Technologies to Enable Autonomous Detection for BioWatch: Ensuring Timely and Accurate Information for Public Health Officials: Workshop Summary (2014)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

H State of the Art for Autonomous Detection Systems Using Immunoassays and Protein Signatures R. Paul Schaudies, Ph.D. GenArraytion, Inc. A white paper prepared for the June 25–26, 2013, workshop on Strate- gies for Cost-Effective and Flexible Biodetection Systems That Ensure Timely and Accurate Information for Public Health Officials, hosted by the Institute of Medicine’s Board on Health Sciences Policy and the Na- tional Research Council’s Board on Life Sciences. The author is respon- sible for the content of this article, which does not necessarily represent the views of the Institute of Medicine or the National Research Council. INTRODUCTION Antibody-based strategies dominated the biological agent detection landscape during the 1980s and well into the 1990s. Antibodies have been used by the U.S. military for the detection of biological warfare agents since the first Gulf War. The ability of immune systems to identi- fy and synthesize immunoglobulins against non-self-structural elements, primarily proteins, with excellent sensitivity and specificity is the back- bone of antibody-based detection and identification approaches. The introduction of nucleic acid–based detection and identification technologies, primarily nucleic-acid amplification technologies, in the later 1980s presented an ability to look at the genetic information in addi- tion to the structural elements of pathogenic organisms. While systems such as BioWatch have focused almost exclusively on nucleic-acid am- plification technologies, the continued development of antibody- and 173 

174 TECHNOLOGIES TO ENABLE AUTONOMOUS DETECTION FOR BIOWATCH protein-based recognition systems warrants a fresh look at the state of maturity of structural recognition approaches to biological identification. This paper highlights some of the more mature and promising technolo- gies that could be adapted by the BioWatch Program for the identifica- tion of harmful biological agents. All of the systems discussed in this paper can meet or exceed the detection requirements established for the next generation of BioWatch. The advantage of antibody-based and other protein-based technologies is speed. Many of these systems provide answers in 3 to 10 minutes and can operate in a continuous fashion at reasonably low cost with minimal infrastructure support. Some of the systems can reach sensitivities ap- proaching that of nucleic-acid amplification. STATEMENT OF THE PROBLEM The Department of Homeland Security (DHS) Office of Health Af- fairs is researching the potential to employ a fully autonomous net- worked biodetection capability that will be deployed, operated, and sustained, both indoor and outdoor, in selected BioWatch jurisdictions throughout the United States to continuously monitor the air for agents of biological concern. This requirement supports the DHS plans to increase the capability of the BioWatch system by augmenting and ultimately re- placing the current collection and biodetection capability with an auton- omous biodetection capability that will improve timeliness, time reso- lution, population coverage, and cost-effectiveness while enabling the program to stay within its fiscal constraints. The autonomous detection capability must be able to (1) rapidly pro- cess and accurately analyze aerosol samples with a high level of confi- dence; (2) automate and integrate the major system functions into the detector, including aerosol sample collection, preparation, analysis, and analytical results reporting; (3) operate in its intended indoor and outdoor environments; and (4) disseminate and archive analysis results and sys- tem operational data via the C3 network, known as the BioWatch Gen-3 Operations Support Service. The requirement is for autonomous biodetectors, for both indoor and outdoor use, that continuously monitor the air for agents of biological concern (24 hours per day, 365 days per year) and all necessary information technology, equipment, consumables, and technical support for the nationwide deployment, operations, and maintenance of this autonomous biodetection capability. 

APPENDIX H 175 Appropriateness to BioWatch Problem The use of antibodies, or immunoglobins, as a structural recognition element in a biological detection architecture takes advantage of one of the body’s premier defense systems. Antibodies represent the body’s ability to discriminate between self and non-self. The standard IgG immunoglobin is a Y-shaped molecule with the short arms representing identical binding elements and a base that can be anchored to a variety of surfaces. The binding region recognizes a small region, generally three to five amino acids in size, which is unique from the host that generated the antibody as part of an immune response. A single immune cell makes a single antibody structure generating “monoclonal” antibodies. The col- lection of individual antibodies from blood of a host represents a multi- tude of these individual antibodies and is termed polyclonal, each individual antibody targeting a unique structure. The antibodies are responsible for binding foreign entities to target them for destruction and removal from the body. The binding is specific and sensitive. Normal ranges of disassociation constants (indicating tightness of fit) range from 10−9 to 10−11. These values are important be- cause they are central in determining the sensitivity and specificity of a detection system utilizing the antibody as a detection element. Biological identification systems rely on the collection and detection of recognizable entities that are unique to the harmful biological agent and not found independent of it. The BioWatch Program is designed to protect against a catastrophic event caused by bacteria or viruses. Toxins are also a concern because they are potent at very low levels and repre- sent a challenge to nucleic acid–based detection systems. The argument that nucleic acids may contaminate the preparation has merit, but anti- body-based systems are more effective and consistent than nucleic acid– based systems for the detection and identification of toxins. In addition, there are some antibody-based systems that provide detection and dis- crimination of microorganisms at levels that support the autonomous de- tection requirements established for BioWatch Gen-3. Sample Collection and Processing Sample collection and processing are the ultimate driver for sensi- tivity. The fundamental objective for biological sample collection and preparation is to extract the target signatures from the environment into the volume required for one or more specific assays. A front-end air- 

176 TECHNOLOGIES TO ENABLE AUTONOMOUS DETECTION FOR BIOWATCH sampling device typically removes particles of defined size and mass into a collection matrix, generally an aqueous liquid. This liquid undergoes extraction(s) of various types to separate the biological signature(s) of interest from the clutter in the sample. As tests are run on smaller vol- umes, the concentration of target must increase, given a fixed level of detection. If a given test has a sensitivity of 100 signatures, then the con- centration (signatures/ml) is indicated in Table H-1. The signature may be a genetic sequence, an epitope for antibody binding, or a mass spec- trometer target. If the copy number required for detection decreases by a factor of 10, then all of the concentration numbers also decrease by that factor. This simple table highlights the requirement for efficient collection and pro- cessing for analysis of biological targets. Signature copy number from a single organism can be larger than the actual copy number of the organism in multiple ways. Most of the anti- bodies are directed against antigenic determinants that are present on the surface of the cell. Generally there are hundreds to thousands of copies of these structures on a single bacterium or virus. If the target organisms are sonicated to release the soluble contents, hundreds to thousands of membrane fragments can be released from a single organism (each be- having as a separate entity), providing many more independent binding opportunities and dramatically increasing the sensitivity of the system. In this way antibody detection and identification can compete with sensitiv- ities achieved by nucleic-acid amplification technologies. TABLE H-1 Required Concentrations Copy Number Required Volume Copy Number/ml Required 100 1.0 ml 100 100 1.0 ul 100,000 100 1.0 pl 100,000,000 100 1.0 fl 100,000,000,000 NOTE: fl = femtoliter; ml = milliliter; pl = picoliter; ul = microliter. 

APPENDIX H 177 Antibody–Antigen Interaction The vast majority of detection technologies that employ antibodies as structural recognition elements involve initial immobilization of a capture antibody on a solid substrate. This substrate can be a membrane, as in traditional lateral flow assays. Alternatively, an optically transparentmatrix can be used for evanescent wave utilization of suspended substrates, such as a bead used in magnetic separations or optical laser discrimination. Figure H-1 illustrates a basic approach, referred to as a sandwich assay, for the utilization of antibodies. The primary, or capture, antibody binds the target in such a way that the target is then associated with the capture substrate. Both antigen and reporter antibody binding events are critical to the fidelity of the detection system. Nonspecific binding is caused by a label, or reporter molecule, being localized with the capture antibody, and it creates noise in the system. It is generally accepted that a specific signal must be at least three times the intensity of the background to be a specific binding event. The HMGB1 Elisa Kit is a 2-step sandwich ELISA. FIGURE H-1 Visual representation of the sandwich-style assay in which the capture antibody (pink) is immobilized to a substrate. The red HMGB1 repre- sents the target that binds the capture antibody and is subsequently bound by the blue reporter antibody. NOTE: ELISA = enzyme-linked immunosorbent assay; HMGB1 = high-mobility group box 1 protein. SOURCE: http://dc184.4shared.com/doc/goS0-D6B/preview.html. 

178 TECHNOLOGIES TO ENABLE AUTONOMOUS DETECTION FOR BIOWATCH Reporter Molecules The reporter molecule is a readable signal that represents what is measured by the system and is intended to be a function of amount of target present in the sample. A wide variety of reporter systems are used in antibody-binding detection methods and represent a significant varia- ble in the sensitivity of various systems. The types of reporter systems can be divided into three broad categories. The first is optical reflectance, such as the colloidal label on secondary antibodies utilized in most lat- eral flow devices. The second is fluorescence whereby a reporter mole- cule is excited at one wavelength and then emits light at a different wavelength. The spectral and temporal excitation and emission differ- ences allow various methods of collection to enhance the signal-to-noise difference. The third method is an enzymatic reaction whereby a non- visible substrate is converted into a detectable product. The enzymatic products vary with different systems. PLATFORMS This section presents a variety of different platforms, all of which have been incorporated into a product that can reach TRL 6-plus between 2016 and 2020. Luminex The Luminex platform can accommodate both antibody-based and nucleic acid–based biological detection schemes. It is presented first be- cause it is relatively mature and is incorporated into the current BioWatch Program. The Gen-2 BioWatch Program has used the Luminex LX-200 platform to run a polymerase chain reaction (PCR) nucleic-acid amplification–based test. The LX-200 is also integrated in an autonomous PCR-based Gen-3 prototype that has undergone substan- tial field testing. A modification of the Gen-3 prototype includes the newer MagPix instrument that employs magnetic beads and in which a robust charge-coupled device (CCD) camera replaces the flow cytometer in the LX-200. For antibody-based detection, Luminex uses a bead-based platform in which the capture antibodies are covalently linked to color- coated beads. The BioWatch Program uses the LX-200 system, which incorporates a laser and a complex fluidics system. For future systems 

APPENDIX H 179 the newer MagPix instrument is recommended as it eliminates the flow cytometer laser and complex fluidics. Figure H-2 illustrates the Luminex open architecture platform that allows for the incorporation of new as- says into an existing menu of tests. Figure H-3 illustrates the detection mechanism for the MagPix sys- tem, which uses LEDs for bead optical interrogation and immobilizes the beads with a magnet to allow for increased image capture time, thereby improving sensitivity. Figure H-4 shows the Magpix system, which has a relatively small footprint and has user-friendly interfaces. The MagPix uses a ruggedized hardware platform that does not re- quire a company-trained individual to conduct operations. Manufacturer provider performance is provided in Table H-2. FIGURE H-2 Illustration of the multiplexing capability of existing Luminex antibody-based assays for biothreat targets. FIGURE H-3 Illustration of the optical imaging system of MagPix. NOTE: CCD = charged-coupled device; LED = light-emitting diode. 

180 TECHNOLOGIES TO ENABLE AUTONOMOUS DETECTION FOR BIOWATCH FIGURE H-4 Image of the complete MagPix system with computer interface. TABLE H-2 MagPix Performance Multiplexing capability Up to 50 individual analytes ® Reagent compatibility MagPlex magnetic microspheres Dynamic range 3.5 logs Microplate type 96-well plate Reading speed 96 wells ≤60 mins Sample temperature control 35°–60°C (95°–131°F) Sample volume uptake 20–200 ml Probe piercing Yes Auto adjust-probe height Yes Daily startup ≤15 minutes Sensitivity Approximately 106 copies of DNA or single-digit picogram levels of protein, 102–103 copies of bacteria. Dynamic range 3.5 logs Strengths The open architecture allows for sig- nificant flexibility for generation of additional tests in any laboratory. Significant field experience and de- sign engineering experience. Weakness Only a fraction of the beads incubated with target are analyzed, which dilutes the signal. 

APPENDIX H 181 PathSensors, Inc. PathSensors, Inc. (PSI) manufactures reagents and instrumentation for the detection of pathogens in indoor and outdoor environments, with a focus on biodefense and agri-food applications. Detection of pathogens is based on the CANARY® technology originally developed at Massa- chusetts Institute of Technology’s Lincoln Laboratory (Science, 2003) and licensed exclusively to PSI for these applications. The CANARY® technology shown in Figure H-5 differs significant- ly from the standard immobilized antibody. This is the only technology discussed in this paper that exploits the natural mechanism of antibody response when a specific antigen is detected. The antibodies are expressed in B lymphocytes that are the natural display mechanism within the host. Upon formation of a dimer with the antigen serving as the bridge, the antibodies stimulate a cascade of intracellular events. In this process, the cascade is linked to the generation of photons, which are detected using a photomultiplier tube. The process forces antigen and antibody together by centrifugal force and the response is as specific as the antibody and very rapid. The complete response occurs in less than 5 minutes. P a th o g e n P a th o g e n - s p e c if ic B io c h e m ic a l a n t ib o d ie s s ig n a lin g p a th w a y B io lu m in e s c e n t P r o t e in L u m in e s c e n c e FIGURE H-5 Principle of the CANARY® Biosensor. B-cell lines are engineered to express membrane-bound antibodies that trigger the release of intracellular calcium upon specific interaction with their target. These intracellular signaling processes rapidly culminate in the activation of bioluminescent proteins to gener- ate photons of light within seconds after contact between the biosensor and its cognate pathogen. SOURCE: Pathfinder CEO. 

182 TECHNOLOGIES TO ENABLE AUTONOMOUS DETECTION FOR BIOWATCH The principal benefits of the CANARY® technology include the speed of detection, which takes as little as 2 minutes, and the sensitivity that is conferred by the intracellular signal amplification, which allows detection of very low levels of target. In analytical assays, CANARY® demonstrates the ability to detect as few as 10–100 colony-forming units (cfu) of target organisms with a time to result of 3 minutes, including sample processing (see Figure H-6). This technology has been fielded in one or more Department of De- fense buildings for over 5 years with excellent results. The monitoring system operates with a trigger from a particle detector. The system has had thousands of collections and tests without any false positives over the 5-year period of operation. The system is fully capable of adding ad- ditional targets by the incorporation of different antibody genes into the progenitor B-cell line. This technology is a definite candidate for BioWatch, either as a trigger or as an integrated detection/identification system. Strengths: A significant strength of this approach is speed, with results within minutes with good sensitivity. Weakness: Reagent stability is a concern because the B cells must be viable in order to respond. Anything in the collected sample that harms the integrity of the B cell could cause a problem. Luminescent Light Output vs. Time for Bacillus anthracis 1000000 (cfu/sample) 0 0 100000 10K 10K 1K 10000 1K RLU 100 100 1000 25 25 10 100 10 10 s s s s s s s s s s s s 0s 0s 0s .0 .0 .0 .0 .0 .0 .0 .0 .0 .0 .0 .0 0. 4. 8. 12 16 20 24 28 32 36 40 44 48 52 56 TIME FIGURE H-6 Sensitivity and speed of the CANARY® technology. In analytical assays, CANARY® enables detection of 10 cfu of B. Anthracis (Sterne) spores in only 3 minutes. NOTE: RLU = relative luminescence unit. SOURCE: Pathfinder CEO. 

APPENDIX H 183 TABLE H-3 Selected Subset of Available B-Cell Lines with Analytical Sensitivities Antigenic Assay Agent Test Panel Target Sensitivity Bacillus Targets Positive against live 10 cfu/sample anthracis anthrax spore Sterne strain, (Sterne) (spores) coat gamma-killed and live <100 cfu/sample antigen Ames strain (Ames) Yersinia Targets F1 Tested against 50 cfu/sample pestis antigen inactivated Yp CO-92, and live Kim5 strains Francisella Targets cell Positive for <100 cfu/sample tularensis wall polysac- inactvated Ft Schu4, LVS charide Smallpox Orthopox Tested against live <100 pfu/sample specific vaccinia virus, (UV-Psoralen inactivated vaccinia inactivated virus (Lister strain) virus) 500 pfu/sample (live virus) Brucella Tested against B. suis, B. 50 cfu/sample spp. abortus, B. melitensis Venezuelan Targets VEE Positive against killed 500,000 equine virus capsid VEE3880, Menall, P676, pfu/sample Strain encephalitis protein TC-83, PTF-39, and variability in LoD (VEE) virus Trinidad donkey Ricin Ricin A chain Tested against active ricin <0.4 ng/sample Botulinum BoNT/AHc Tested against active 16 pg/sample toxin botulinum toxin A against active toxin Bacillus Tested against live spores 50 cfu/sample subtilis spores NOTE: BoNT/AHC = heavy chain of botulinum neurotoxin serotype A; cfu = colony-forming unit; LoD = limit of detection; LVS = live vaccine strain; ng = nanogram; pfu = plaque-forming unit; pg = picogram; spp. = species; UV = ultraviolet. 

184 TECHNOLOGIES TO ENABLE AUTONOMOUS DETECTION FOR BIOWATCH Research International Research International, Inc. (RI) has been involved in the BioWatch program for a number of years. It supplied a portion of the air sampling system to Lawrence Livermore National Laboratory during development of the autonomous pathogen detection system. Since then it has been a subcontractor to Northrop Grumman on the Gen-2 program and has manufactured the entire aerosol collection system for all the units deliv- ered to date. The primary operating costs for a BioWatch Program accrue from the biological assays needed to detect and identify pathogens. The assay chemistries required and the labor involved in frequent servicing are ex- pensive. To minimize these costs one could implement a strategy of monitoring bioaerosol concentrations with an optical biofluorescence instrument that has no consumables and minimal maintenance. Only when this instrument detects a rise in aerosol concentration would collec- tion and wet assay identification be initiated. The TacBioHawk is an integration of the TacBio biological particle detector and the BioHawk aerosol collector and bioidentifier into a single package. It is designed for continuous outdoor monitoring of biothreat agents over a wide range of environmental conditions. The biological particle detector uses ultraviolet-excited fluorescence to monitor biological particle levels in the air and elastic scattering to monitor other types of particles. This particle detection portion of the TacBioHawk is extremely useful for tracking background levels of air- borne materials and providing an alarm if there is an unusual increase in biological aerosol levels. However, it does not identify the type of bio- logical material detected, and for that reason it is correctly characterized as a “trigger.” The BioHawk portion of the system provides identification in the event of a TacBio alarm. When the BioHawk is triggered into operation by the TacBio, an aerosol sample is automatically collected and an eight- agent sandwich format fluoroimmunoassay is performed. The entire pro- cess can be completed with no human intervention within a 20- to 30- minute period. Additional samples can be collected and assayed within 17 minutes. 

APPENDIX H 185 TacBio Biodetector Module The TacBio aerosol biodetector’s mode of operation is based on the well-known fact that non-biological materials primarily absorb and scat- ter impinging ultraviolet (UV) light, whereas biological materials fluo- resce as well as scatter UV light. With the recent development of GaN semiconductor light-emitting diodes, high-brightness UV sources have become available that are energy efficient and more reliable and tem- perature tolerant than current UV lasers (<400-nm wavelength). In addi- tion, very-high-speed analog electronics are used to count the individual photons emitted by each particle, providing a “Geiger counter” operating mode that eliminates most of the drift and aging problems seen with ana- log optoelectronics. The TacBio requires little maintenance and no con- sumables, and it is an ideal first line of defense for monitoring areas that cannot be accessed regularly. Although aerosol data can be transferred to a remote location for analysis and storage using the TacBioHawk’s serial RS-232 data line, it is also stored internally on a removable SD memory card within the unit. This card can store up to 5 years of aerosol alarm data. BioHawk Module The BioHawk® is a portable eight-channel bioassay system with au- tomated sample collection capability that is suitable for high-sensitivity monitoring of live agents and protein toxins. Once activated by the TacBio, the BioHawk® collects aerosol particulates from the surrounding air with its wetted wall cyclone air sampler and automatically transfers the concentrated sample from the air sampler to a bioassay module. This module performs automated sandwich-style antibody assays in eight sep- arate detection channels over a period of 15 to 20 minutes, after which results are communicated to users through a touch-screen liquid-crystal display and by wireless or RS-232 link to personnel at a remote location. System operation may also be remotely controlled in real time with vir- tually every fluid or air handling function being capable of remote ma- nipulation. The bioidentification portion of the instrument can also analyze flu- ids from other sample collection systems or from liquid samples loaded into the instrument through a manual sample port. A built-in micro- computer provides unparalleled operating flexibility and sophistication. 

186 TECHNOLOGIES TO ENABLE AUTONOMOUS DETECTION FOR BIOWATCH Bioassays are performed within a disposable, credit card–sized, plas- tic assay coupon. Unlike other technologies on the market that are based on a “use once and discard” philosophy, this coupon and the reagents stored within it may be used for 10 assay procedures before being dis- carded. Because the coupon can execute up to 8 different assays simulta- neously, a total of 80 individual assays may be performed before the coupon is discarded. This capability can substantially reduce life-cycle cost. While possessing a high level of function and great versatility, the unit is still very easy to use. Most global functions such as air sampling and bioidentification are performed using multistep protocols (recipes) developed by RI and stored in the system’s computer memory. The un- sophisticated user needs only the most fundamental level of training since the internal processes and steps are preset through the built-in computerized recipes. For more sophisticated users, bundled Windows- based software allows the development of customized sample collection and detection protocols. The system is designed to be operated over a very wide temperature range, from −32°C to 60°C. In low-temperature environments sampled air is preheated to above freezing before entering the wetted wall cyclone sampler. At high ambient temperatures, cooling fans circulate air into the transport case to prevent overheating. Bioassay reagents are maintained in a very stable lyophilized (freeze-dried) form until they are needed and are only rehydrated when a bioassay is to be performed, greatly extend- ing the assay coupon’s useful life. Once hydrated, reagents are useable for a period of up to 48 hours. The current version of the TacBioHawk was developed for the armed forces of a NATO country, which specified using only a limited amount of power for the unit. Because air temperatures below freezing require heating the incoming air for operation of the wetted wall cyclone and for operation of the immunoassays, the air sampler was limited to 40 liters per minute (lpm). However, access to mains power will allow operation with a 325-lpm aerosol collector. If this volume proves insufficient, RI’s SASS 4000 aerosol preconcentrator, which operates at 3600 lpm, can be integrated into the system. It would increase overall sensitivity by rough- ly a factor of 10. 

APPENDIX H 187 The characteristics of the TacBioHawk are as follows:  Sensitivity o 8 × 104 cfu-min/m3 for Bacillus anthracis spores with 325 lpm aerosol collection. o Probability of detection of 95 percent for subspecies depend- ent on the particular Department of Defense Critical Rea- gents Program antibodies used.  Specificity o False-positive rate 1 to 2 percent for 17-minute assay with a limit of detection of 5 × 104 cfu/mL for B. anthracis. o Speciation is dependent on quality of C-reactive protein (CRP) antibodies  Flexibility o Current hardware supports simultaneous assays for one to eight agents in the same 2-mL sample. Addition of a second detection module would increase range to 16 agents. A max- imum of 20 agents is possible. o Quantitative information is saved in on-board flash memory and may be automatically transmitted to a remote computer if desired.  Measurement interval o Aerosol trigger: 1 minute updates o Sample collection and agent identification initiated on alarm from aerosol trigger o Sample collection: 5 minutes minimum recommended for first sample after alarm o Agent identification: 30 minutes assay time for first sample after first collection o Sample collection: user definable after initial collection o Agent identification: 17 minutes for each additional sample  Cost o Initial purchase: $125,000 to $150,000 in quantities <10 de- pending on exact configuration. Significant discounts availa- ble for larger volumes. o Operational maintenance: <$1,000/year hardware; ~3 hours/ quarter staff time for cleaning. o Consumables: $250 per assay coupon. Coupon only used if assay is triggered. Once triggered, up to 10 assays may be run. 

188 TECHNOLOGIES TO ENABLE AUTONOMOUS DETECTION FOR BIOWATCH  Automation o Once set up on site, the TacBioHawk is fully automated. On- ly maintenance and replacement of consumables required. o Mean time between failures: >10,000 hours with regular maintenance. o Availability: currently 99.9 percent or better depending on environment.  Operational environments o Can operate in dirty internal environments o Current design can operate −32°C to 60°C. Operation be- tween 50°C and 60°C is permitted up to a total time of 1,000 hours. o Once set up on site, the TacBioHawk is fully automated. On- ly maintenance and replacement of consumables required. Strengths: This system has extensive field testing. Weaknesses: Specificity of antibodies may not provide needed discrimination. Battelle REBS Biosensor Technology Battelle has patented an integrated collection and identification spec- troscopic technology, the Resource Effective Bio-Identification System (REBS), which provides a portable platform that is a lightweight, low- cost, networked, battery-operable system with near-real-time identifica- tion of environmental aerosols, surface contamination, and waterborne biological warfare agents (BWAs) (see Figure H-7). The process uses Raman spectroscopy for BWA identification and has undergone exten- sive government field testing over several years. By using a spectroscop- ic method, REBS significantly reduces the amount of consumables and overall life-cycle cost, reduces the amount of time to identify agents, and includes an agile software-updatable identification capability for emerg- ing threats. REBS autonomously processes samples from collection to detection/collection subsystems through to identification, thus removing errors that are often encountered in manual sample processing methods. REBS delivers an archive confirmatory sample via a removable vial compatible with genetic and other confirmatory methods. REBS delivers performance for approximately $0.05 per sample. This drastic reduction in identification cost versus that of traditional systems allows for contin- 

APPENDIX H 189 uous collection and identification of BWAs while eliminating the burden of reagent usage. Probability of Correct Identification REBS demonstrated a probability of correct identification above 97.5 percent for the samples supplied by a government sponsor for an identi- fier technology readiness evaluation. During limit-of-identification test- ing at Battelle Memorial Institute, all of the aerosol challenge samples were identified correctly. Based on the number of trials in this test, the 95 percent confidence bounds on the probability of correct identification are 96.5–100 percent. These tests included four BWAs, including gram- negative and gram-positive bacteria, virus, and protein simulants. Battelle has carried out extensive testing to establish the probability of correct identification, first based on Raman spectroscopy of single particles and, second, on the consensus of several identified particles FIGURE H-7 Battelle’s Resource Effective Bio-Identification System (REBS) provides collector, detector, and identifier functions within a single, portable system. SOURCE: Provided by Battelle. 

190 TECHNOLOGIES TO ENABLE AUTONOMOUS DETECTION FOR BIOWATCH within an aerosol collection. The probability of correct identification for a single particle has consistently performed to greater than 90 percent for any agent within an extensive spectral database. However, a system alarm utilizes several particles to form a consensus on the identity of a collected aerosol sample. Based on the consensus result, the system had been configured to identify at least three particles to confirm the aerosol identity. By accumulating information over these three particles, the probability of correct identification for a sample is above 97.5 percent for all agents tested. Detection and Identification False-Positive Rates REBS has been extensively tested concerning its detection and iden- tification false-positive rates in realistic environments. The REBS false- positive rate has been measured over several years in both indoor and outdoor locations throughout the United States. In government-sponsored tests, REBS demonstrated 1,629 hours of outdoor operation without a reported false-positive, at Ft. Bliss in El Paso, Texas, during July 2009. To date these tests correspond to 9,774 distinct samples with a 10-minute analysis interval. More recently, three REBS systems were tested in the Boston subway system from October 2012 through April 2013. In this test REBS demonstrated 2,454 hours of operation in the subway system without a reported false-positive. These tests correspond to 5,860 distinct samples with a 15-minute analysis interval. No samples in any trial have been identified as one of the targets within the system database, and therefore the probability of false- positives currently has an average of 0.0 percent with an upper bound of 0.001 percent (80 percent confidence interval). Naturally occurring bio- logical particulates are routinely detected in these tests, but none has been misidentified as Bacillus anthracis, Francisella tularensis, Vene- zuelan equine encephalitis, Clostridium botulinum toxin, or any other material within the system database. Size and Weight REBS is 1.5 cubic feet and weighs 38 pounds. Its dimensions are 18 × 12 × 12 inches. REBS requires electrical power from at least one source. Potential power sources include military-style batteries, 120- to 240-VAC (volts alternating current) at 50- to 60-Hz (hertz) facility pow- er, or 21- to 32-VDC (vehicle dynamics control) vehicle power. 

APPENDIX H 191 System Communications REBS is locally controlled via a simple push-button keypad. Alarms, warnings, battery life, system status, configuration, and meteorological conditions are viewable by a liquid crystal display interface adjacent to the keypad. Remote control and monitoring are achieved using a wired or wireless network communication protocol and remote computer graph- ical user interface. Expandable Threat Identification REBS detects and identifies at the genus, species, and strain levels based on the Raman signatures generated by a laser field interacting with the molecules and structural conformations of biological materials that comprise BWAs. Battelle currently maintains more than 120 spectral signatures in an on-board database, which is easily and cost-effectively expandable to include new threat materials through a simple software patch update via wireless connectivity. As a result, hardware modifica- tions, system recalibrations, perishable assays, and costly reagents are not required to expand the functionality of REBS. Battelle has constructed a library of signatures for various spores, bacteria, viruses, toxins, common battlefield interferants, chemical warfare aerosols, and some explosives. Expanding threat identification signatures takes as little as 24 hours and has been tested in government-sponsored technology readiness evalua- tions. Once the signature is determined, it can be sent wirelessly to fielded systems, enabling instantaneous enhanced threat warning capabilities. Analysis Interval The limit of identification for the REBS system is dependent on the analysis time interval. An increase in the time available for analysis in- creases the sensitivity of the system. An analysis time of 20 minutes is approximately two times more sensitive than the same system configured for a 10-minute analysis time. The additional analysis time allows for more materials to be collected and more of the sample to be analyzed. The ability to change the REBS duty cycle represents a unique capabil- ity; the duty cycle can be adjusted via software configuration changes. This attribute can be desirable for the various situations where the threat likelihood may be higher than normal. In government-sponsored testing, REBS was configured to collect and analyze samples with a 20-minute 

192 TECHNOLOGIES TO ENABLE AUTONOMOUS DETECTION FOR BIOWATCH duty cycle. In this test REBS demonstrated a 23-minute time to result at a BWA simulant concentration of approximately 200 particles per liter of air. System Cost and Consumables REBS is a reagentless BWA identification technology that eliminates the need for expensive and perishable reagents and consumables. REBS is significantly less expensive to operate than any other equivalent tech- nology, based on logistical burden expense reduction alone. Moreover, the REBS ability to add new threat agents via software offers a compel- ling value over similar technologies that are assay- or wet chemistry– based. Strengths: Cost per analysis is low, rapid cycle time, nondestructive analysis of sample. Weakness: Customers are not familiar with technology, therefore a sig- nificant barrier to entry. MesoScale The MesoScale technology is a mature system with many systems in operation. The electrochemiluminescence system offers significant sensi- tivity and was once a system of choice for the U.S. Army. The system provider was purchased, and the older systems are no longer fully sup- ported by the new owner. Antibody-based assays immobilize the capture antibody in a black plastic well, and then standard sandwich assays are performed with a proprietary reporter light-generating reaction. Electrochemiluminescence detection uses labels that emit light when electrochemically stimulated (see Figure H-8). Background signals are minimal because the stimulation mechanism (electricity) is decoupled from the signal (light). Labels are stable and nonradioactive and offer a choice of convenient coupling chemistries. They emit light at ~620 nm (nanometer), eliminat- ing problems with color quenching. Few compounds interfere with electrochemiluminescent labels, so one can use large, diverse libraries with confidence. Multiple excitation cycles of each label amplify the sig- nal to enhance light levels and improve sensitivity. 

APPENDIX H 193 FIGURE H-8 Representation of the electrochemical light-generating reporter system for MesoScale technology. NOTE: MSD = Meso Scale Discovery; Ru(bpy) = ruthenium tris difulleride; TPA = tripropylamine SOURCE: Company website. FIGURE H-9 Multispot plates, which offer arrays within the well for increased throughput and assay multiplexing, are available in 24-, 96-, and 384-well for- mats, with up to 100 spots per well. SOURCE: Company website. The system offers multiplexed detection technology by spatially sep- arated spots on the bottom of a 24-, 96-, or 384-well format microtiter plate (see Figure H-9). Sensitivities of the MesoScale systems are generally superior to standard ELISA assays. The sensitivities achieved by currently available systems are adequate for the BioWatch requirements. 

194 TECHNOLOGIES TO ENABLE AUTONOMOUS DETECTION FOR BIOWATCH Strengths: Sensitive reproducible tests. Proven technology. Weakness: Not currently configured for BioWatch operations. TECHNOLOGY FOR FUTURE APPLICATIONS Quanterix Corporation Simoa Technology Quanterix has developed, and is commercializing, a novel antibody- based technology that makes some fundamental changes when compared with existing approaches. The figure below provides a cartoon diagram of the various steps from sample collection to readout of results. Antibodies are covalently attached to microspheres and then incubat- ed with sample in such a way that the ratio of bead to target is slightly less than unity. The goal is to have only one target per bead. All reactions are in solution phase, which speeds the binding reaction compared to antibodies fixed on a planar surface. The beads occupied with target are then incubated with a secondary reporter antibody, which is covalently linked to an enzyme. The conjugated enzyme bead complex is rinsed in a solution contain- ing an optically inert substrate. The bead–substrate mixture is floated across a plane of microwells, each of which can hold only one bead. The wells are then sealed and the enzyme converts substrate into fluorescent product. Each fluorescent well represents a single target molecule. The process is diagramed in Figure H-10. The result is a quantified number of targets with sensitivities up to over 1,000-fold demonstrated with exist- ing technology. The current instrument, pictured in Figure H-11, is a floor standing unit. While the biodefense industry may not be a current market for Quanterix, the technical approach offers some significant improvements over existing technologies. This type of approach should be considered for post-2020 fielding. 

APPENDIX H 195 (a) (b) (c) FIGURE H-10 The Quanterix antibody-based technology. (a) Single protein molecules are captured and labeled on beads using standard ELISA reagents. (b) Tens of thousands of beads—with or without immunoconjugate—are mixed with enzyme substrate and loaded into individual femtoliter-sized wells. The microwells are sealed with oil. (c) Fluorophore concentration in the small sam- ple volume of wells containing the target analyte rapidly reach detectable limits using conventional fluorescence imaging and can be digitally counted. The per- centage of beads containing labeled immunocomplexes can be computed at low concentration because they follow a Poisson distribution; at higher concentra- tions the intensity of the aggregate signal provides an analog measurement. SOURCE: Company website. 

196 TECHNOLOGIES TO ENABLE AUTONOMOUS DETECTION FOR BIOWATCH FIGURE H-11 Quanterix Simoa HD-1 analyzer. SUMMARY All of the systems presented have the capability to meet the majori- ty of the autonomous detection system requirements. The most signifi- cant challenge will be the ability to provide sufficient specificity for organism discrimination. A clear example of this is that a Food and Drug Administration–approved antibody-based test for Bacillus anthracis, sold by Tetracore, Inc., uses the harmless vaccine strain for a positive control. This challenge may be overcome by the speed in which these technologies provide information to the customer community.