ct values of <30 (unpublished data) and Ct values of positive cattle in cohort 1 during the seasonal period of BTV transmission typically ranged from 19–28, whereas during the interseasonal period Ct values among cattle in this cohort ranged from 30–38 [29], [30]. Only one of the 29 calves enrolled during January 2013 (cohort 2) became qRT-PCR positive (Figure 2). This calf was strongly positive (Ct 20) in June and Ct values increased by August, 2013 with subsequent seroconversion, indicating an active infection occurred in this calf beginning in June of 2013. Of the sentinel cattle that were enrolled during the interseasonal period of 2014, only 4/30 cattle had very weak Ct values (37–39), All 15 calves enrolled during 2014 were seropositive (colostral antibodies) but remained negative via qRT-PCR. Together, evaluation of the various sentinel animals confirmed the lack of virus transmission to sentinel cattle during the interseasonal periods of both 2013 and 2014.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Seasonal patterns of BTV infection among parous female C. sonorensis midges and sentinel cattle collected on a dairy farm during the period of August 2012-August, 2013.

The point estimate (solid line) of infection rate per 1,000 tested Culicoides; 95% confidence intervals (shaded region); maximum temperature (°C - red dash line); minimum temperature (°C – blue dash line). Shaded regions at the base of the graph are representative of historically described seasonal (red shading) and interseasonal (green shading) periods of infection.

https://doi.org/10.1371/journal.pone.0106975.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. Seasonal detection of BTV RNA among parous female C. sonorensis midges; blue histogram represent mean cycle threshold (Ct) values obtained from qRT-PCR assays.

Panel A. Seasonal detection of BTV RNA among sentinel cattle enrolled in two cohorts; Panel B. Cattle in cohort 1 (red histogram) were enrolled during the seasonal period of infection and monitored throughout the year (July, 2012-August, 2013) while cattle in cohort 2 (green histogram) were enrolled during the interseasonal period and monitored throughout the season (January 2013-August, 2013). In each figure, the scale on the y-axis is inverted where 15 represents the upper limit and 40 the lower limit and error bars are indicative of the maximum and minimum Ct values for each date. The solid line indicates the threshold for interpretation of a positive result (light grey; Ct range 15–30). Below the black line, Ct results are interpreted as indeterminant (medium grey; Ct range 30–35) and negative (dark grey; Ct range 36–40).

https://doi.org/10.1371/journal.pone.0106975.g002

Adult C. sonorensis were collected during the daytime on three separate dates during the interseasonal period of both 2013 and 2014: 1. 18 Feb 2013 (32 parous females); 2. 21 Feb 2014 (3 males, 3nulliparous females, 2 parous females); 3. 21 Mar 2014 (2 males, 17 parous females). Of the parous females collected, two pools were positive by qRT-PCR assay (Ct values of 26 and 27 respectively) and one pool had a high Ct value (38) indicating a very weak or negative result. Culicoides collected from the emergence traps included 8 males, 3 nulliparous females on 14 March 2014 and 5 males, 13 nulliparous females on 21 Mar 2014. All midge pools collected via the emergence traps were negative for BTV RNA via qRT-PCR assay. Furthermore, evaluation by qRT-PCR of nulliparous female and male midges collected throughout the year showed no evidence of BTV viral RNA.

Discussion

The objective of this study was to describe the interseasonal maintenance and seasonal/interseasonal dynamics of virus infection of vector midges and cattle resident on a dairy farm in Northern California with endemic but highly seasonal BTV infection. For the first time, this study documented the presence of BTV RNA in pools of parous female C. sonorensis midges collected during the interseasonal period in 2013 and again in 2014, which potentially explained the mechanism for overwintering of BTV on this farm. The detection of BTV in pools of parous female midges collected host-seeking during the day in the interseasonal periods of both years most likely reflects one of the following mechanisms: 1) long-lived female midges that were infected during the prior seasonal period of virus transmission, which were less active in the fall and re-emerged in midwinter during a transient period of higher temperature; or 2) uninfected female midges that emerged during February-March from cohorts oviposited in the fall to feed on a BTV-infected animal, although animal surveillance at the time identified no active virus circulation among an extensive cohort(s) of sentinel cattle. Other less likely but potential explanations for the presence of virus-infected parous female midges in mid-winter include: 3) uninfected larval stages that remained viable throughout the winter to emerge as host-seeking females in midwinter that then fed on viremic hosts and then oviposited, or 4) vertical transmission of BTV via immature stages to produce infected host-seeking female midges in the winter. Weighing the evidence, we consider that mechanism 1 represents the most probable explanation.

This finding is globally relevant to temperate areas and to future incursive zones of BTV activity facilitated by climate change. Previously Nevill (1971) proposed potential methods for the interseasonal maintenance of BTV on the high veldt of South Africa, including: 1. transovarial transmission of BTV in vector insects; 2. a complicated overwintering cycle involving an unidentified intermediate host; 3. prolonged infection of ruminant livestock; 4. prolonged survival of adult Culicoides insects; 5. an ongoing and slow/low-level cycle of infection between cattle and Culicoides midges. Comments with regards to each of these potential explanations and our observations include:

1. Transovarial transmission of the virus from infected midge females to their F1 progeny. Although this phenomenon has been shown to occur at various rates in some arboviruses transmitted by mosquitoes and ticks, it has not been proven definitively for BTV with any Culicoides species [31], [32], [33]. Similarly, we did not identify BTV infection of either male or nulliparous female C. sonorensis collected during the interseasonal period although sample sizes were small.
2. A complicated overwintering cycle that involves some unidentified intermediate host such as reptiles or birds. This possibility seems unlikely in California because C. sonorensis appears to feed principally on ruminants. Analysis of blood-fed C. sonorensis midges collected during this project showed that these vectors fed only on cattle (Bos taurus) and Black - tailed deer (Odocoileus hemionus columbianus) (unpublished data).
3. Prolonged infection of ruminant livestock. Extensive studies in recent years confirm that the duration of BTV infection in livestock is considerably less than that of the approximately 6 - month interseasonal period. Although viral RNA persists for several months following BTV infection of ruminants, duration of the viremia that is infectious to vector midges is typically <60 days [27], [28], [34], [35], [36].
4. Prolonged survival of adult Culicoides insects infected following the period of active BTV transmission. Female midges are infected for life, but become infectious only after an extrinsic incubation period that is temperature dependent. The lifespan of BTV-infected vector midges in the field is unknown, but is less than 90 days under laboratory conditions [2], [37], [38]. However, much remains to be learned regarding temperatures at potential refugia where adult midges may rest in nature during the interseasonal period and their impact on midge survival.
5. An ongoing and slow/low-level cycle of infection and transmission between cattle or other ruminants and Culicoides midges throughout the interseasonal period.

During both 2013 and 2014, only parous female midges collected in CO₂ traps used for our interseasonal collection during daylight hours were strongly positive for BTV viral RNA. BTV was not detected in either male or non-parous female midges and midges were not collected after sunset. The collection of substantial numbers of parous females only during daylight hours on a limited number of days during the interseasonal period contrasts with the “catch” obtained using conventional methods, i.e. light traps, that are typically set during dusk and run throughout the night when temperatures are lowest and few insects are active. BTV viral RNA was detected among sentinel cattle at the site through February, 2013, and again in February, 2014, but Ct values of RNA from positive cattle blood collected during winter months were high (Ct range 30–38) indicating few RNA copy numbers as compared to blood samples collected during the July – November period of active BTV transmission (range of Ct values, 19–28 in cattle blood, 25–29 in vector midges) in 2012. In contrast, the Ct values of the pools of parous midges collected in February, 2013, and March, 2014 were strongly positive (Ct <30). In our experience, infectious virus typically is associated with Ct values ≤30 in cattle blood.

A marked change in midge population age structure occurred with the transition from the late seasonal to interseasonal periods of BTV activity, with the majority of midges collected after late November, 2012 being parous females (70% of collections, as contrasted with 26–38% in prior months). Furthermore, concurrent sampling of dairy waste-water lagoons confirmed that midge larvae were not detected in edge mud after December, 2012 and that midge pupae (indicative of impending emergence of adult midges) were not found until July, 2013 [18]. In 2014, adult emergence was detected with emergence traps on two separate dates, but midges collected at those times tested negative for BTV viral RNA. This finding confirmed the low-level emergence of midges throughout the winter. However, we found no evidence of vertical transmission of BTV in these insects or in nulliparous females collected and tested during the active transmission period. Thus, vertical transmission would seem to play little or no role in the interseasonal maintenance or seasonal transmission of BTV infection on the farm.

In contrast, the presence of BTV in 2 pools of overwintering parous female C. sonorensis midges implicated adult parous midges in the interseasonal maintenance of BTV on the farm. The low Ct (26, 27) values of these midge pools is consistent with the presence of infectious BTV (Ct≤30), which indicated that BTV persisted in the vector at least through midwinter (although it does not necessarily imply fully disseminated infection) [39]. Gravid mosquitoes can survive >6 months within refugia in California and still transmit arboviruses [40], [41]. Adult C. sonorensis, once infected with BTV, apparently remain infected for life [1], [42]. Although the maximal lifespan of BTV-infected parous female midges in the field remains unknown, the resumption of gonotrophic activity of BTV-infected adult midges following the interseasonal period could re-establish the transmission cycle in endemic temperate areas. Alternatively, but less likely, the adult C. sonorensis midges could have been infected by feeding on a viremic (recently infected) animal during the winter, completed a gonotrophic cycle, and were then collected when attempting to re-feed.

In summary, our findings strongly suggest that BTV infected adult C. sonorensis midges contribute to the maintenance of BTV through the interseasonal “overwintering” period in temperate regions, most likely by their prolonged survival until the next seasonal transmission period.

Acknowledgments

We gratefully acknowledge producers, staff, and herd managers of the farms from which the samples were obtained.