*Review* **Veterinary Diagnostic Approach of Common Virus Diseases in Adult Honeybees**

**Julia Dittes 1,2,\*, Heike Aupperle-Lellbach 3, Marc O. Schäfer 4, Christoph K. W. Mülling <sup>2</sup> and Ilka U. Emmerich <sup>5</sup>**


Received: 20 September 2020; Accepted: 19 October 2020; Published: 21 October 2020

**Abstract:** Veterinarians are educated in prevention, diagnosis and treatment of diseases in various vertebrate species. As they are familiar with multifactorial health problems in single animals as well as in herd health management, their knowledge and skills can be beneficial for the beekeepers and honeybee health. However, in education and in practice, honeybees are not a common species for most veterinarians and the typical veterinary diagnostic methods such as blood sampling or auscultation are not applicable to the superorganism honeybee. Honeybee colonies may be affected by various biotic and abiotic factors. Among the infectious agents, RNA-viruses build the largest group, causing covert and overt infections in honeybee colonies which may lead to colony losses. Veterinarians could and should play a more substantial role in managing honeybee health—not limited to cases of notifiable diseases and official hygiene controls. This review discusses the veterinary diagnostic approach to adult bee examination with a special focus on diagnosis and differential diagnosis of the common virus diseases Acute Bee Paralysis Virus (ABPV)-Kashmir Bee Virus (KBV)-Israeli Acute Paralysis Virus (IAPV)-Complex, Chronic Bee Paralysis Virus (CBPV) and Deformed Wing Virus (DWV), as well as coinfections like *Varroa* spp. and *Nosema* spp.

**Keywords:** honeybee veterinary medicine; acute bee paralysis; chronic bee paralysis; deformed wing virus; *varroa* infestation; nosemosis

#### **1. Introduction**

In contrast to human medicine, one of the most prominent features of veterinary medicine is that a veterinarian has to deal with a large variety of species. Whereas dogs, cats and horses are common patients in a veterinary practice, the honeybee (*Apis mellifera* L.) is only rarely in contact with veterinarians as a patient. Official veterinarians and diagnostic laboratories are responsible for the analysis of honey samples and for animal disease control in case of, e.g., American foulbrood, but the general disease prevention and control are done by the beekeepers. The majority of the establishments for veterinary education in Europe accredited by ESEVT (European System of Evaluation of Veterinary Training) offer elective or no training in honeybee veterinary medicine. Only in 24 of 68 surveyed schools honeybee veterinary medicine is part of the curriculum as a separate subject [1]. This is a very limited consideration in the curriculum and veterinary practice for a food-producing animal with such a tremendous importance for global agriculture [2].

However, in recent years, the honeybee has aroused increasing interest among veterinarians. While only 9 veterinarians held a specialization about honeybees in Germany in 2014 [3], five years later, there were 17 bee veterinarians or nearly twice as many experts in this field [4]. Honeybee diseases are on the rise [5] and at the same time, a decline of honeybee colonies and beekeepers in Europe can be seen [6]. More and more, veterinarians publish specialist literature to deal with honeybee diseases [7,8]. At the supranational level, the European Parliament in 2008 called for the Commission "to incorporate into its veterinary policy, research into, and actions to tackle bee disease" [9], and in April 2011, a European Union Reference Laboratory for Bee Health was designated to coordinate diagnostic methods, disease monitoring and expert training [10,11].

In honeybee colonies, a large number of individuals together form a superorganism. Both this unit as well as the individual bee have to be investigated carefully. General principles of herd health management and hygiene concepts are well-known to veterinarians and thus, can be applied to the honeybee. Nevertheless, dealing with diseases in honeybees is different from the normal approach in veterinary medicine. Usual examination techniques, known from mammals, are not applicable to insect species, although the general approach and procedures are similar. Food control is not possible as the feed is provided by the environment and not by the beekeeper or the veterinarian [7]. The honeybee is a food-producing animal and assuring the quality of honey is an additional objective for veterinarians and beekeepers.

A variety of biotic and abiotic factors have an impact on honeybee colonies. Belsky et al. presented a broad overview of these stressors: habitat and climate changes, weather, the density of apiaries and food resources, as well as transportation of colonies, are external aspects influencing the honeybee [12]. Equally, the bees depend on intrinsic factors such as genetics and queen longevity [13]. The intensive agriculture with monocultures decreases the plant diversity and thus, the food supply for the honeybees. Between mass flowering of, e.g., rapeseed (*Brassica napus*) and sunflower (*Helianthus annuus*), pollen harvest severely declines [12] and limits adequate nutrition. Such an insufficient protein diet weakens the bees in defending against pathogens [14]. Various bacteria, microsporidia, viruses and pests cause bee diseases, among which viruses have become more relevant during recent years [5].

This review presents basic information on selected virus diseases and the process of colony examination and diagnostics from the veterinary perspective, with a focus on adult honeybees.

#### **2. Virus Diseases in Honeybees and Contributing Factors**

Viruses, mostly positive single-strand RNA viruses, are the largest class of honeybee infecting pathogens [13]. Over 20 bee viruses have been identified to date, including the Acute bee paralysis virus (ABPV), Kashmir bee virus (KBV), Israeli acute paralysis virus (IAPV), forming the ABPV-KBV-IAPV complex, Black queen cell virus (BQCV), Chronic bee paralysis virus (CBPV), Deformed wing virus (DWV) and Sac brood virus (SBV). Detailed reviews about the most important ones can be found in the literature [15–18]. Information on symptoms and transmission routes of ABPV-KBV-IAPV, CBPV and DWV, which is relevant to the veterinary diagnostic approach, will be presented in this review. Table 1 provides a summary of taxonomy, symptoms, affected castes and main transmission routes.


**Table 1.** Overview of selected honeybee viruses (taxonomy, symptoms, affected casts of bees and transmission routes), modified after Vidal-Naquet [19].

Legend: RNA = Ribonucleic acid, w = worker, d = drone, q = queen, c = direct contact, o = oral–fecal, vec = vector-borne, ver = vertical, v = venereal, to = transovarial.

#### *2.1. Acute Bee Paralysis Virus—Kashmir Bee Virus—Israeli Acute Paralysis Virus—Complex*

ABPV, KBV and IAPV are three similar icosahedral viruses of the family *Dicistroviridae*. During the last years, this complex has been reported in association with colony collapse disorder (CCD), a phenomenon of severe colony weakening without visible worker bee mortality [7,20,21]. ABPV and IAPV cause symptoms such as trembling, paralysis, inability to fly, darkening and loss of hair from thorax and abdomen, affecting mostly individual bees and not the whole colony [22]. Hou and Chejanovsky describe experimental symptoms of IAPV such as disorientation, shivering wings, crawling and progressive paralysis until death in infection experiments [20]. ABPV and KBV are a reason for a sharp decline in the adult bee population [22].

Transmission of these viruses is possible vertically as well as horizontally. Furthermore, the *Varroa* mite plays an important role as a mechanical vector [15]. It transfers the virus while feeding from hemolymph of the bees and increases the viral load in the colony [15].

#### *2.2. Chronic Bee Paralysis Virus*

CBPV is an unclassified RNA virus that causes two different syndromes in honeybees. Type A, the paralytic form, is characterized by trembling, disorientation and ataxia. Type B affected bees show black, hairless and greasy shining abdomens. Both can occur in the colony at the same time and lead to massive worker bee losses [23].

Comprehensive information on this virus can be found in our case report presenting an overt infection with CBPV in two colonies [24].

#### *2.3. Deformed Wing Virus*

DWV is a picorna-like virus, that is often involved in winter losses of honeybee colonies. Two main variants, A and B, can be differentiated [25–28]. In most cases, it can be detected in colonies as a covert infection (see definition of covert under Section 3 in this review) without causing clinical symptoms [15]. Transmission routes are vertical (via eggs and sperm) or horizontal through larval food, trophallaxis or cannibalism of pupae during hygienic behavior [18]. However, *Varroa destructor* plays the key role in transmission as well as virulence and pathology of the virus [15]. Obvious symptoms are seen in the colony only if DWV replicates in the mite before being transferred to the bees. Clinical signs are deformed wings, bloated, discolored, shortened abdomens, hypoplastic glands and pupal death [29]. Infected bees are not viable and die within less than 67 h after emergence [15].

#### **3. Veterinary Diagnostic Approach**

Asmentioned above, honeybee health is often influenced bymany different factors [7]. Without clinical findings, the appearance of a pathogen in a colony does not constitute a disease. For that reason, the terms overt and covert infections were introduced by de Miranda and Genersch to describe honeybee virus diseases [30]. The descriptive terms overt and covert are widely used in insect virology. Overt infections are characterized by obvious clinical findings related to the virus infection and a high virus production rate. Acute and chronic forms are differentiated. In covert infections, low titers of virus particles are present in the absence of clinical symptoms. Vertical transmission allows virus persistence over several generations and competent virus particles can turn into overt infections due to various influencing factors. Persistent infections with low-level virus production can be distinguished from latent infections without virus production [30].

From the veterinary perspective, a holistic diagnostic work-up of medical issues in honeybee colonies is important, because the environment, the colony, the pathogens and every bee are each just a link in the chain leading to occurrences of infections [7]. Figure 1 shows a detailed plan for diagnostics and management in general. Starting from the environmental observations, followed by an examination of the hive, the colony and the bees, samples are taken, and relevant laboratory diagnostics carried out and further illnesses investigated. The resulting problem list leads to a prognosis and a management plan. The main goal is the healthy colony formed by fit individuals.

**Figure 1.** Diagnostic approach in bee diseases, © Julia Dittes.

#### *3.1. Medical History, Appeareance of the Hive and Environment*

Like in any other species, clinical examination starts with a detailed anamnesis. Gathering information from the beekeeper gives a first overview about the acute problem, which often is superficial and skewed by incorrectly interpreted data [31]. Asking detailed questions can help: When did the beekeeper first observe symptoms or changes in the colonies? How frequently are the hives inspected during the season? How is *Varroa* infestation monitored and controlled? Is the beekeeper migrating the colonies? Maybe the beekeeper can also offer information about weather, crops in the region around the apiary or the density of apiaries in the vicinity—information needed for epidemiological evaluation.

The location and the environment around the beehive have an immense relevance to the occurrence of health problems and should therefore be carefully observed. The weather, the food range, the landmarks and further factors influence the bees in a similar way as a stable or field does to cattle, for example. A complete hazard analysis is shown by Formato and Smulders [10]. Vitally important are accessible water, identifiable food sources, climate and wind [31]. The hive model, its material and the hive's general condition should be noticed. Fecal spots as well as traces of predators and pests may be seen during external examination.

Figure 2A,B show two apiaries in different locations. The size of the colony and the offered space in the hive should be evaluated [7,31].

(**A**) (**B**)

**Figure 2.** (**A**) Apiary in April 2019: colonies with two brood chambers in wooden Zander hives and polystyrene Segeberger Classic hives, standing in pairs of two. © Julia Dittes. (**B**) Apiary with Zander hives with one or two brood chambers and honey supers on top, standing next to each other. © Heike Aupperle-Lellbach.

#### *3.2. Examination of the Alighting Board and Observation of the Entrance Hole*

First pathologic findings can be identified during observation of the alighting board and entrance hole. The mass of bees flying in and out—depending on temperature, weather, time of day and state of colony development—provides a first impression about the strength of a colony. Dead bees as well as bees and larvae with alterations of the phenotype, carried out of the hive, may be found in front of the entrance. Veterinarian and beekeeper should take their time to observe the entrance hole carefully (Figure 3A,B). The way the bees fly, the manner of landing and the behavior of guarding bees offer a lot of information, because unusual flying behavior may be a neurologic symptom of virus diseases. Furthermore, the ground in front of the hive should be investigated for waste, feces, dead bees and larval mummies [7]. These observations, together with an assessment of the environment and hive from the outside, enable a first evaluation of the colony condition [31].

**Figure 3.** (**A**) Observation of the entrance hole enables a first evaluation of colony strength and activity according to season, weather and environment.©Julia Dittes. (**B**) Alighting board of a Chronic Bee Paralysis Virus-infected honeybee colony with some black hairless individuals (yellow circles). © Silvia Heisch.

#### *3.3. Clinical Examination of the Superorganism "Honeybee Colony" and Observations of Living Bees*

The "internal" clinical examination of the bee colony and of individual bees are the next steps. It is important to have a look at the adult bees of all castes (workers, drones and the queen), as well as the brood in different stages and the hive material. Honeybees are social insects that can survive within the colony as a superorganism only. The colony is functional if each single individual performs its tasks. In the following, we will focus on the clinical findings in adult bees.

Clinical examination of the bees and the frames should be performed together with the beekeeper. This way, it is possible to analyze the beekeeper´s operations while handling the bees [7]. General features to assess after opening the hive are the colony behavior, the odor and the colony strength [7,31]. Normally, a honeybee colony smells like wax. If vinegar or foul smellis perceived, there islikely a disease. The intercombspaces, where bees are visible, can be counted to give a first idea of colony size. This might be confirmed later by estimating the number of bees, brood cells and food, e.g., using the Liebefeld method [32,33]. The bees are carefully examined to find body deformities, wing abnormalities like V-wings or K-wings, changes in color and size, phoretic *Varroa* mites as well as behavioral, neurological, social or digestive symptoms. Signs of CBPV disease are hairless, black bees and/or neurological symptoms such as trembling, circling or paralysis [23,24]. Abnormal wing position and wing form are shown in case of CBPV or DWV. Phoretic *Varroa* indicate an intense infestation rate with the mite. Phenotype changes, e.g., crippled bees, high mortality and paralysis symptoms, can also be a sign of intoxication. Table 2 lists different phenotype changes in honeybees, which can be seen in a sample of dead bees as well, and their possible causes.

While examining the bees, the frames are assessed: the color and brightness of the wax correspond to the age of the combs. Old wax, having been used in several brood cycles, is much darker than newly produced wax. The frames themselves can be clean or show fecal spots in case of digestive problems [34]. Pollen and honey stores are evaluated to get an idea about the alimentation of the colony.

It is advisable to have a look at how easily the frames can be removed from the hive by the beekeeper. If they are not easy to remove, propolis foraging activity or insufficient surveillance by the beekeeper could be the reason [7].

For all symptoms, it is determined whether they affect a small number of individuals or numerous bees, which would indicate a severe problem. Furthermore, it is relevant which castes of bees show the symptoms.

During the colony investigation, the queen is sought. It is observed whether she shows any abnormalities and whether the symptoms are equal to those of the workers. The queen does not always have the same virus load and symptoms as her colony [35].


**Table 2.** Phenotype changes in honeybees and their possible causes.

Legend: DWV = Deformed wing virus, CBPV = Chronic bee paralysis virus.

#### *3.4. Taking Samples for Laboratory Analysis*

To investigate the phenotype of bees and pathological changes in more detail and to perform a laboratory analysis, bee samples have to be taken. Most suitable are living symptomatic bees, euthanized. However, in case of increased forager mortality, freshly dead bees from the soil in front of the hive could be appropriate as well, but it has to be considered that there can be false-negative results in virus diagnostics because RNA is unstable in the environment. Different diagnostic methods require a specified number of bees, a fact that should be considered when taking the samples. To assess the size of a sample, the following reference value can be considered: 100 mL are equivalent to approximately 330 honeybees and 31 g. The required number of bees is caught and euthanized by freezing (about 15 min, −20 ◦C) [8]. Other ways of euthanizing are using 96% ethanol, carbon dioxide or sampling after asphyxiation with sulfur. The sample with dead bees is sent chilled in an air-permeable case to the laboratory. In general, it is advisable to contact the responsible laboratory for information on size and condition of the sample.

#### 3.4.1. Examination of Dead Bees

Dead bees are examined, and their phenotype is described. They can be sorted according to their size and symptoms: Bees can be smaller than normal and show shortened abdomens (Figure 4A) when infected with several viruses, e.g., Deformed Wing Virus (DWV). More indicative for DWV, associated with a severe *Varroa* infestation load, are crippled wings in freshly hatched honeybees (Figure 4B). Williams et al. describe a ranking in six categories, according to the severity of wing abnormities [36]. Hairless black abdomens are seen in case of a CBPV infection [24] or as a result of genetics, alimentation within honeydew flow period or for mechanical reasons. Robbery and fighting as well as maturing in foragers may result in breaking hairs and black abdomens, which, however, only affects individuals, whereas genetics, alimentation and CBPV affect the whole colony. If the abdomens of several bees seem to be bloated, pressure on the abdomen may lead to a light-brown fluid leaking from the gut. In such cases, the bee´s gastro-intestinal tract can be pulled out for further investigation. Often, an extended proboscis is seen (Figure 4A), which is a more unspecific sign and not necessarily evidence for an intoxication. A summary of phenotype changes is listed in Table 2 above.

**Figure 4.** (**A**) sample of dead bees of a honeybee colony infected with CBPV in 2019: inhomogeneous size of worker bees, shortened abdomen (orange arrows), extended proboscis (blue arrow); © Julia Dittes. (**B**) freshly hatched honeybee infected by DWV; © Heike Aupperle-Lellbach.

#### 3.4.2. Examination of Debris

A plastic drawer is put under the hive to sample debris. Examination of debris provides significant information about the health, development status and strength of a honeybee colony. Building materials such as wax scales, cell lids and drops of propolis, as well as cell components like pollen, sugar and melicitose crystals or drops of diluted food [37], can be found (Figure 5). The respective amounts of these components are indicators for the strength of the colony.

**Figure 5.** Photographs of a plastic drawer after three days under the hive of an *Apis mellifera carnica* colony in September with debris containing a lot of *Varroa* mites (**A**), separated legs and wings of bees (**B**), pollen (blue cycle) and wax cylinders (red cycle) (**C**). © Julia Dittes.

Furthermore, traces of predators can be seen: Parts of bees are a sign of wasps or mice in the hive (Figure 5B). Mice leave 3–8 mm long feces on the drawer. Smaller, dark brown feces belong to the greater wax moth, *Galleria mellonella*, which can also be detected by its creamy-white to grey larvae [7].

Wood chips or straw from the feeding trough or condensation water can be found, too.

Debris examination is additionally used for *Varroa* infestation control. The mite can be seen with the naked eye. Their number, counted after the drawer was under the hive for a defined time, can be used to quantify the infestation [7].

The validity of the information gathered from debris depends on the weather, the season and the time the drawer was under the hive [37]. In relation to these conditions, it constitutes an important part of examination.

#### *3.5. Laboratory Diagnosis*

#### 3.5.1. PCR to Detect Viral Diseases

This step has to be adjusted to the clinical observations in each specified case. Honeybee viruses are mainly detected by Real-Time Reverse Transcription (RT)-Polymerase Chain Reaction (PCR). In the following, the methods are described for CBPV, ABPV and DWV in the way they are performed at the Friedrich-Loeffler-Institute, the Federal Research Institute for Animal Health. A sample of 50 symptomatic bees is an appropriate sample size to submit to the laboratory.

From each bee sample, ten bees are homogenized using the gentleMACSTM Dissociator. Total RNA is purified from 150 μL of clarified bee homogenate using the RNeasy Mini Kit (Qiagen, Venlo, NL) according to the manufacturer's instructions. For each of the viruses, a one-step real-time RT-PCR is subsequently performed in duplicate using the AgPathIDTM One-Step RT-PCR Kit (Applied BiosystemsTM,, Waltham, MA, USA) in a 96-well reaction plate with 2.5 μL RNA in a final volume of 12.5 μL.

For CBPV detection, it contained 320 nM of forward and reverse primer (qCBPV 9: 5 - CGC AAG TAC GCC TTG ATA AAG AAC -3 ; qCBPV 10: 5 - ACT ACT AGA AAC TCG TCG CTT CG -3 ), and 200 nM of the CBPV probe (CBPV 2 probe: 5 - FAM- TCA AGA ACG AGA CCA CCG CCA AGT TC -BHQ1 -3 ) [38].

For ABPV detection, it comprised 800 nM of forward and reverse primer (ABPV1: 5 - CAT ATT GGC GAG CCA CTA TG -3 ; ABPVRn: 5 - CTA CCA GGT TCA AAG AAA ATT TC -3 ), and 90.4 nM of the ABPV probe (ABPVnTaq: 5 - FAM- ATA GTT AAA ACA GCT TTT CAC ACT GG -BHQ1 -3 ) [39].

For DWV-A detection, it contained 350 nM of forward and reverse primer (F-DWV\_4250: 5 - GCG GCT AAG ATT GTA AAT TG -3 ; R-DWV\_4321: 5 - GTG ACT AGC ATA ACC ATG ATT A -3 ), and 100 nM of the DWV-A probe (Pr-DWV\_4293: 5 - FAM- CCT TGA CCA GTA GAC ACA GCA TC -BHQ1 -3 ) [40]. For DWV-B detection, it contained 1.2 μM of forward and reverse primer (F-VDV1\_4218: 5 GGT CTG AAG CGA AAA TAG -3 ; R-VDV1\_4290: 5 - CTA GCA TAT CCA TGA TTA TAA AC -3 ), and 400 nM of the DWV-B probe (Pr-VDV1\_4266: 5 - FAM- CCT TGT CCA GTA GAT ACA GCA TCA CA -BHQ1 -3 ) [40].

The thermal cycling conditions are 10 min at 45 ◦C (RT = reverse transcription), 10 min at 95 ◦C (RT inactivation, initial denaturation, activation of DNA Polymerase), followed by 41 amplification cycles at 94 ◦C for 15 s and 60 ◦C for 45 s. The results are expressed as the mean of the two replicates for each reaction. Figure 6 shows an amplification diagram for evaluation. The assay is combined with an internal control assay in a duplex real-time RT-PCR—with the exception of DWV-B detection. This is carried out as a simplex approach, since another primer–probe combination interferes with the PCR approach and hinders or prevents the amplification of the virus fragment. DWV-B therefore is detected in parallel to the closely related DWV-A virus, which also helps in differentiating DWV disease. The performed internal controls are, on the one hand, a universal internal control system based on IC2-RNA [41], and, on the other hand, the detection of β-actin in the extracted samples [42].

Both internal controls monitor that the RNA extraction was successful in all samples (and extraction controls), as well as confirm their transcription into cDNA and the amplification of those during real-time RT-PCR. Furthermore, this allows to see whether a uniform effective RNA extraction took place.

**Figure 6.** Amplification of a sample of a honeybee colony investigated for CBPV and ABPV in duplicate, result: ABPV = Ct 35.23, CBPV = Ct 14.33; green = positive controls, blue = CBPV, violet = ABPV, where the sample lines cross the blue straight line, Ct is determined, © Julia Dittes.

#### 3.5.2. Monitoring the *Varroa* infestation

The *Varroa* mite is one of the main stressors, probably globally "the greatest threat" [7] for honeybee colonies. If there is no control, colonies with a high mite load are weakening until they ultimately collapse. Additionally, *Varroa destructor* serves as a mechanical and biological vector for various honeybee viruses and suppresses the immune response of honeybees [43]. Therefore, it should be monitored carefully.

To evaluate the mite load in a colony, different methods are described in the literature. It is possible to examine the debris, the brood and adult bees. Drone brood can be investigated by removing the lids of the cells and studying the larvae and combs individually (Figure 7A). In case of a larger number of samples, the brood cells can be washed out into a sieve system where mites and brood are gathered separated from each other [44].

**Figure 7.** (**A**) *Varroa* infestation control via examination of drone brood: red-eyed larva with *Varroa* mite, © Julia Dittes. (**B**) DWV-infected adult honeybee with phoretic *Varroa* mites on thorax, © Ilka Emmerich.

Adult bees play a role in assessing the mite load (Figure 7B). Examination of the natural mite fall in the debris is a common method to determine the mite load. The natural mite fall is the number of mites per day falling naturally from the bees down to the bottom board. A drawer is placed under the hive over a period of 2 to 5 days (Figure 8A). It can be lined with a sticky paper towel. The drawer is examined, and the mites are counted (Figure 8B). The oval, reddish-brown *Varroa* mites are 1.2 to 1.7 mm [7] and can easily be detected in the debris (Figure 8C). Finally, the number of mites per day has to be calculated.

**Figure 8.** *Varroa* infestation control via debris examination: (**A**) A drawer with a sticky paper is put under the hive, (**B**) counting the number of *Varroa* mites, (**C**) close-up of *Varroa mites* on the drawer. © Jens Emmerich.

Bak et al. compared the flotation and the powdered sugar shake method in 2009 with similar results for both [45]. About 300 young adult bees (is equivalent to 100 mL) are sampled from a frame with uncapped brood for either flotation or powdered sugar method. For flotation, the bees are shaken in alcohol solution for five minutes, after freezing or killing them in alcohol. The mites are removed from the bees and can be counted (Figure 9).

The second method leaves the bees alive. Dusted with icing sugar, they are shaken gently two times for two minutes. The grooming behavior of the bees is stimulated and the mites' feet do not stick any longer, thus, the mites are dislodged from the bees (Figure 10). Either the sugar is sieved out or dissolved in water, so that the separated mites can be counted [45,46]. If the mite infestation rate is under 5%, the colony is slightly affected, if the value is above 10%, an immediate treatment should be applied [7].

**Figure 10.** *Varroa* infestation control via iced sugar shaking method, (**A**) iced sugar-dusted bees in a shaking bucket, (**B**) dusted bees back in the hive after the procedure, (**C**) dusted mites on a paper towel. © Ilka Emmerich.

#### 3.5.3. Detection of Nosemosis

Fecal spots at or in honeybee hives can be a sign for nosemosis, but they do not have to be [34]. Dysentery can be caused by different pathogens such as CBPV, *Malpighamoeba mellificae* and *Nosema*species, or be a symptom of stress. In 21 examined honeybee colonies with fecal spots, only 43% could be proven to be *Nosema*-positive [34].

*Nosema*-species can occur in a honeybee colony as a covert infection without harming it. But, it can serve as an additional and weakening stressor to the honeybees. In some cases, it can be a precursor of other diseases by opening infection routes due to epithelial damage in the midgut [47].

For investigation of nosemosis, abdomens of 20 freshly dead bees are separated in a mortar (Figure 11A). They are crushed with the pestle while adding 5 mL of aqua purificata (Figure 11B). Then, water is added to form a solution of 1 mL per bee. A drop of that sample is put onto a microscope slide and covered with a coverslip without trapping air bubbles (Figure 11C). The slide is examined under the microscope at 400× magnification [48]. This solution is suitable for detection of both *Nosema* and *Malpighamoeba mellificae*.

**Figure 11.** Detection of Nosemosis: (**A**) abdomens of 20 honeybees in a mortar, (**B**) crushing the abdomens with aqua purificata, (**C**) drop of the sample on a microscopic slide. © Julia Dittes.

*Nosema* spores are oval-shaped and about 4 to 7 μm in length and 2 to 4 μm in width. The two species, *N. apis* and *N. ceranae,* cannot be distinguished by this method [7]. For further differentiation, a PCR would be required [49].

An estimation of the number of spores can be carried out using the scheme described by Ritter in 1996 [50]. The spores seen per visual field are counted and classified into three categories: Less than 20 spores per visual field signify a slight infestation, between 20 and 100 spores per visual field indicate a moderate infestation and more than 100 spores constitute a severe infestation [50]. The three grades are presented in Figure 12. For a more specific result, a hemacytometer should be used [48].

#### 3.5.4. Intoxications

If a massive number of dead bees is found in front of a hive or crippled bees are seen in the colonies, the beekeeper will often suspect a bee poisoning incident with a bee damage as a result of exposure to toxic plant protection products. However, there are some facts which have to be considered. Are all colonies in an apiary affected by increased mortality or is it only one or two? Is the apiary near fields making an application of a pesticide possible? And, are there other symptoms occurring?

In Germany, the Institute for Bee Protection in the Federal Research Centre for Cultivated Plants, Julius-Kühn-Institut (JKI), in cooperation with the national plant protection services examines samples of honeybees with suspected poisoning and plants. The bee sample undergoes a pollen analysis with pollen from the pollen basket (corbicula) or bee hairs to narrow down the incident to plants, which the bees visited. A *Nosema*-detection is performed, because *Nosema*-infected bees are more susceptible to intoxications. Finally, a bioassay with larvae of *Aedes aegyptii* L. is done to detect the presence of bee toxic insecticides. In 2018, 141 samples were sent in with the suspicion of bee intoxication, but only 61 samples were suitable for investigation. In 19 of these 61 incidents, bee toxic insecticides have been detected [51]. Further information can be found on the JKI website [52].

**Figure 12.** Microscopic investigation of bee samples of two honeybee colonies to estimate nosemosis infestation (400×), orange circles = Nosema spores (oval-shaped, 4–7 μm in length, 2–4 μm in width), upper row = visual fields, lower row = close-ups. (**A**) Sample of a colony with a moderate infestation (between 20 and 100 spores per visual field), (**B**) sample of a colony with a slight infestation (less than 20 spores per visual field), (**C**) sample of a colony with a severe infestation (more than 100 spores per visual field), © Julia Dittes.

The numbers clearly illustrate the relevance of sending in appropriate bee material. An appropriate sample for this investigation should consist of 1000 bees (is equivalent to 100 g). The sample should be taken under witness during the first 24 h after seeing symptoms and immediately sent chilled to the investigating institution [53]. The basic veterinary knowledge about correct sampling can be valuable here for taking an appropriate bee sample and, if possible, a suspicious plant sample as well as taking and storing retention samples in case of subsequent questions. A detailed photographic documentation of the incident is advisable.

#### *3.6. List of Medical Issues, Diagnoses and Prognosis*

After examinations, the next step is to list medical issues and establish a final diagnosis. According to the list of medical issues, a prognosis is derived, and a treatment plan created. An example in case of an overt CBPV infection can be found in Dittes et al. [24]. Investigations show that asymptomatic healthy colonies may have higher virus loads than diseased colonies [54,55]. Therefore, the laboratory results must precisely be linked to the information gathered by anamnesis, clinical findings and observations to evaluate the relevance of pathogens. The results form the basis for the choice of therapeutic measures.

If a notifiable disease is diagnosed, the relevant authorities have to be informed. There are notifiable diseases to theWorld Organization for Animal Health (OiE) and to the veterinary authorities in each country, depending on local regulations. The Terrestrial Animal Health Code lists infections with *Paenibacillus larvae* and *Melissococcus plutonius* and infestations with *Acarapis woodi*, *Varroa* spp., *Aethina tumida* and *Tropilaelaps* spp. in Article 1.3.8 [56]. In Germany, infections with *Paenibacillus larvae* and infestations with

*Aethina tumida* and *Tropilaelaps* spp. are notifiable. Currently, virus diseases form a part to the notifiable diseases only in Romania [7].

#### *3.7. Outcome Control*

An infected honeybee colony should be monitored closely during disease management to recognize problems and success as soon as possible and intervene in case of complications. It is advisable to use a logbook to record detailed information on every colony during disease management as well as during the whole season. Veterinarians are familiar with the management and interpretation of patient records as well as drawing conclusions out of them. A detailed logbook is important to gain basic information to choose and apply appropriate therapy concepts for disease control. The main point to prevent DWV infection and disease, as well as further *Varroa*-associated diseases, is to control the *Varroa* mite infestation. Further studies have to show recommendable therapy measures to control other viral diseases e.g., CBPV in honeybee colonies with overt infections.

#### **4. Conclusions**

A holistic approach to bee disease diagnosis is important to establish a correct and comprehensive diagnosis and save colonies. Disease outbreaks in honeybees are often tied to more complex interactions than in other species. Especially, the biological form of life in a superorganism has to be considered during the entire course of examination and management. Furthermore, the diagnostic techniques and possible ways of treatment are more limited than in the "normal" veterinary patient. There are no vaccinations and less veterinary medical products available for honeybees. The lifetime of an individual bee and the dependence on the season have to be considered.

A detailed and careful investigation is the base of diagnosis, and treatment decisions are to be made focusing on the colony as a whole rather than the individual animal. Depending on the colony location, the number of colonies in an apiary and beekeeping strategies as well as further influencing factors, diagnostics and the management plan have to be adapted to the individual requirements.

Veterinarians can be a valuable asset to the beekeeper because of their ability to prevent, diagnose and treat diseases in various species and in populations. In the future, veterinarians should recognize the important role and also the opportunities they can have within the honeybee sector. Taking the chance offered, the veterinarians may be part of education of beekeepers, part of sanitary audits and part of bee health overall in the veterinary practice, in addition to their official function in Veterinary Authorities.

**Author Contributions:** Conceptualization, C.K.W.M., H.A.-L. and I.U.E.; methodology, J.D., I.U.E. and M.O.S.; validation, J.D. and I.U.E.; formal analysis, J.D.; investigation, J.D., I.U.E. and M.O.S.; resources, C.K.W.M.; writing—original draft preparation, J.D.; writing—review and editing, I.U.E., H.A.-L., C.K.W.M. and M.O.S.; visualization, J.D., I.U.E. and H.A.-L.; supervision, C.K.W.M., I.U.E. and H.A.-L.; project administration, C.K.W.M. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest, although Heike Aupperle-Lellbach and Marc Schäfer are working in laboratories that regularly offer the laboratory diagnostics. The company (LABOKLIN GmbH & Co KG) had no role in the design, execution, interpretation or writing of the study.

#### **References**


**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Honey Bee Virus Transmission via Hive Products**

**Dominik Schittny 1, Orlando Yañez 1,\* and Peter Neumann 1,2**


Received: 27 May 2020; Accepted: 10 July 2020; Published: 21 July 2020

**Abstract:** The global trade of honey bee hive products has raised concern about pathogen transmission. However, the efficacy of hive products as virus vehicles is poorly understood. Here, we investigated the transmission capacity of hive products for Deformed wing virus genotype A (DWV-A) in a fully-crossed hoarding cage experiment and estimated the transmission risk by screening commercial products. Western honey bee workers were provided with honey, pollen and wax either contaminated with high (~2 <sup>×</sup> 109), medium (~1.7 <sup>×</sup> <sup>10</sup>8), low (~8 <sup>×</sup> <sup>10</sup>6) or zero (control) DWV-A genome copies. For 10 days, mortality was monitored. Then, virus titers were quantified in bee heads and 38 commercial products using RT-qPCR. For honey and pollen, a positive association between DWV-A concentration and mortality was observed. High concentrations always resulted in infections, medium ones in 47% of cases and low ones in 20% of cases. No significant difference was observed between the tested products. In commercial honey and pollen, 7.7 <sup>×</sup> <sup>10</sup>2–1.8 <sup>×</sup> 105 and 1.4 <sup>×</sup> 103–1.3 <sup>×</sup> <sup>10</sup><sup>4</sup> DWV-A copies per gram were found, respectively. The results show that DWV-A transmission via hive products is feasible. The risk of introducing novel viruses and/or strains should be considered in trade regulations by including virus analyses for health certificates of hive products

**Keywords:** honey bee; virus; DWV-A; hive products; honey; pollen; wax

#### **1. Introduction**

The international trade of honey bee products increased over the past decades, thereby enhancing chances for the spread of bee diseases [1,2]. Indeed, many honey bee pathogens have already been discovered in traded hive products [2]. To protect the health of humans, animals and plants, most countries joined the Terrestrial Animal Health Code (26th edition) released by the World Organization for Animal Health (OIE) in 2017, in which the trade of animal products is also regulated. However, amongst the honey bee pathogens, viruses are not covered by the terrestrial code due to a lack of specific criteria (OIE 2016), even though they are frequently associated with honey bee products and may potentially cause harmful effects [3–5]. This seems surprising because it seems most likely that viruses are also spreading as a side effect of the worldwide trade of bee products. Regarding the transmission of viruses, the international movement of live honey bees arguably plays the main role in the global dispersal of bee viruses [6], facilitating the transmission between colonies. Imported bee packages containing viral agents may act as a source of infection for other colonies in the surrounding area (horizontal transmission). The viral particles can be horizontally transmitted in different ways, such as when an infected bee drifts from its own colony to another [7], contact between bees during robbing or while foraging in common food sources [8], and also by human activity, when contaminated material is shared between colonies and apiaries [8]. The international trade of queens, in addition, allows the introduction of viral agents inside the recipient colonies. It has been shown that queens can hold many viruses at the same time and are able to transmit them vertically to their offspring [3,4,9–11]. Viruses have also been detected in honey bee products such as honey, pollen and royal jelly freshly

extracted from colonies [3–5], as well as pollen pellets recently brought to colonies [5]. Moreover, the infectivity of viruses carried by frames containing honey and pollen (bee bread) has also been shown when colonies became infected after receiving such contaminated frames [5]. Still, the role of honey bee hive products for the transmission of viruses in the trade scenario has not been explored in detail.

In addition to the risk of introducing novel viruses, novel strains of already established ones may pose additional threats. Due to genetic recombination, significant changes in the viral genome may occur resulting from the insertion of gene fragments from another viral strain during coinfection of a host cell [12]. Indeed, recombination is increasingly recognized as a major driver of virus evolution [13]. While the high mutation rates observed in RNA viruses often generate deleterious mutations [14], recombination events purge those deleterious mutations [15] and can often result in adaptations for the virus, such as expanding the host range, evasion of host immunity and changes in virulence [16,17].

Although the emergence of recombinants from deformed wing virus (DWV) genotypes were associated with increasing virulence to western honey bees, *Apis mellifera* [18–24], the potential of genetic recombination is not fully understood in the context of honey bee viruses.

Ubiquitous DWV is amongst the most harmful pathogens of honey bees [6,25–28]. It can cause clear clinical symptoms, such as crippled wings and a reduced host lifespan [27,29,30], and is a known key driver of honey bee colony losses [31,32]. DWV is a positive sense single-stranded RNA virus (family Iflaviridae; genus Iflavirus) [26,33] and is a recent global epidemic in honey bees [32]. The latter is probably driven by the ectoparasitic mite, *Varroa destructor*, because it is a very efficient vector of DWV. It generates a disease epidemic within the honey bee colony, which then dwindles until it dies [26,34,35]. *V. destructor* have also reduced the genetic diversity of DWV [36], promoting the spread and global distribution of DWV genotype A (DWV-A). The emergence of new genotypes, such as DWV genotype B (DWV-B, also known as Varroa destructor virus-1), has raised concern about their differences in virulence [22,37–39]. DWV is infecting honey bees, wild bees and probably other arthropods [5,40] (reviewed in [41]). A better knowledge of the transmission of this particular honey bee virus is therefore of importance, due to the considerable concern for both apiculture and nature conservation efforts.

In this study, we investigate whether or not honey bee hive products can, in principle, act as matrices for DWV-A transmission and how efficient they are. To achieve this goal, fully-crossed laboratory hoarding cage experiments were conducted and complemented with a survey of DWV-A titers in commercial bee products. Since royal jelly and propolis are known to have antimicrobial properties [42], our study focused on honey, pollen and wax.

#### **2. Materials and Methods**

#### *2.1. Study Set Up*

In Bern, Switzerland, twelve queenright local honey bee colonies, *A. mellifera*, were screened for DWV-A infections. Adult workers (*N* = 30) were collected from middle frames of each colony in March and April 2015, pooled and tested for DWV-A using RT-qPCR [43]. The three colonies with the lowest infection levels were chosen and tested again in June prior to the experiment, when they had a mean of 1.6 <sup>×</sup> 103, 3.9 <sup>×</sup> <sup>10</sup><sup>3</sup> and 6.1 <sup>×</sup> 103 virus copies per bee, respectively.

One sealed worker brood frame was taken from each experimental colony and placed in an incubator at 34 ◦C and 70% relative humidity (RH) until adult emergence [44]. After 48 h, freshly emerged workers were randomly distributed between the experimental hoarding cages. The fully-crossed hoarding cage experiments [44] were conducted from June to August 2015 and designed to test whether or not honey bee products spiked with DWV-A are able to induce an infection in honey bees. To see if the efficacy in DWV-A transmission is different among the tested honey bee products, three treatments with different initial amounts of DWV-A (high, medium and low) were used for each product.

Each series of experiments, corresponding to each honey bee product, consisted of four treatments: three treatments where the honey bee product had been spiked with three different concentrations of DWV-A (5 <sup>×</sup> 109/5 <sup>×</sup> 108/5 <sup>×</sup> 10<sup>6</sup> copies per ml for honey; 1 <sup>×</sup> 109/1 <sup>×</sup> 107/1 <sup>×</sup> 106 copies per g for pollen; 2.5 <sup>×</sup> 108/2.5 <sup>×</sup> <sup>10</sup>6/2.5 <sup>×</sup> <sup>10</sup><sup>5</sup> copies per cm<sup>2</sup> for wax; Table 1) and a non-spiked treatment with UV-sterilized products as control. All treatments consisted of five repetitions. Each cage was equipped with a honey solution feeder (3 mL syringe), a pollen paste feeder (modified centrifugation tube) and a piece of wax foundation (4 cm2). Then, 30 newly emerged workers were introduced into each cage and kept for 10 days at 30 ◦C and 70% RH in the incubator. Dead individuals were removed from the cages daily, recorded and stored at −20 ◦C. Honey and pollen consumption was also controlled daily and feeders refilled if required. After two days, the spiked honey and pollen products have been consumed in all cases, resulting in an average consumption of 1 <sup>×</sup> 108 and 3.3 <sup>×</sup> <sup>10</sup><sup>7</sup> copies per bee for high, 1 <sup>×</sup> 10<sup>7</sup> and 3.3 <sup>×</sup> 105 copies per bee for medium and 1 <sup>×</sup> 10<sup>5</sup> and 3.3 <sup>×</sup> 104 copies per bee for low DWV treatment, respectively. Then the feeders were replaced with sterilized food. For the wax product, the spiked piece of wax was available for the entire 10 days. At day 10, all remaining bees were stored at −20 ◦C.

**Table 1.** Initial Deformed wing virus genotype A (DWV-A) genome copy numbers in the different treatments. The number of DWV-A copies per contaminated honey, pollen and wax are given. For comparison, the average number of DWV-A genome copies consumed per bee individual (virus copies per cage divided by number of bees in cage) is also listed.


#### *2.2. Cage Experiment*

#### 2.2.1. Bee Product Preparation

Only honey and pollen that tested negative for DWV-A by RT-qPCR [43] was used for the experiments. Additionally, these were irradiated with UV light for 120 min. During the UV-treatment, the honey (25 g) was mixed by slowly rotating the honey containing tube each 30 min. A 50% (*w*/*w*) honey solution was prepared using the UV light-treated honey and Milli-Q water (Millipore Corporation, Billerica, MA, USA). The solution was mixed and stored at −20 ◦C until used in the experiments.

The pollen grains were crushed to a powder using a stone mortar. The grained pollen was spread on a sheet of paper and irradiated with UV light for 120 min. Every 30 min, the pollen powder was mixed using a sterilized spatula. The pollen paste was prepared with the following proportions: 40% UV-light treated pollen, 50% powder sugar and 10% MilliQ water. The pollen paste was then wrapped into aluminum foil and frozen at −20 ◦C until usage.

The wax was provided as small pieces of organic wax foundation, which was cut in square pieces with an edge length of 20 mm. The pieces of foundation were displayed on a sheet of paper and each side was irradiated with UV light for 30 min. After irradiation, the wax pieces were stored at −20 ◦C.

#### 2.2.2. Propagation of DWV-A

DWV-A was propagated using standard methods [45]. Red-eyed worker pupae were microinjected with 2 <sup>×</sup> <sup>10</sup><sup>7</sup> virus copies in 2 <sup>μ</sup>L PBS solution (Phosphate Buffered Saline; pH 7.4) between the 2nd and 3rd integuments using a 50 μL micro syringe (Hamilton Microliter™ Syringes, Reno, Nevada, USA) and 30-gauge disposable needles. Pupae were incubated at 30 ◦C and collected after 6 days. Each pupa was macerated individually in 500 μL PBS and homogenized with 100 μL chloroform using strong vortex. After centrifugation at 13,000 rpm for 10 min, the supernatant was collected and stored at −20 ◦C. DWV-A was quantified using RT-qPCR [43] and diluted to 1 <sup>×</sup> 107 copies per <sup>μ</sup>L to make a stock solution for the spiking of the bee products.

#### 2.2.3. Spiking Bee Products with DWV-A

In a pilot study, the average amount of honey and pollen consumed by a single bee per day was estimated. On average, one bee consumed 20+/−3.1 mg of honey solution and 16+/−3.6 mg of pollen paste per day. The DWV-A concentration in the bee products was estimated to apply the required number of virus copies within the first two days of the experiment for honey and pollen.

The honey solution was spiked according to the different DWV-A concentrations, keeping the ratio of honey and water 1:1 (*w*/*w*). In the case of pollen, the DWV-A solution replaced the water in the recipe for the pollen paste (10%). A homogenous spiking of wax was not possible because the virus RNA would degrade at temperatures where the wax would melt. Therefore, the virus solution was applied on the surface of the pieces of wax foundation and allowed to rest overnight for the water to evaporate. Since bees do not consume wax, the desired virus amount was applied on the piece of wax foundation, which was left in the cage for the entire duration of the experiment.

#### 2.2.4. Cage Construction

The cages were clear polystyrene cups with a diameter of 63 mm and an inner volume of 75 cm3 (RIWISA AG Kunststoffwerke, Hägglingen, Switzerland). The cups were turned upside down with the lid acting as the bottom of the cage. Each cage was fitted with three holes for ventilation (large hole in lid covered with a mesh fabric permeable to air) as well as for holding the pollen (2 mL micro-centrifuge tubes) and honey (2 mL plastic syringes) feeders.

#### 2.2.5. Detection of DWV-A

To test for overt virus infection, only bee heads (*N* = 10 per cage) were considered for analyses [46]. The heads were removed using a scalpel sterilized using EtOH and flaming after each cut. Heads from the same treatment were homogenized individually in 100 μL TN-Buffer (10 mM Tris HCl, 10 mM NaCl) using a metal bead (5 mm diameter) and an electronic crushing shaker machine (Retsch Mixer Mill MM 300, Haan, Germany). A NucleoSpin® RNA II kit (Macherey-Nagel, Oensingen, Switzerland) was used for RNA extraction following the manufacturer's recommendations and using 50 μL of pooled bee head homogenates from the same cage. The purified RNA was then eluated using 60 μL of RNase free water (Macherey-Nagel, Oensingen, Switzerland).

Reverse transcription was performed using standard protocols [32]. The concentration of RNA was measured using a spectrophotometer (Witec NanoDrop® ND 1000 Spectrophotometer, Sursee, Switzerland). Then, 1 μg of RNA and 0.75 μL of 100 mM hexamer primer (Microsynth AG, Balgach, Switzerland) were heated at 70 ◦C for 5 min and then cooled down to 4 ◦C. To obtain a final volume of 25 μL, a master mix, consisting of 5 μL M-MLV Reaction Buffer (Promega, Fitchburg, Wisconsin, USA), 1.25 μL 2.5 mM dNTP Mix (Bioline, London, UK) and 1 μL 200 u/μL M-MLV reverse transcriptase (Promega, Fitchburg, Wisconsin, USA), was added and heated to 37 ◦C for 60 min, before cooling down to 4 ◦C. A total of 10 fold dilutions were used for the quantification assays.

Each sample was tested for DWV-A using quantitative PCR (qPCR) [32]. The qPCR was conducted using the KAPA SYBR® Fast Universal qPCR kit (KAPA Biosystems, Wilmington, Massachusetts, USA). Briefly, a 12 μL total volume reaction consist of 6 μL of 2x reaction buffer, 0.24 μL forward and reverse primers each (10 μM, Table 2), 2.52 μL water and 3 μL of cDNA. The qPCR was run using an Illumina® Eco Real-Time PCR System (Illumina, San Diego, CA, USA. The amplification conditions were initiated by heating to 95 ◦C for three minutes in order to activate the polymerase. Then, during each of 40 repeating cycles, the samples were heated to 95 ◦C for three seconds and cooled down to 57

◦C for 30 s. For melting, curve analysis samples were heated to 95 ◦C for 15 s, cooled down to 55 ◦C for another 15 s and heated from 55 ◦C to 95 ◦C while the strand dissociation was recorded.


**Table 2.** PCR primers used for the relative virus quantification of DWV-A.

Standard curves prepared from DWV-A and *A. mellifera* β-Actin gene were used for virus quantification and normalization, respectively. The standard curve dilutions (10−<sup>2</sup> to 10−<sup>5</sup> ng/reaction) were prepared from purified PCR products (Table 2). Two kinds of negative controls were applied (1: RNA-extraction control without bee sample to check for possible contamination in the reagents; 2: PCR negative control, using water instead of cDNA template).

#### *2.3. Survey of DWV-A in Commercial Honey and Pollen Products*

Honey (*N* = 34) and pollen (*N* = 5) products were acquired from a variety of local Swiss grocery stores. The honeys originated from all continents except Antarctica and the pollen originated from Spain. For extracting RNA from honey and pollen, 120 mg of these bee products and 200 μL of TN-Buffer were mixed thoroughly by using a shaker (Retsch Mixer Mill MM 300, Haan, Germany) and a metal bead (2 mm diameter) that was put inside the sample tube. After shaking, the sample was centrifuged for 5 min at 14,000 rpm. 50 μL of the supernatant were then used for the RNA extraction following the NucleoSpin® RNA II kit (Macherey-Nagel, Oensingen, Switzerland) protocol. Reverse transcription as well as qPCR [32] were conducted as described above.

#### *2.4. Statistical Analyses*

The statistical analyses were performed using the NCSS statistical software version 10. Since the Kolmogorov-Smirnov test rejected normality in all the used data sets (Test value = 0.289 for virus copies data, 0.305 for delta DWV-A data, 0.214 for cage infection data and 0.246 for mortality data), non-parametric tests were used. The sample size was the number of cages per treatment (*N* = 5) because considering each bee individually would result in pseudo replications [39]. For all tests a critical *p*-value of 0.05 was used.

Using the log rank test with a Bonferroni correction, we checked each bee product experiment for significant differences in mortality between treatments of different DWV-A concentrations. Bees from all five cages underlying the same treatment were integrated for the survival analyses. There were 150 bees per treatment and the survival was recorded for 10 days.

The Kruskal-Wallis multiple-comparison z-value test with Bonferroni correction was used to test for differences in DWV-A titers between high, medium and low DWV-A treatments as well as control treatments with sterilized products. As before, all bees from the same treatment were integrated to one population.

To compare the DWV-A transmission efficacy between different honey bee products a Δ DWV-A level was calculated. The average number of DWV-A copies taken up by the bees during the experiment was subtracted from the detected DWV-A copies in the heads. Thus the Δ-DWV-A value was a measure for DWV-A replication within the bees. For the analysis, all bees receiving virus copies via the same honey bee product were integrated. In addition, here the Kruskal-Wallis multiple-comparison *z*-value with Bonferroni correction was used.

#### **3. Results**

#### *3.1. Cage Experiment*

#### 3.1.1. Survival

With the exception of one outlier, death rates of less than 15% mortality during the ten-day experiments were observed in all control cages. In the case of the honey assays, the mortality was significantly higher at the high DWV-A treatment compared to all lower DWV-A concentrations and the control (Log Rank test, *p* < 0.0001 per each pair-wise comparison). In the pollen assays, the control cages had significantly lower death rates than those of high (*p* < 0.0001) and medium (*p* = 0.0005) DWV-A treatment. Concerning wax, no significant differences in mortality between different treatments were found (Figure 1).

**Figure 1.** Kaplan-Meier survival plots of caged honey bee workers over the experimental period for each bee product ((**A**) honey, (**B**) pollen and (**C**) wax) with high, medium and low initial DWV-A concentrations. Survival for each treatment was pooled from 5 cages with 30 bees each. Significant differences are indicated by different letters (a,b).

#### 3.1.2. DWV-A Infection Levels

While the Kruskal-Wallis test indicated significantly higher DWV-A titers in the heads of bees from high DWV-A treatments compared to bees from control treatments in honey (*z* = 2.83), pollen (*z* = 3.64) and wax (*z* = 2.89), the high DWV-A treatments also showed significantly higher DWV-A titers compared to the low DWV-A treatments in the case of honey (*z* = 2.78) and pollen (z = 2.83) treatments. The critical significance level of the z-value, concerning Bonferroni correction, was 2.64 with the *p*-value set at 0.05. As in the survival curves, the correlation between initial DWV-A concentration and mortality or virus titers, respectively, was the strongest in pollen and the weakest in wax (Figure 2).

According to the detected amount of DWV-A in their heads, a significant bimodal distribution (Kolmogorov-Smirnov normality, test value = 0.289), was found (Figure 3). The group with the lower DWV-A showed between 5 <sup>×</sup> 103 and 5 <sup>×</sup> 105 (median = 2.2 <sup>×</sup> 104) copies per bee, while the group with the higher DWV-A showed between 1 <sup>×</sup> 109 and 2 <sup>×</sup> 10<sup>11</sup> (median = 2.0 <sup>×</sup> 1010) copies per bee. Bees from the higher DWV-A group had significantly more viruses (between one to five orders of magnitude) than the initial fed amount (from 3.3 <sup>×</sup> 10<sup>4</sup> to 1 <sup>×</sup> 10<sup>8</sup> copies per bee; Kruskal-Wallis test, *z*-value = 5.65). Therefore, virus replication can be considered to occur in those bees.

Regarding the frequency of infection, the number of bee cages that showed an infection as defined above was different depending on the treatment. High DWV-A treatment resulted in an infection of all cages independent of the bee product. At the medium DWV-A treatment, there were two out of five infected cages in honey and pollen each, while there were three in wax. The largest difference was seen at low virus concentrations. There was one infected cage in the case of honey, two in the case of wax while no cages were infected in the case of pollen.

**Figure 2.** DWV-A copies detected per bee head in the different treatments ((**A**) honey, (**B**) pollen and (**C**) wax). Medians, upper and lower quartiles and maximum and minimum are shown. Significant differences are indicated by different letters (a,b).

**Figure 3.** Frequency distribution of DWV-A copy numbers per bee head, all tested hive products pooled together. The highest amount of initially provided DWV-A during the experiment was 1.0 <sup>×</sup> 10<sup>8</sup> copies per bee (red line). There is a significant bimodal distribution (Kolmogorov-Smirnov normality, test value = 0.289) with one group showing high DWV-A titers and another group showing low DWV-A titers. The virus titers of the two groups are significantly different (Kruskal-Wallis test, *z*-value = 7.8444 with *p*-value set at 0.05) and there was significantly more virus detected in the high virus titer group than was initially provided during the experiment (*z*-value = 5.6537 with *p*-value set at 0.05). This implies that bees with a high DWV-A titer got infected and that virus replication took place. The box plots at the lower part of the figure show differences between the detected virus titers of different treatments.

The comparison between the different bee products showed no significant differences (*z* = 0.38 for honey and pollen, 1.88 for honey and wax and 1.50 for pollen and wax) in virus titers between the different honey bee products (Figure 4).

**Figure 4.** Quantification of replicative DWV-A in bees exposed to DWV-A-spiked honey, pollen and wax. Δ-DWV-A represents the initial amount of DWV-A provided in the products subtracted from the detected DWV-A amount in the bee heads, as a measure for DWV-A replication. For each box, data from high, medium and low initial DWV-A were pooled. There were no significant differences among the bee products.

#### *3.2. Survey of DWV-A in Commercial Honey and Pollen*

In all honey and pollen samples, the detected amount of DWV-A was low. In honey, between 7.6 <sup>×</sup> 102 and 1.8 <sup>×</sup> 105 virus copies with a median of 1.2 <sup>×</sup> 104 virus copies per gram were detected. In pollen, between 1.4 <sup>×</sup> 103 and 1.2 <sup>×</sup> <sup>10</sup><sup>4</sup> virus copies with a median of 3.5 <sup>×</sup> <sup>10</sup><sup>3</sup> virus copies per gram were detected (Figure 5).

**Figure 5.** Survey results for commercial honey and pollen. DWV-A copies per mg in 34 commercial honey and 5 pollen samples are shown. Even though a few honey samples showed higher DWV-A titers, no significant difference between the two products was found (Kruskal-Wallis multiple-comparison *z*-value test, *z*-value = 1.9324, *p* > 0.05).

#### **4. Discussion**

Our results clearly show that DWV-A transmission via hive products is feasible. The data also show that mortality increases when honey bees are fed with higher titers of DWV-A via honey and pollen, but not via wax. DWV-A infection was detected in all cages from the high treatment (fed with at least 3.3 <sup>×</sup> 107 copies per bee), irrespective of the tested bee product. Only very small amounts of DWV-A were detected in the commercial bee products.

#### *4.1. DWV-A Transmission Experiment*

#### 4.1.1. Survival and DWV-A Titers in Caged Bees

The survival plot shows that only the high DWV-A treatment affects mortality in the honey treatments. Moreover, this treatment resulted in high DWV-A infections in all cages. This suggests an association between mortality and DWV-A infection by oral consumption of high DWV-A titers (1.0 <sup>×</sup> 108 virus copies per bee) via honey. In contrast, mortality was not different from the controls at the medium and low treatments, despite the occurrence of high DWV-A titers in some of those cages. Similar to honey, the high DWV-A treatment affects mortality in the pollen assays. Moreover, mortality in the medium treatment was also higher than in controls. On the other hand, no significant difference in mortality was found between the low treatment and the control, which is consistent with the low DWV-A titers detected in the bee heads from the pollen low treatment.

In the case of wax assays, no differences in mortality were found among different DWV-A treatments and controls. One possible reason could be that wax does not serve as a food resource, so the oral pathway is somehow different in comparison to honey and pollen.

#### 4.1.2. Data Structure of Detected DWV-A in Bee Heads

Looking at the detected DWV-A titers in the bee heads, there was one pattern that could be observed across all bee products and treatments, including the controls. All bees could be divided into either a high (median <sup>=</sup> 2.0 <sup>×</sup> 1010 copies per bee) or a low (median <sup>=</sup> 2.2 <sup>×</sup> <sup>10</sup><sup>4</sup> copies per bee) DWV-A titer group (Figure 3). There was a distinct gap between 5 <sup>×</sup> 105 and 1 <sup>×</sup> 109 DWV-A copies per bee, with not a single sample in this range. The highest DWV-A amount fed in the experiment was 1 <sup>×</sup> 108 copies per bee considering that each bee had fed the same (average) amount of spiked bee product. If that assumption is true, all bees from the group with high detected virus titers (>1 <sup>×</sup> 10<sup>9</sup> copies/bee) had at least one order of magnitude more virus copies in their heads than the maximal amount of virus that they had taken up during the experiment. Therefore, all bees with a DWV-A amount of at least 1 <sup>×</sup> <sup>10</sup><sup>9</sup> copies can be considered as having experienced DWV-A replication and thus an overt infection. In contrast, there were samples in the group of low detected DWV-A titers that showed considerably less DWV-A copies than have been fed, indicating that infection does not always occur. Since up to 5 × 10<sup>5</sup> DWV-A copies per bee were found even in bees from the negative control, these virus titers are likely to represent DWV-A cover infection levels.

#### 4.1.3. Product Comparison

The honey treatments were spiked with more virus copies compared to pollen and wax, because the different physical properties of the hive products only enabled a certain maximum of treatment solution to be absorbed. Hence, a direct comparison of the tested hive products for viral titers was not feasible. Instead, we compared the DWV-A transmission efficacy between the products using a Δ-DWV-A level as a measure for virus replication within the bees. A significant difference of transmission efficacy between honey, pollen and wax was not found. Even though the obtained results from different bee products are mostly similar, there might be a trend that wax transmits DWV-A easier than honey and pollen. This can be seen in Figure 4, where the median of wax is the highest. That implies that there was more virus replication in the wax treatments, compared to the two other bee products. Indeed, the high levels of the detected virus in the low DWV-A treatment of wax were not significantly different from the high DWV-A treatment, which stands in contrast to the observations of the other products. This might be a clue for a lower DWV-A threshold for an infection via wax. Another indicator can be found when looking at the number of individual cages in which DWV-A replication occurred. In the case of wax, replication took place in two out of five cages in low DWV-A treatment and in three out of five cages in medium DWV-A treatment (number of infected cages: honey low = 1; pollen low = 0; honey medium = 2; pollen medium = 2). However, the impression that wax might be the most efficient matrix of all tested products could also have a methodical reason: in contrast to the honey and pollen assays, where bees consumed the spiked products during 48 h, the spiked pieces of wax were in contact with the bees for 10 days. The amount of virus was similar in the wax compared to the pollen and honey assays, but the virus was highly concentrated on the surface due to an inhomogeneous DWV-A distribution. Apart from that, a different transmission route, such as topical transmission [49], may produce different effects in the host parasite interactions since wax does not serve as food for honey bees.

#### *4.2. Survey of DWV-A in Commercial Honey and Pollen*

Detected amounts of DWV-A in commercially available honey and pollen were relatively low (1.8 <sup>×</sup> 10<sup>5</sup> and 1.2 <sup>×</sup> 104 copies per gram, the highest values, respectively). However, Mazzei and colleagues [50] found up to 3.0 <sup>×</sup> 106 virus copies per gram in pollen samples freshly collected by honey bee workers. This amount is 230-fold higher compared to our highest value in pollen. One reason for this difference could be the quick degradation of DWV-A particles at room temperature [51]. In the same line, Graystock and colleagues [52] only detected DWV-A in 2 out of 25 samples of pollen provided as food for bumblebee colonies.

#### Transmission Risk Under Realistic Conditions

Overall, the transmission risk of DWV-A via bee products under realistic conditions is not very high. Nevertheless, it is possible and should be considered. In the case of pollen, the cage experiment shows that a transmission and infection, defined by detection of high DWV-A titers, can only take place if the supplied DWV-A concentration lies between 1 <sup>×</sup> <sup>10</sup><sup>6</sup> and 1 <sup>×</sup> 107 copies per gram. However, in all samples tested in the survey of commercial pollen none reached a concentration in that range. Therefore, the risk of DWV-A to produce infection when transmitted via pollen is considered minimal.

By contrast, honey appears more efficient as a potential matrix for DWV-A transmission. With one out of five cages infected at a concentration of 5 <sup>×</sup> <sup>10</sup><sup>6</sup> virus copies per milliliter, a transmission is possible even at lower concentrations. Since we had no test point at lower concentrations, the minimal infectious dose of DWV-A via honey was not clearly defined. However, there were 3 out of the 33 honey samples that showed concentrations of more than 1 <sup>×</sup> 105 virus copies per milliliter in the commercial honey survey. Thus, those three samples were very close to the concentration that induced infection in one out of five cages. Based on these data, the risk of DWV-A transmission via honey under realistic conditions is 3/33 × 5 which is around 2% with the transmission threshold set at 1 × 10<sup>5</sup> DWV-A copies per milliliter. It is known that viral particles exposed to environmental conditions (e.g., humidity and heat) may quickly deteriorate [51,53,54]. This variable has not been thoroughly tested in this study. Nevertheless, it has been shown that viruses contained in frames with honey and bee bread remain infective even after been stored by six months at room temperature [5]. Since honey and bee bread stored in frames may represent a different kind of matrices than commercial processed honey and pollen products, data regarding the infectability of DWV in such commercial processed hive products stored for longer periods of time are needed. Until then, even though the estimated infection risk of 2% appears to be rather small, it is only a question of time before an infection occurs, especially if bees consume large amounts of imported honey. Moritz and Erler [55] observed a positive correlation between the amount of honey produced per colony and the amount of honey imports. Thus, they are considering the possibility of relabeling imported honey to locally produced honey. This illegal relabeling might be disguised by feeding the imported honey to local bee colonies. In that case, local bees would consume huge amounts of imported and possibly contaminated honey, which would increase the chance of infection dramatically.

For wax as a matrix, we have no data for realistic DWV-A concentrations of commercial products. That makes it impossible to evaluate a realistic transmission risk. However, wax is often heated during commercial processing and this may inactivate viruses, thereby decreasing the role of wax foundation as a virus carrier in the global trade scenario. By contrast, wax showed some transmission potential in the cage experiments and should thus also be considered as a potentially effective DWV-A vehicle, e.g., as a result of the common apicultural practice of exchanging non-processed combs between colonies.

As DWV is ubiquitous in most continents, a major concern should be the introduction of novel, potentially more virulent strains. Genetic recombination between different strains may indeed result in new viral variants with enhanced virulence. For example, distribution of DWV-B (also known as VDV-1) has expanded over the last decade [56] and recombination events with DWV-A appear to be frequent [18–23]. Moreover, the selection of a virulent recombinant variant between DWV-A and DWV-B has been already reported [23]. However, due to the potent role of *V. destructor* as a virus vector [35], it appears as if a recombinant DWV variant with lower virulence has been selected to optimize its transmission by *V. destructor* [57]. Therefore, more studies are necessary to assess the real impact of DWV genetic recombination on the emergence of potentially more virulent strains.

Risk is composed of the probability of occurrence and the damage caused if that undesired event actually occurs [58]. Therefore, the potential of hive products for virus transmission reported here should be considered, even though the probability of occurrence is probably rather low. Indeed, the potential damage resulting from novel viral strains could obviously be severe. Therefore, the potential of hive products for virus transmission should be considered in daily beekeeping practice. For example, it appears very risky to feed cheap honey to colonies, which is unfortunately a common practice of

some beekeepers. A sustainable way to limit the risk effectively would be to include virus analyses in health certificates for import and export of hive products.

#### **5. Conclusions**

Our data clearly show that transmission of DWV-A via hive products is feasible. Since there are currently no regulations aiming to limit the spread of viruses due to global trade, the results are calling for respective mitigation measures, i.e., health certificates for hive products.

**Author Contributions:** Conceptualization, O.Y. and P.N.; methodology, D.S. and O.Y.; formal analysis, D.S.; resources, O.Y. and P.N.; data curation, D.S.; writing—original draft preparation, D.S.; writing—review and editing, D.S., O.Y. and P.N.; supervision, O.Y. and P.N.; project administration, P.N.; funding acquisition, P.N. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Swiss Federal Food Safety and Veterinary Office (FSVO; project number 1.15.01) and the Vinetum Foundation.

**Acknowledgments:** Appreciation is given to Kaspar Roth for technical support.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
