The known unknowns of Powassan virus ecology

Abstract

Powassan virus (POWV; Family: Flaviviridae, Genus: Flavivirus) is the sole North American member of the tick-borne encephalitis sero-complex. While associated with high rates of morbidity and mortality, POWV has historically been of little public health concern due to low incidence rates. However, over the last 20 yr, incidence rates have increased highlighting the growing epidemiological threat. Currently, there are no vaccines or therapeutics with tick habitat reduction, acaricide application, and public awareness programs being our primary means of intervention. The effectiveness of these control strategies is dependent on having a sound understanding of the virus’s ecology. In this Forum, we review what is currently known about POWV ecology, identify gaps in our knowledge, and discuss prevailing and alternative hypotheses about transmission dynamics, reservoir hosts, and spatial focality.

Introduction

Viruses of the tick-borne encephalitis sero-complex (Family: Flaviviridae; Genus: Flavivirus) are distributed across Eurasia, North America, and the Indian subcontinent and are the causative agent of severe central nervous system infections and hemorrhagic disease (Kemenesi and Bányai 2019). First isolated from a fatal case of encephalitis in Ontario, Canada in 1958, Powassan virus (POWV) is the only member currently circulating in North America (McLean and Donohue 1959, Ebel et al. 2001). Historically, POWV has been of little public health concern. Between 1958 and 1998 there were only 27 human cases identified in North America (Gholam et al. 1999). Such low incidence rates were reflective of the known ecology of the virus. POWV is maintained in an enzootic transmission cycle between small to medium-sized woodland mammals and Ixodes cookei and Ixodes marxi ticks (Mclean and Larke 1963, Main et al. 1979). While these tick species will parasitize humans, they are nidicolous and often exist in close proximity to their hosts, groundhogs, and red squirrels, respectively (Mclean and Larke 1963, Ko 1972, Eisen 2022). Consequently, human encounters are infrequent. Yet over the last 20 yr, there has been a sharp increase in the number of human infections; 165 reported human cases between 2016 and 2021 in the United States (Centers for Disease Control and Prevention 2021). In 1997, a novel subtype of POWV was described, POWV lineage II or more commonly referred to as deer tick virus (DTV), which now accounts for the vast majority of human infections (Telford et al. 1997, Ebel et al. 1999). While these 2 lineages are serologically indistinguishable, they are genetically (~84% nucleotide and ~93% amino acid similarity) and ecologically distinct (Beasley et al. 2001, Kuno et al. 2001). Specifically, DTV is strongly associated with I. scapularis ticks which have a catholic feeding preference, actively quest, and routinely bite humans. The recent increase in infections can be partially explained by improved physician awareness and diagnostics; however, the most significant contributor is the increase in both the population size of I. scapularis ticks and their primary reproductive host, white-tailed deer (Odocoileus virginianus) (Spielman 1994, Fish 2021). Over the last century there has been a 100-fold increase in white-tailed deer populations as a result of reduced predation and human land use patterns (Côté et al. 2004). This population boom in the primary reproductive host of I. scapularis has had a correspondingly positive impact on tick populations. Accordingly, not only have DTV infections increased but those of other I. scapularis transmitted pathogens including Borrelia burgdorferi (the causative agent of Lyme disease), Babesia microti, and Anaplasma phagocytophilum (Centers for Disease Control and Prevention 2019, Swanson et al. 2023).

In lieu of vaccines or therapeutics, public awareness campaigns, acaricide treatments, and tick habitat reduction efforts are the primary strategies to reduce disease risk (Eisen 2021). For these strategies to be effective a sound understanding of the virus’s ecology is required. While much can be inferred from decades of work performed on tick-borne encephalitis virus in Europe and Russia, the ecologies of both the vector and vertebrate host(s) are different. In addition, environmental, climatic, and landscape differences could change the overall ecology of POWV. It is therefore critical that a more comprehensive understanding about POWV ecology be elucidated. In this Forum, we examine what is known about POWV transmission dynamics, reservoir hosts, and spatial focality as well as discuss gaps in knowledge. Because DTV is responsible for the preponderance of human cases, this forum will primarily focus on this lineage. Further, we challenge existing dogma and present alternative perspectives that should be taken into account when designing experiments and interpreting past and future studies examining DTV ecology.

Transmission Dynamics

The 3 main modes of tick-borne virus transmission are via horizontal transmission between tick and host, direct cofeeding transmission between ticks feeding in close proximity, or vertical transmission from female adult tick to offspring (Jones et al. 1987, Labuda et al. 1993b, Danielová et al. 2002). While the relative contribution of each of these modes of transmission to DTV maintenance in the field remains largely unknown, recent field, laboratory, and modeling studies can provide insight in DTV transmission dynamics.

Ixodes scapularis has been implicated as the main tick vector for DTV (Telford et al. 1997, Ebel et al. 2000), with infection rates of I. scapularis in the field typically ranging between 0% and 6% (Brackney et al. 2008b, Anderson and Armstrong 2012, Aliota et al. 2014, Knox et al. 2017, Price et al. 2021, McMinn et al. 2023). Field isolation of both POWV lineage I and tick-borne encephalitis virus (TBEV) from alternate hard-tick species other than their main vectors provides a first indication that multiple tick species may contribute to tick-borne virus transmission in the field (Thomas et al. 1960, L’Vov et al. 1974, Pukhovskaya et al. 2018, Chitimia-Dobler et al. 2019, Hart et al. 2023, Lange et al. 2023). However, further confirmation of the ability of a tick species to transmit a virus has to be obtained through vector competence experiments in the laboratory. A recent laboratory study on the vector competence of I. scapularis, Dermacentor variabilis, and Amblyomma americanum confirmed the ability of these 3 tick species to become infected, disseminate, and transmit DTV to vertebrate hosts (Sharma et al. 2021). The rate of infection of emerged nymphs that were allowed to feed on infectious mice during the larval stage ranged between 20% and 30% (Sharma et al. 2021). Moreover, transstadial transmission and horizontal transmission from infected nymphs to naive mice was close to 100% (Costero and Grayson 1996, Grubaugh et al. 2016, Sharma et al. 2021). The ability of different hard-tick species to become infected and transmit DTV is in line with evidence from other tick-borne viruses such as TBEV and Thogoto virus for which different tick species have been implicated as potential vectors (Kazimírová et al. 2017, Talactac et al. 2018, Ličková et al. 2020). Thus, the combined evidence from other tick-borne viruses as well as the ability of different tick species to transmit DTV under laboratory conditions makes us question the single vector paradigm for DTV, and may suggest a more complex ecology of DTV with multiple tick species that could contribute to transmission (Sharma et al. 2021).

While horizontal transmission of DTV has been studied in the laboratory, the contribution of the 2 other potential modes of transmission remains mostly unknown. Studies of TBEV have suggested that direct transmission from infected nymphs to larvae when feeding in close proximity on a host is likely the main mode of transmission (Randolph et al. 1999, 2000, Randolph 2004), and can occur in absence of a systemic infection of the host (Labuda et al. 1993a, Randolph et al. 1996). The success of cofeeding transmission requires spatio-temporal concordance between larval and nymphal feeding. Consequently, 2 ecological conditions must be met: synchrony in seasonal questing patterns of larvae and nymphs and co-occurrence on the same host. In Europe, larvae and nymphs of Ixodes ricinus, the primary vector of TBEV, co-occur from spring to autumn and are regularly found co-infesting the same rodent host (Craine et al. 1995, Richter et al. 2002, Kurtenbach et al. 2006, Burri et al. 2011, Kiffner et al. 2011). It is therefore possible for cofeeding transmission to play a role in maintaining the virus in nature. By contrast, in North America, the immature life stages of I. scapularis have different phenologies with nymphs active in late spring/early summer and larvae active in late summer (Wilson and Spielman 1985, Ogden et al. 2005). As a result, even if DTV can be transmitted via cofeeding like TBEV and Thogoto virus, which remains unknown, the likelihood of cofeeding transmission of DTV in nature may be low due to these ecological constraints (Jones et al. 1987, Labuda et al. 1993a, 1993b). There remains the possibility that immature ticks feeding in close proximity to adult ticks on white-tailed deer (Odocoileus virginianus) could permit cofeeding transmission; however, data on immature and adult co-occurrence is unavailable. Another area that is often underappreciated is the possibility of intrastadial cofeeding transmission (between individuals of the same life stage). Future field and laboratory studies examining (i) the co-occurrence frequency of immature/adult ticks on white-tailed deer and (ii) the potential for intrastadial cofeeding transmission of DTV while feeding on a nonviremic host will be fundamental to understanding the potential contribution of cofeeding transmission in DTV maintenance. Finally, while ecological barriers currently exist for interstadial immature cofeeding transmission, climate predictions project a warming climate which could change the phenology of I. scapularis ticks and cofeeding potential in the future.

Vertical transmission is another potential mode of DTV transmission that requires further investigation. One study observed that 1 in 6 (16.6%) batches of F2 larvae derived from infected F1 females could transmit POWV to hamsters, although this study did not determine what percentage of larvae within an egg clutch were infected (Costero and Grayson 1996). Analysis of larvae collected from trapped rodents revealed low rates of TBEV infection (<1%); however, the larvae were tested in pools and exact rates of vertical transmission could only be estimated (Danielová et al. 2002). Based on the available data we expect that DTV vertical transmission may be inefficient as well, but this requires further experimental confirmation. Specifically, testing individual larvae derived from infected females will be necessary to establish accurate vertical transmission rates. Such studies can be performed in controlled laboratory studies or by bringing field-collected engorged females back to the laboratory and screening the offspring. Extending these studies to include variations in temperature and humidity will be critical to determining if environmental factors can influence the overall success of vertical transmission.

Modeling studies can provide additional insights in the relative importance of different modes of transmission. Nonaka et al. (2010) incorporated tick seasonality into an epidemiological model and found that cofeeding transmission was one of the critical modes of transmission to sustain DTV (Nonaka et al. 2010). Moreover, these models revealed that cofeeding transmission by itself could sustain DTV transmission, while other modes of transmission such as horizontal and vertical transmission were not sufficient. These findings highlight the importance of unraveling the contribution of different modes of DTV transmission to better understand potential targets to disrupt the transmission cycle. It will be important that these models are regularly revised to incorporate emerging data. For instance, how would the model predictions change if other tick species such as D. variabilis or A. americanum, which were recently shown to be competent vectors for DTV, were factored into the model (Sharma et al. 2021)?

Reservoir Hosts

For an animal to be considered a reservoir host of a given pathogen it must satisfy certain criteria, such as the ability to become infected without significant adverse health outcomes while sustaining adequate pathogen loads for subsequent transmission. In addition, there must be geographic and ecologic overlap between the pathogen, vector, and host. Transmission and maintenance cease to exist if any of these are absent. Predicated on the ecology of related Eurasian TBEVs, the most likely reservoir host for DTV would be a rodent (Michelitsch et al. 2019b). Rodents reside in the woodland leaf-litter in close proximity to ticks and have a high reproductive rate ensuring the availability of naive hosts. Given that they are habitat generalists, are in high abundance, and overlap with I. scapularis populations and DTV foci, the white-footed mouse (Peromyscus leucopus) was initially identified as a possible reservoir host. This was further bolstered by the fact that several other pathogens including B. burgdorferi, B. microti, and A. phagocytophilum are maintained in an enzootic transmission cycle between white-footed mice and I. scapularis (Stafford et al. 1999, Holman et al. 2004). Serosurveys supported these assumptions as ~3% of white-footed mice were found to be POWV seroreactive in known DTV foci in Wisconsin and Massachusetts while another rodent species, the meadow vole (Microtus pennsylvanicus), was nonreactive (Ebel et al. 2000). However, the virus has never been recovered from mice in the field. Experimental assessment of DTV infections of white-footed mice revealed that animals quickly cleared infection and DTV infection of white-footed mouse-derived cell lines resulted in severely restricted viral growth (Izuogu et al. 2017, Mlera et al. 2017). Together, these data suggest that white-footed mice are frequently exposed to POWV lineage I or DTV, but are unlikely the reservoir host. Subsequent surveys have found other mammalian species including, southern red-backed vole (Myodes gapperi), spotted skunks (Spilogale putorius), porcupine (Erethizon dorsatum), and snowshoe hares (Lepus americanus) among others to be POWV seropositive; however, it is unclear if they were reactive to lineage I or DTV (McLean et al. 1962, Main et al. 1979, Zarnke and Yuill 1981, Deardorff et al. 2013, Dupuis et al. 2013). Interestingly, experimental infections of many of these species, including woodchuck (Marmota monax), opossum (Didelphis marsupialis), gray fox (Urocyon cinereoargenteus), red fox (Vulpes fulva), striped skunk (Mephitis mephitis), racoon (Procyon lotor), fox squirrel (Sciurus niger), and snowshoe hares (Lepus americanus), reveal inter- and intra-species variability in susceptibility with minimal viral replication and low viremias, thereby eliminating them as potential reservoir hosts (Kokernot et al. 1969, Zarnke and Yuill 1981, Nemeth et al. 2021). A recent study analyzed the remnant blood meals of questing I. scapularis nymphs and found a strong positive association between ticks that had previously fed upon shrews (Blarina brevicauda or Sorex spp.) and DTV prevalence (Goethert et al. 2021). Further DTV RNA was detected in the brain of 1 shrew collected from the field sites. Together these data implicate shrews as a likely reservoir host for DTV; however, experimental infections of shrews by needle and tick confirming these findings will be necessary.

Due to their public health relevance, short generation times, and ease of use, mosquito systems have informed much of our understanding about vector competence and horizontal transmission dynamics of arboviruses. However, given the physiologic, behavioral, and anatomical differences between mosquitoes and ticks, it is worth reconsidering applying mosquito-arbovirus transmission dynamic dogma to tick-arbovirus systems. Mosquitoes are temporarily associated with their hosts imbibing relatively small volumes of blood (~1–3 µl) (Ogunrinade 1980). Within hours of feeding, mosquitoes undergo diuresis rapidly removing excess water, secrete a storm of digestive enzymes, and encase the blood meal in a peritrophic matrix all of which could adversely affect the ability of an arbovirus to establish infection (Plawner et al. 1991, Brackney et al. 2008a, Kato et al. 2008). Consequently, arboviruses must achieve relatively high titers (>10⁴) in the vertebrate host for infection of the mosquito to occur (Nguyet et al. 2013). Ticks on the other hand, associate and feed on their host for days, consume large volumes of blood (larvae [~1 µl], nymphs [~25 µl], and adults [~0.5–1.4 ml]), and digestion occurs intracellularly with components of the blood being actively transported into digestive vesicles/endosomes (Koch and Sauer 1984, Sojka et al. 2013). As discussed above, experimental studies have yet to identify a vertebrate host that supports POWV or DTV viremias higher than 10³ (Kokernot et al. 1969, Zarnke and Yuill 1981, Nemeth et al. 2021). Consequently, none of these vertebrates would be considered likely reservoirs for POWV when applying arbovirus-mosquito transmission dogma. Yet, given the differences between ticks and mosquitoes, such conclusions may be flawed. In fact, it was demonstrated that Dermacentor andersoni ticks can become infected with POWV lineage I at rates of 1–5% when allowed to feed on rabbits with viremias <10^2.5 (Chernesky 1969). Such infection rates are consistent with those observed in I. scapularis collected from known DTV foci (Brackney et al. 2008b, Anderson and Armstrong 2012, Aliota et al. 2014, Knox et al. 2017, Price et al. 2021, McMinn et al. 2023). This raises the possibility that low-level viremias are sufficient to sustain DTV transmission in nature. In fact, similar low-level viremias have been found in the purported reservoir hosts of TBEV in Europe (Michelitsch et al. 2019a). Given that numerous vertebrate species can support low levels of viremia, it raises the possibility that there is not a single reservoir host (white-footed mouse or shrew), but rather a collective mammalian reservoir. For instance, the aforementioned study identifying shrews as a likely DTV reservoir also identified shrews as the most common host of larval ticks at their study sites (Goethert et al. 2021). However, another study found that the white-footed mouse was the most frequent host of immature ticks with little interactions with shrews at other sites (Ginsberg et al. 2021, Goethert et al. 2021). If multiple reservoirs do exist, then the relative abundance of any given rodent species at a DTV foci dictates which species serves as the reservoir host. In fact, this seems to be the case for TBEV in Europe with many potential small mammal species contributing to the maintenance of the virus in nature (Weidmann et al. 2006, Knap et al. 2012, Pintér et al. 2014). Alternatively, DTV may be maintained in nature without a reservoir host through cofeeding and vertical transmission as discussed above.

Reservoir hosts for a specific pathogen are often thought of and discussed as a monolith; however, it is well established that individual differences can dictate one’s response to infection from susceptibility to pathogenesis. Variables such as age, reproductive status, immune and nutritional status, and sex have all been shown to contribute to infection outcomes (Klein and Flanagan 2016). Despite such variables being well known, they are often not considered when assessing the suitability of a species to serve as a potential reservoir host. It is intriguing to consider that POWV transmission dynamics are more complex than simply tick X and reservoir host Y. Could the age and/or sex of an identified reservoir host individual significantly influence their likelihood of transmitting the virus back to a naive tick vector? Could seasonal resource availability alter host nutritional status thereby impacting viral loads and subsequent transmission? These are possibilities that should be considered when evaluating reservoir host suitability and enzootic transmission cycles, but that remain vastly understudied.

Spatial Focality

The majority of clinical cases of DTV in the United States have been detected in states in the upper Midwest and the Northeast (Centers for Disease Control and Prevention 2018). This is likely due to the geographic distribution of I. scapularis and high abundance in both regions. Sequencing of DTV isolates from both regions revealed that viruses form 2 genetically distinct clades (Pesko et al. 2010, Anderson and Armstrong 2012, Bondaryuk et al. 2021), which have likely been separated for more than a millennium (McMinn et al. 2023). Thus far, very little evidence has been found for movement of DTV between both regions, but a recent study found a DTV isolate from New Jersey clustering within the Midwest clade (McMinn et al. 2023). This indicates that long-distance movement of DTV between the Midwest and the Northeast may happen infrequently, but the underlying drivers of long-distance movement remain unknown. Understanding the drivers of long-distance movement is important to assess the risk for DTV emergence into new regions, especially with range expansion of I. scapularis ticks.

Recent phylogenetic studies have significantly increased the DTV genomic dataset and revealed important new insights in the patterns of emergence and spread in the United States (McMinn et al. 2023, Vogels et al. 2023). Phylogeographic reconstructions revealed that DTV was likely introduced into the Northeast no later than the mid-1970s (Vogels et al. 2023). It was estimated that the virus first established in southern parts of New York and Connecticut, followed by emergence towards more northern regions. The timing of introduction is similar to the introduction of other tick-borne pathogens such as Borrelia burgdorferi and Babesia microti in the Northeast (Eisen and Eisen 2018), which has been associated with the reintroduction of white-tailed deer in the 20th century (Spielman 1994), followed by population expansions of I. scapularis (Fish 2021). Further range expansions of I. scapularis may pose a risk to further spread of DTV to new regions.

Phylogeographic analyses cannot only provide important insights in the patterns of emergence and spread, but they can also help to understand local transmission dynamics of pathogens. DTV is maintained in highly localized transmission foci with tight clustering by location, and limited mixing between foci in the Northeast (Vogels et al. 2023). The isolation between foci is particularly clear in Connecticut, where foci tightly cluster by location and have been stable for over 10 yr (Anderson and Armstrong 2012). The highly focal geographical distribution of DTV is similar to what has been reported for TBEV in Europe (Labuda and Randolph 1999, Randolph et al. 1999), and provides clues on how DTV may be maintained in transmission foci. Limited mixing between nearby locations supports the hypothesis that cofeeding transmission from infected nymphs to larvae could be important for DTV maintenance, as small mammalian hosts typically do not move across long distances. Infrequent dispersal of DTV could also provide support for the hypothesis that vertical transmission from adult female to larvae is rare, and does not frequently result in the establishment of new transmission foci (Vogels et al. 2023). With adult ticks dispersing over longer distances when feeding on deer, more mixing of DTV would be expected if vertical transmission would play an important role in forming new transmission foci. Phylogeographic analyses could, thus, provide support to hypotheses on the relative contribution of different modes of transmission.

Novel insights in the geographic distribution of DTV raise important new questions on the ecology of the virus that are relevant to better inform control programs. Although several stable DTV transmission foci have been identified in both the Midwest and Northeast (Brackney et al. 2008b, Robich et al. 2019, McMinn et al. 2023, Vogels et al. 2023), it is unclear what size these foci are. Research on TBEV in Europe has suggested that the virus can be maintained in microfoci as small as 250 m² (Dobler et al. 2011, Weidmann et al. 2011, Boelke et al. 2019). These microfoci may be surrounded by larger macrofoci in which there is an increased risk for human exposure as infected ticks can be dispersed outside of the microfocus through animal movement (Dobler et al. 2011). Furthermore, it is unknown how frequently new foci will form and what the underlying drivers of long-distance dispersal are. Historic intercontinental spread of POWV has been linked to the fur trade, but it is unclear what the role of animal trade and natural dispersal (e.g., birds or deer) is in the long-distance movement of DTV (Leonova et al. 2009, Ebel 2010).

Future Research Priorities

DTV is a growing public health concern yet significant knowledge gaps exist in our understanding of the basic ecology of this neglected pathogen. To address this, future research priorities should focus on both the broad geographic distribution of DTV as well as focal transmission dynamics. Sequence analysis clearly demonstrates that DTV exists in geographically isolated transmission cycles, yet the number and size of these foci are undefined (Fig. 1). Further it is unclear how new foci are formed. We suspect that infected ticks are hitching rides from deer or birds, but once introduced how does the virus establish a stable transmission cycle and how frequent are these events? At a more refined scale, it is unclear how efficient different modes of transmission are (Fig. 1). We know that POWV, and presumably DTV, can be transovarially transmitted and that TBEV can be transmitted via cofeeding, but their relative contribution to maintenance is unknown (Labuda et al. 1993b, Costero and Grayson 1996). Further, it is still unclear which, if any, animal species serve as reservoir hosts for DTV. Addressing these knowledge gaps will be critical to elucidating DTV ecology and better defining risk and informing more effective control strategies.

Funding

This publication was made possible by CTSA Grant Number UL1 TR001863 from the National Center for Advancing Translational Science (NCATS), a component of the National Institutes of Health (NIH) awarded to C.B.F.V. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of NIH.

References