Varroa destructor feeds primarily on honey bee fat body tissue and not hemolymph

Abstract

The parasitic mite Varroa destructor is the greatest single driver of the global honey bee health decline. Better understanding of the association of this parasite and its host is critical to developing sustainable management practices. Our work shows that this parasite is not consuming hemolymph, as has been the accepted view, but damages host bees by consuming fat body, a tissue roughly analogous to the mammalian liver. Both hemolymph and fat body in honey bees were marked with fluorescent biostains. The fluorescence profile in the guts of mites allowed to feed on these bees was very different from that of the hemolymph of the host bee but consistently matched the fluorescence profile unique to the fat body. Via transmission electron microscopy, we observed externally digested fat body tissue in the wounds of parasitized bees. Mites in their reproductive phase were then fed a diet composed of one or both tissues. Mites fed hemolymph showed fitness metrics no different from the starved control. Mites fed fat body survived longer and produced more eggs than those fed hemolymph, suggesting that fat body is integral to their diet when feeding on brood as well. Collectively, these findings strongly suggest that Varroa are exploiting the fat body as their primary source of sustenance: a tissue integral to proper immune function, pesticide detoxification, overwinter survival, and several other essential processes in healthy bees. These findings underscore a need to revisit our understanding of this parasite and its impacts, both direct and indirect, on honey bee health.

Keywords: Varroa; apiculture; fat body; honey bee; insect physiology.

Fig. 1.

V. destructor shows consistent preference for the underside of the metasoma of adult host bees, an area predominated by fat body tissue just beneath the cuticle. (Left) Diagram showing frequency of Varroa found in each location on 104 parasitized worker bees in five trials (st, sternite; tg, tergite). Varroa were found on the underside of metasoma as opposed to locations on the mesosoma or head (generalized linear model, GLM: Χ₂² = 6.5, P < 0.001). Mites strongly preferred the third segment of the metasoma to any of the other 23 locations (GLM: Χ₂² = 4.5, P < 0.001). Mites were also found preferentially on the Left side of the host (χ): χ² = 24.02, P < 0.001.

Fig. 2.

Feeding site of Varroa on adult honey bee imaged via low-temperature scanning electron microscopy. Images representative of 10 worker bees with attached mites prepared for imaging of which all 10 showed a wound in the intersegmental membrane. (A–F) Representative images of 10 bees parasitized by Varroa. Location of the mite shown with white arrow (A). The mite is wedged beneath the third tergite of the metasoma (B). When removed, a detailed impression of the mite can be observed in the intersegmental membrane in addition to a wound where the mouthparts of the mite would be (black arrow in C). Note, the ambulacra, or foot pads, of the mite (white arrows) remained attached to the membrane when the mite was extracted (C and D). Higher magnification of the wound reveals distinct grooves in the wound matching the modified chelicera of the mouthparts of the mite, colorized for clarity [moveable digit (yellow), corniculus (green)] (F). (A and C) Reproduced with permission from ref. 47.

Fig. 3.

Varroa briefly parasitizing worker bees were used to pinpoint the precise location of the feeding site, revealing the ultrastructural morphology of the feeding wound, bacteria at the feeding site, and damage to the fat body after only hours of association with the host bee. Images captured via transmission electron microscopy. (A and B) Histological cross-section of a worker bee with Varroa attached between the third and fourth segments of the metasoma. Fat body tissue is shown beneath the intersegmental membrane (A). The wound caused by the feeding mite in the membrane of the bee is clearly visible as a large mound with a hole intersecting the membrane (arrowhead indicating the hole) (B). FB, fat body; M, intersegmental membrane; Mu, muscle tissue; St, sternite; Te, tergite; V, Varroa. (C and D) Feeding wound at higher magnification, showing a hole with irregular edges where mouthparts of the mite have penetrated the membrane (arrowhead). The black arrow indicates bacteria at the feeding site, and the white arrow indicates exposed contents of fat body cells likely due to the extraoral digestive processes of the feeding mite (C). Higher magnification reveals further detail distinguishing two morphologically distinct bacteria (D). (E and F) Higher magnification of degraded cell contents (E). Remnants of condensed chromatin from the nuclei of cells can be observed (blue arrowheads) in addition to the crystalline lipid droplets found abundantly in fat body trophocytes (red arrowheads) (F). (A and C) Reproduced with permission from ref. 47.

Fig. 4.

Host tissue collected from bees with fluorescently stained internal tissues showing the fidelity of each biostain, Nile red (lipophilic) and Uranine (hydrophilic), for each host tissue (fat body and hemolymph, respectively). Varroa are shown as well. Columns show honey bee tissue and a Varroa specimen fed on a biostained host bee in brightfield with successive columns showing fluorescence from these samples associated with Uranine, Nile red, and the two biostains imaged together. All scale bars represent 1 mm. (A–D) Gut tissue (A). Note: the high degree of fluorescence shown by both biostains (B–D). (E–H) Hemolymph tissue (E). Note: Uranine biostain shows high biochemical affinity for hemolymph (F and H). Barely discernible levels of Nile red (G) are likely a result of this biostain reacting to circulating lipophorin and cells present in the hemolymph. (I–L) Fat body tissue (I). Note: Nile red biostain shows high biochemical affinity for fat body (K and L). There are small amount of hemolymph present throughout fat body tissue that likely contributes to the low levels of Uranine fluorescence visible (J). (M–P) Photochemically cleared Varroa female (M). Note: very little fluorescence associated with the hemolymph can be seen (N and P); however, fat body fluorescence is intense with signal emanating primarily from the lobes of the digestive system (O and P).

Fig. 5.

Mean fluorophore levels detected in Varroa and in biostained honey bee tissues. Fluorescence values reported in arbitrary fluorescence units (AFU). (A–C) Mean fluorophore levels detected in Varroa after 24 h of exposure to stained host bees (A). Levels of the fat body fluorophore (Nile red) are higher than that of the hemolymph (Uranine) fluorophore and appear in the same proportion as in the fat body of the host bees (B). Proportion test (prop test): χ² = 3.62e-28, P = 1, n = 10. This proportion differs significantly from that of the hemolymph of these bees (C), providing further evidence that mites are not consuming this tissue in significant amounts (prop test: χ² = 197.33, P < 0.001, n = 10).

Fig. 6.

Varroa were fed on nurse bees given only one of the two fluorescent biostains and imaged via confocal laser scanning microscopy, verifying that low signal associated with hemolymph was not a result of the fat body fluorophore (Nile red) obscuring the fluorescence of the hemolymph fluorophore (Uranine). Mites that fed exclusively on bees with biostained hemolymph (A) showed fluorescence only marginally above the control (B). Mites fed on bees with fluorescently stained fat body show such high levels of Nile red in the digestive system that the shape of the gut can be clearly observed via fluorescence imaging (C). (C) Reproduced with permission from ref. 47.

Fig. 7.

Mites fed honey bee fat body tissue survived longer and produced more eggs than mites provisioned with hemolymph. High mortality was observed across treatments, likely because of the artificial setting. After 3 d, mites receiving 0%:100% and 25%:75% hemolymph:fat body as their diet maintained survivorship at 60%, while the 100%:0% hemolymph:fat body and the starvation control had already exhibited full mortality. Final sample size consisted of 15 mites per treatment. (A and B) Survivorship curve showing starvation control and all five host tissue diets (A). (B) Representation of the same data with levels combined that show no difference in survivorship. Note: mites provisioned hemolymph and mites given no food showed no difference in survivorship. However, survivorship differed substantially between the hemolymph treatment and all treatments given any level of fat body (χ² = 16.1, P < 0.001). (C) Egg production differed between treatment diets (ANOVA: P < 0.004). A positive linear relationship was observed between egg production and the amount of fat body in the diet of the mite (R² = 0.7894). (D) Average survivorship of mites differed by diet. Survivorship and the ratio of fat body by volume adhere to a strong positive linear relationship (R² = 0.9634).