Shortly after I graduated high school, commercial apiaries started to report massive losses of honeybees. Honeybees are probably the most economically valuable insects in the world, and are responsible for pollinating most of the food we eat. Here in the United States there’s an entire industry built up behind honeybees, with most US honeybees being transported to California to pollinate almonds at some point in the year.
Unfortunately there are a lot of wrong-headed things out there in the press. One common idea I see spread through facebook meme, such as the image to the left, is that biotech crops are responsible for killing the bees. This is a hypothesis that’s been pretty thoroughly researched in a variety of ways. Industry data very strongly indicates this is not the case: a recent meta analysis performed by Monsanto published in PLOS ONE reviewed experiments done by a wide variety of researchers and concluded that there were no effects on survival of bees on Bt crops. Academic research is consistent with the industry data, from a 2005 review on the nontarget effects of Bt crops in the Annual Review of Entomology:

Neither Bt cotton nor Bt maize requires bees for pollination, but cotton nectar is attractive to them and produces a useful honey. Maize pollen may be collected when other pollen sources are scarce. Pre-release honey bee biosafety tests have been conducted for each Bt crop registered in the United States, including Cry9C maize and Cry3A potatoes. Each test involved feeding bee larvae and sometimes adults with purified Cry proteins in sucrose solutions at concentrations that greatly exceeded those recorded from the pollen or nectar of the GM plants in question. In each case, no effects were observed. The rationale for requiring larval and not adult bee tests is questionable, because adult bees ingest considerable quantities of pollen in their first few days post emergence. Larvae, particularly later instars, also consume pollen along with jelly secreted by nurse adult bees, but only recently have there been attempts to quantify pollen ingestion by individual larvae. Other studies with bees fed purified Bt proteins, or pollen from Bt plants, or bees allowed to forage on Bt plants in the field have confirmed the lack of effects noted by the U.S. Environmental Protection Agency (EPA). Post-release monitoring programs are now underway to assess impacts of North American GM crops on pollinators under commercial field conditions.

Admittedly, there could be better data on this subject. For instance some of the research that’s been done has been done without some essential controls, like this German group which fed honeybee larvae a mixture of Bt pollen without determining that active Bt proteins were present in the pollen. The data, however, is generally against the idea that Bt crops harm bees.

What isn’t harming the bees?

Other groups have suggested neonicotinoids, but the problem with this is that we don’t have a a lot of the neccessary data we’d need to know for sure if this is the case. We’ve discussed this in the Biofortified comments before. The Xerces Society for Invertebrate Conservation put out a very detailed report on the subject, which concluded:

The failure of foraging bees to return to their hives has led many people to suggest that a link exists between CCD and the behavioral disruptions observed with sublethal exposure to neonicotinoid insecticides. As of yet, no single insecticide or combination of insecticides have been linked to CCD, though many chemicals have been found in hives. Researchers that compared gene expression in honey bees from healthy colonies and from collapsed colonies found no link between expression of genes that code for proteins associated with the detoxification of insecticides and collapsed colonies. This suggests that insecticide exposure, whether to neonicotinoids or another class, is not a primary factor in CCD….
…While neonicotinoids and other agrochemicals do not appear to be the direct cause of CCD, they may be a contributing factor to already stressed colonies. It is increasingly important that future studies focus on interactions of multiple factors suspected of contributing to CCD.

My favorite TV show, Doctor Who even jumped in on the commotion by claiming that the bees had returned home to an alien planet:

There aren’t a lot of easy answers on the topic, and pesticides are probably one of the better ideas being tossed around in the popular media. There are some good reasons to think that some pesticides, especially neonicotinoids, may be a big contributing factor. However, the losses began well after crops expressing Bt proteins were adopted and a very long time after neonicotinoids were adopted so these have never struck me as likely cuplrits. While the Doctor Who episode featured the disappearance as a mostly tongue-in-cheek thing, religious fanatics have suggested that this was a sign of the apocalypse. Cellphones have also been another popular crackpot theory.  So it’s unlikely that these are factors in colony collapse.
Personally, I wish I could say definitively that GMOs or neonicotinoids were responsible for Colony Collapse Disorder because then I could say that we knew why the bees are disappearing. Instead, there are a lot more factors to consider. While it would be a simple task to fill Biofortified with an unfruitful discussion of what colony collapse isn’t, I think it would be a much better idea to discuss what colony collapse disorder is, talk about some of the factors involved in the decline of honeybees, and engage in follow-up discussions in the comments.
I’m going to break this article into 5 sections:

1. History of Beekeeping
2. Honeybee Biology and Apid Ecology
3. Parasites Pathogens and Pesticides
4. Nutrition
5. The Big Picture

While I’d like to discuss each in detail, the list is going to be very incomplete. For each section, I’m going to give a snapshot: I just want readers to be aware that this is a huge area of research.

History of Beekeeping

Earler, I said that I didn’t think Bt crops or neonicotinoid insecticides were direct causes because they were relatively new things. There are reasons for this. The first thing you need to realize when reading up on this subject is that we’ve been keeping bees for awhile. About 15,000 years ago, some enterprising caveman discovered that bees stockpiled delicious honey and we began braving sheer cliffs and angry insects to collect this honey. Awhile later, another enterprising individual discovered that you could build boxes to keep bees in and by about 5,000 years ago the very basic foundations of modern beekeeping were laid out. If one of those ancient beekeepers were kidnapped by The Doctor and transported to a modern apairy, they probably wouldn’t be completely lost. The materials have changed, but the methods of beekeeping haven’t really changed all that much.
For as long as we’ve been keeping good records, we’ve recorded losses. One of the articles announcing the Colony Collapse problem appeared in PLOS Biology in 2007, and described these ancient losses in quite a bit of detail:

Some winter losses are normal, and because the proportion of colonies dying varies enormously from year to year, it is difficult to say when a crisis is occurring and when losses are part of the normal continuum. What is clear is that about one year in ten, apiarists suffer unusually heavy colony losses. This has been going on for a long time. In Ireland, there was a “great mortality of bees” in 950, and again in 992 and 1443. One of the most famous events was in the spring of 1906, when most beekeepers on the Isle of Wight (United Kingdom) lost all of their colonies. American beekeepers also suffer heavy losses periodically. In 1903, in the Cache valley of Utah, 2000 colonies were lost to a mysterious “disappearing disease” following a “hard winter and cold spring”. More recently, there was an incident in 1995 in which Pennsylvania beekeepers lost 53% of colonies.
Often terms such as “disappearing disease” or “spring dwindling” are used to describe the syndrome in which large numbers of colonies die in spring due to a lack of adult bees. However in 2007, some beekeepers experienced 80–100% losses. This is certainly the extreme end of a continuum, so perhaps there is indeed some new factor in play.

Furthermore, the original USDA action plan reviews some other serious threats to beekeeping that happened at the same time. While worrying is a legitimate reaction, I’m not entirely convinced that this is a new phenomenon. It’s entirely possible that similar things have happened before. However, there are a few more confounding factors which prevent us from being able to say that the same problems are coming around again. There are new things around, like neonicotinoids and pyrethroids (specifically fluvalinate) which weren’t around then and could be contributing today. There are also other new things, like a shrinking environment, constant travel and increased global spread of disease organisms, which are probably contributing factors. However, given the fact that similar things have happened in the past it’s unlikely that we’re dealing with a completely new phenomenon.

Honeybee Biology and Apid Ecology

Honeybees aren’t your average insect, and there are many quirks of their biology that make them unusual if you compare them to more common livestock like sheep or pigs. A lot of these factors make them more susceptible to parasites and pathogens. Honeybees are weird in that they’re difficult to think of as a collection of multiple organisms, but in my opinion, are more accurately described as a superorganism. Superorganisms are a collection of entities that function as a single organism. While honeybee colonies consist of thousands of individuals, the colony functions as a single creature.
Honeybees are very highly evolved eusocial organisms with a strict division of labor, both in terms of reproduction and everyday nest maintenance. Unlike humans and ants, honeybees need more than a male and a female to found a nest. The colony begins when a drone mates with a virgin queen, and after she returns a big chunk of the colony splits off to form a new colony. The queen is incapable of raising larvae by herself, and if you were to put a mated queen in a beehive there would be no colony. Instead, she needs workers to build wax combs in which to lay her eggs. These workers have their own division of labor based on age: young workers clean the nest and tend the young, middle age workers  forage and older workers forage and defend the colony. To reproduce, the bees need all these groups. They need drones and queens to reproduce, and the workers to do the work. Without all of these working in unison, there is no bee colony.
These social roles are very important, and a large part of honeybee immunity revolves around social roles. Bees keep the colonies quite clean, and even work together to seal the hive from invaders with propolis. Besides sealing the colony, propolis also has antimicrobial properties. It’s not the cure for cancer, but it likely stops the spread of bacteria and viruses.  While bees within colonies tend to stick to themselves, there are circumstances where stranded workers may enter another hive… especially if their home colony is suddenly transported out of foraging range. During later times of the year, stronger colonies may raid weaker colonies.
With this in mind, the 2007 USDA action plan defines CCD thusly:

Symptoms of CCD include: (i) sudden loss of the colony’s adult bee population with very few bees found near the dead colonies; (ii) several frames with healthy, capped brood with low levels of parasitic mites, indicating that colonies were relatively strong shortly before the loss of adult bees and that the losses cannot be attributed to a recent infestation of mites; (iii) food reserves that have not been robbed, despite active colonies in the same area, suggesting avoidance of the dead colony by other bees; (iv) minimal evidence of wax moth or small hive beetle damage; and (v) a laying queen often present with a small cluster of newly emerged attendants

We’ll discuss those parasites later.
**One problem that I see consistently in online discussions of Colony Collapse Disorder is that many assume that any loss of bee colonies is the same thing as CCD. There are many reasons a honeybee colony can die, and the definition of CCD above eliminates some of the other possibilities. For example, under the second symptom (ii) there must be low levels of Varroa mites which means that the colonies aren’t killed by Varroa mites. Later definitions also remove colonies with damaging levels of Nosema. Pesticide poisoning also results in large numbers of dead bees around the hive, so direct pesticide poisoning does not match up with these symptoms.
While it’s tempting to assume that these symptoms are explained by only the older workers being afflicted, I would caution against that interpretation because it can be explained equally well by a slow-acting agent acquired as larvae that finally kills the bees well after they’ve emerged as adults. The fact that only newly emerged adult bees are still present means that newly emerged adult bees are still present.
The other thing that you need to realize is that we get a lot of pollination services from native bees, and those native bees  we rely on for pollination are in trouble as well. In fact invertebrates as a group aren’t doing too well, but bees seem to be particularly hard hit. Ecosystem fragmentation and loss of plant diversity are definitely major contributors, but cross-contamination of pathogens between native and our introduced honeybees is probably a contributing factor. In other words, we’re losing vital pollinators on two fronts, and it’s too early to know for sure if the two events are related.

Parasites, Pathogens and Pesticides

Some of these guys we’ve already mentioned, but have yet to discuss in detail. Honeybees, like virtually every other organism on the planet, have their own set of parasites and pathogens which can wreak havoc on bee colonies. Some will immunosuppress the adults, others weaken the entire colony, while others target larvae. The list of pathogens is quite long, and some are bigger problems in some parts of the world and in some parts of the country. I could easily devote an entire post to their biology, but here are just four along with the problems they cause:

• Deformed Wing Virus (DWV) is a virus which infects honeybees, and is spread among the workers in a colony. Under certain circumstances, which I’ll describe later, the virus can replicate out of control and cause damage in honeybee colonies. The damage results from lots of honeybees having deformed body parts, and stunted wings. A good chunk of these bees die after soon after they emerge.
• Nosema are single celled fungal parasites. They inject their cell’s cytoplasm into the intestinal cell of the bee and replicate rapidly. The parasite is spread through the feces, and often you can tell if a colony has a Nosema infection by looking for larger than normal amounts of feces around the outside of the hive. Infected bees are puffy and white, and have trouble absorbing nutrients from their food. Infected bees are quite easy to tell apart from noninfected bees.
• Paralytic viruses is a broad category that includes several different viruses, among them Acute Paralysis Virus and Israeli Acute Paralysis Virus. These viruses do pretty much exactly what it sounds like, they slowly render the bee immobile and unable to fly. The role of these viruses is controversial because some research indicates that they’re involved, while other groups downplay their role.
  
• Varroa destructor is a parasitic mite that’s closely related to ticks. They attach to the outside of honeybee adults and larvae and suck their hemolymph, knocking their immune system out to facilitate this process. V. destructor also vectors viral pathogens. The damage from V. destructor infestation is both from transmission of disease and from removal of hemolymph.

These are just four pathogens that are major problems in honeybees, but they’re not neccessarily culprits in the Colony Collapse problem. There are also insects that play a role in weakening honeybee colonies. Hive beetles will burrow through honeycomb, consuming pollen and bee larvae. Waxworms are caterpillar larvae which burrow through frames of beeswax that are in storage. While these organisms generally steer clear of hives implicated in CCD, it’s possible they could still play a role by tracking pathogens into colonies.
Bees fight these pathogens off in three ways. First, as already mentioned, they line most of their colonies with propolis which has antimicrobial properties. Second, they practice a form of socialized medicine where they will expel diseased larvae and adults from the colony. They can also practice a so-called ‘behavioral fever’ where the bees vibrate their bodies to overheat intruders. Some bees have also learned to use this against predators, as shown in the video at the bottom of this section.
A third form of defense is the bee’s immune system. They have a full compliment of weapons at their disposal, from antibiotic proteins that form holes in bacterial cells and blow them up, to RNAinterference that destroys nucleic acide sequences of viruses, to hemocytes that encapsulate invaders. However, they tend to rely more on sanitation than their immune system. One thing worth mentioning isn’t what they have, but what they lack. The husbandry of other livestock has certainly benefited from the fact that they can be used as models for humans, but bees aren’t so lucky. Honeybees are insects, and insects lack an antibody production system. Because they don’t produce antibodies, they can’t be vaccinated*.
Bees are exposed to quite a few pesticides, believe it or not. These not only include pesticides that they’re exposed to while foraging (i.e. those neionicotinoids), but also pesticides that they’re exposed to in order to combat V. destructor. There are sophisticated sampling methods to sample for V. destructor and there are damage thresholds that have been set. Thus, V. destructor can be managed in an Integrated Pest Management (IPM) setting.
Fluvalinate is a synthetic pyrethroid that’s more toxic to mites than to bees, and this is used to control V. destructor in bee colonies. Organophosphate insecticides, namely Coumaphos, are used when pyrethroid insecticides fail because of resistance. When frames are stored, waxworms are sometimes controlled with napthalene, the active ingredient in mothballs.

Nutrition and Stress

A lot of people imagine aparies as places that are surrounded by miles of quiet little meadows like what you see on TV, but sadly this is not the case. Honeybees are likely impacted by the fragmentation of habitat just like native bees, and their diets are often supplemented by nectar substitutes like high fructose corn syrup. While high sugar diets in humans contribute to obesity by increasing calories, the diets of bees consist largely of sugary liquids like nectar. The problem is that wild nectar contains lots of sugar in addition to other things like amino acids that are lacking in HFCS, and other substitutes aren’t economically viable.
Furthermore, bees are exposed to a lot of variable conditions when they travel. They need a high diversity of pollen because pollen can vary widely in it’s nutritional composition. When the honeybees are loaded onto trucks and transported cross-country it’s questionable whether they’d be exposed to the types of pollen that are, nutritionally speaking, best for them.These would presumably be in protein rich pollen substitutes, but I’m unaware of any studies that have evaluated whether these will nourish well enough to support an immune response.

The Big Picture

Pathogens, pesticides and all the other factors I mentioned above do not exist in a vacuum. Instead, everything interacts with everything else. Pathogens co-infect bees in the field and a lot of the research that’s underway is aimed at looking at the bee pathogens which are correlated with Colony Collapse. Unfortunately, there are no pathogens which are correlated with Colony Collapse. There are no pesticides which are exclusively connected with Colony Collapse, either. There are also no genes which are consistently upregulated in Colony Collapse colonies. Instead, it looks like there are a bunch of smaller incidents which involve a lot of factors acting in unison.
The reason why there’s a lot of focus on diseases in CCD research is because CCD bees tend to have higher numbers of bee pathogens, and lots of different pathogens. The problem, however, is that there are no diseases that are exclusively associated with CCD. Furthermore, a lot of the pathogens that are in healthy colonies are also found in CCD colonies. Researchers are just starting to understand these interactions. A recent PLOS ONE paper shows how complicated these interactions can be. In the figure below, the circles represent different pathogens, and the black lines represent correlations. The bigger the circle, the greater the proportion of bees in the colonies that are infected. The thicker the lines, the stronger the correlations between pathogens.

Some pathogens, like Nosema apis (N. apis), Kashmir Bee Virus (KBV), and Acute Bee Paralysis Virus (ABPV) are detected more commonly in collapsing colonies which is denoted by the size of the circles. In collapsing colonies, how often different pairs of pathogens are detected in the colonies are shown by the thickness of the lines. Unfortunately, at this time, we’re not sure whether these pathogens are causing the collapse of the colonies or if they’re there because the colonies are collapsing.
All of these pathogens have been shown to be able to kill bee colonies, but whether they’re actually killing the colonies is hard to determine. There are instances where some of these pathogens, like Deformed Wing Virus (DWV) can be present but not cause symptoms and there are conditions which cause the same pathogens to become very problematic.
One such interaction is between Varroa destructor and DWV. To ensure a meal, the mites have to immunosuppress their hosts. In other animals, this often creates openings that allow other pathogens to establish. Nazzi et. al 2012 did a pretty good series of tests looking at how infestation of V. destructor changes infection by Deformed Wing Virus. They found that by artificially infesting bees infected with DWV with V. destructor, they could get the number of DWV genome copies to increase by quite a bit. Then, they looked at expression of an immune factor connected to the NF-KB pathway, Dorsal, as a product of viral infection and found that only viral infection but not mite feeding caused a decrease in the production of RNA that codes for this protein. Artificial silencing of Dorsal resulted in an increase of DWV reproduction. As more mites feed on the bees, the more susceptible the bees become to DWV. Because the viruses are able to replicate easier, they’re able to better drive down the transcription of factors that play a role in defense against viruses. The results from their study pretty strongly imply that feeding of the mites causes a loop that results in DWV replicating out of control, and higher levels of DWV cause the bees to die at a faster rate. Because some factors involved in antiviral immunity are downregulated, it also implies that bees that are both infested with V. destructor and infected with DWV are more susceptible to infection by other pathogens.
The V. destructor-DWV interaction is worth mentioning, not because it’s important for CCD but because it describes the type of research that needs to be done. V. destructor is frequently sampled for and relatively easy to detect, which means that this parasite is easily controlled. Similar interactions between honeybee pathogens may go undetected, and V. destructor is present at some level in most colonies. While V. destructor may be absent from collapsing colonies, the diseases it vectors are still present and we don’t understand how these diseases interact.
Finally, not all honeybee diseases are well studied. There are some organisms, notably Malpighamobea mellificae, which cause diseases in honeybees but haven’t been studied since the 1960s. M. mellificae replicates in the bee’s Malpighian tubules (kind of like their kidneys) and keep these from functioning properly. They can cause problems in beehives, especially around springtime when the bees are recovering from hard winters. The infected bees get a form of dysentary, and often flee the colony in a manner similar to Collapsing hives. Since there’s no research on the pathogen, it’s unclear if this pathogen plays any role in Colony Collapse (Evans and Schwarz, 2011).

Conclusions

There are a lot of factors involved in colony collapse disorder, and it’s unlikely that there is ‘One True Cause of Colony Collapse’. There are a lot of things that correlate with the symptoms and timelines, but we have to start to separate these. Above, I’ve shown a lot of things that correlate with collapsing colonies… but at the same time, so did my high-school graduation. A correlation can shed light on some details of the problem, but this isn’t the same thing as causation. We know that similar incidents have happened before, but they were a lot more localized than the current set of events. Pesticides, pathogens and environmental factors are likely to be involved but we’ve really got no idea what role these factors play at the current time. There’s good progress being made, but it’s still too early to know why it’s happening this time around.
While we might not know exactly why honeybee colonies are dying, there are a lot of good entomologists trying to figure out what factors are involved in Colony Collapse. The stuff I’ve described above is merely a snapshot of a small subset of interactions which entomologists have to dig through to get to the heart of the problem. Throughout this entire piece, I’ve tried to stick to open-access literature so interested readers can have access to good information on this stuff. If you’re interested in further reading, there are a variety of institutions which have good information. The Xerces Society does a lot of great work on studying the native bees. The USDA has a great webpage on Colony Collapse Disorder and puts out annual open-access reports on what progress has been made on CCD over the past few years.
.
* The use of the term ‘vaccination’ is controversial within insect immunology because some (including myself) think it implies antibody production. Instead of producing antibodies, insects rely on proteins that detect broad classes of pathogens. They can modify how specific and sensitive their immune response is by upregulating certian components these after exposure, and in some cases ‘store’ bits of double-stranded RNA. There’s research ongoing to determine if exposure to bits of pathogen could be used as a prophylactic measure in the way we use vaccines.
**This paragraph was added by edit on 3/14/2013 after my first comment below.