Engineered bacteria boosting honey bee health – a silver bullet?

The finding that engineered bacteria could help against varroa and the Deformed Wing Virus (DWV) made a big buzz. These bacteria, Snodgrassella alvi, live naturally in the bee gut. Leonard et al. modified them to produce double-stranded RNA (dsRNA), which should interfere with crucial viral and varroa genes. Sounds crazy. But let’s go step by step.

The research on the bacteria in the bee gut has advanced quite a bit in the past years. However, to my knowledge until now there weren’t any practical applications of this research. On the other hand, Garbian et al. attempted to use the RNA interference (RNAi) against varroa already in 2012. Yet, the production of dsRNA is expensive and it degrades rapidly in the environment. That’s why Leonard and colleagues chose to take the “detour” via S. alvi. The bacteria would produce the dsRNA in the bee gut. By this, they hoped that the distribution within the colony would be more effective as well.

Feeding bees with engineered bacteria

S. alvi was modified with plasmids, small circular pieces of genetic information. These plasmids are used in “classic” genetic engineering for transferring genes between organisms. These techniques aren’t very popular in Europe, but this is a different story. I’ll come back to this later, however. The approach is very simple: the researchers fed the bacteria to honey bee workers. In these lab experiments, the bacteria produced dsRNA in the bee gut and the molecules distributed throughout the whole body of the bees. Most importantly, the DWV titers were lower in bees which fed the engineered bacteria compared to the untreated control. In addition, the researchers introduced varroa mites to the lab cages and found that the mites died earlier on treated bees.

Finally, the dsRNA activated also some immune genes in the bees and this effect was persistent during the 15 days of the trial. Leonard et al. expect that the molecules will be transmitted between the bees in the colony. However, data on this are not available yet. This is the point some researchers criticize the experiments. Professor Randolf Menzel from the University in Berlin says: “At the moment we only have limited data from lab experiments with a few bees. We don’t know if this approach will have sufficient efficacy against varroa and virus infections in full colonies.” In addition, it’s not clear if the dsRNA is specific. This is an important point: the authors mention effects on the bee physiology and behaviour. If this method proofs efficient in full colonies, this would be a point of special interest during clinical studies in the registration process.

Ecological consequences of using engineered bacteria unknown

In addition to these questions, Menzel is also sceptical about the environmental security of this approach. “We don’t know if these bacteria are able to survive outside the bee gut and if other insects are hosts for them as well.”, he says. Professor Robert Paxton from the German Center of integrated Biodiversity Research is less concerned about the bacteria themselves but about the plasmids. Bacteria use plasmids to exchange genetic information – also between different species. “Before using this promising approach we need large empirical studies to make sure that no genes get into the environment.”, Paxton points out.

Besides this, he is quite positive: “In the next years, this approach promises to solve the two main problems in beekeeping; varroa mites and DWV.”. However, Paxton also stresses that we need more information. For instance, it’s still unknown if the bacteria also get into the brood. Varroa reproduces in sealed brood cells and damages the developing bee. If the engineered bacteria and the dsRNA they produce remain in the adult bees, a persistence of 15 days isn’t enough. Current treatments acting only on the mites on adult bees stay in the colony for at least six weeks – two full brood cycles. If this would be necessary with this approach needs careful testing and evaluation.

Some clinical remarks

Besides the length and persistence of the treatment, there are some other clinical questions to solve before this could really become “a microbiome silverbullet for honey bees”. The microbiome is the community of bacteria in the bees’ gut in this case. In the lab studies, the bees got food with the engineered bacteria. I didn’t find any indication of a dose/efficacy relationship. How many bacteria must I feed to a colony or per bee to reach a sufficient efficacy? These are laboratory tests, which are necessary before even approaching a colony. Further, is feeding also the right application mode in practical beekeeping?

In addition, I got a bit alarmed when I read “We show that engineered S. alvi can stably recolonize bees and produce dsRNA […] and repress host gene expression, thereby altering bee physiology, behaviour and growth.” In a graph, the authors show that the feeding activity of the bees in the cage increases, as well as their weight when fed with the bacteria. That’s not necessarily a good sign. Higher food consumption may also mean stress, more need for energy because of the immune reaction or other factors. Do the bees weigh more because of larger fat or protein reserves or is it just water? The experiments lasted only for 15 days, without defining the age of the bees they used for these experiments. These are questions to answer to ensure the safety for the honey bee colony.

Two last caveats

I still have many more questions, but want to limit myself to two last points. First, there is the issue of the engineered bacteria themselves. As I mentioned before, genetic modification isn’t very popular in Europe. I already saw some comments from beekeepers saying “No way!” to this method. Most of the discussion is about genetically modified crops, not that much about medicinal uses of the technique. However, there would be a whole lot of regulatory and legislation work to do before this approach would end in a usable product for beekeepers. Who would accept it, probably, if the advantages are really that big as promised.

The last caveat… is somewhat more complex. The RNAi technique was already tested against varroa without the bacteria detour by Garbian et al. in 2012. In the State Institute of Apiculture of Hohenheim they wanted to test the method. And discovered that it wasn’t the RNAi which acted against the varroa mite, but the lithium chloride in the buffer. I now wonder – is the premise of this whole new publication correct? I’m not a microbiologist or expert in molecular biology. I don’t know if the detour with the bacteria changes something. But, I cannot help in asking if the described effects are a correlation or a causal relationship. That, my dear reader, is the question here.

Edit February 19th, 2020

Over on LinkedIn, I got a comment from Strong Microbials Inc., a company producing probiotics. They reacted to my doubts about the premise of this paper, that it’s really the dsRNA having an effect on Varroa and DWV. This is what they said:

One of the problems with dsRNA is producing, storing and delivering significant quantities to make an effect on the target organism. That’s why stabilized dsRNA in Garbian et al 2012 and Ziegelmann et al 2017 contained lithium containing buffer. dsRNA is not a stable molecule produced outside the bee.  The innovation that is showcased in Sean Anderson’s work is his use of engineered bacteria to reliably, continuously produce dsRNA on demand inside the bee. The science behind using gut symbionts to produce dsRNA to influence their host is solid and has been used for decades in other organisms, such as nematodes C. elegans. It’s first application to honeybees, but otherwise very right. Our understanding of microbial engineering makes this possible. Microbes are great at producing what we ask of them!

I don’t know the work of Sean Anderson and asked them for the citations, to read it myself. All my other objections on the method remain valid though. At the current state, this is about 5-10 years from a registered bee health product. At least in Europe.