Adaptation, multiplexing, and communication contribute to the advancement of the animal biodetection industry.
Pam Hullinger is the chief veterinary officer and leader of the Food and Agricultural Security Program at Lawrence Livermore National Laboratory (Livermore, CA). In 2001, she was one of approximately 300 U.S. veterinarians selected to assist the United Kingdom's rapid response to a devastating outbreak of foot-and-mouth disease. She can be reached at email@example.com.
When the United Kingdom suffered an outbreak of foot-and-mouth disease (FMD) in 2001 that left its sheep and cattle populations devastated, the British government enlisted the help of several U.S. veterinarians to assist with the FMD eradication efforts. Hindsight has revealed the inefficient methods of clinical detection employed in the outbreak response. In an effort to avoid a similar issues in future outbreaks and build a more efficient and cost-effective system, the U.S. Department of Agriculture (USDA) and the Department of Homeland Security (DHS) has spearheaded a project to look beyond the current clinical IVD platform to improve strategic response to endemic outbreaks.
To learn more about how various animal disease detection and response factors affect veterinary assay development, IVD Technology editor Richard Park spoke with Pam Hullinger, chief veterinary officer at Lawrence Livermore National Laboratory (LLNL; Livermore, CA). In this interview, Hullinger discusses the advantages of multiplexing assays, as well as the benefits of strong communication between laboratory research staff and field staff. She also talks about developing a stronger foreign animal defense veterinary infrastructure, adapting human biotechnology for veterinary animal diagnostics, and how the struggle to secure proper funding can affect long-term goals.
IVD Technology: During the past few years, what mandates or directives have been handed down to Lawrence Livermore in terms of efforts to develop biodefense diagnostics, especially in the area of veterinary and animal testing?
Pam Hullinger: Lawrence Livermore has been specifically funded to develop diagnostic assays for FMD and highly pathogenic avian influenza (HPAI). We were asked to develop assays that would provide a method for early FMD identification and would differentiate the FMD virus from other endemic or domestic diseases that may exhibit similar clinical signs.
We were also charged with developing new assay technologies that could differentiate infected from vaccinated animals to help identify vaccinated animals that may have become infected with FMD after vaccination. With HPAI, we focused on the development of rapid hemoglutanin-typing assays. From a foreign animal disease standpoint, these diseases—FMD for livestock industries and HPAI for poultry industries—were areas where we've received support and direction for diagnostics assay development.
In 2001, you traveled to the United Kingdom to assist with an FMD outbreak. What were the results of your time there?
I spent roughly three months working on the 2001 FMD outbreak in the United Kingdom. During that time, I gained a personal appreciation for how devastating this disease could be in a country with a strong veterinary infrastructure, similar to the United States. I was charged with trying to determine whether herds of cattle or flocks of sheep were actually infected with FMD solely based on the evaluation of clinical signs and, thus, needed to be destroyed. At the time, the control policy was to kill all the animals on the infected premises. In order to stop the disease, the identification of one infected animal on the premises was the justification to condemn the entire herd or flock.
Because the lesions were sometimes very subtle in sheep and often mimicked other potential disease conditions or trauma from grazing rough pasture, it was virtually impossible to determine whether these animals were truly infected with FMD without the use of diagnostic tests. In the face of an outbreak and a situation where we did not have the opportunity to do diagnostics testing on-site or off-site, all decisions were influenced by the necessity to err on the side of caution. So essentially, if one animal had any lesions resembling FMD in his mouth, the whole flock was destroyed.
Retrospectively, sample analyses determined that many of the flocks of sheep that were destroyed were killed unnecessarily because adequate diagnostics testing wasn't available to support the veterinarians in the field. It was a tragic loss of animal life, but it also complicated or prevented the responding agencies from doing their job as effectively as possible. Instead of directing resources towards euthanizing animals that did not pose a threat, those resources could have been directed more appropriately to control the outbreak more efficiently.
Until the 2001 outbreak, FMD hadn't overwhelmed any developed, disease free countries that would have been interested in devoting significant funding for research in this area. I was hoping we could produce something positive out of this tragedy and develop better tools for future outbreak responses or to enhance disease control in endemic countries that won't result in the destruction of so many healthy animals.
At that time, I was working for the California Department of Food and Agriculture (Sacramento, CA), and Alex Ardans, the director of the California Animal Health and Food Safety Laboratory at the University of California, Davis, had a relationship with the staff at LLNL. He had several conversations with LLNL regarding the technologies used to provide multiplexed assays for human diagnostics. After reviewing their biotechnological research, he wondered whether LLNL could create a multiplexed assay targeted towards agricultural diseases that would provide a better means to screen for FMD and potentially help differentiate it from some of the other agricultural diseases.
That's where the initial concept for this project developed. After talking to people about their experiences in handling the U.K. FMD outbreak, we asked, “How can we do this better?” For this project, we also wanted to create a test that could potentially be used in routine endemic disease surveillance or testing for other endemic diseases, such as bovine viral diarrhea (BVD), but could also screen for FMD in the background.
Learning from Adversity
As a result of your work on FMD in the United Kingdom, what sort of knowledge did you bring back to the United States and implement into monitoring and testing livestock here in the United States?
One of the things we learned was the need for scalable surge capacity. Often, diagnostic labs are set up to run a certain amount of tests per day based on a routine caseload or sample submissions. But in an outbreak scenario, the number of samples you receive scales on the orders of hundreds to thousands. One of the challenges of the 2001 U.K. FMD outbreak was they weren't prepared to scale their diagnostic test capacity, which is why we weren't allowed to take testing samples in the field. So another facet of this project was to provide some robotic solutions that allowed a multiplex assay to be completed in a semi-automated process so that throughput could be enhanced during an outbreak response.
If you are conducting surveillance for FMD as well as for diagnostics testing during an outbreak response, there are a wide variety of tests available. You would run multiple tests attempting to address different questions depending on what you are searching for and what you are attempting to prove. In many cases, the tests would be run in a specific sequence. Many people feel that the primary value of a multiplexed assay, once validated and deemed appropriate for that purpose, is to use that test daily for diagnosing domestic diseases that resemble FMD as part of a targeted surveillance program.
Let's say you obtain a negative FMD polymerase chain reaction and decide that it is not consistent with what is reportedly observed in the animals in the field, you may decide to run a second test. Perhaps in the second test, you discover the cows had BVD or another disease that could cause compatible lesions. This discovery makes you feel more comfortable calling that animal negative for FMD now that you have a positive diagnosis for another disease that could cause a similar clinical presentation.
A number of different tests are available for use in testing for the various disease agents in the prototype-multiplexed assay. Such tests provide parallel testing for multiple disease agents. It allows researchers to know exactly how much time it will take to set up and run all the individual nucleic acid detection assays necessary. With this multiplex prototype, you can essentially set up one assay and receive results in about four hours for eight different disease agents. If you ran all of those tests independently, it would take much longer.
What are some of the current efforts at Lawrence Livermore to develop biodefense diagnostics, especially in the area of veterinary and animal testing?
In addition to the initial cross-species prototype assay that was developed as a version 1.0, we've developed two species-specific assays: a bovine-specific panel and a porcine-specific panel. Both assays have undergone the initial analytical characterization against a subset of diseased agents they are designed to detect.
Now we need to obtain diagnostic performance data for them, specifically with regards to FMD. Ben Hindson is the lead author on a recently published article that describes the tests that we conducted on the initial prototype multiplex assay at the Institute for Animal Health (Pirbright, United Kingdom) with our collaborators. In the report, Hindson details our work evaluating this assay on an archived set of historical FMD samples they've received from around the world. Pirbright Research Laboratory is a world reference lab in Europe, so when FMD is suspected in nearby countries, the samples are sent to their laboratory for confirmatory testing. They provided a rich archival library of real samples that we could use to measure a diagnostic assay's performance.
Based on our analytical characterization, we think both assays are very promising; however, spiked samples in a lab are very different from samples found in the field that can include dirt, cow saliva, and chewed grass, as well as other elements that could inhibit test performance. Until we can understand the assays' performance in the field, both tests are definitely still in the R&D phase.
What is the eventual long-term goal of your hard work and R&D efforts?
The goal of these assays is to transition them to the USDA and possibly implement them into the National Animal Health Laboratory Network (NAHLN).
USDA has been charged with the mission space for foreign animal disease surveillance and response. Although the work is funded through the Department of Homeland Security (DHS), it has always been clear that the research and technology is intended for USDA.
The agency controls how and when foreign animal diseases are tested in the United States. LLNL has tried to work very closely with them from the beginning to help assure that what we are developing produces an assay USDA believes will work for their purposes in multiple scenarios.
What is the process involved in developing veterinary and animal diagnostics? Do you and your colleagues at Lawrence Livermore identify specific pathogens and develop completely new technologies for those agents or do you adopt current technology to detect those agents?
The work we've done to date has largely involved applying techniques developed for human biothreat detection to agricultural problems. In many cases, the funding in the agricultural biodefense area is not as well supported as the funding for the human biodetection side. It's a great opportunity to leverage the existing infrastructure and technologies that have been developed for human disease defense research. So we don't necessarily have to support the full R&D of the technology itself, only the adaptation of that technology to agricultural diseases, which makes it more affordable for our needs.
The first step when developing or adapting an assay is to understand the problem and the purpose for which a new or novel assay or technology is needed. In this case, you might define the problem as a lack of adequate tools to conduct surveillance for FMD in populations of animals displaying common clinical signs. Then, you determine what kind of tool you would desire to address the issue and then you would determine the assay requirements. When designing an assay, it is important to understand the sample matrix, what type of tissue you are analyzing, and what will be the target population for the test. Based on the disease, you need to determine how much nucleic acid will be present in the sample, and what are the limits of detection in order to find what you're searching for in a clinical sample.
If you're testing an animal with early lesions for FMD, you will find plenty of virus samples to test. However, that's less of a factor if you're testing an animal in a preclinically infected state where there is likely less virus present. The animal might look normal and the viral load in its oropharyngeal cavity or nares would be much lower than that expected in a clinical animal. You want an assay that's going to have a lower detection limit and allow you to amplify an agent if it's present, which indicates that the clinically normal animal is not in a preclinically shedding state.
Has the technology proven to be easily transferable between human biothreat technologies and agricultural biodefense?
Yes. The lab has conducted quite a bit of work with bead-based assays and developed significant expertise in multiplexing technologies, as well as various nucleic acid, antibody, or protein detection technologies. So a number of panels developed for human biothreat agents have been leveraged for agricultural studies.
Essentially, detecting nucleic acid is detecting nucleic acid, regardless of whether it comes from a human or an animal. Researchers can avoid many of the pitfalls that occur in a normal R&D process by understanding how to properly, efficiently, and effectively design and develop the probes and primers and generate the product lengths that work in the conditions that are conducive to multiplex assay development.
What do Lawrence Livermore researchers look for and try to accomplish when developing veterinary and animal diagnostics?
We strive to develop tests that are sensitive, specific, and provide rapid, reliable results for the assay's intended use. We also design them to be cost-effective and easily transferable to the end users. To be successful, an assay needs to be validated from both an analytical perspective and a field, diagnostic, performance perspective.
The test also needs to be well understood, so that when it's being used for its intended purpose, the results will be interpreted correctly. Such tests won't ever be perfect, but they need to produce results that can be interpreted in the context of the populations being tested. So when researchers have a negative result, they can ask, “How many other samples do I need to take from that population to increase my certainty that that flock or herd is negative?”
Developing Multiplex Assays
What characteristics would you consider important for a successful veterinary or animal diagnostic? What are the minimum standards (e.g., durability, speed, portability, etc.) that you're looking for?
The multiplex assays we've developed at this point are intended more for lab-based testing and are not currently field-deployable. It would be ideal to analyze test samples on a farm without having to leave the farm environment, and there are a number of people working on IVD technologies that might help researchers achieve just that.
Another goal is to create a test that's also cost-effective and disposable. It would allow researchers to take the test to the animal, run the assay on the farm, and dispose of the device. That's not likely to be a multiplex assay, but a singleplex assay or a slightly modified multiplex assay could potentially be developed for field use.
Durability, repeatability, and reproducibility are important IVD characteristics, as well as having a platform that can be successfully operated in the hands of many end users. One of the components of the multiplex assay development project demonstrated at NAHLN laboratories had a 92% assay success rate when the assay was operated by trained end users. In other words, 92% of the tests generated actionable results. So the technology isn't so complex that it cannot be easily transitioned to end-user labs.
How do the characteristics of a successful veterinary and animal diagnostic differ according to the agents that they're designed to detect, or can one diagnostic be used to detect multiple agents? How has the development of multiplex technologies influenced the way LLNL has developed its animal and veterinary diagnostics?
Multiplex diagnostics has been our primary focus. Multiplex diagnostic assays can certainly have all the performance characteristics of singleplex assays. Once we meet that goal, we will have more options for screening a number of different disease agents in a more cost-effective manner.
The best way to assemble a multiplex assay is to first determine where you want to use the test. Are you sampling blood at a slaughterhouse? Are you sampling air in a sales yard? Are you conducting surveillance on a fluid sampling produced at a pig barn?
Once you understand what your sample is, then you can ask, “What type of disease agents would appear in that sample if an animal was infected?” Different agents will be shed in different manners depending on the disease, so you wouldn't expect to find Bluetongue virus in a fecal sample, but you might expect to find FMD. Researchers need to know what diseases are present at detectable levels in different veterinary samples. Then, you need to design a multiplex to target disease agents that would appear in that sample matrix of interest.
The staff in the field that is actually charged with implementing surveillance programs must have strong communication with the staff in the labs that is developing these technologies, because the laboratory staff needs a clear understanding of the desired performance characteristics in the field in order to develop an assay that meets the needs of the end user. Sometimes that communication doesn't always occur early enough in the development stage and results in challenges that require a return to the drawing board.
Collaborating with Agencies and Industry
Does LLNL work with IVD manufacturers in developing veterinary and animal diagnostics? How does Lawrence Livermore go about seeking such collaborations?
The collaboration process with endemic disease assays is different than foreign animal disease assays. The federal government directs our foreign animal disease biodetection development, and we have to work within the confines of the work sponsor. However, our relationships with commercial partners are key to many of our successes in diagnostic assay development. There are certainly projects that lend themselves to the development of a close working relationship with commercial partners, and we are always looking for opportunities to transition our developed technologies to venues where they can be potentially commercialized.
At what point would Lawrence Livermore or other government agencies be inclined to work with industry partners? Is it something that would happen in the early stage of a project or later?
It could work both ways. In some cases, there could be opportunities to work with a private company from the beginning of a project to help develop something to satisfy customer needs or the relationship could occur later in the project when opportunities for commercialization become available. The collaboration can happen anywhere in the process.
What has been the degree of coordination among the various government agencies on developing veterinary and animal diagnostics?
The Department of Energy isn't necessarily directly involved in directing the research programs. It is in the chemistry and biology division, where we largely do work for others, that is primarily focused on global security. The interagency coordination is primarily with our sponsors, DHS and USDA.
What efforts are being made at Lawrence Livermore to continue looking for ways to make veterinary and animal diagnostics better and faster? What future challenges do you foresee in developing such diagnostics in the future?
The biggest challenge has always been finding sustained funding to support activities in veterinary diagnostics. In many cases, a lot of support becomes available due to certain events. For instance, HPAI poses a potential threat to enter the United States, so funding has been made available to support efforts for prevention and control of that disease. The 2001 FMD outbreak in the United Kingdom stimulated more interest in research for that disease, and a subsequent upsurge renewed interest after the 2007 outbreak as well. There's also ongoing FMD eradication projects conducted in South America, but finding sustained funding streams for these projects can pose the biggest challenges.
Because we continually strive to find better ways of providing accurate, timely information to decision makers who are charged with responding to these events, we meet such challenges by leveraging work that's been conducted on the human biothreat side. It provides us with access to technologies that would be difficult to find core funding for during the early agricultural biodefense planning stage. It also involves talking to potential work sponsors, such as DHS, USDA, and the state departments of agriculture to understand their challenges and visions for the types of tools and technologies they would like to have available in the face of an outbreak.
We've certainly had some very positive experiences working with the Institute for Animal Health at Pirbright, as well as the National Centre for Foreign Animal Disease in Winnipeg, Manitoba. It's important to strategize with our international counterparts to try to effectively target our research so we're not creating the same wheel. If LLNL is not involved in generating a specific assay, we often take part in a larger coordinated effort to help optimize the limited resources of each of these facilities to ensure that the greatest amount of work is completed in the shortest period of time. There is only so much space available in the containment facilities for large animals, so the more collaborative effort involved in developing these technologies, the better.
What are the advantages of developing new modeling tools, such as MESA, that assist in controlling diseases?
MESA is an acronym for Multi-Scale Epidemiological and Economic Simulation and Analysis. It is a decision support system used to simulate animal disease spread patterns, which allows us to evaluate the theoretical benefits of different control strategies and potential counter measures. From a diagnostic perspective, it can be very useful in helping us explore the potential cost benefits of developing new disease detection technologies that inform investment funding for R&D.
The analyses are also useful for estimating diagnostic test capacities necessary to support various disease outbreaks. The information MESA offers includes a look at how an outbreak might unfold given a precise introduction scenario of a specific agent in a localized area and a determination of how that outbreak might unfold and what type of surge capacity is necessary in that time frame in order to meet the diagnostic test demands of that emergency event.
MESA has an intraherd model that tracks the disease as it spreads within the herd or flock. A contact network links the different herds and flocks together and creates an average rate of dispersal based on animal movement patterns as well as the movement of inanimate objects (or fomites) that could cause indirect disease transmission, such as tracking infected mud from one farm to another.
In the intraherd model, detection depends on the development of clinical signs within the herds or flocks. With this in mind, we can theoretically explore the possibility of a new technology that allows us to detect disease in preclinically infected animals so a veterinarian in the field would be able to detect disease faster. After the disease is detected and confirmed, control measures are set in motion. Such measures would mimic biosecurity and quarantine different zones around the infected premises based on the national response plan. The model also would perform an economic assessment based on the disease's impact.
The parameters in the model try to capture the biological variability of disease. Every time you run a simulation, you will receive a different outcome because you retrieved a sample from different distributions. Such simulations help you understand the spectrum of likely outcomes given a specific scenario introduction, as well as how quickly we need to test samples given a precise outbreak scenario.
Our focus is working on foreign animal disease diagnostics. The lab's expertise lies in the technologies and the technology platforms. As a result, we rely heavily on the partnerships we form with federal agencies and academic contributors, to provide the detailed disease expertise needed to successfully develop such technologies. It definitely has to be a partnership in order to be successful.