Medical Device Link.

Originally Published IVD Technology April 2004

In Person

Bill Colston, PhD, is associate program leader of the chemical and biological nonproliferation program at Lawrence Livermore National Laboratory (Livermore, CA) and adjunct professor in the department of hematology and oncology at the University of  California, Davis Medical Center (Davis, CA). He can be reached at colston1@llnl.gov.

Not only can IVDs serve to alert a physician to a pathogen’s attack on a patient’s body, they can also warn a nation of an environmental invasion by a bioterrorism agent. Special events like the Olympics, mass transit systems like the subways in use in major cities, and even massive office buildings could potentially serve as effective targets for an attack with an agent of bioterrorism. A pressing awareness of the threat of such an attack spread across the globe following the anthrax attacks of 2001. Since then, many governmental agencies and research institutions have been able to accelerate their work on biothreat detection systems. Lawrence Livermore National Laboratories (LLNL; Livermore, CA) is one such institution.

To shed some light on the research and development process for these systems, IVD Technology editor Richard Park spoke with Bill Colston, PhD, associate program leader of the chemical and biological nonproliferation program at LLNL and adjunct professor in the department of hematology and oncology at the University of California, Davis Medical Center (Davis, CA). In this interview, Colston discusses the development of IVDs for environmental monitoring and use by first responders, the influence of governmental agencies on the development of these systems, and collaborations with industry. 

IVD Technology: Since the 2001 anthrax attacks, what mandates or directives have been handed down to LLNL for the development of biodefense diagnostics?

Bill Colston: We have had an active program in biodefense since 1997. We had research funding instruments for environmental monitoring in the field for making onsite measurements; and we also had done a lot of work developing systems for detecting class A bioagents previous to 2001. 

After 2001, the Department of Homeland Security (Washington, DC) was created. At that point, after the anthrax attacks, biodefense issues became a little more realistic and pressing. So, the scope of what we were working on broadened a bit. The government wanted long-term national defense solutions that could be applied immediately. 

Earlier this year, ricin was found at the U.S. Capitol. Do you feel a continued sense of urgency and pressure from Homeland Security to deliver immediate results?

Yes. I think we still maintain balance in the program between working on and worrying about things that may present threats three to five years down the road, and putting systems in place to respond to immediate needs presented by the ricin incident and others. So, yes, we are even now called on to respond to things like the ricin attack.

Research and Development

What are some of the current efforts at Lawrence Livermore to develop biodefense diagnostics?

Without getting into too many specifics, we have very active programs in portable detection. We are developing tools for first responders, the Secret Service, and others to use as portable instrumentation for biodiagnostics performed in the field. 

We have smoke alarms that sit in the field or in a facility and continually monitor the air. In addition, we have done a lot of work in collaboration with the Centers for Disease Control and Prevention’s (CDC; Atlanta), Lab Response Network by developing validated assays for biodefense. We’re also working actively with the U.S. Department of Agriculture (USDA; Washington, DC) and others to address some of the agricultural terrorism issues with new diagnostic and surveillance mechanisms.

What is the process involved in developing biodefense diagnostics? Do you identify specific agents and develop completely new technologies, or do you adapt technologies that you had in the past to detect those agents?

The answer to that is both. Sometimes we have integrated technologies that are adapted for use for diagnosis or detection of biothreat agents. In other cases, we have had to develop new research tools and new assays or detectors because the necessary technologies just weren’t available. If the technology’s out there, we will certainly use it and adapt it before developing something entirely new.

For example, we developed an environmental monitor. It’s similar to a smoke alarm and sits in the environment. It monitors the air. Three different companies were involved in building that device. 

The core technology we use for detection was developed by a company called Luminex (Austin, TX). In 1999 when they came out with their first instrument, LLNL felt like that was a big paradigm shift in the way biothreat agent diagnostics could be done because they were measuring a hundred things with every sample as opposed to one sample per agent. So we were some of the very first users of that technology. We built a lot of platforms around it. As a result, we’ve ended up on the leading edge of meeting the need for environmental monitoring.

Are these monitors in use right now?

Yes. The technology is very mature. It will be deployed this year. In fact, we’re on a route to commercialize it so that, ultimately, we’ll hand this off to a company.

Can this system screen for multiple different types of agents that could be released into the air?

Yes. And it’s scaleable for measuring all types of agents, too. So, you can measure viruses, toxins, and spores, all with the same technology. The biotech industry has begun to use platforms that are highly multiplexed, like the Affymetrix (Santa Clara, CA) chip or the bead-based approaches. We made use of this type of application early on. Multiplexing allows for the use of built-in controls and it reduces cost.

How do you keep track of the agents that are out there right now and stay abreast of new developments in terms of genetically altered agents that could come out of nowhere?

There are ways to go about detecting things that don’t depend on knowing the sequence you’re looking for. Those methods rely more on mechanism-based detection techniques, where you’re looking for common variance factors or common portions of a substance or looking at those response factors that lead you back to the class of virus that you might be interested in. 

When developing diagnostics, do LLNL researchers consider what possible changes can be made to known viruses and other biothreat agents to make them harder to detect? 

Absolutely. It is one of the tougher problems to go about addressing. So, in terms of things we work on in the near versus mid- versus long term, it’s certainly on the mid- to long-term scope of things. However, we always keep in mind that our products may need to be adapted to be capable of dealing with those types of threats.

Will cost-related challenges include not only funding the continued development of these technologies, but also once they’re developed, who’s going to eventually pay for them?

The cost of biodefense detection systems is going to become more and more relevant to the development of these systems, especially as more and more systems become deployed and the market becomes defined. Because if you have to bear the cost of environmental monitoring for lots of sensors, first-responder tools, or facility monitors, often having a device that costs too much to maintain just means it won’t be put into as many places and it won’t be properly maintained as well. So I think cost becomes a huge driver, particularly again in terms of the disposable and reusable costs. 

If you’re going to put first-responder tools at every fire station, which probably isn’t the right thing to do, but if you wanted to do that, $10,000 is a lot for a fire department to pay for anything. So, if a detection tool costs more than a fire truck, it’s not something they’re going to invest in and it’s not going to be on their priority list. So the market requirements are going to drive the application of these products as well as research and development aspects.

Another good example of the effect of costs is the system that was deployed at the Olympics. In that case, we had a high-confidence, centralized laboratory with technicians working 24 hours a day. Samples were collected remotely by couriers and then brought back to the lab for analysis by technicians. The cost for that amount of manpower is very high. If you scale the operation up and add a lot more collectors, the costs become almost exponentially larger. 

We improved upon the original system with the development of the new autonomous pathogen-detection system. In the new system, the sampler, sample preparation, and detector are all in the same box, so the manpower requirements go down a lot. So, as you scale up detection operations by adding more systems, the cost increases more or less linearly rather than exponentially. 

Governmental Support

Is LLNL still under the Department of Energy or has it been shifted over to Homeland Security?

The defense diagnostics aspect of what we do is now handled under Homeland Security. Our entire chemical and biological warfare nonproliferation program was moved to the Department of Homeland Security. It was fortunate and a good fit for the department because LLNL already had efforts in place, had a lot of the research team, and had been thinking about the problem for a while. So we were able to hit the ground running.

Has that shift from Energy to Homeland Security changed the culture or the attitudes of the researchers in your department?

From the very beginning, we were working with companies and trying to get products out to the field quickly. So, if anything, that move made the efforts even a little bit more applied than they were before. I think it shifted the focus a little bit from more-long-term research to things could be put out in six months or a year because there was a perception that the country needed an immediate response to its biodefense needs.

Do you feel that the coordination among the various government agencies is working to further the development of biodefense diagnostics and that will it improve?

Yes. I think it’s actually getting better, primarily due to the formation of the Department of Homeland Security. Now there is a central player. They have a clear mission to work on this; and they serve as a place to go to coordinate among DOD, HHS, CDC, and NSF and everyplace else. I’ve been involved in multiple activities where they have tried to canvas what’s being done and to prevent duplication of effort.

Setting up a new department is a big task. As the department develops the infrastructure it needs to meet its mission, it will become more and more clear where the overlap is, and what their role is versus the military and versus others. 

It was much more amorphous before Homeland Security was established because it wasn’t clear exactly who owned the mission of civilian protection from biothreat. A lot of people were investing in it, either directly or indirectly, but no one was really mandated to do civilian protection, at that point, for biodefense.

Is it really clear now that Homeland Security is the main coordinating body in terms of developing biodefense diagnostics?

You have to be careful in how you use [the term] diagnostic in this sense. Certainly, I think Homeland Security is responsible for environmental monitoring, and that’s largely what we’ve been talking about—response mechanisms, places in time, places for triage, things like that. Certainly, Health and Human Services has a big role in developing new clinical devices for this area. I think they’re still in the early stages of negotiation, but the fact that they’re talking a lot and coordinating on a strategic level is good.

You mentioned that LLNL has been collaborating with the CDC in terms of validating assays; what sort of work has that involved?

We had developed unique capabilities in terms of building high-confidence nucleic acid assays. CDC is an unbiased player in all of this, and it is the agency that validates most of these products when the assays go out. So, from the beginning, we tried to include them in the process of how these assays were developed, how they were validated, how they would be used at CDC’s laboratory response network.

One of the philosophies at LLNL is to involve the stakeholders in the project investment from the beginning, as opposed to developing something and then trying to make it fit what the stakeholders’ requirements are.

What projects has LLNL collaborated on with USDA? Developing detection systems for crops, water supply, livestock would probably present a different set of challenges when compared with detection in human or environmental samples.

We’ve been working with USDA to look at the agricultural bioterrorism issue. We’ve been determining what their requirements are and whether some of the technologies in development for other areas will be of use. Much of our work with USDA will be affected by the cost drivers. The cost that the government is willing to sustain to protect a human population may not be reasonable for protecting millions of cattle. Other issues are how we will go about building a surveillance architecture for that, how the workflow is handled, how the population is sampled, and what the response mechanisms will be. All of these are different for agricultural than they are for human areas. 

What sort of collaborations have you had with the Department of Defense? 

We spend a lot of time collaborating with U.S. Army Medical Research Institute of Infectious Diseases (USAMRIID; Fort Detrick, MD) on some of these technologies. I think our relationship and collaborative work have gotten better since Homeland Security was established. In the past, we have also had direct funding from DOD agencies such as DARPA (Defense Advanced Research Projects Agency; Arlington, VA). They have been long-standing customers of LLNL for awhile. 

What standards for diagnostics in biodefense have been generated, if any at all?

They began to look at that issue last year. Part of what they’re working on is what type of infrastructure they will require for testing of real agents. For instance, a lot of testing right now is done at military facilities. I think it’s obvious to everyone that those facilities are not going to be enough to handle all the new devices and requirements that are coming out. I don’t think we have a good set of standards in metrics yet for testing most biothreat agents, but certainly the development of those is underway. 

A lot of our standards and metrics come from real-world incidents where we had to respond to something, and we found that there were gaps in our capabilities. Those then drive the need for higher-confidence detectors; and in some cases, more-robust platforms or higher-throughput platforms, as the case may be.

Do you think that application-specific standards might be developed for civilian, versus military, versus agricultural, versus environmental tests? 

Yes. I think it’s safe to say that there will be no universal set of standards, and there shouldn’t be. The development of standards has to be driven by the requirements for the application. Different media have different constraints, whether you’re measuring something in blood or whether you’re measuring something in a subway or an airport; it’s a very different type of scenario. 

Collaborating with Industry

Please go into some more detail about which IVD manufacturers LLNL has worked with and what biodefense technologies it has developed with those manufacturers.

We work with a lot of different manufacturers, so this isn’t an inclusive list. Cepheid (Sunnyvale, CA) is a company that we have worked with. Our history with that company is interesting, and it’s a model that we like to follow here. Much of its technology, the real-time PCR technology, was developed here at LLNL. One of the inventors of that technology went off and helped form the company, Cepheid, and then their Smartcycler system actually became integrated in some of our biodetection systems here at LLNL. 

An extension of that technology was licensed to and commercialized by Smiths Group (London), and they are making an instrument called the Bio-Seeq, which is a handheld PCR unit designed for first responders. 

Another company that we’ve had significant work with is Illumina (San Diego). They have a highly multiplexed bead-based format where they measure hundreds of thousands of things on the ends of fiber optics. It’s a similar format to the Affymetrix chip, but the beads allow it to be very flexible. 

We have also worked with companies that aren’t necessarily involved in the detector end, but rather are involved in either air samplers or fluidics. When we begin a new project, we look through the project requirements. We do a gap analysis to determine where we will need to look outside of LLNL for collaborators. We try to do a good job of determining when portions of the work have already been done and of including those innovators in the research on the components that we’re developing.

Does LLNL actively watch for systems that are being developed in the private sector that could possibly be adapted and used for biodefense?

Absolutely. We have active collaborations with a number of diagnostic manufacturers; and when new technologies or companies become available, we become aware of them. We usually make the effort to go and kick the tires to determine the status and the maturity of the technology, what the gaps are, and whether it’s useful for applications that we care about. 

How does LLNL go about seeking such collaborations? 

If there’s a need for a particular technology and we have funding through Homeland Security or other sources to address it, we actively go out and try to form collaborations with companies to meet whatever the requirements are. Sometimes, it actually ends up going the other way where companies approach us because they need some expertise here at LLNL. Since we’re a government industry, we follow all the conflict of interest requirements. So, when we do develop new technologies here, it’s obviously open for any company to license or bid on. 

What areas of expertise are unique to LLNL? 

Because we’ve been in the biodefense business for a while, we have a better sense of what the mission and the real requirements are for these products. There are certainly a lot of products that came out of LLNL like the Joint Genome Institute and the Human Genome Project that have fairly unique capabilities for determining the sequences of genes for nucleic acid probe development. As I mentioned earlier, a lot of the very rapid PCR instrumentation came out of here initially. 

We and other government institutions have fairly unique access to select agents and other things that a lot of times companies don’t have. We have archives of environmental samples for assay backgrounds; and we have all the near neighbors. In other words, we have a wide range of archives. 

If a system is looking at an anthrax sample, for instance, there are many bacteria that are not anthrax, but which could cross react with anthrax, that might lead to a false-positive result. So whenever you’re validating something for false positives, you want to make sure you run it against as many near neighbors as you can. And, that’s either done computationally or on the bench, depending on what you’re working with. 

Does LLNL offer money to the private sector to develop biodefense diagnostics?

Yes. We’re not a sponsor, in the sense that we don’t offer grants. However, when we collaborate, some of the money is subcontracted out to companies if they have unique capabilities. In some cases, we’re just buying an instrument from them. In other cases, it’s more of active collaboration where we’re paying for a piece of the development to be done by the company.

You mentioned that you do license technologies developed at LLNL to companies. How exactly does that work?

We have a commercialization segment within LLNL. I think that when we come up with a new technology, it’s posted in the Commerce Business Daily (CBD). I think that’s where we essentially start with the licensing process. Then companies come to our licensing department, and we proceed from there.

Did the events of 2001 improve the market for biodefense products, and do you think that this potential will increase well into the future?

That question calls for a complex answer because the civilian market for these products is still unknown. Most of the funding comes through the federal government. It is still unclear whether the technologies are going to be paid for at the state, city, or federal level. Certainly, I can say there’s been a much bigger interest in biodefense from diagnostic companies now than before the 2001 attack. When we approached companies before, some were interested in collaborating with us, but it wasn’t their primary market. So it was hard to get their attention completely. Now I think that, on a perception level, that’s changed. I think many companies view this as a new market area. And I think that’s actually helped accelerate the science behind the technology.

Previous to 2001, and even currently, the military has been the primary customer for these types of devices. They have a very defined market. They have bids and proposals that they send out, and there are companies that understand what that market looks like. Now it has moved into the civilian sector, detectors have different requirements. I still think the market is unknown in terms of how many instruments will be deployed and, again, who will pay for them. 

Improving Technology

What characteristics does LLNL consider important for a successful biodefense diagnostic?

It’s highly dependent on the application. From the beginning, we and the Department of Homeland Security involve the stakeholders in determining the requirements for a specific application. For example, a tool developed for first responders that allows them to make onsite measurements of biothreat agents should be extremely robust, very simple to use (which means there has to be a lot of automation), portable, and relatively fast. 

Those characteristics are very different from those required for an atonomous measurement where a system will be installed in one place. In that case, size doesn’t matter that much, but the system has to measure things continuously for days or weeks on end. Again, the system requirements really depend on the application. Homeland Security has defined sets of requirements based on system models. We develop technologies against those requirements.

Do the specificity levels have to be higher for equipment used by those first responders who have to go to a scene where there is a possible bioterrorism attack and where they must determine any immediate effect? Perhaps their requirements for specificity and sensitivity are different from those for equipment that’s used in a laboratory.

Yes. We specify selectivity and sensitivity whenever we term requirements. The issue with first responders to the anthrax attacks was not so much sensitivity as false-positive rates. They found that, initially, they were using tests for applications that they weren’t specifically designed for. They were getting false-positives, then responding by evacuating a facility or raising an alarm. It only took a few false alarms to lead people to think that the detectors weren’t good enough. 

False positive is probably one of the biggest obstacles we deal with for biothreat detection. For environmental measurements made in large public places like in an airport where you’re trying to provide information to policymakers so that they can respond properly, the policymakers and the public have to have high confidence in the result. Current strategies rely on a confirmation by a very high quality lab, where one or two layers of confirmation are used to say whether a result is a false positive when it comes in.

Do successful biodefense diagnostics differ according to the agents that they are designed to detect or can one diagnostic be used to detect multiple agents?

Certainly there are a number of autonomous platforms that measure multiple things from the same sample. Some of the tradeoffs you incur from multiplexing are that when you measure nucleic acids, you can’t measure toxins. So, if a detection technology is based solely on nucleic acid detection, many biothreat compounds of interest, like ricin, may be missed. On the other hand, nucleic acid assays are much more specific and sensitive than immunoassays. So you’re always dealing with a trade off. You actually need some combination of both, and they need to be applied in ways that are relevant to the application or scenario that you’re interested in. 

Will biodefense diagnostics continue to be increasingly multiplexed and able to handle not just anthrax, but anthrax, smallpox, plague, ricin, etc.?

Absolutely. In fact, not only is the number of things that can be detected an advantage of multiplexed systems, but multiplexing also offers an advantage in terms of cost. If you’re going to deploy a system in every city or give one to every first responder, the cost of the reagents begins to add up and to be a big deal. So the multiplexing capability allows for a significant reduction of cost. 

Many aspects of cost influence system development. There’s the upfront cost of buying the instrument itself, there’s the cost of maintenance (which is a really important issue when you’re leaving a system to run for the long term), and then there’s the disposable or reusable cost. The medical industry deals with this on a regular basis. If you’re developing an environmental monitor, you also want to reduce the cost of the reusables and consumables as much as possible if you’re making measurements every hour, seven days a week.

What efforts is LLNL making to make biodefense diagnostics better as well as faster?

I think that faster detection is a requirement that is going to become very important, both for first responders and for facility monitoring where an answer needs to be delivered in minutes rather than hours. We have some seed programs in that area looking at reagentless detection scenarios, using spectroscopic techniques, or new technologies that have very fast, very sensitive assays as opposed to the conventional PCR or immunoassays. LLNL is also devoting a lot of funding to resolving the issue of emerging or genetically engineered threats, and, how you go about dealing with those threats as opposed to an instance in which you know what it is you’re looking for. Those are probably two of the biggest research areas that we’re focused on.

Another important consideration is how the sample prep will be done. A lot of false positives, or in some cases false negatives, can be due to the biological effects of the chemical backgrounds in which the tests are conducted. For instance, if a detection system is working in a very dirty environment like the subway, and there are substances in the air that inhibit the assay, those substances either need to be cleaned up to make sure that the assay is working, or a positive or negative control needs to be built in so that the result can be read properly. How you handle and process the media is a big challenge.

What requirements and characteristics need to be considered for civilian biodefense applications as opposed to military applications? 

The military has well-defined concepts of their operations. For instance, they will be watching for a cloud of stuff coming toward a group of soldiers, and they have to protect the critical infrastructure. They have some toleration for false positives because they have built-in response mechanisms. That’s a much different scenario than if you’re trying to protect a civilian event like the Olympics. 

If you found a test to be positive for a biothreat agent at the airport, the governor would have to come in, evacuate everyone, and cancel flights; it is a much different response architecture than what you deal with in the military. And that really drives the requirements for these detectors to be different than what the military can accept. In these cases, I don’t think sensitivity is so much the driver as specificity.

What future challenges do you foresee arising from events that require the use of biothreat diagnostics?

When we’re talking about diagnostics, I think one of the bigger challenges is going to be developing rapid triage tools for dealing with a large attack and sorting out who’s been infected and who hasn’t. In the case of SARS, which obviously wasn’t a terrorist threat but an emerging pathogen, they quarantined on the order of 100:1 or 200:1 people for every person that actually had the disease. Such a large-scale quarantine had a huge economic impact, among other things. So, the need for mass-triage devices is obvious. 

It’s a different concept of operation than the way hospitals are used to dealing with things. Hospitals would be overwhelmed very quickly from sizeable attacks, especially if an infectious agent was used, because it’s difficult to protect the infrastructure as it exists today in the emergency rooms. They just don’t have enough beds or proper procedures in place for dealing with infectious disease attacks like they have for chemical attacks. So, the ways in which the hospitals’ responses change will drive the development of the detectors as well. 

Another challenge that we will have to face is dealing with unknowns, either genetically-modified substances, emerging threats, or things that come about as the result of an accident like the IL-4 incident that occurred in Australia where they ended up creating an antibiotic-resistant strain of cowpox virus. 

Copyright ©2004 IVD Technology